cme/moc developing a therapeutic game plan for the

PeerView .com /BXY900

Richard S. Finn, MDDavid Geffen School of Medicine at UCLA Los Angeles, California

Robin K. (“Katie”) Kelley, MDHelen Diller Family Comprehensive Cancer Center University of California, San Francisco San Francisco, California

Amit Singal, MD, MSUT Southwestern Medical Center Dallas, Texas

Chair

Faculty Faculty

Participate in interactive questions, download activity slides, and obtain your instant CME/MOC credit online.

This CME/MOC activity is jointly provided by Medical Learning Institute, Inc. and PVI, PeerView Institute for Medical Education.

CME/MOC

Developing a Therapeutic Game Plan for the Management of Hepatocellular Carcinoma: Expert Insights From the Patient CaseBook

What’s Inside

3

11

CaseBook 1: Starting With a Strong Offensive Line—Selecting Treatments for Patients With Newly Diagnosed Advanced HCC

CaseBook 2: The Game Changer—Navigating the New Wave of Second-Line Treatment Options in Advanced HCC

CaseBook 3: Going Back to the Playbook—Expanding Options With Clinical Trials for Every Patient With HCCSymposium Summary and Audience Q&A

21

30

https://www.PeerView.com/BXY900

PeerView .com /BXY900 2Go online to complete the post-test and evaluation for CME/MOC credit

Activity Information

Media: Enduring MaterialAccredited Activity Release Date: March 06, 2020Accredited Activity Expiration Date: March 05, 2021Time to Complete Activity: 90 minutes

Activity Description The therapeutic management of hepatocellular carcinoma (HCC) has significantly evolved in recent years with the approvals of multikinase, antiangiogenic, and checkpoint inhibitors. Moreover, the treatment landscape is poised to broaden as emerging strategies focus on unmet clinical needs (eg, the use of combination regimens with immunotherapy and targeted agents, dual checkpoint inhibition, and systemic treatment in early disease). This PeerView onDemand activity, based on an engaging CaseBook symposium, features a collection of patient case scenarios detailing critical clinical decision points along the HCC disease spectrum, including ideal sequencing of targeted agents and immune checkpoint inhibitors for patients with advanced HCC and optimal timing of transition from locoregional to systemic therapy for those with early-/intermediate-stage HCC. Each CaseBook segment has been paired with expert analysis of the clinical evidence on available therapeutic strategies and ongoing research on novel combinations in clinical trials.

Target Audience This activity has been designed to meet the educational needs of medical and gastrointestinal oncologists and other clinicians involved in the management of patients with hepatocellular carcinoma.

Educational Objectives Upon completion of this activity, participants should be better able to:• Summarize the safety and efficacy evidence on available multikinase inhibitors for newly-

diagnosed patients with HCC• Review the available treatment options, such as targeted and immune checkpoint therapies,

and their associated clinical data for patients with advanced HCC who have progressed after first-line treatment

• Describe ongoing clinical trials that are studying combination strategies, including immunotherapy with targeted therapy, dual checkpoint inhibitor therapy, or others for patients with HCC

• Examine the potential use of systemic treatments, including multikinase and checkpoint inhibitors, in earlier HCC disease settings

• Develop safe and effective treatment plans across multiple lines of therapy for patients with advanced HCC based on guideline recommendations and validated evidence

Providership, Credit, and SupportThis CME/MOC activity is jointly provided by Medical Learning Institute, Inc. and PVI, PeerView Institute for Medical Education.

This educational activity is supported by medical education grants from Exelixis, Inc., Genentech, and Merck & Co., Inc.

Physician Continuing Medical EducationThis activity has been planned and implemented in accordance with the accreditation requirements and policies of the Accreditation Council for Continuing Medical Education (ACCME) through the joint providership of Medical Learning Institute, Inc. and PVI, PeerView Institute for Medical Education. The Medical Learning Institute, Inc. is accredited by the ACCME to provide continuing medical education for physicians.

The Medical Learning Institute, Inc. designates this enduring material for a maximum of 1.5 AMA PRA Category 1 CreditsTM. Physicians should claim only the credit commensurate with the extent of their participation in the activity.

MOC Statement

Successful completion of this CME activity, which includes participation in the evaluation component, enables the participant to earn up to 1.5 MOC points and patient safety MOC credit in the American Board of Internal Medicine’s (ABIM) Maintenance of Certification (MOC) program. Participants will earn MOC points equivalent to the amount of CME credits claimed for the activity. It is the CME activity provider’s responsibility to submit participant completion information to ACCME for the purpose of granting ABIM MOC credit.

Participation information will be shared with ABIM through the ACCME’s Program and Activity Reporting System (PARS). Blinded individual or aggregated participant data may be shared with the funder of the activity.

Faculty DisclosuresChairRichard S. Finn, MD Professor of Clinical Medicine Division of Hematology/Oncology David Geffen School of Medicine at UCLA Los Angeles, California

Richard S. Finn, MD, has a financial interest/relationship or affiliation in the form of: Consultant and/or Advisor for: AstraZeneca; Bayer Corporation; Bristol-Myers Squibb; C stone; Eisai Inc.; Eli Lilly and Company; F. Hoffmann-La Roche/Genentech, Inc.; Merck & Co., Inc.; Novartis Pharmaceuticals Corporation; and Pfizer Inc.

FacultyRobin K. (“Katie”) Kelley, MD Associate Professor of Clinical Medicine Helen Diller Family Comprehensive Cancer Center Division of Hematology/Oncology University of California, San Francisco San Francisco, California

Robin K. (“Katie”) Kelley, MD, has a financial interest/relationship or affiliation in the form of: Consultant and/or Advisor for Bayer; Bristol-Myers Squibb; Exelixis, Inc.; and QED Therapeutics without compensation. Consultant and/or Advisor with consulting fees and/or travel support paid to self from F. Hoffmann-La Roche/Genentech, Inc. and Ipsen Biopharmaceuticals, Inc. Grant/Research Support from Adaptimmune; Agios, Inc.; AstraZeneca; Bayer; Bristol-Myers Squibb; Eli Lilly and Company; EMD Serono, Inc.; Exelixis, Inc.; Merck & Co., Inc.; Novartis Pharmaceuticals Corporation; Partner Therapeutics Inc.; QED Therapeutics; and Taiho oncology, Inc. Research Support paid to institution. Internal Displacement Monitoring Centre (IDMC) member with service fees paid to self from F. Hoffmann-La Roche/Genentech, Inc.

Amit Singal, MD, MS Associate Professor of Medicine Medical Director of the Liver Tumor Program Clinical Chief of Hepatology UT Southwestern Medical Center Dallas, Texas

Amit Singal, MD, MS, has a financial interest/relationship or affiliation in the form of: Consultant and/or Advisor for Bayer Corporation; Bristol-Myers Squibb; Eisai Inc.; Exact Sciences; Exelixis, Inc.; F. Hoffmann-La Roche; Genentech, Inc.; Glycotest; Merck & Co., Inc.; and Wako Diagnostics. Other Financial or Material Support from Target PharmaSolutions, Inc. for Steering Committee.

Planning Committee DisclosuresThe planners from Medical Learning Institute, Inc., the accredited provider, and PeerView Institute for Medical Education, the joint provider, do not have any financial relationships with an ACCME-defined commercial interest related to the content of this accredited activity during the past 12 months unless listed below.

Content/Peer Reviewer DisclosuresThe following Content/Peer Reviewers have nothing to disclose:

Natalie I. Vokes, MDPamela Ash, RN, MSN, CBCN

Disclosure of Unlabeled UseThis educational activity may contain discussions of published and/or investigational uses of agents that are not indicated by the FDA. The planners of this activity do not recommend the use of any agent outside of the labeled indications. The opinions expressed in the educational activity are those of the faculty and do not necessarily represent the views of the planners. Please refer to the official prescribing information for each product for discussion of approved indications, contraindications, and warnings.

DisclaimerParticipants have an implied responsibility to use the newly acquired information to enhance patient outcomes and their own professional development. The information presented in this activity is not meant to serve as a guideline for patient management. Any procedures, medications, or other courses of diagnosis or treatment discussed or suggested in this activity should not be used by clinicians without evaluation of their patient's conditions and possible contraindications and/or dangers in use, review of any applicable manufacturer's product information, and comparison with recommendations of other authorities.

Method of ParticipationThere are no fees for participating in or receiving credit for this accredited activity. For information on applicability and acceptance of continuing education credit for this activity, please consult your professional licensing board.

A statement of credit will be issued only upon receipt of a completed activity evaluation form and will be emailed to you upon completion. You will receive your certificate from [email protected]. If you have questions regarding the receipt of your emailed certificate, please contact via email at [email protected].

In order to receive credit, participants must view the activity and complete the post-test and evaluation form. A score of 70% or higher is needed to obtain MOC credit. There are no pre-requisites and there is no fee to participate in this activity or to receive MOC credit. Statements of Credit are awarded upon successful completion of the post-test and evaluation form.

About This CME/MOC ActivityPVI, PeerView Institute for Medical Education, and Medical Learning Institute, Inc. are responsible for the selection of this activity’s topics, the preparation of editorial content, and the distribution of this activity. Our activities may contain references to unapproved products or uses of these products in certain jurisdictions. The preparation of PeerView activities is supported by educational grants subject to written agreements that clearly stipulate and enforce the editorial independence of PVI and Medical Learning Institute, Inc.

The materials presented here are used with the permission of the authors and/or other sources. These materials do not necessarily reflect the views of PeerView or any of its partners, providers, and/or supporters.



Dr. Finn: On behalf of PeerView, I welcome you all, and thank you for joining me and my colleagues. I’m very fortunate to be joined by two very, very good friends and very eloquent speakers in the liver cancer space, Katie Kelley from UCSF, a local and a ‘49ers fan, I suspect, and Amit Singal from UT-Southwestern, probably a Cowboys fan, but it doesn’t matter.

But thank you all for joining us tonight for this CME event. And we’ll be talking about advances in liver cancer. And having been in the space for some time, it’s amazing to think that there are CME activities for liver cancer, considering the lack of progress in the field for so long. But that has clearly changed in the last 2 to 3 years and is changing approximately every 6 months now.

HCC Mortality in the United States Is Increasinga,1,2

Top 15 Causes of Cancer Death in the United States (2000-2014)

+ 2.7%

-2.5%

-2.5%

-1.9%

-2.6%

-0.9%

-2.3%

-1.1%

-2.5%

-0.8%

-0.6%

-2.3%

Annual Change, % (2000-2014) +3 -3 -2 -1

Lung Colon and rectum Breast Pancreas Prostate Leukemia Liver (#7) Lymphoma Bladder Brain Esophagus Ovary Kidney Myeloma Stomach

0 +1 +2

+0.1%

+0.4%

a Slide courtesy of Amit Singal, MD, MS. 1. http://seer.cancer.gov. Accessed October 21, 2019. 2. Global Burden of Disease Liver Cancer Collaboration et al. JAMA Oncol. 2017;3:1683-1691.

So, I will be opening the evening by discussing first-line options and practice-changing data. So, I take this slide from Amit. He put this together very nicely and really highlights why we’re speaking tonight. While there has been a lot of progress in the last few years, really, if we look at the cancer problem in the United States, we’ve made a lot of progress in lung, breast, and colon cancer—new treatments there, earlier detection, and improved outcomes.

CaseBook 1: Starting With a Strong Offensive Line—Selecting Treatments for Patients With Newly Diagnosed Advanced HCC

But if we look at the cancers that are increasing and leading causes of cancer death, liver cancer has actually grown more than other malignancies. And that reflects not only the burden of disease and the problem of underlying liver disease in the United States, which largely goes unnoticed, I think, until it becomes more advanced, and then even advanced liver disease and then looking for liver cancer in those patients with liver disease, but also reflects the fact that we haven’t had real progress in our therapeutic approach to this disease for some time.

Patient Case 1: 59-Year-Old Man With NASH Cirrhosis

59-year-old man with NASH cirrhosis

Child–Pugh A; AFP 178 ng/mL

No ascites or encephalopathy

ECOG PS 0

Previously treated with two rounds of TACE Follow-up MRI shows multifocal HCC (LR-5) with three

lesions, largest 6.5 cm with invasion into the right portal vein What are his options?

And that has changed, and we’ll start off our data presentation with a case. A 59-year-old gentleman with nonalcoholic steatohepatitis, what is also commonly called fatty liver disease, which will lead to cirrhosis. And while we have seen effective treatments for hepatitis C, which has been driving the incidence of liver cancer so long in the United States and presumably will be tailing off over time. But diabetes, fatty liver, obesity, and eventual cirrhosis will remain a leading cause of liver cirrhosis and liver cancer.

So this patient is Child–Pugh A, which is well compensated, and has an AFP of 178, but keep in mind, that is not used to diagnose the disease. The disease is typically going to be diagnosed on imaging, with a hypervascular lesion with delayed washout in a patient with cirrhosis.

This patient has no stigmata of end-stage liver disease, such as ascites or encephalopathy, a good performance status, and presents initially with focal disease in the liver, which is treated with chemoembolization with good disease control.

However, we know that chemoembolization is not a curative procedure, and eventually patients will progress beyond what can be controlled with chemoembolization. This patient develops now multifocal disease, the largest being 6.5 cm with invasion of the right portal vein. So this patient still has cancer confined to the

Richard S. Finn, MDDavid Geffen School of Medicine at UCLA Los Angeles, California



liver, but it is advanced because of the involvement of the portal vein. So, what are his treatment options at this time?

Phase 3 SHARP Trial: Sorafenib Versus Placebo in Advanced HCC1

a Majority of patients were Child–Pugh A. 1. Llovet JM et al. N Engl J Med. 2008;359:378-390.

Prob

abili

ty o

f Sur

viva

l

Months Since Randomization

0.00

0.25

0.50

0.75

1.00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

HR = 0.69 (95% CI, 0.55-0.87) P = .00058

Sorafenib OS = 10.7 mo Placebo OS = 7.9 mo

Stratification • Macroscopic

vascular invasion (portal vein) and/or EHS

• ECOG PS ≤2 • Geographic

region

Sorafenib 400 mg orally

twice daily continuous dosing

(n = 299) R N = 602a

Placebo 2 tablets orally

twice daily continuous dosing

(n = 303)

Sorafenib

Placebo

299 290 270 249 234 213 200 172 140 111 89 68 48 37 24 7 1 0 303 295 272 243 217 189 174 143 108 83 69 47 31 23 14 6 3 0

Number at Risk Sorafenib

Placebo

0

So, since 2008, sorafenib has been the standard of care for these patients. The SHARP study, which you’re all very familiar with, established a role for systemic treatment in the management of liver cancer. Keep in mind, before these data, no systemic drugs were shown to improve survival in liver cancer, and that gave an opportunity for locoregional treatments to fill that void.

Chemoembolization was something that had been done for many years, and even in patients who had characteristics beyond those that showed a benefit with chemoembolization, it continued to be done.

Then comes the SHARP study, which shows that for patients such as the one we just talked about, who has advanced disease, sorafenib improves survival. This study had a median of 3 months, but a risk reduction, as represented by the hazard ratio, of 0.69, or a 31% decrease in the risk of death.

And this study ushered in a new phase, a new way of thinking about liver cancer—that systemic treatment can change the natural history. However—and I remember being interviewed at the time of the SHARP study—what’s the projection for the future of liver cancer? “Oh, this is great. We’re going to have new drugs every few years.” We had trials, a lot of trials, but we did not have any new drugs for about a decade.

GIDEON Trial: Tolerability in Child–Pugh A and B1

a Includes Child–Pugh C, 74 patients; NE, 493 patients. b Treatment emergent. c Calculated based on 365.25 days per year. d Patients with dosing data. e Higher amount of Child–Pugh B patients discontinued treatment compared with Child–Pugh A patients. 1. Marrero JA et al. American Society of Clinical Oncology (ASCO 2013). Abstract 4126.

Total (N = 3,202)a (100)

Child–Pugh A (n = 1,968) (61.5)

Child–Pugh B (n = 666) (20.8)

All Drug Related All Drug Related All Drug Related AE, %b

All grades 85.3 66.0 84.0 68.5 88.6 64.4 Grade 3/4 31.7 23.5 32.4 25.6 31.5 21.9 SAEs 43.3 9.3 36.0 8.8 60.4 14.1

AE rate, event per patient-yearb,c

Any AE 1.61 1.24 1.44 1.17 2.14 1.55 Diarrhea 0.58 0.51 0.54 0.48 0.71 0.62 HFSR 0.51 0.50 0.55 0.54 0.42 0.41 Fatigue 0.45 0.29 0.38 0.27 0.62 0.34 Discontinuation due to AEs, %b 31.4 28.9 40.1 Daily dose, mg (median)d 688.0 677.0 741.5 Treatment duration, wk (median)d 15.0 17.6 9.9

Sorafenib safety results during treatment are generally consistent across patients irrespective of liver functione

And we did gain some real-world experience with sorafenib over that time, which assured us of the safety of the drug. We know that hand-foot syndrome, GI toxicity, such as diarrhea, anorexia, weight loss, are the most common things. And the incidence of these side effects seems pretty stable between Child A and Child–Pugh B patients. But the expected activity of sorafenib in Child–Pugh B patients is lower, meaning that survival is very much dictated by performance status as well as their Child–Pugh status.

And in clinical research versus clinical practice… In research, we concentrate on the Child–Pugh A patients, because it’s in those patients that we can show the anticancer activity of a systemic drug. For patients who have more decompensated liver disease, their risk of death is from cirrhosis, and it might be very hard in the context of a clinical trial to show that we can change an outcome or improve an outcome as related to their cancer.

However, in the real world, not every patient we see clearly fits into a clinical trial paradigm. They have some decompensated liver disease, and they may be Child–Pugh B. They might not be performance status 0 or 1. But data from this observational study, the GIDEON study, at least gives us some confidence that the drug is safe and tolerated similarly as in the Child–Pugh A population.

• Lenvatinib is an oral multikinase inhibitor that targets VEGFR1-3, FGFR1-4, PDGFRα, RET, and KIT1-4

• There have been 4 failed phase 3 trials in frontline HCC in the past 10 years5-8

Lenvatinib: REFLECT Study

1. Matsui J et al. Int J Cancer. 2008;122:664-771. 2. Matsui J et al. Clin Cancer Res. 2008;14:5459-5465. 3. Tohyama O et al. J Thyroid Res. 2014;2014:638747. 4. Yamamoto Y et al. Vasc Cell. 2014;6:18. 5. Cheng AL et al. J Clin Oncol. 2013;31:4067-4075. 6. Johnson PJ et al. J Clin Oncol. 2013;31:3517-3524. 7. Cainap C et al. J Clin Oncol. 2015;33:172-179. 8. Zhu AX et al. J Clin Oncol. 2015;33:559-566.

In Vitro Kinase Inhibitory Profiles3

IC50, nmol/L Lenvatinib Sorafenib VEGFR1 4.7 21 VEGFR2 3.0 21 VEGFR3 2.3 16 FGFR1 61 340 FGFR2 27 150 FGFR3 52 340 FGFR4 43 3,400 RET 6.4 15 KIT 85 140 PDGFRα 29 1.6 PDGFRβ 160 27 BRAF 8,700 310 RAF1 1,600 46

So the first new drug in the frontline setting—and actually now we have only two drugs that are currently approved in first line—was lenvatinib. Lenvatinib, like sorafenib, is a small molecule inhibitor of the VEGF receptors, very potent, similar to sorafenib.



It differentiates itself, as it does have increased kinase activity against the fibroblast growth factor receptor, and FGFs have been pursued for some time as a mechanism of resistance to VEGF—that if you develop resistance to VEGF-driven angiogenesis, perhaps the FGF family can drive that. But also, FGF is a growth factor that has been described as important in proliferation in liver cancer. So there are some differences in the kinase profile, and, in a small phase 2 single-arm study, lenvatinib seemed to have a little higher response rate and very provocative survival.

Phase 3 REFLECT Trial: Lenvatinib Versus Sorafenib in the First Line1

Global, Randomized, Open-Label, Phase 3, Noninferiority Study

• Primary endpoint: OS • Secondary endpoints: PFS, TTP, ORR, QOL, and PK lenvatinib exposure parameters • Tumor assessments were performed according to mRECIST by the investigator

• Patients with unresectable HCC • No prior systemic therapy

for unresectable HCC • ≥1 measurable target lesion per mRECIST • BCLC stage B or C • Child–Pugh A • ECOG PS ≤1 • Adequate organ function • Patients with ≥50% liver occupation, clear

bile duct invasion, or portal vein invasion at the main portal vein were excluded

Sorafenib 400 mg twice daily

(n = 476)

Lenvatinib 8 mg (body weight <60 kg) or 12 mg (body weight ≥60 kg)

once daily (n = 478)

1. Kudo M et al. Lancet. 2018;391:1163-1173.

Stratification • Region

(Asia-Pacific or Western)

• MVI and/or EHS (yes or no)

• ECOG PS (0 or 1) • Body weight

(<60 kg or ≥60 kg)

N = 954

R

So that led to the REFLECT study. The REFLECT study was a very large—over 950 patients—open label, so everybody knew what they were getting—trial of lenvatinib versus sorafenib in patients with advanced liver cancer, Child–Pugh A.

They used this term Barcelona B or C. And for those who aren’t familiar, the staging system in liver cancer tends to be the BCLC staging system. And those who have C are patients who have metastatic disease or vascular invasion, whereas those who are Barcelona B—that’s not Child–Pugh B, but Barcelona B—are those who have typically liver-confined disease and are in good shape and are the ones in which chemoembolization has been shown to improve survival. But again, like in this case, even though you’re Barcelona B and you get TACE, eventually you will stage migrate to C, or you will become TACE refractory.

Unique to this study, which was different from other phase 3 studies in advanced liver cancer, is that it excluded patients who had more than 50% of their liver involved by tumor or clear bile duct invasion, not the most common thing in liver cancer, but invasion of the main portal vein, right?

So when we say main portal vein, we’re talking about the portal vein that is outside the liver, right? The portal vein comes in, and then it branches to the right and left in subsequent branches. It’s very common for liver cancer to grow into those vessels, and it starts in a small branch, and eventually migrates to the main portal vein. But those patients who had main portal invasion were excluded, and they generally carry a very poor prognosis.

And here you see the stratification factors. And lenvatinib, unlike sorafenib, is based by weight. Less than 60 kg, 8 mg, or 12 mg for

those who are greater than 60 kg. And that is different than the doses used in thyroid cancer or kidney cancer.

REFLECT Primary Endpoint: Noninferior OS Compared With Sorafenib1


So this study was powered for noninferiority, and it met that endpoint. So we saw lenvatinib did 13.6 months survival; sorafenib, 12.3 months; and sorafenib performed better here than in the SHARP study, which was about 11 months. And this hazard ratio of 0.92.

And here you see the confidence interval, the upper limit being 1.06. If it was up less than 1.08, that allowed us to declare this a positive study for noninferiority.

REFLECT Secondary Endpoint: Kaplan–Meier Estimate of PFS1


Now, interestingly, the primary endpoint of OS was noninferior, but if we look at some of the secondary endpoints, we see that lenvatinib performed better. Here, looking at progression-free survival, 3.7 months to 7.4 months, hazard ratio of 0.66.



Lenvatinib Had a Significantly Increased Objective Response Rate Compared With Sorafenib1

Parameter mRECIST by

Independent Review RECIST v1.1 by

Independent Review

Lenvatinib (n = 478) Sorafenib (n = 476) Lenvatinib (n = 478) Sorafenib (n = 476)

ORR, n (%) 194 (40.6) 59 (12.4) 90 (18.8) 31 (6.5)

95% CI 36.2-45.0 9.4-15.4 15.3-22.3 4.3-8.7

Odds ratio (95% CI) 5.01 (3.59-7.01) 3.34 (2.17-5.14)

BOR, n (%)

CR 10 (2) 4 (1) 2 (<1) 1 (<1)

PR 184 (38) 55 (12) 88 (18) 30 (6)

SD 159 (33) 219 (46) 258 (54) 250 (53)

Durable SD 84 (18) 90 (19) 163 (34) 118 (25)

PD 79 (17) 152 (32) 84 (18) 152 (32)

NE/unknown 46 (10) 46 (10) 46 (10) 43 (9)


And interestingly, responses were higher with lenvatinib. For a decade, we know we could improve survival—survival being the most important endpoint in cancer studies, I would argue—but we could improve survival with sorafenib without a high response rate.

Here, with lenvatinib, if we look at overall response, here, looking at independent review, using modified RECIST, which again is a way to look at liver tumors that takes advantage of the unique characteristics of cancers in the liver, liver cancers in the liver—that they’re hypervascular.

So instead of measuring the total size of the tumor, the sum of the longest diameter of the tumor, we’re measuring the sum of the longest diameter of the enhancing component. So by modified RECIST, the response rates were about 40% with lenvatinib, 12% with sorafenib. By conventional RECIST, response rates were still fairly high, lenvatinib about 19%, and, as we saw in historical studies with sorafenib, single-digit response rates, about 7%.

Adverse Event Profile Differs Between Sorafenib and Lenvatinib1


AE Lenvatinib (n = 476), n (%) Sorafenib (n = 475), n (%) Any Grade Grade 3/4 Any Grade Grade 3/4

Hypertension 201 (42) 111 (23) 114 (30) 68 (14) Diarrhea 184 (39) 20 (4) 220 (46) 20 (4) Decreased appetite 162 (34) 22 (5) 127 (27) 6 (1) Decreased weight 147 (31) 36 (8) 106 (22) 14 (3) Fatigue 141 (30) 18 (4) 119 (25) 17 (4) Palmar-plantar erythrodysesthesia 128 (27) 14 (3) 249 (52) 54 (11) Proteinuria 117 (25) 27 (6) 54 (11) 8 (2) Dysphoria 113 (24) 1 (0) 57 (12) 0 (0) Nausea 93 (20) 4 (1) 68 (14) 4 (1) Decreased platelet count 87 (18) 26 (6) 58 (12) 16 (3) Abdominal pain 81 (17) 8 (2) 87 (18) 13 (3) Hypothyroidism 78 (16) 0 (0) 8 (2) 0 (0) Vomiting 77 (16) 6 (1) 36 (8) 5 (1) Constipation 76 (16) 3 (1) 52 (11) 0 (0) Elevated AST 65 (14) 24 (5) 80 (17) 38 (8) Rash 46 (10) 0 (0) 76 (16) 2 (0) Alopecia 14 (3) 0 (N/A) 119 (25) 0 (N/A)

Now, the adverse event profiles of these two drugs have some overlap and some differentiation. Both of them cause hypertension, but lenvatinib is a much more potent angiogenic inhibitor, and therefore we see more grade 3/4 and all-cause hypertension. Both drugs cause anorexia and diarrhea, similarly, and weight loss, as well as fatigue.

Hand-foot syndrome, we know, is a problem with sorafenib. It’s one of the more common side effects. Certainly, higher grade is

more common with sorafenib than lenvatinib in this study, and even all grades is less with lenvatinib.

However, more potent and more frequent with lenvatinib is proteinuria, again, representing this VEGF effect and on-target effect. And therefore, urinalyses need to be performed are on a regular basis to monitor that, just as we do with other VEGF-targeted drugs.

New Directions With Immunotherapy

So, immunotherapy has impacted every area of oncology, right? And certainly, liver cancer has been touched.

Combining VEGF Inhibition and PD-1/PD-L1

• Bevacizumab (anti-VEGF) is an antiangiogenic agent with additional immunomodulatory effects

• In combination, bevacizumab may further enhance atezolizumab’s efficacy by reversing VEGF-mediated immunosuppression to promote T-cell infiltration into the tumor

Atezolizumab Promotes T-cell

activation by allowing B7.1 co-stimulation1

Bevacizumab Promotes

DC maturation2,11,12

Bevacizumab Normalizes the tumor vasculature, increasing T-cell infiltration2-6

Bevacizumab Decreases the activity of immunosuppressive cells (MDSCs and Tregs)2,3,7-10

Atezolizumab Restores anti-cancer immunity1 with activity further enhanced through VEGF-mediated immunomodulatory effects

Activated T cells

DCs

Tumor antigens

Tumor cells

1. Chen DS, Mellman I. Immunity. 2013;39:1-10. 2. Hegde PS et al. Semin Cancer Biol. 2018;52:117-124. 3. Wallin JJ et al. Nat Commun. 2016;7:12624. 4. Goel S et al. Physiol Rev. 2011;91:1071-1121. 5. Motz GT et al. Nat Med. 2014;20:607-615. 6. Hodi FS et al. Cancer Immunol Res. 2014;2:632-642. 7. Gabrilovich DI, Nagaraj S. Nat Rev Immunol. 2009;9:162-174. 8. Roland CL et al. PLoS One. 2009;4:e7669. 9. Facciabene A et al. Nature. 2011;475:226-230. 10. Voron T et al. J Exp Med. 2015;21:139-148. 11. Gabrilovich DI. Nat Med. 1996;2:1096-1103. 12. Oyama T et al. J Immunol. 1998;160:1224-1232.

And initially the approvals of nivolumab and pembrolizumab, which we’ll hear about from Katie, were based on single-arm phase 2 studies in the second-line setting. However, as many of you are aware—and it’s a great privilege to share these data with you, which is probably one of the first times, I think, in the United States, as it was only presented at the ESMO Asia meeting in November 2019—it is this combination of VEGF inhibition and PD-L1 inhibition.

And while this slide focuses on the roles of bevacizumab and atezolizumab, I think in many ways there is a generic effect here for PD-1 inhibition and VEGF- or VEGFR-targeted agents.

Now, bevacizumab, you know, has been around for a long time, and conventional thinking and the base of the development was that we starve tumors of blood. Angiogenesis is a hallmark of cancer, and, therefore, we can have a cytostatic effect. It



normalizes the vasculature, which probably explains its synergy with chemotherapy in lung cancer, colon cancer, and other malignancies.

But our understanding of the microenvironment and the role of VEGF and the endothelium has evolved over time. And there are preclinical data that suggest that bevacizumab can change the immune microenvironment decreasing the amount of immunosuppressor cells, increasing those that have immune presentation cells.

And then, as you know, drugs like atezolizumab or nivolumab, pembrolizumab, durvalumab, among others, by interfering with a PD-1/PD-L1 interaction can activate T cells to attack the tumor. And so this paved the way for the combination now of bevacizumab and atezolizumab in many malignancies. And here we’re going to be looking at data in liver cancer.

Atezolizumab Plus Bevacizumab in Advanced HCC: Response1

a Four patients were unevaluable. b Region data from one patient are missing. c Baseline AFP data from five patients are missing. d EHS and MVI baseline data from two patients are missing. 1. Pishvaian MJ et al. European Society for Medical Oncology 2018 Congress (ESMO 2018). Abstract LBA26.

ORR Overall, n/n (%)a

CR PR

23/73 (32) 1/73 (1)

22/73 (30) SD 33/73 (45) PD 13/73 (18) By region, n/n (%)b

Asia excluding Japan Japan/USA

12/41 (29) 10/31 (32)

By etiology, n/n (%) HBV HCV Nonviral

11/36 (31) 10/23 (43) 2/14 (14)

By baseline AFP, n/n (%)c

<400 ng/mL ≥400 ng/mL

12/41 (29) 11/27 (41)

By EHS/MVI, n/n (%)d

EHS and/or MVI MVI negative EHS negative Neither EHS nor MVI

18/64 (28) 13/32 (41) 9/22 (41) 5/8 (63)

Max

imum

SLD

Red

uctio

n Fr

om B

asel

ine,

%

100

80

60

40

20

0

-20

-40

-60

-80

-100

CR PR SD

NE PD

So the first study was this single-arm phase 1, which then became a phase 2 study, and its most mature data presentation had an overall response rate of 32%. Now, that is a number we have not seen in liver cancer data sets. Here, with two monoclonal antibodies, and you can see on the waterfall plot here, even those who had responses, and then a number of patients, 45% of patients, had stable disease as their best response.

Phase 1b: Atezolizumab Plus Bevacizumab in Advanced HCC: Response1 (Cont’d) Other Responses

DCR (CR+PR+SD), n/n (%) 56/73 (77)

≥16 wk 48/73 (66)

≥24 wk 34/73 (47)

Median DOR (range), mo NR (1.6+ to 22.0+)

≥6 mo, n/n (%) 12/23 (52)

≥12 mo, n/n (%) 6/23 (26)

Ongoing response, n/n (%) 19/23 (83)

Median follow-up, mo 7.2

SLD Change Over Time and Duration of Response per INV-Assessed RECIST v1.1

1. Pishvaian MJ et al. ESMO 2018. Abstract LBA26.

Cha

nge

in S

LD F

rom

Bas

elin

e, %

Time, mo

PR/CR (n = 23) SD (n = 33) PD (n = 13)

First new lesion NE (n = 1)

First PD

Median PFS was 14.9 mo

And here you see the spider plot for this, this data set, and the median PFS was about 15 months. So, like we saw with other IO

agents, responses tend to be very durable, and here the response rate was higher than we saw with single agents.

Phase 1b: Atezolizumab Plus Bevacizumab in Advanced HCC: Safety1

1. Pishvaian MJ et al. ESMO 2018. Abstract LBA26.

Grade 3/4 TRAEs (≥5% of Patients); n = 103 n (%) Hypertension 10 (10)

Most Common AEs (≥20% of Patients); n = 103 n (%) Decreased appetite 29 (28) Fatigue 21 (20) Rash 21 (20) Pyrexia 21 (20)

Grade ≥3 Atezolizumab AESIs Requiring Systemic Corticosteroids n (%)

Pneumonitis 2 (2) Encephalitis autoimmune 1 (1) Drug-induced liver injury 1 (1) Colitis 1 (1) AST increased 1 (1) Gamma-glutamyltransferase increased 1 (1) Diabetes mellitus 1 (1) Pancreatitis 1 (1)

And the safety did not include anything new. Obviously, we were dealing with a cirrhotic population. We get concerned about underlying liver function. But really, in this selected patient population—it was Child–Pugh A—all these patients, importantly, had endoscopies to make sure they did not have varices. So they had to be scoped within 6 months.

We did not see a high incidence of GI bleeding, and generally, we saw a side effect profile that could be explained mostly by bevacizumab and, to some degree, IO events from atezolizumab. And here you see a spattering of events that have been described with these drugs.

Updated Results: Atezolizumab Plus Bevacizumab1

Atezolizumab + Bevacizumab (n = 60)

Atezolizumab (n = 59)

PFS 5.6 mo 3.4 mo

HR (80% CI) = 0.55 (0.40-0.74); P = .0108

Any grade TRAEs 68% 41%

Grade 3-4 TRAEs 20% 5%

Atezolizumab + Bevacizumab (n = 104)

ORR 36% (37 patients), with 75% ongoing

Any grade TRAEs 88%

Grade 3-4 TRAEs 39%

1. Lee M et al. ESMO 2019. Abstract LBA39.

Now, the most recent data that had come out to support this combination was this study that looked at single-agent atezolizumab versus the combination. We know 10 years ago many of us did studies with bevacizumab in liver cancer, 2006 to 2009. And it didn’t look all that active. It was mostly cytostatic. In retrospect, maybe if bevacizumab went versus placebo, maybe we would have had a positive study. But needless to say, it was not pursued for registration at that time.

However, we needed good data with single-agent atezolizumab to give us a sense whether the combination is synergy, or maybe it’s just atezolizumab’s activity in liver cancer.



Needless to say, the combination here gave a response rate of 36%, with long duration. And if we look at PFS, the atezolizumab PFS was 3.4 months, and that was increased to 5.6 months with the combination, suggesting that there is an additional benefit with bevacizumab compared with atezolizumab.

And there were some more adverse events when the combination was used as compared to single-agent atezolizumab. But again, when we saw in the other cohort, and as we’ll see in the phase 3 study, those are generally manageable.

Phase 3 IMbrave150 Trial: Atezolizumab Plus Bevacizumab Versus Sorafenib in Untreated Patients1

Atezolizumab + bevacizumab

Sorafenib

• Primary endpoints: PFS and OS

1. https://clinicaltrials.gov/ct2/show/NCT03434379. Accessed January 23, 2020.

R

• Locally advanced or metastatic and/or unresectable HCC

• No prior systemic therapy for HCC • ≥1 measurable untreated lesion • ECOG PS 0-1 • Esophagogastroduodenoscopy

(EGD) within 6 months • Adequate hematologic

and end-organ function • Child–Pugh A

N = ~480

So the IMbrave150 study was the definitive study of atezo and bev versus sorafenib. This was an open-label study, giving two IV drugs versus an oral drug. This was a very classic, typical phase 3 study, good performance status, Child–Pugh A, Barcelona B or C, but advanced cancer. And again, patients had endoscopies within 6 months. And it had two endpoints, progression-free survival and OS.

And many studies now in the frontline setting have included progression-free survival. Because we now have so many second-line agents, we’re seeing an increase in overall survival for patients with newly diagnosed liver cancer. And it has raised the question of whether we will we be able to show OS given crossover and things like that.

Phase 3 IMbrave150: PFS1,a

a Data cutoff: August 29, 2019; median survival follow-up: 8.6 months. b Assessed by IRF per RECIST 1.1. c 197 patients (59%) in the atezolizumab + bevacizumab arm vs 109 (66%) in the sorafenib arm had an event. d HR and P were from Cox model and log-rank test and were stratified by geographic region (Asia vs rest of world, including Japan); AFP level (<400 vs ≥400 ng/mL) at baseline and MVI and/or EHS (yes vs no) per IxRS. e The 2-sided P value boundary is .002. f 96 patients (29%) in the atezolizumab + bevacizumab arm vs 65 (39%) in the sorafenib arm had an event.

1. Cheng A-L et al. ESMO Asia 2019. Abstract LBA3.

Confirmed PFSb

Median PFS (95% CI), moc

Atezolizumab + bevacizumab 6.8 (5.7-8.3) Sorafenib 4.3 (4.0-5.6)

HR, 0.59 (95% CI: 0.47-0.76)d,e

P < .0001d,f

Now, here we see the progression-free survival endpoint. It met its endpoint, sorafenib, 4.3 months, and atezolizumab plus bevacizumab, 6.8 months. That’s a hazard ratio of 0.59, and it was

statistically significant. And this is the first time we have a positive study for improvement over sorafenib using an active control arm.

Phase 3 IMbrave150: OS1,a

a Data cutoff: August 29, 2019; median survival follow-up: 8.6 months. b 96 patients (29%) in the atezolizumab + bevacizumab arm vs 65 (39%) in the sorafenib arm had an event. c HR and P were from Cox model and log-rank test and were stratified by geographic region (Asia vs rest of world, including Japan); AFP level (<400 vs ≥400 ng/mL) at baseline and MVI and/or EHS (yes vs no) per IxRS. d The 2-sided P value boundary based on 161 events is .0033. 1. Cheng A-L et al. ESMO Asia 2019. Abstract LBA3.

Median OS, mo (95% CI)b

Atezolizumab + bevacizumab NE Sorafenib 13.2 (10.4, NE)

HR, 0.58 (95% CI: 0.42-0.79)c

P = .0006c,d

OS

As nice as the PFS data are, what’s really the most important, I think, is the overall survival data. Here you see sorafenib performed at 13.2 months, very typical data set in the modern era for sorafenib. However, at the time of follow-up, still, the median survival for atezolizumab plus bevacizumab has not been reached. You can see these curves separate early and remain separated over time.

Phase 3 IMbrave 150: Response Rate and Duration of Response1

IRF RECIST 1.1 IRF HCC mRECIST Atezo + Bev

(n = 326) Sorafenib (n = 159)

Atezo + Bev (n = 325)a

Sorafenib (n = 158)

Confirmed ORR, n (%) (95% CI)

89 (27) (23, 33)

19 (12) (7, 18)

108 (33) (28, 39)

21 (13) (8, 20)

CR 18 (6) 0 33 (10) 3 (2)

PR 71 (22) 19 (12) 75 (23) 18 (11)

Stratified P valueb < 0.0001 < 0.0001 SD, n (%) 151 (46) 69 (43) 127 (39) 66 (42) PD, n (%) 64 (20) 39 (25) 66 (20) 40 (25)

DCR, n (%) 240 (74) 88 (55) 235 (72) 87 (55) Ongoing response, n (%)c 77 (87) 13 (68) 84 (78) 13 (62) Median DOR, months (95% CI)

NE

6.3 (4.7, NE)

NE

6.3 (4.9, NE)

Event-free rate at 6 months, n (%) 88 59 82 63

a IRF HCC mRECIST–evaluable population was based on patients who presented with measurable disease at baseline per HCC mRECIST criteria. b Stratification factors included geographic region (Asia vs rest of world, including Japan), AFP level (<400 vs ≥400 ng/mL) at baseline and MVI and/or EHS (yes vs no) per IxRS. c Denominator is patients with confirmed CR/PR. Data cutoff, August 29, 2019; median survival follow-up, 8.6 mo. 1. Cheng A-L et al. ESMO Asia 2019. Abstract LBA3.

The median still was not reached here at 13 months, and you can see many patients are censored and remain on study. That translated into a hazard ratio of 0.58; so that’s a 42% reduction in the risk of death with this combination versus sorafenib. And I’ll just throw out that the response rate was 27%, very similar to the phase 1B/phase 2 study with the combination.



Phase 3 IMbrave150: Safety1

a Safety-evaluable population. 1. Cheng A-L et al. ESMO Asia 2019. Abstract LBA3.

Safetya

≥10% frequency of AEs in either arm and >5% difference between arms

And here you see the safety data. And in light orange is all grades, and then the darker color would be grade 3/4, a similar pattern with atezolizumab/bevacizumab. Light blue would be all grades, and then grade 3/4 in darker blue. And clearly, there’s more hypertension, maybe, with atezolizumab/bevacizumab. But other toxicities are fairly mild and very similar to sorafenib, and if not better.

And if we look at delays in quality of life, this certainly favors the combination. If we look at time to detrimental change in quality of life, 3.6 months with sorafenib, over 11 months with atezolizumab/bevacizumab.

0

20

40

60

80

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Phase 3 IMbrave150: Patient-Reported Outcomes1,2,a

Data cutoff, August, 2019; median survival follow-up, 8.6 months. a Pre-specified secondary endpoint that was not formally tested. TTD was defined as the time from randomization to first decreased from baseline of ≥10 points in the global health status/QOL scale of EORTC QLQ-C30 maintained for two consecutive assessments or one assessment followed by death from any cause within 3 weeks. b In the PRO-evaluable population, defined as patients who had a baseline and ≥1 other PRO assessment (atezo + bev, n = 309; sorafenib, n = 145). c HR from Cox model and stratified by geographic region (Asia vs rest of world, including Japan), AFP level (<400 vs ≥400 ng/mL) at baseline and MVI and/or EHS (yes vs no) per txRS. 1. Galle PR et al. ASCO GI 2020. Abstract 476. 2. Cheng A-L et al. ESMO Asia 2019. Abstract LBA3.

• Atezolizumab + bevacizumab also delayed TTD in patient-reported appetite loss, fatigue, pain, and diarrhea vs sorafenib

• Lower proportion of patients on atezolizumab + bevacizumab experienced clinically meaningful deterioration in each of these symptoms vs sorafenib

• Atezolizumab + bevacizumab also delayed TTD in patient-reported appetite loss, fatigue, pain, and diarrhea vs sorafenib

• Lower proportion of patients on atezolizumab + bevacizumab experienced clinically meaningful deterioration in each of these symptoms vs sorafenib

Baseline QOL scoreb

Mean (SD)

Better QOL

Sorafenib 68.79 (21.20)

Max 100

Atezo + Bev 71.04 (21.07)

0 Min

QOL Median TTD (95% CI), mo

Atezolizumab + bevacizumab 11.2 (6.0, NE) Sorafenib 3.6 (3.0-7.0)

HR, 0.63 (95% CI: 0.46-0.85)c

336 239 208 181 157 134 121 99 78 58 40 32 20 14 7 5 NE 165 93 60 39 31 22 22 14 12 7 4 4 2 1 NE NE NE

No. at Risk Atezo + Bev

Sorafenib

Det

erio

ratio

n-Fr

ee R

ate

(QO

L), %

Time, mo

And here you see, by several measurements, preserving role function and physical functioning significantly improved with the combination versus sorafenib.

Phase 1b Trial: Lenvatinib Plus Pembrolizumab in Unresectable HCC1

a Acute respiratory distress syndrome (n = 1); intestinal perforation (n = 1); bacterial peritonitis (n = 1). b Two TEAEs leading to discontinuation (acute respiratory distress syndrome and acute respiratory failure) were reported in the same patient. c Patients with postevaluable tumor assessment. d Zero CR confirmed. e Seven PR confirmed. 1. Ikeda M et al. ASCO 2018. Abstract 4076.

Parameter, n (%) Lenvatinib + Pembrolizumab

Part 1 (n = 6) Part 2 (n = 24) Overall (N = 30) TEAEs 6 (100) 24 (100) 30 (100)

Treatment-related TEAEs 6 (100) 22 (91.7) 28 (93.3) TEAEs grade ≥3 5 (83.3) 13 (54.2) 18 (60.0)

Serious AEs 2 (33.3) 6 (25.0) 8 (26.7) Fatal AEsa 0 3 (12.5) 3 (10.0)

Dose modifications

LEN or PEM dose interruptions due to TEAEs 5 (83.3) 13 (54.2) 18 (60.0)

LEN dose reductions due to TEAEs 5 (83.3) 13 (54.2) 18 (60.0)

Discontinuation of LEN or PEM due to TEAE(s)b 0 5 (20.8) 5 (16.7)

Lenvatinib + Pembrolizumab Part 1 (n = 6)

Part 2 (n = 24)

Overall (N = 30)

BOR, n (%) CRd 0 1 (5.0) 1 (3.8) PRe 4 (66.7) 6 (30.0) 10 (38.5) SD 2 (33.3) 13 (65.0) 15 (57.7) PD 0 0 0

ORR (including unconfirmed responses), n (%) 4 (66.7) 7 (35.0) 11 (42.3)

95% CI 22.3-95.7 15.4-59.4 23.4-63.1 ORR (excluding unconfirmed responses), n (%) 3 (50.0) 4 (20.0) 7 (26.9)

95% CI 11.8-88.2 5.7-43.7 11.6-47.8

Summary of TEAEs: Safety Analysis Set

Summary of Tumor Response: Investigator Assessment by mRECIST–Efficacy Analysis Setc

So there are other combinations that look very exciting. Lenvatinib and pembrolizumab. Lenvatinib, the TKI we talked about in the REFLECT study, and pembrolizumab, a PD-1 antibody. Here, in this phase 1B study, initially 30 patients, we had a response rate of 42%. And in [ORR (excluding unconfirmed responses)], this is still staying strong, around 30%. And really, there were no new toxicities with this combination.

Again, seeing the common things that we see with lenvatinib and also occasional things that are associated with pembrolizumab. But the response rate certainly suggests synergy, but not synergy necessarily for the toxicity.

Phase 1b Trial: Updated Results of Lenvatinib Plus Pembrolizumab in Unresectable HCC1

a Includes unconfirmed PRs (2 patients). 1. Llovet J et al. ESMO 2019. Abstract 747P.

Response Parameter, n (%) Lenvatinib +

Pembrolizumab (n = 67) mRECIST Per Investigator

ORR (CR + PR + unconfirmed PR or CR)a 30 (44.8)

CR 4 (6.0)

PRa 26 (38.8)

Stable disease 25 (37.3)

Progressive disease 6 (9.0)

Unknown/not evaluated 6 (9.0)

And these data were updated at ESMO, and here response rates by modified RECIST, still very high, over 40%. CR rate of 6%. And even stable disease, 37%, which can lead to survival advantage.



Treatment until disease progression or intolerable

toxicity

Phase 3 LEAP-002 Trial: First-Line Lenvatinib ± Pembrolizumab in Advanced HCC1

a 12 mg (for participants with screening body weight ≥60 kg) or 8 mg (for participants with screening body weight <60 kg). 1. https://clinicaltrials.gov/ct2/show/NCT03713593. Accessed January 14, 2020.

• Primary endpoints: OS and PFS • Secondary endpoints: ORR, DOR, DCR, and safety

Lenvatinib 8 or 12 mga orally once daily +

placebo

Lenvatinib 8 or 12 mga orally once daily +

pembrolizumab 200 mg IV every 3 weeks

R

• BCLC stage C or B disease not amenable to LRT or refractory to LRT and not amenable to a curative treatment approach

• Child–Pugh A • ECOG PS 0-1

N = 750

So this has led to the phase 3 LEAP study, which is actively enrolling. This is a double-blind placebo-controlled study of lenvatinib and pembrolizumab versus lenvatinib and placebo, powered, again, for OS and progression-free survival.

Phase 3 COSMIC-312 Trial: Cabozantinib ± Atezolizumab Versus Sorafenib in Advanced HCC1


Cabozantinib 40 mg orally once daily +

atezolizumab 1,200 mg IV every 3 weeks

(n = 394)

Sorafenib 400 mg orally twice daily

(n = 192)

Cabozantinib 60 mg orally once daily

(n = 64)

• Primary endpoints: PFS and OS

Study in Adults With Advanced HCC Who Have Not Received Prior Systemic Anticancer Therapy in the Advanced Setting

R

• Histologic or cytologic diagnosis of HCC not amenable to curative treatment

• Measurable disease per RECIST 1.1 • BCLC stage B or C; Child–Pugh A • ECOG PS 0-1

N = ~640

There's ongoing interest in these combinations. A number of studies in frontline, and clearly, the first choice for any patient with advanced liver cancer in the frontline setting is a clinical trial, I would argue. And we have cabozantinib and atezolizumab versus cabozantinib alone versus sorafenib. This study is actively accruing. Katie, I know, is very involved with that. And there are several abstracts at this meeting that are building on our understanding of liver cancer in the frontline setting.

Patient Case 1 Follow-Up: 59-Year-Old Man With NASH Cirrhosis

59-year-old man with NASH cirrhosis

Child–Pugh A; AFP 178 ng/mL

No ascites or encephalopathy ECOG PS 0 Previously treated

with two rounds of TACE Follow-up MRI shows multifocal

HCC (LR-5) with three lesions, largest 6.5 cm with invasion into the right portal vein

First-Line Options for This Patient Sorafenib Lenvatinib Emerging option: atezolizumab/

bevacizumab (IMbrave150 results) Phase 3 clinical trials

Checkpoint inhibitor + TKI (ie, LEAP-002 or COSMIC-312)

PD-1/L1 inhibitor + CTLA-4 inhibitor (ie, HIMALAYA or CheckMate -9DW —will be discussed in next presentation)

So in follow-up to this patient we discussed who now has portal vein invasion, progressing after TACE who is still in good shape, we

have several options available. Sorafenib and lenvatinib, you could argue, are both equally effective here, given the OS endpoint. Bevacizumab and atezolizumab, not yet approved, but certainly would be a good option. And certainly, if the patient qualified, a clinical trial is something I would strongly recommend. Katie, I hand it over to you.



Incidence of Subsequent Lines of Therapy

1. Yau T et al. ESMO 2019. LBA38_PR. 2. Finn RS et al. J Clin Oncol. 2019:JCO1901307. [Online ahead of print]

Type, n (%) Pembrolizumab (n = 278)

Placebo (n = 135)

Any

Yes 116 (41.7) 64 (47.4)

No 162 (58.3) 71 (52.6)

Approved anticancer therapy

Yes 88 (31.7) 43 (31.9)

Anti-PD1/PD-L1 agents 19 (6.8) 14 (10.4)

Others 69 (24.8) 29 (21.5)

No 190 (68.3) 92 (68.1)

Nivolumab (n = 371)

Sorafenib (n = 372)

Any subsequent therapy, n (%) 181 (49) 196 (53)

Systemic therapy, n (%) 140 (38) 170 (46)

TKI 132 (36) 86 (23)

Chemotherapy 15 (4) 25 (7)

Investigational agent 10 (3) 40 (11)

IO 7 (2) 76 (20)

Other 2 (1) 4 (1)

Local therapy, n (%) 63 (17) 61 (16)

Radiotherapy, n (%) 52 (14) 38 (10)

Surgery, n (%) 10 (3) 14 (4)

CheckMate -459: Subsequent Therapy1 KEYNOTE-240: Post-Study Anticancer Therapy2

Around half of patients received downstream therapies after first- or second-line therapy in CheckMate -459 and KEYNOTE-240 • Opportunity for individualized treatments in later lines

Dr. Kelley: And so I’ll give the second part of this talk today with the title, “The Game Changer: Navigating the New Wave of Second-Line Treatment Options in Advanced HCC.” And, really, the reason for the title “The Game Changer” is that the idea of second-line or later-line therapies in advanced HCC is really a new one.

For the longest time, systemic therapy for HCC has really been a one-and-done kind of scenario, where, after sorafenib, patients would generally progress to end-stage cancer and end-stage liver disease and be done with their intent at palliative therapy, anticancer therapy, at least.

And what we see now with both better therapies for patients with liver disease, thanks to our colleagues in hepatology, but also more active therapies in the first line and now, for the first time, options for downstream therapy, we're seeing that, on this slide, I'm showing here that there are actually a substantial number of patients going on to receive downstream therapies.

And here are a couple of examples. This is from the CheckMate -459 trial of first-line sorafenib versus nivolumab, where about half of patients went on to receive second-line therapy after progression on this trial from both treatment arm, as well as the control arm.

CaseBook 2: The Game Changer—Navigating the New Wave of Second-Line Treatment Options in Advanced HCC

And likewise, in a second-line population, KEYNOTE-240 of pembrolizumab versus placebo, which I’ll show you in a moment, again, almost 50% of patients went on to receive third-line therapy.

So again, this is really quite startling and striking and encouraging to realize that about half of patients go on to receive downstream therapies and have an option to have improvement in their cancer outcomes even after first line.

Patient Case 2: 54-Year-Old Woman With Chronic HBV

Asian woman with chronic HBV on entecavir with undetectable viral load

Treated with TACE to a 6-cm right lobe liver tumor 6 months ago

Now MRI shows enlarging 8-cm right lobe tumor, HCC (LR-5), with two satellite lesions and new right branch portal vein tumor thrombus

Child–Pugh A ECOG PS 0 AFP level 142 ng/mL

Started on lenvatinib 12 mg daily – Required dose reduction to 8 mg/d

for fatigue and HTN – Demonstrated initial tumor regression

on restaging imaging for 6 months – Developed lung metastases

after 6 months, AFP rise to 233 ng/mL

What are the options?

So that brings me to our next case, patient number 2. This is a patient from my clinic, a 54-year-old with chronic hepatitis B. She was Asian and had good viral control on entecavir, received TACE for an initial 6-cm right lobe liver lesion, but unfortunately soon after, showed enlargement in that initially treated lesion, as well as two satellite lesions and a new branch portal vein tumor thrombus, making her BCLC C, which Rich so nicely explained. She had preserved liver function, Child–Pugh A, and good performance status, and an elevated AFP of 142.

So at that time, following upon the recent REFLECT trial data, she was started on lenvatinib 12 mg daily. She did require dose reduction to 8 mg for fatigue and hypertension, but had a really nice response with regression and prolonged benefit for about 6 months.

Thereafter, she did have new lung metastasis and a rise in her AFP, prompting the question, “What is the appropriate next line of therapy?” or, in this case, “What’s the best second-line therapy?” as, again, the intro to our next part of this talk.

Robin K. (“Katie”) Kelley, MDHelen Diller Family Comprehensive Cancer Center University of California, San Francisco San Francisco, California



Patient Case 2: Multiple Options Beyond First-Line Therapy1

1. Adapted from Llovet JM et al. Nat Rev Clin Oncol. 2018;15:599-616.

First Line

SecondLine

PD

Nivolumab Pembrolizumab

PD-1 inhibitors also currently approved as second-line options

Ramucirumab

OS HR = 0.71 (vs placebo)

Child–Pugh A ECOG PS ≤1

AFP ≥400 ng/mL

Cabozantinib



Regorafenib



Tolerant to sorafenib (~85%)

Lenvatinib

OS HR = 0.92 (vs sorafenib)

Child–Pugh A ECOG PS ≤1 No invasion of

main portal vein

Sorafenib



Higher benefit in HCV infection and lack of EHS

Advanced stage (BCLC stage C: portal invasion and/or EHS) Intermediate stage (BCLC stage B: multinodular) progressing on LRTs

Atezolizumab + bevacizumab (not FDA approved)

OS HR = 0.58 (vs sorafenib)

Child–Pugh A; ECOG PS 0-1 Measurable untreated lesion

No prior systemic therapy for HCC Esophagogastroduodenoscopy (EGD) within 6 mo

We do now have multiple options for first-line therapy, as well as second-line and downstream therapies, in advanced HCC. And so I’ll cover these sequentially, which is really in their order of evolution clinically, with regorafenib, cabozantinib, and ramucirumab being antiangiogenic therapies, and then two immunotherapies now with FDA approval in the United States, nivolumab and pembrolizumab, based on single-arm phase 2 data.

Phase 3 RESORCE Trial: Regorafenib Versus Placebo in the Second-Line Setting1

• Primary endpoint: OS • Secondary endpoints: PFS, TTP, response rate, and DCR • 152 centers in 21 countries in North and South America, Europe, Australia, and Asia • All patients received BSC • Patients were treated until disease progression, death, or unacceptable toxicity

• HCC with documented radiologic progression during sorafenib treatment

• Tolerance of sorafenib ≥400 mg/d for ≥20 of past 28 days

• N = 573

Placebo (n = 194)

Regorafenib 160 mg orally, 3 weeks on, 1 week off (4-week cycle)

(n = 379) R

1. Bruix J et al. Lancet. 2017;389:56-66.

Stratification • Geographic region

(Asia vs rest of the world)

• MVI • EHD • ECOG PS (0 vs 1) • AFP (< or ≥400 ng/mL)

2:1

So, starting with regorafenib. So regorafenib is another multikinase inhibitor, quite similar to sorafenib structurally, with the addition of a fluorine moiety, which actually confers greater potency in its target inhibition.

And so in the phase 3 RESORCE trial, the patients with HCC with progression after sorafenib—again, the only treatment with efficacy at that time—patients who had tolerated sorafenib at a dose of about 400 mg/day or higher—so half of standard dose or better—for at least 20 of the past 28 days were eligible for enrollment if they had progression on sorafenib and tolerance on sorafenib. Patients were randomized 2:1 to either regorafenib at a dose of 160 mg/day, 3 weeks on, 1 week off versus placebo, with a primary endpoint of overall survival.

Regorafenib Improves Overall Survival in the Second-Line Setting1

a Based on mRECIST. 1. Bruix J et al. Lancet. 2017;389:56-66.

0

20

40

60

80

100

0 3 6 9 12 15 18 21 24 27 30 33

Prob

abili

ty o

f OS,

a %

Time Since Randomization, mo No. at Risk Regorafenib 379 316 224 170 122 78 54 34 21 10 4 0 Placebo 194 149 95 62 37 26 16 8 5 3 1 0

Placebo

Regorafenib

Regorafenib (n = 379)

Placebo (n = 194)

Events, n (%) 233 (6) 140 (72)

Censored, n (%) 147 (39) 54 (28)

mOS, mo (95% CI) 10.6 (9.1-12.1) 7.8 (6.3-8.8)

HR (95% CI) 0.63 (0.50-0.79); P < .0001 (2-sided)

And this shows very strong survival improvement for regorafenib compared with placebo, with a median overall survival of 10.6 months for regorafenib versus 7.8 months for placebo, a hazard ratio of 0.63, leading to approval of regorafenib as a second-line therapy in the United States and in multiple other parts of the world around 2017. And this was a huge landmark in the field, as the first second-line therapy for advanced HCC.

Regorafenib Survival Benefit Consistent Across Subgroup Analyses1


And we look further into the data and see that this benefit was quite consistent across clinically relevant prognostic subgroups, including region, which is a proxy for hepatitis B or C status in most studies, regardless of AFP level. So low AFP and high AFP both benefitted similarly; this is important to remember as we look to ramucirumab ahead. Also, according to macrovessel invasion or not. And here, again, are the viral results.



Treatment-Emergent AE Profile1


Regorafenib (n = 374), n (%) Placebo (n = 193), n (%)

Any Grade Grade 3 Grade 4 Any Grade Grade 3 Grade 4

Any AE 346 (93) 173 (46) 14 (4) 100 (52) 31 (16) 1 (1)

HFSR 196 (52) 47 (13) N/A 13 (7) 1 (1) N/A

Diarrhea 125 (33) 9 (2) 0 18 (9) 0 0

Fatigue 110 (29) 24 (6) N/A 37 (19) 3 (2) N/A

Hypertension 87 (23) 48 (13) 1 (<1) 9 (5) 6 (3) 0

Anorexia 88 (24) 10 (3) 0 12 (6) 0 0

Increased blood bilirubin 70 (19) 24 (6) 1 (<1) 7 (4) 4 (2) 0

Abdominal pain 34 (9) 5 (1) N/A 5 (3) 0 N/A

Increased AST 48 (13) 16 (4) 3 (1) 15 (8) 9 (5) 1 (1)

So, as I mentioned, regorafenib has similar structural features as sorafenib, and as one might expect, it also has similar toxicity profile. The predominant side effects are hand-foot skin reaction, diarrhea, fatigue, hypertension, anorexia, and we see the grade 3 rates in particular for hand-foot skin reaction are about 13%, likewise for hypertension. So these are class effects that need management, just as for sorafenib.

Phase 3 CELESTIAL Trial: Cabozantinib Versus Placebo After Prior Sorafenib1

• Primary endpoint: OS • Secondary endpoints: PFS and ORR

• Advanced HCC; Child–Pugh A

• ECOG PS 0-1 • Prior sorafenib • Progression on

≥1 prior therapy • ≤2 prior therapies

N = 760

Placebo

Cabozantinib 60 mg/d

R

1. Abou-Alfa G et al. N Engl J Med. 2018;379:54-63.

Stratification

• Disease etiology (HBV, HCV, other)

• Region (Asia, other)

• Presence of MVI and/or EHS of disease (yes or no)

2:1

All patients had previously received sorafenib, and 27% had received two previous systemic anticancer regimens for advanced HCC

So, turning to the next second-line agent to be approved in this context, cabozantinib. So, cabozantinib is another multikinase inhibitor, but it does have a distinct target profile from both regorafenib and sorafenib, including the same antiangiogenic kinases, like VEGFR2 and -3, but also one key difference is that cabozantinib inhibits MET or hepatocyte growth factor receptor, which is thought to be implicated in antiangiogenic resistance.

And so cabozantinib was studied versus placebo in the phase 3 CELESTIAL trial, also in Child–Pugh A, advanced HCC patients who had had prior sorafenib. It did not require any particular tolerability to sorafenib. So patients who had failed sorafenib due to intolerable side effects were eligible, unlike RESORCE.

And another difference from the RESORCE regorafenib trial was that the CELESTIAL trial allowed either one or two prior therapies, so it included a mixture of second- and third-line patients. Patients were, again, randomized 2:1 to cabozantinib versus placebo, and the primary endpoint, again, in this registration trial was survival.

Cabozantinib Improves Overall Survival in the Second- or Third-Line Setting1

a Critical P ≤ .021 for second interim analysis. 1. Abou-Alfa G et al. N Engl J Med. 2018;379:54-63.

a

And so here we see that cabozantinib had a significant overall survival improvement over placebo, 10.2 months versus 8.0 months, with a hazard ratio of 0.76. We did present a subanalysis of these data for the strict second-line population, which showed an even more pronounced benefit in the second-line setting.

Cabozantinib Improves Progression-Free Survival in the Second- or Third-Line Setting1


• ORR: cabozantinib 4% vs placebo 0.4% (P = .0086) • PR or SD: cabozantinib 64% vs placebo 33%

Similarly, progression-free survival is also improved, 5.2 months versus 1.9 months for placebo, with a hazard ratio of 0.44.

No. of patients OS PFS

HR HR

Overall 707 0.76 0.44

Region Asia Other regions

175 532

1.01 0.71

0.46 0.45

Race Asian Non-Asian

241 423

0.86 0.75

0.43 0.47

EHS and/or MVI Yes No

598 109

0.73 0.99

0.45 0.46

Etiology HBV HCV Other

267 156 284

0.69 1.11 0.72

0.31 0.61 0.48

Prior lines of therapy 1 prior regimen 2 prior regimens

509 192

0.74 0.90

0.43 0.58

Cabozantinib Survival Benefit Across Subgroup Analyses1


0.5 Favors cabozantinib

0.25 1 2 Favors placebo

0.5 Favors cabozantinib

0.25 1 2 Favors placebo

And we see relatively symmetric survival and, as it turns out, also PFS benefit across, again, the clinically relevant subgroups and prognostic subgroups, including both patients with one or two prior lines of therapy showing PFS improvement.



All-Causality Grade 3 or 4 AEs1,a

a Grade 3/4 events reported in ≥5% of patients in either treatment group. 1. Abou-Alfa G et al. N Engl J Med. 2018;379:54-63.

AE, % Cabozantinib (n = 467)

Placebo (n = 237)

Any grade 3 or 4 AE 68 36

HFSR 17 0

Hypertension 16 2

Increased AST 12 7

Fatigue 10 4

Diarrhea 10 2

Asthenia 7 2

Decreased appetite 6 <1

Anemia 4 5

Grade 5 TRAEs • Cabozantinib (n = 6):

hepatic failure, esophagobronchial fistula, portal vein thrombosis, upper GI hemorrhage, pulmonary embolism, and hepatorenal syndrome

• Placebo (n = 1): hepatic failure

So, turning to tolerability and toxicity, again, as an antiangiogenic tyrosine kinase inhibitor, as we expect to see, hand-foot skin reaction is also a significant side effect of cabozantinib, with grade 3 or 4, 17%; diarrhea, 10%, grade 3 or 4; and hypertension, again, being the noteworthy class effects, as expected in this context.

197 172 121 87 56 37 26 14 4 0 0 0 0 0 95 76 50 36 19 12 4 1 0 0 0 0 0 0

119 93 65 46 35 22 12 8 4 2 1 1 1 0 131 90 44 28 18 14 8 4 3 1 0 0 0 0

Ramucirumab Improves OS in the Second-Line Setting in Patients With AFP ≥400 ng/mL1,2

a REACH trial did not show benefit of ramucirumab in patients regardless of AFP level. 1. Zhu AX et al. Lancet Oncol. 2015;16:859-870. 2. Zhu AX et al. Lancet Oncol. 2019;20:282-296.

0

20

40

60

80

100

0 3 6 9 12 15 18 21 24 27 30 33 36 39

OS,

%

Time, mo

Censored observations Ramucirumab Placebo

0

20

40

60

80

100

0 3 6 9 12 15 18 21 24 27 30 33 36 39

OS,

%

Time, mo No. at Risk Ramucirumab Placebo

Censored observations Ramucirumab Placebo

AFP ≥ 400 ng/mL REACHa: RCT Subgroup Analysis REACH-2: RCT

Ramucirumab (n = 197)

Placebo (n = 95) Difference P

mOS, mo 8.5 7.3 1.2 – HR (95% CI) 0.710 (0.531-0.949) .0199



mOS, mo 7.8 4.2 3.6 – HR (95% CI) 0.674 (0.508-0.895) .0059

No. at Risk Ramucirumab Placebo

So, turning to the third agent in our second-line list of therapies now available, ramucirumab. Ramucirumab is not a tyrosine kinase inhibitor, but rather an antiangiogenic VEGFR2 monoclonal antibody given intravenously. And I think it’s really important to note that ramucirumab’s story is another landmark in the field of HCC, because it has now become the first agent with a biomarker.

And this evolved in that ramucirumab was first studied in a trial called REACH, a phase 3 study looking at all patients across AFP values, serum AFP, and showed a trend toward benefit, but was not statistically significant for its primary endpoints.

Subanalyses according to alpha-fetoprotein cutpoints—looking at various cutpoints, 100, 200, 400 ng/mL—showed that patients with increasing levels of AFP had a trend towards improved survival and progression-free survival, as well as other efficacy endpoints. This does potentially stand to reason in that AFP is a marker for proangiogenic—or an increased level of angiogenic activity—genomic signaling in patients with high AFP in terms of their genetic signatures of the tumors.

And so looking at the subanalysis of the REACH phase 3 trial, it was really intriguing to see this benefit in the high-AFP group

compared with those same high-AFP patients treated with placebo.

And that prompted the launching of a second phase 3 trial, REACH-2, just in patients with high AFP, greater than 400 ng/mL. And sure enough, in the second trial, selected for the biomarker of AFP, ramucirumab did achieve a significant survival benefit—admittedly modest, but there is a notable tail on this curve, and the hazard ratio was 0.71.

Ramucirumab Improves OS in the Second-Line Setting in Patients With AFP ≥400 ng/mL1,2

1. Galle P et al. 12th Annual Conference of the International Liver Cancer Association 2018 (ILCA 2018). Abstract O-001. 2. Zhu A et al. Ann Oncol. 2018;29 (suppl 5):Abstract LBA-001.



Death, n (%) 246 (77.8) 190 (84.1) — — mOS, mo 8.1 5.0 3.1 —

HR (95% CI): 0.694 (0.571-0.842) .0002

To increase the power of this analysis, here we see a combined analysis of the patients with high AFP from the initial REACH-1 study, as well as all patients with REACH-2, and shows a more pronounced benefit when we have a greater sample size with overall survival of 8.1 months for patients with high AFP treated with ramucirumab versus 5.0 months for those treated with placebo.

So, collectively, the data from REACH-1 and REACH-2 have led to approval of ramucirumab as another second-line option, with the first biomarker selected requirement for eligibility for this therapy of AFP greater than 400 ng/mL.

AE,a n (%) Ramucirumab (n = 197) Placebo (n = 95)

Any Grade Grade ≥3 Any Grade Grade ≥3 Hypertension 49 (24.9) 25 (12.7) 12 (12.6) 5 (5.3) Bleeding/hemorrhage events 48 (24.4) 10 (5.1) 12 (12.6) 3 (3.2)

GI hemorrhage events 12 (6.1) 7 (3.6) 5 (5.3) 2 (2.1) Epistaxis 27 (13.7) 1 (0.5) 3 (3.2) 0

Proteinuria 40 (20.3) 4 (2.0) 4 (4.2) 0 Arterial thromboembolic events 5 (2.5) 3 (1.5) 1 (1.1) 1 (1.1) Venous thromboembolic events 2 (1.0) 0 2 (2.1) 1 (1.1) GI perforation 2 (1.0) 2 (1.0) 2 (2.1) 2 (2.1) Fistula 1 (0.5) 0 0 0

Ascites 35 (17.8) 8 (4.1) 7 (7.4) 2 (2.1) Hepatic encephalopathy 8 (4.1) 6 (3.0) 0 0

Infusion-related reactionsb 17 (8.6) 0 3 (3.2) 0

Adverse Events of Special Interest1

a AEs of special interest according to MedRA preferred terms or consolidated categories. b Infusion-related reactions occuring within 24 hours of infusion. 1. Zhu AX et al. Lancet Oncol. 2019;20:282-296.

Another example of antiangiogenic class effects of adverse events here. We see rates of hypertension in the ramucirumab group of about 12.7% for grade 3 or higher compared with placebo. And also, as expected for antiangiogenic therapies, a slight increased rate of bleeding, although, again, not so different than placebo,



noting that patients with advanced HCC are susceptible to bleeding events such as variceal bleeds.

And so the safety profile was actually—I would take this as—reassuring in terms of the bleeding risk of this agent. But hypertension is particularly important to manage as it is for the antiangiogenic tyrosine kinase inhibitors.

Hepatic encephalopathy is also noteworthy, something that is, again, rare, but an important side effect to manage and highlighting, again, the importance of us taking care of patients with liver disease and being aware of management of their chronic underlying liver disease in parallel with the cancer.

Patient Case 2 Follow-Up: 54-Year-Old Woman With Chronic HBV

Asian woman with chronic HBV on entecavir with undetectable viral load

Treated with TACE to a 6-cm right lobe liver tumor 6 months ago

Now MRI shows enlarging 8-cm right lobe tumor, HCC (LR-5), with two satellite lesions and new right branch portal vein tumor thrombus

Child–Pugh A ECOG PS 0 AFP level 142 ng/mL Discontinued lenvatinib

Selecting a Second-Line Targeted Therapy Option for This Patient Consider AE profile, AFP levels,

liver function, comorbidities, and prior treatment history – Regorafenib

Did not receive prior sorafenib – Ramucirumab

AFP levels <400 ng/mL – Cabozantinib

Allowed ≤2 lines of prior therapy

Favorable data in patients with HBV

So, turning back to our case again, the patient who had been treated with lenvatinib for advanced HCC as first line, how do we treat her now that she’s progressing with new lung metastases and right portal vein tumor thrombus?

So the second-line choices for this patient—you know, how do I approach a decision like this? Consider the adverse event profile of the drug, AFP levels, liver function, comorbidities, and prior treatment history. And so in this case, regorafenib was not an appropriate choice, in that she was treated with first-line lenvatinib and had not received first-line sorafenib. So I did not choose regorafenib for her. Her AFP level, although elevated, was less than 400 ng/mL, so not an appropriate candidate for ramucirumab at this point.

And in this case, I treated her with cabozantinib, because the trial allowed up to two prior lines of therapy, and one of which was not necessarily sorafenib, and did show a trend towards benefit in hepatitis B patients.

Patient Case 2 Follow-Up: 54-Year-Old Woman With Chronic HBV (Cont’d)

• Patient started on cabozantinib 60 mg daily owing to progression on first-line lenvatinib – CELESTIAL trial included second- and third-line patients after

various first- and second-line therapies • AFP currently <400 ng/mL but could consider ramucirumab if it rises

to >400 ng/mL in future

So she started on cabozantinib 60 mg daily, again, based on the fact that CELESTIAL had a permissive eligibility, allowing second- and third-line patients after various prior lines of therapies. And we can consider ramucirumab should her AFP rise higher in the future.

Patient Case 3: 69-Year-Old Man With HCV

69-year-old man with chronic HCV was treated and cured with direct-acting antiviral therapy 2 years ago

He then presents to his PMD with right upper quadrant pain He was found to have an 8-cm mass and satellite lesions in the right lobe; he was

hypervascular with delayed washout, and he has right portal vein thrombosis Hemoglobin 12.3; MCV 90; platelets 95; bilirubin 1.8; Alb 3.1;

creatinine 0.9; AFP 78 ng/mL History: HCV, T2DM, hypertension No ascites; no encephalopathy; PS 1

So that brings me to the next patient case, who is a 69-year-old man with hepatitis C who was previously treated and cured with direct-acting antiviral therapy a few years ago, now presenting with right upper-quadrant pain, found to have also another case with a large right lobe liver mass, this time with synchronous satellite lesions and a right portal vein tumor thrombus that was enhancing and compatible with malignancy.

He had signs of portal hypertension, with low platelets, at 95, an elevated bilirubin, and a mildly elevated alpha-fetoprotein, has history of hepatitis C, as I mentioned, with portal hypertension, some diabetes and hypertension. Performance status of 1.



Patient Case 3: 69-Year-Old Man With HCV (Cont’d)

Staging: no extrahepatic disease; BCLC C Started on sorafenib 400 mg twice daily; dose reduced to 400 mg daily

after 3 weeks because of fatigue, weight loss, and anorexia Continued for 2 months but developed worsening pain and rising AFP,

prompting restaging imaging Tumors found on the left lobe, and the right lobe masses

have grown Extension of PVT

What do we do next?

He is BCLC C because of the vascular invasion, although there was no extrahepatic disease. He started on first-line sorafenib but required dose reduction because of toxicity.

He continued for 2 months but had worsening pain and rising AFP throughout that first 2 months and had early interval restating imaging and showed prompt progression on his first scans, including extension to the contralateral lobe and enlargement of the right lobe mass, extension of the portal vein tumor thrombus.

So by contrast to the prior lady, who had been treated with lenvatinib and had a nice response for at least 6 months, this patient had an immediate progression, prompting the question of “What do we do next?”

Phase 1/2 CheckMate -040 Trial: Nivolumab in HCC1,2

1. https://clinicaltrials.gov/ct2/show/NCT01658878. Accessed January 14, 2020. 2. El-Khoueiry AB et al. Lancet. 2017;389:2492-2502.

• Primary endpoints: Safety, tolerability, and ORR

• Other endpoints: CR, DCR, DOR, TTR, TTP, TTP rate, PFS, OS, OS rate, biomarkers, and PK

Nivolumab • Noninfected/HBV/HCV • Dose escalation

(n = 48) • Dose expansion

(n = 214)

Nivolumab Child–Pugh B

Nivolumab + ipilimumab

R

Nivolumab Sorafenib

Nivolumab ± ipilimumab + cabozantinib

R

• HCC not amenable to curative resection • Child–Pugh ≤6, ≤7 for dose escalation, Child–Pugh B for cohort 5 • Progressed on ≥1 prior line of systemic therapy, intolerant to

sorafenib, or refused sorafenib (dependent on cohort) N = 620

So that brings me to a discussion of the role of immunotherapy in the second and later lines of HCC treatment now. The first large trial to really examine the efficacy of this new realm of immunotherapy in oncology in the HCC context was the phase 1/2 CheckMate -040 trial of nivolumab, which, as everyone recalls, nivolumab is an anti–PD-1 monoclonal antibody, which has shown efficacy across a variety of tumor types.

And the phase 1/2 CheckMate -040 trial started with a monotherapy nivolumab arm, which I’ll show you in a moment, but has now expanded to a variety of different cohorts, and I’ll cover a couple of these as we go.

Nivolumab Dose Expansion: Treatment-Emergent Adverse Events1

1. El-Khoueiry AB et al. Lancet. 2017;389:2492-2502.

Uninfected (n = 113)

HCV (n = 57)

HBV (n = 50)

Total (N = 214)

Any Grade

Grades 3-4

Any Grade

Grades 3-4

Any Grade

Grades 3-4

Any Grade

Grades 3-4

Patients with any TRAE, n (%) 9 (8) 4 (3) 5 (10) 4 (8) 2 (4) 1 (2) 16 (7) 9 (4) Symptomatic TRAEs reported in >4% of all patients

Rash 16 (14) 2 (1) 9 (18) 0 8 (16) 0 33 (15) 2 (1) Pruritus 18 (15) 0 14 (28) 1 (2) 13 (25) 0 45 (21) 1 (<1) Diarrhea 19 (16) 2 (1) 5 (10) 0 3 (6) 1 (2) 27 (13) 3 (10) Decreased appetite 6 (5) 0 2 (4) 1 (2) 3 (6) 0 11 (5) 1 (<1) Fatigue 34 (30) 2 (1) 8 (16) 1 (2) 7 (14) 0 49 (23) 3 (1) Nausea 10 (8) 0 6 (12) 0 1 (2) 0 17 (8) 0 Dry mouth 9 (8) 0 2 (4) 0 2 (4) 0 13 (6) 0

Laboratory-value TRAEs reported in >4% of all patients

↑ AST 9 (8) 4 (3) 6 (12) 5 (10) 1 (2) 0 16 (7) 9 (4) ↑ ALT 7 (6) 2 (1) 7 (14) 3 (6) 3 (6) 0 17 (8) 5 (2)

So in the first part of the study, dose expansion, I think it’s really important to highlight in a hepatitis C patient population—again, with a hepatologist sitting next to me—that we really have to think about safety in a different way and make sure that our drugs are safe in the context of liver dysfunction, particularly in the context of the risk of immune reactivation or immune hepatitis in a liver disease population.

And so the CheckMate -040 trial was, I think, very prescient and wise to design their dose-expansion phase 1 in three distinct cohorts defined by viral status: hepatitis C, hepatitis B, and uninfected or nonviral, and to first take a very close look to make sure that immunotherapy with PD-1 inhibition was safe in this compromised population.

And I think the take-home points from this busy toxicity table, which you don’t need to read all the details, are that the toxicity was quite similar to other tumor types, and most importantly, in this laboratory value section, highlighting transaminitis, the rates of grade 3 or higher transaminitis are quite low, less than 5%. So, it is very reassuring that we’re not seeing a higher rate of immune hepatitis in this population.

Outcomes With Nivolumab in HCC1

1. El-Khoueiry AB et al. Lancet. 2017;389:2492-2502.

Uninfected, SOR naïve: ORR 23%

Uninfected, prior SOR: ORR 21%

HBV: ORR 14%

HCV: ORR 20%

Median duration of response >9.9 months, occasional CRs; led to FDA approval in 2017 for advanced HCC after failure of sorafenib

From the phase 1b expansion of over 200 patients, the spider plots of response are shown here. And I think these are quite striking for a couple reasons. One, we see that there’s a dramatic proportion of patients with downward-trending spider plots, and these are measured in weeks—which, here is, for example, the 1-year mark of 50 weeks. So these are lines.



The second thing that’s dramatic, besides the downward trend in all these lines, is also the duration of these lines, and that some of these lines go down to the 0, or 100% shrinkage rate. So what we’re seeing is a substantial number of patients with regression that lasts a long time, including a few with complete or 100% responses.

And that was true in the uninfected patients who had not had prior sorafenib, about 80 patients, and the objective response rate was 23% in that cohort, quite durable. In the uninfected patients, meaning nonviral with prior sorafenib, the objective response rate was 21%; for hepatitis C, 20%; and for hepatitis B, 14%. And the median duration of response was about 10 months or longer, depending on which data cut and cohort we’re looking at. And the strength of this data coupled with its safety led to FDA approval in 2017 for advanced HCC after failure of sorafenib.

Phase 3 CheckMate -459 Trial: Nivolumab Versus Sorafenib in Advanced HCC1

• Primary endpoint: OS • Secondary endpoints: ORR, PFS, and biomarkers

• Advanced HCC – Not eligible for surgical and/or

locoregional therapies – Progression after surgical

and/or locoregional therapies • Child–Pugh A

N = 743 Sorafenib

Nivolumab

R


These data also prompted the recently reported CheckMate -459 trial, nivolumab versus sorafenib in advanced HCC. And I should clarify that this was a first-line study, so I’m straying into Rich’s territory here, because it followed upon second-line data. And this was a randomized 1:1 trial, open label, of nivolumab versus standard-of-care sorafenib in Child–Pugh A first-line advanced HCC patients with a survival endpoint.

CheckMate -459: Survival1

a Based on Kaplan-Meier estimates. b Stratified Cox proportional hazards model; HR is nivolumab over sorafenib. c P value from log-rank test; final OS boundary: .0419 for a 2-sided nominal P value. 1. Yau T et al. ESMO 2019. Abstract LBA38_PR.

0

20

40

60

80

100

0 3 6 9 12 15 18 21 24 27 30 33 36 39

PFS,

%

Time, mo

12-mo rate 22% 14%

24-mo rate 14% 6%

0

20

40

60

80

100

0 3 6 9 12 15 18 21 24 27 30 33 36 39

OS,

%

Time, mo

Nivolumab (n = 371)

Sorafenib (n = 372)

HR (95% CI)b

Median PFS (95% CI),a mo

3.7 (3.1-3.9)

3.8 (3.7-4.5)

0.93 (0.79-1.10)

Nivolumab (n = 371)

Sorafenib (n = 372)

HR (95% CI)b P

Median OS (95% CI),a mo

16.4 (13.9-18.4)

14.7 (11.9-17.2)

0.85 (0.72-1.02) .0752c

OS PFS

371 326 271 235 211 187 165 146 129 104 63 39 17 0 372 328 274 232 196 174 155 133 115 80 47 30 7 0

No. at Risk Nivolumab Sorafenib

371 193 110 84 68 56 46 36 32 19 11 7 1 0 372 207 87 55 32 26 21 14 10 6 5 0 0 0

No. at Risk Nivolumab Sorafenib

12-mo rate 60% 55%

24-mo rate 37% 33%

Nivolumab Sorafenib

Nivolumab Sorafenib

And these were the data presented at ESMO, showing actually that the study was negative—did not achieve a statistically significant survival benefit in the overall population. There as a trend towards improvement, 16.4 months for nivolumab versus 14.7 months

for sorafenib, noting, again, that both of these cohorts did quite well compared with our historical data, which has always been generally less than a year, until quite recently, speaking again to the value of multiple lines of therapy. Progression-free survival was also similar between the two arms.

CheckMate -459: Response1

a Per blinded independent central review using RECIST v1.1; defined as CR + PR. b Defined as CR + PR + SD + non-CR/non-PD. 1. Yau T et al. ESMO 2019. Abstract LBA38_PR.

Nivolumab (n = 371) Sorafenib (n = 372) ORR,a n (%) 57 (15) 26 (7) Best overall response, n (%) CR 14 (4) 5 (1) PR 43 (12) 21 (6) SD 130 (35) 180 (48) Non-CR/non-PD 16 (4) 9 (2) PD 136 (37) 105 (28) Not evaluable 32 (9) 52 (14) DCR,b n (%) 203 (55) 215 (58) Median duration of disease control (95% CI), mo 7.5 (6.5-10.7) 5.7 (5.6-7.4) Median time to response (range), mo 3.3 (1.6-19.4) 3.7 (1.5-11.1) Median DOR (range), mo 23.3 (3.1 to 34.5+) 23.4 (1.9+ to 28.7+)

• Improvement in ORR was observed with nivolumab compared with sorafenib (odds ratio [95% CI], 2.41 [1.48-3.92]) ─ Higher CR rate was observed with nivolumab compared with sorafenib

But what we can see is that, similar to the phase 2 data, there was a subset of patients with prolonged, deep responses, about 16% overall, with 4% having complete responses; 12%, partial responses; and the median duration of response, 23 months. So, while the overall population did not benefit enough to achieve a positive study, there is a subset of patients with deep, durable, meaningful responses, again, showing out of this data set.

• FACT-Hep is a disease-specific questionnaire that measures the effects of HCC and its treatment on HRQOL

• Completion rates were ≥70% at all time points through week 113 • Clinically meaningful differences between treatment arms

were observed for Fact-Hep total in favor of nivolumab through week 113a

Effect of Nivolumab Versus Sorafenib on HrQOL1

Physical well-being

(PWB) 7 items

Emotional well-being

(EWB) 6 items

Social/family well-being

(SWB) 7 items

Functional well-being

(FWB) 7 items

1. Yau T et al. ESMO 2019. Abstract LBA38_PR.

5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101 105 109 113Overall BL -20

-15

-10

-5

0

5

10

15

20

Favors sorafenib

Favors nivolumab

LS M

eans

(95%

Cl)

Visit Week

Health-related quality of life shows that the patients receiving nivolumab had favorable changes in their health-related quality of life compared with those receiving sorafenib across a variety of domains.



• Nonrandomized, multicenter, open-label, phase 2 trial assessing the PD-1 inhibitor pembrolizumab 200 mg every 3 weeks

• Study enrolled patients (N = 104) with HCC either intolerant to or showing radiographic progression after treatment with sorafeniba

• Primary endpoint: objective response

Phase 2 KEYNOTE-224 Trial: Pembrolizumab in Previously Treated HCC1

a ECOG PS 0-1; adequate organ function, Child–Pugh A. 1. Zhu AX et al. Lancet Oncol. 2018;19:940-952.

So reinforcing the data from nivolumab, we also have phase 2 data looking at pembrolizumab in patients with previously treated HCC, pembrolizumab being another anti–PD-1 monoclonal antibody. KEYNOTE-224 looked at about 100 patients with second-line HCC and also examined their objective response.

KEYNOTE-224: Response1

1. Zhu AX et al. Lancet Oncol. 2018;19:940-952.

Median DOR not reached Objective response: 17%

And we see, very similar to the nivolumab monotherapy data, that about 17% of patients had a prolonged, durable response.

AE, n (%) Grade 1-2 Grade 3 Grade 4 Grade 5 Fatigue 18 (17) 4 (4) 0 0 Pruritus 12 (12) 0 0 0 Diarrhea 11 (11) 0 0 0 Rash 10 (10) 0 0 0 Nausea 8 (8) 0 0 0 Asthenia 7 (7) 0 0 0 Increased AST 7 (7) 7 (7) 0 0 Decreased appetite 6 (6) 1 (1) 0 0 Myalgia 6 (6) 1 (1) 0 0 Hypothyroidism 6 (6) 0 0 0 Increased ALT 5 (5) 4 (4) 0 0 Arthralgia 5 (5) 0 0 0 Maculopapular rash 5 (5) 0 0 0 Hyperbilirubinemia 3 (3) 1 (1) 1 (1) 0 Dyspnea 4 (4) 1 (1) 0 0 Anemia 2 (2) 1 (1) 0 0 Adrenal insufficiency 1 (1) 2 (2) 0 0

KEYNOTE-224: Selected Adverse Events1

1. Zhu AX et al. Lancet Oncol. 2018;19:940-952.

Again, also, reinforcing and reassuring that the safety data from nivolumab was similar in this other antibody, we see an acceptable rate of increased ALT and other transaminitis of about 5% or less. So, again, reassuring that there’s no increased rate of immune-related hepatitis.

Phase 3 KEYNOTE-240 Trial: Pembrolizumab Versus BSC as Second-Line Therapy1

1. Finn RS et al. J Clin Oncol. 2020;38:193-202.

OS,

%

Pembrolizumab reduced the risk of death by 22% and improved PFS over placebo “These differences did not meet significance per the prespecified statistical plan”

N = 413

Pembrolizumab (n = 278)

R Placebo (n = 135)

Time, mo

100 90 80 70 60 50 40 30 20 10

0 0 4 8 12 16 20 24 28 32

Median (95% CI), mo 13.9 (11.6-16.0) 10.6 (8.3-13.5) Primary endpoints: OS and PFS

Prespecified P = .0174 required for statistical significance

Events HR (95% CI) P

Pembrolizumab 183 0.781 (0.611-0.998) .0238 Placebo 101

No. at Risk Pembrolizumab

Placebo 278 135

237 113

190 84

152 65

110 42

57 23

16 8

1 1

0 0

Pembrolizumab Placebo

So that second-line data prompted the KEYNOTE-240 randomized phase 3 trial of pembrolizumab versus best supportive care as a second-line agent. And so pembrolizumab versus placebo was performed, randomized 1:1.

And we see here that though these curves do show some separation and improvement in the primary endpoint of median overall survival, 13.9 months for pembrolizumab versus 10.6 months for placebo, hazard ratio of 0.78, the study did not meet the prespecified P value required for significance under the statistical plan. So, while there was an improvement in outcomes, it was not statistically significant under the study design, in part due to a dual primary endpoint and in part due to the sample size.

KEYNOTE-2401

1. Finn RS et al. J Clin Oncol. 2020;38:193-202.

• DOR, median (range): pembrolizumab, 13.8 mo (>1.5+ mo to >23.6 mo); placebo, not reached (2.8 mo to >20.4 mo)

Pembrolizumab Placebo 0

10

20

30

OR

R, %

(95%

Cl)

13.8 (7.7-19.5) P = .00007

18.3 (14.0-23.4)

4.4 (1.6-9.4)

Parameter No. (%)

Pembrolizumab (n = 278)

Placebo (n = 135)

Objective response 1 (18.3) 6 (4.4)

95% Cl 14.0 to 23.4 1.6 to 9.4

Estimated treatment difference 13.8

95% Cl 7.7 to 19.5

P .00007

Best overall responses

CR 6 (2.2) 0 (0)

PR 45 (16.2) 6 (4.4)

SD 122 (43.9) 66 (48.9)

≥23 weeks 37 (13.3) 20 (14.8)

PD 90 (32.4) 57 (42.2)

Not evaluable 7 (2.5) 3 (2.2)

Not assessable 8 (2.9) 3 (2.2)

DCR 173 (62.2) 72 (53.3)

Nonetheless, we again see that a subset of patients had durable responses, including, in this study, about 18% [objective response], and the median duration of response was 13.8 months.



Phase 1/2 CheckMate -040 Trial: Nivolumab in HCC1,2

1. https://clinicaltrials.gov/ct2/show/NCT01658878. Accessed January 14, 2020. 2. El-Khoueiry AB et al. Lancet. 2017;389:2492-2502.

• Primary endpoints: Safety, tolerability, and ORR

• Other endpoints: CR, DCR, DOR, TTR, TTP, TTP rate, PFS, OS, OS rate, biomarkers, and PK

Nivolumab • Noninfected/HBV/HCV • Dose escalation

(n = 48) • Dose expansion

(n = 214)

Nivolumab Child–Pugh B


R

Nivolumab Sorafenib

Nivolumab ± ipilimumab + cabozantinib

R

• HCC not amenable to curative resection • Child–Pugh ≤6, ≤7 for dose escalation, Child–Pugh B for cohort 5 • Progressed on ≥1 prior line of systemic therapy, intolerant to

sorafenib, or refused sorafenib (dependent on cohort) N = 620

So, returning back to the CheckMate -040 trial in the last few minutes of this section, and I’ll show you here the data we have so far on the combination of nivolumab plus ipilimumab in the second-line setting, the first IO plus IO, or immuno-oncology doublet, in this context.

CheckMate -040: Nivolumab Plus Ipilimumab—Study Design1

• Primary endpoint: Safety and tolerability using NCI CTCAE v4.0; ORR and DOR based on investigator assessment

• Secondary endpoints: DCR, PFS, OS, TTP, TTR • Other: BOR and ORR based on BICR-assessed tumor responsea

1. Yau T et al. ASCO 2019. Abstract 4012.

Arm A: NIV01 + IPI3

every 3 wk x 4

Arm B: NIV03 + IPI1

every 3 wk x 4 R

• Advanced HCC, sorafenib treated, intolerant, or progressors

• Uninfected, HCV infected, or HBV infected

• CP score A5-A6 • ECOG PS 0-1

Nivolumab 240 mg IV every 2 wk Flat dose

Arm C: NIVO3 every 2 wk +

IPI1 every 6 wk

Treatment until disease progression or intolerable

toxicity

So nivolumab plus ipilimumab has been studied in second-line patients who have been previously treated with sorafenib or intolerant or with progression on sorafenib, again, looking at patients according to viral status (nonviral, HCV, or HBV) and Child–Pugh A patients.

CheckMate -040: Nivolumab Plus Ipilimumab—Efficacy Results1

Arm A NIVO1/IPI3 Q3W

(n = 50)

Arm B NIVO3/IPI1 Q3W

(n = 49)

Arm C NIVO3 Q2W/

IPI1 Q6W (n = 49)

ORR by BICR using RECIST v1.1, n (%) 16 (32) 15 (31) 15 (31)

BOR, n (%) CR 4 (8) 3 (6) 0 PR 12 (24) 12 (24) 15 (31) SD 9 (18) 5 (10) 9 (18) PD 20 (40) 24 (49) 21 (43) Unable to determine 3 (6) 4 (8) 4 (8)

DCR, n (%) 27 (54) 21 (43) 24 (49) Median TTR (range), months 2.0 (1.1–12.8) 2.6 (1.2–5.5) 2.7 (1.2–8.7)

Median DOR (range), months

17.5 (4.6 to 30.5+)

22.2 (4.2 to 29.9+)

16.6 (4.1+ to 32.0+)

OS,

%

Time, mo

0102030405060708090

100

0 3 6 9 12 15 18 21 24 27 30 33 36 39

Arm A mOS (95% Cl) = 22.8 mo (9.4-NE) Arm B mOS (95% Cl) = 12.5 mo (7.6-16.4) Arm C mOS (95% Cl) = 12.7 mo (7.4-33.0)

• Similar ORR, DCR, and DOR were observed across treatment arms ─ Consistently high ORR (>30%) was achieved in all treatment arms ─ In total, 7 patients had complete response (4 in arm A and 3 in arm B)

• Arm A: NIVO1/ IPI3 Q3W × 4 followed by nivolumab 240 mg IV Q2W flat dose

• Arm B: NIVO3/ IPI1 Q3W × 4 followed by nivolumab 240 mg IV Q2W flat dose

• ORR is defined as CR + PR • SD does not include 2 patients in arm A and 1 patient in arm B who

were reported as non-CR/non-PD • DCR is defined as CR + PR + SD + non-CR/non-PD

1. El-Khoueiry AB et al. ILCA 2019. Abstract O-13.

So these data were presented at the 2019 ILCA meeting and looked at three different dose levels of nivolumab and ipilimumab, the two IO agents, in combination. And we see a quite encouraging

increase in objective response rate using a central response review, about 30% in each arm.

These responses are quite durable, lasting around 17 months or longer in each cohort. And we see here they are also coupled with a really encouraging survival signal, including in arm A with the highest dose of ipilimumab reaching 22.8 months in a second-line population, again.

CheckMate -040: Nivolumab Plus Ipilimumab—Adverse Events1

• Hepatic select TRAEs in 13 pts (27%) in arm A, 12 pts (24%) in arm B, and 8 pts (17%) in arm C

• Systemic corticosteroids used to treat select TRAEs in 25 pts (51%) in arm A, 12 pts (24%) in arm B, and 11 pts (23%) in arm C

1. Yau T et al. ASCO 2019. Abstract 4012.

Arm A NIVO1/IPl3 Every 3 wk

(n = 49)

Arm B NIVO3/IPl1 Every 3 wk

(n = 49)

Arm C NIVO3 Every 2 wk/IPl1

Every 6 wk (n = 48)

n (%) Any Grade Grade 3-4 Any Grade Grade

3-4 Any Grade Grade 3-4

Skin 29 (69) 4 (8) 24 (49) 2 (4) 21 (44) 1 (2)

Endocrine 16 (33) 1 (2) 9 (18) 1 (2) 10 (21) 1 (2)

Hepatic Aspartate aminotransferase increased Alanine aminotransferase increased Blood alkaline phosphatase increased Blood bilirubin increased Hepatitis Drug-induced liver injury Liver function test abnormal Transaminases increased

13 (27) 10 (20) 8 (16) 1 (2) 3 (6) 2 (4) 1 (2) 1 (2)

0

11 (22 8 (16) 4 (8)

0 0

2 (4) 1 (2) 1 (2)

0

12 (24) 10 (20) 7 (14) 2 (4)

0 0 0 0

1 (2)

6 (12) 4 (8) 3 (6) 1 (2)

0 0 0 0

1 (2)

8 (17) 6 (13) 4 (8) 3 (6) 2 (4)

0 0 0 0

2 (4) 2 (4)

0 0 0 0 0 0 0

Gastrointestinal 12 (27) 3 (6) 1 (2) 1 (2) 8 (17) 1 (2)

Pulmonary 4 (8) 1 (2) 0 0 0 0 Hypersensitivity/infusion reaction 4 (8) 0 0 0 1 (2) 0 Renal 0 0 0 0 1 (2) 1 (2)

When we look at the safety, however—that’s again one of the important questions before we bring forward a new combination in HCC—I think it’s important to highlight here that the arm A with the higher-dose ipilimumab did have an increased rate of hepatic adverse events, in particular, and systemic corticosteroids were used to treat treatment-related adverse immune events in 51% of patients in this cohort. So a very important signal of efficacy coupled by increased immune toxicity that requires careful patient selection.

Phase 3 CheckMate -9DW Trial: Study Design1


• Primary endpoint: OS • Secondary endpoints: ORR, DOR, and time to symptom deterioration

• Advanced HCC based on histological confirmation with ≥1 RECIST 1.1 measurable previously untreated lesion

• Child–Pugh 5 or 6

• ECOG PS 0-1

N = 726

Sorafenib or lenvatinib


R

These encouraging data have launched the phase 3 CheckMate -9DW trial of nivolumab plus ipilimumab in combination versus first-line sorafenib or lenvatinib, according to physician’s choice.



Durvalumab Plus Tremelimumab Data1

1. Kelley RK et al. ASCO 2017. Abstract 4073.

Antitumor Activity

HBV+ (n = 11)

HCV+ (n = 9)

Uninfected (n = 20)

All (N = 40)

Confirmed ORR, % (95% CI)

0 (0.0-28.5)

11.1 (0.3-48.2)

30.0 (11.9-54.3)

17.5 (7.3-32.8)

CR + PR, (confirmed + unconfirmed), % (95% CI)

9.1 (0.2-41.3)

11.1 (0.3-48.2)

40.0 (19.1-63.9)

25.0 (12.7-41.2)

DCR at week 16, % (95% CI)

45.5 (16.7-76.6)

44.4 (13.7-78.8)

70.0 (45.7-88.1)

57.5 (40.9-73.0)

Investigator-Assessed Response

Most common AEs were fatigue, pruritus, and elevated liver enzymes

And lastly, before we conclude this section, I’ll share some additional second-line data on another immuno-oncology combination, durvalumab plus tremelimumab. Durvalumab is an anti–PD-L1 inhibitor. Tremelimumab is a CTLA-4 inhibitor like ipilimumab.

And we have looked at this combination, again, in a variety of different viral contexts, and in a phase 1B context with about 40 patients, the objective response rate was about 17% for confirmed, 25%, [confirmed and] unconfirmed.

• Unresectable HCC not eligible for LRTs

• BCLC stage B or C

• Child–Pugh A

• No prior systemic therapy

N = ~1,200

Phase 3 HIMALAYA Trial: Durvalumab Plus Tremelimumab Versus Sorafenib1


• Primary endpoint: OS • Other endpoints: TTP, PFS, ORR, DCR, DOR, and QOL

Durvalumab

Sorafenib

Durvalumab + tremelimumab

Durvalumab + tremelimumab

Regimen 1

Regimen 2 R

And this has prompted an ongoing randomized phase 2, as well as the phase 3 HIMALAYA trial, with three different IO arms: durvalumab, durvalumab plus tremelimumab with two different dose levels, compared with the standard of care, sorafenib, at that time. This was, again, in the first-line context based on the second-line data. And the study is completed or in accrual, and we expect results in the near future.

Ongoing Trials of Combination Treatments With Checkpoint Inhibitors1

1. http://www.clinicaltrials.gov. Accessed January 7, 2020.

Combinations of IO + Anti-VEGF Atezolizumab + bevacizumab (IMbrave150; phase 3) Anti–PD-L1 + anti-VEGF

Nivolumab + lenvatinib (phase 1b) Anti-PD1 + TKI

Sintilimab + bev biosimilar (ORIENT-32; phase 2/3) Anti-PD1 + anti-VEGF

Atezolizumab + cabozantinib (COSMIC-312; phase 3) Anti–PD-L1 + TKI

Nivolumab + sorafenib (phase 2) Anti-PD1 + TKI

Avelumab + axitinib (VEGF Liver 100; phase 1b)

Anti–PD-L1 + TKI Pembrolizumab + lenvatinib (LEAP-002; phase 3) Anti-PD1 + TKI

Pembrolizumab + regorafenib (KN-743; phase 1b)

Anti-PD1 + TKI

Spartalizumab + sorafenib (phase 1b) Anti-PD1 + TKI

Camrelizumab + apatinib (phase 3) Anti-PD1 + TKI

Sorafenib + pembrolizumab (phase 2) Anti-PD1 + TKI

Avelumab + regorafenib (phase 1/2) Anti–PD-L1 + TKI

Combinations of Two Checkpoint Inhibitors (IO + IO) Nivolumab + ipilimumab (CheckMate -9DW; phase 3) Anti-PD1 + anti-CTLA4

Durvalumab ± tremelimumab (HIMALAYA; phase 3)

Anti–PD-L1 + anti-CTLA4

Relatlimab ± nivolumab (phase 1/2) Anti-LAG3 ± anti-PD1

Other Combinations

Camrelizumab + FOLFOX (phase 3) Anti-PD1 + chemotherapy

Durvalumab ± bevacizumab + TACE (EMERALD 1, phase 3)

Anti–PD-L1 ± anti-VEGF + TACE

Nivolumab + TACE (phase 1) Anti-PD1 + TACE

Here is a selection of many ongoing trials of combination treatments with checkpoint inhibitors now in play.

Patient Case 3 Follow-Up: 69-Year-Old Man With HCV

Consider checkpoint inhibitor immunotherapy as second- or third-line therapy • In patients who have not received immunotherapy in first line • If rapid progression and/or intolerance to TKI first-line therapy • If increasing degrees of hepatic dysfunction

– Prospective cohort and retrospective case series show acceptable safety and efficacy of nivolumab in Child–Pugh B HCC

• If contraindications to antiangiogenic therapy (eg, nonhealing wounds, active venous thromboembolism, bleeding complications)

So, lastly, to turn back to our case, following up with this patient who had rapid progression on sorafenib, in him, I chose to consider and treat him with checkpoint inhibitor immunotherapy. In particular in patients who have not received immunotherapy in first line, we now look to immunotherapy in second line. I also consider it in patients such as this, who have rapid progression or intolerance to TKI early in first-line therapy.

In patients with increasing degrees of hepatic dysfunction, including Child–Pugh B, we have very little safety data for patients to be treated with lenvatinib and cabozantinib, ramucirumab, in fact, and there are, on the other hand, prospective cohort and retrospective case series showing acceptable safety and efficacy of nivolumab, the PD-1 inhibitor, in Child–Pugh B HCC. And I also consider checkpoint inhibitor immunotherapy as second- or third-line therapy in patients, of course, who have contraindications to antiangiogenic therapy.


Amit Singal, MD, MSUT Southwestern Medical Center Dallas, Texas

21Go online to complete the post-test and evaluation for CME/MOC credit


• A substantial proportion of patients receive at least second-line systemic therapies for advanced HCC

• Choice of second-line therapy depends on which agent(s) were received in first line and the efficacy and tolerability of the first-line regimen – First-line TKI: Consider second-line TKI, ramucirumab, or IO therapy

depending on response and tolerability to first-line TKI along with AFP, liver function, and comorbidities

– First-line IO or IO combination: Consider second-line TKI or ramucirumab

• Future research is needed to guide choice and sequence of second-line or greater therapies after changing first-line treatment landscape

Conclusions

And my concluding points to take home, a substantial proportion of patients receive at least second-line therapy. At least 50% of patients who are eligible for first-line trials are now getting to second-line therapy. And another proportion of those, 50% perhaps, are also going to third-line therapy.

The choice of second-line therapy depends on which agent they received in first line, and so it’s a dramatically changing landscape, and we don’t know yet how to position agents after atezolizumab/bevacizumab, which is the subject of future conferences to discuss. And it also depends on how well they did on the first-line therapy, and their tolerability in the first line.

So my sense in the clinic is, for patients who received a first-line TKI, to consider a second-line TKI in patients who had prolonged benefit with their first-line TKI and tolerated it well. For patients who did not tolerate it and had rapid progression, I’m more likely to choose a second-line IO agent. In patients who received IO therapy in the first line, of course, we would consider a second-line TKI. And that’s it. Thank you.

Dr. Finn: Thank you so much, Katie. That was a great talk. A lot of data.

CaseBook 3: Going Back to the Playbook—Expanding Options With Clinical Trials for Every Patient With HCC

Patient Case 4: 32-Year-Old Man With Chronic HBV

32-year-old man with chronic HBV, on tenofovir, and compensated cirrhosis Undergoing HCC surveillance and found to have liver mass MRI shows 3.5-cm LR-5 lesion in segment VI Child–Pugh A; Bili 0.7, Alb 4.0, INR 1.0 Platelet count 217 ECOG PS 0

Dr. Finn: It’s my pleasure to introduce Amit, again, who is a hepatologist, unlike Katie and I, who are oncologists.

Dr. Singal: Yeah, so hopefully I can fill the big shoes that preceded me.

Dr. Finn: Tough, tough shoes to follow.

Dr. Singal: Yeah. So I think, you know, from the excellent presentations by both Rich and Katie, I think it’s clearer that there have been a lot of advances in the systemic space. And I’ve been tasked with doing something a little bit different. We unfortunately haven’t had those same advances in the early and intermediate space, but there are several trials that are ongoing.

And, you know, what I’ve been asked to do is really highlight some of these clinical trials that are ongoing that are promising and will hopefully yield similar improvements in survival for these patients who present at earlier stages.

So to highlight this, I want to, once again, similar to the prior presentations, start with a case. So this is a young patient who has chronic hepatitis B who’s on treatment for his hepatitis B and has compensated cirrhosis.

He’s undergoing HCC surveillance, and is found to have a liver mass. And fortunately, that was found early, so the liver mass is 3.5



cm, and it’s a LI-RADS 5. So on imaging, it’s consistent with HCC, therefore doesn’t need to undergo a biopsy.

His liver function is compensated, as we discussed. He’s a Child–Pugh A patient, and his platelet count is preserved at 217, suggesting he does not have portal hypertension. He has good performance status.

Modified BCLC Staging System 20181

1. Galle PR et al. J Hepatol. 2018;69:182-236.

HCC in cirrhotic liver

Solitary 2-3 nodules ≤3 cm

Transplant candidate

Resection

No Yes No

Transplant Ablation Chemoembolization Systemic therapy

>5 y 3 mo >2.5 y

Ablation BSC

Optimal surgical candidate

≥10 mo

Yes

Very early stage (0) Single <2 cm

Preserved liver function PS 0

Early stage (A) Single or 2-3

nodules <3 cm Preserved liver function

PS 0

Intermediate stage (B) Multinodular, unresectable

Preserved liver function PS 0

Advanced stage (C) Portal invasion/

extrahepatic spread, Preserved liver function

PS 1-2

Terminal stage (D) Not transplantable

End-stage liver function

PS 3-4

So I think when you take a look at the BCLC staging system, he clearly falls into this early-stage category, which Rich alluded to earlier. So he would be somebody who would fall into this BCLC stage A, with a unifocal lesion, and he would be an optimal surgical candidate and therefore would undergo resection.

And I think one of the things that I like to highlight when I have this slide up is that really early detection is one of the best ways that we can optimize survival. So patients who are found early and can undergo these curative therapies have a survival that exceeds 5 years and often reaches 10-plus years.

Patient Case 4: 32-Year-Old Man With Chronic HBV (Cont’d)

Patient undergoes robotic liver resection without complication Returns to clinic, at which time you reinforce risk

of recurrence and need for continued surveillance

So this patient undergoes a robotic liver resection. He does quite well and goes home early. And he comes to your clinic, and he then discusses what does he need to do? He’s now cured from his HCC but wants to talk about the risk of recurrence and the need for continued surveillance.

• Recurrence: 70%-80% at 5 years

HCC Recurrence Post Resection1-5

1. Ikai I et al. Cancer. 2004;101:796-802. 2. Vauthey JN et al. J Clin Oncol. 2002;20:1527-1536. 3. Shi M et al. Ann Surg. 2007;245:36-43. 4. Katz SC et al. Ann Surg. 2009;249:617-623. 5. Tabrizian P et al. Ann Surg. 2015;261:947-855.

Factors Associated With Outcome No. of nodules (5-y OS, %) Single: 57 ≥3: 26 Ikai I et al

Size of tumor (5-y recurrence, %) <5 cm: 32 >5 cm: 43 Vauthey JN et al

Tumor-free margin (5-y OS, %)

2 cm: 74.9 1 cm: 49.1 Shi M et al

Blood loss (median OS, mo)

<1L: 68 <2L: 49

>1L: 18 >2L: 13 Katz SC et al

• No proven adjuvant therapy to decrease HCC recurrence

N = 661

Median time to recurrence: 22 months

1-year: ~30-35% 2-year: ~50%

1.0-

0.9-

0.8-

0.7-

0.6-

0.5-

0.4-

0.3-

0.2-

0.1-

0.0-

0 12 24 36 48 60

661 347 213 140 102 72

And so, as I discussed, when you undergo resection, this really is associated with long-term survival. The 5-year survival rates with resection approach 70%; however, you leave behind that cirrhotic liver, and one of the downsides of resection is really the high recurrence rate, which mirror that of the survival. So the recurrence at 5 years approaches 70% to 80%.

And here you can see in this table, there are some factors that are associated either with survival or with recurrence. But you can see that patients who have multifocal tumors, larger tumors, have a positive margin or have greater blood loss are associated with either higher recurrence or poorer survival.

And you can see from this figure, when you take a look at a large cohort of patients who underwent resection from a high-volume center, you can see that the recurrence rate at 1 year, about one-third of patients have recurrence. And when you reach that 2-year mark, the recurrence rate is over 50%, with a median time to recurrence of 22 months. So this really highlights that the survival is good, but there is a need for adjuvant therapy that can help decrease HCC recurrence.

Phase 3 STORM Trial: Adjuvant Sorafenib for HCC After Resection or Ablation1

Sorafenib is not an effective intervention in the adjuvant setting for HCC

0

25

50

75

100

0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64

Ove

rall

Surv

ival

, %

Time, mo

HR 0.940 (95% Cl, 0.780-1.134) p = 0.26 (one-sided)

Sorafenib (n = 556) Placebo (n = 558)

Sorafenib (n = 556) Placebo (n = 558)

HR 0.995 (95% Cl, 0.761-1.300); p = 0.48 (one-sided) 0

25

50

75

100

0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60

Rec

urre

nce-

Free

Sur

viva

l, %

Time, mo

HR 0.940 (95% Cl, 0.780-1.134); p = 0.26 (one-sided)

1. Bruix J et al. Lancet Oncol. 2015;16:1344-1354.

So Rich talked about the SHARP trial and the excitement that came after the SHARP trial, and he talked about how I think anyone who was in the field at that time was thinking that this was the first of many trials that would lead to a multitude of first-line options.

And I think in parallel with that, we also thought that this could be an adjuvant therapy to help reduce recurrence after resection or



ablation. And this basically led to the phase 3 STORM trial, which evaluated sorafenib as an adjuvant therapy after resection or ablation.

And so this randomized controlled trial either placed people on sorafenib or placebo, and you can see that there were over 1,000 patients, with a primary outcome of recurrence-free survival. And unfortunately, there was no difference in recurrence-free survival, with a hazard ratio of 0.94, and a median recurrence-free survival of just over 33 months for both groups.

As a secondary endpoint, they also looked at overall survival, and similarly, you see no difference in overall survival. So, unfortunately, despite sorafenib’s effectiveness in the advanced space, this did not appear to be an effective intervention in the adjuvant setting.

Phase 3 CheckMate -9DX Trial: Adjuvant Nivolumab in High-Risk Resected HCC1


• Primary endpoint: RFS • Other endpoints: OS and TTR

R

Nivolumab

Placebo

• First diagnosis of HCC with curative resection or ablation

• Nonviral-related HCC, HBV-HCC, or HCV-HCC • Child–Pugh 5 or 6 • ECOG PS ≤1 • No evidence of tumor metastasis or co-existing

malignant disease • No prior therapy for HCC • No prior liver transplantation and not on waitlist

for transplantation N = 530

So, of course, like everything else, we’re now reevaluating everything with the excitement that comes with the checkpoint inhibitors. And so there are now a multitude of trials that are reevaluating whether we can identify an adjuvant therapy that can help reduce the risk of recurrence. And so, over the next four slides, you’re going to see four ongoing trials that are evaluating different agents in this setting.

So, first, starting with the CheckMate -9DX trial, this is taking patients who undergo resection or ablation, but are thought to be high risk for recurrence. And so high risk of recurrence can be defined by a multitude of different criteria. So this can be multifocal lesions, so three lesions, the largest being less than 5, that undergo resection.

You can have microvascular invasion, or it could be poorly differentiated. Or, if it’s undergoing ablation, you can be between this 3- to 5-cm size cut off. So these patients are all regarded to be at higher risk of recurrence but have no evidence of residual disease after that treatment.

And these patients otherwise have good liver function and good performance status and are being randomized to either nivolumab or placebo, with a primary endpoint of recurrence-free survival.

Phase 3 KEYNOTE-937 Trial: Adjuvant Pembrolizumab1

1. https://clinicaltrials.gov/ct2/show/NCT03867084. Accessed February 14, 2020.

• Primary endpoints: RFS and OS

Pembrolizumab

Placebo

R

• Diagnosis of HCC by radiologic criteria and/or pathologic confirmation

• No radiologic evidence of disease prior to enrollment

• Child–Pugh 5 or 6 • ECOG PS 0 • AFP <400 ng/mL

N = 950

Similarly, there’s the phase 3 KEYNOTE-937 study, which has very similar inclusion/exclusion criteria. So high risk of recurrence after undergoing surgical resection or ablation. 950 patients who are being randomized to pembro or placebo, and this trial has co-primary outcomes of recurrence-free survival and overall survival.

Phase 3 EMERALD-2 Trial: Adjuvant Durvalumab ± Bevacizumab1


Trial of Durvalumab as Monotherapy or in Combination With Bevacizumab as Adjuvant Therapy in Patients Who Are at High Risk of Recurrence After Curative Hepatic Resection or Ablation

R

Durvalumab + bevacizumab

Durvalumab + placebo

Placebo

• Histologically or cytologically confirmed HCC and successfully completed curative therapy (resection or ablation)

• Imaging to confirm disease-free status within 28 days prior to randomization

• ECOG PS 0-1 • Child–Pugh 5 or 6

N = 888

• Primary endpoints: RFS • Secondary endpoints: OS, TTR, among others

The next trial is the phase 3 EMERALD-2 trial. This is evaluating adjuvant durvalumab plus/minus bevacizumab. So once again, high-risk candidates after surgical resection or ablation. Patients have to be disease free 28 days prior to randomization, and they’re being randomized to one of three arms: durvalumab plus bevacizumab, durvalumab alone, or placebo—primary endpoint being recurrence-free survival.

Phase 3 IMbrave050 Trial: Adjuvant Atezolizumab and Bevacizumab1


• Primary endpoints: RFS • Second endpoints: OS, TTR

Atezolizumab + bevacizumab

Active surveillance

R

• Histologically or cytologically confirmed HCC and undergone curative resection or ablation

• High risk for HCC recurrence after resection or ablation

• ECOG PS 0-1 • Child–Pugh Class A

N = 662

And finally, the IMbrave050 trial, which is evaluating adjuvant atezolizumab and bevacizumab. Similar inclusion/exclusion



criteria, 660 patients being randomized to atezolizumab plus bevacizumab or placebo, with a primary endpoint of recurrence-free survival.

So the fact that we have four ongoing trials all actively enrolling does highlight that this is an area that has been identified as an area of need, a lot of excitement, and I think that we’re going to see hopefully positive data, or at least exciting data, come out of these trials soon.

Phase 2 Trial: Perioperative Nivolumab ± Ipilimumab in Resectable HCC1,2

1. Kaseb AO et al. ASCO GI 2019. Abstract 185. 2. Kaseb AO et al. ASCO GI 2020. Abstract 486.

• Primary endpoint: Safety • Secondary endpoints: ORR, pCR, and TTP

R Resectable HCC

N = 26 Nivolumab (arm B) 240 mg every 2 wk +

ipilimumab 1 mg/kg for 6 wk

(n = 3)

Nivolumab (arm A) 240 mg every 2 wk

for 6 wk (n = 5)

Continue adjuvant

immunotherapy for up to 2 y

after resection

Surgical resection

within 4 wk

of last cycle

• pCR was demonstrated in five of 20 patients (25%): two in arm A and three in arm B • Five in arm B and one in arm A experienced grade 3 or higher toxicity prior to surgery • Surgery aborted in five patients (three for disease progression)

In addition to adjuvant therapy, there’s also been excitement for the possibility of neoadjuvant therapy. And so this is a phase 2 study that’s ongoing by Kaseb and colleagues. And so this is evaluating neoadjuvant nivolumab plus/minus ipilimumab in patients who are deemed to be resectable.

You have 26 patients who have been randomized to either nivolumab or nivolumab plus ipilimumab at the dose you can see here. Both of those arms are given for 6 weeks, and then the surgical resection is done within 4 weeks of the last cycle. And then they continue adjuvant therapy for up to 2 years after the resection.

The data that were presented before was on the first five or six patients, and there was a pathologic complete response in about one-third of patients. The updated data continue to show that you do see a pathologic response in a high proportion of patients, 25% in the update, two in the first arm, and three in arm B.

And then you can see that overall the therapy does appear to be well tolerated, but I think the thing that does bear observation is how many patients could have disease progression during this time period that would otherwise prevent those patients from undergoing resection.

And in the update, it appears that surgery was aborted in five patients, and three of them for disease progression. And so for those of you who may remember, this is a potential concern. So when TACE was evaluated as a neoadjuvant therapy, this was observed in a proportion of patients in whom then surgical resection was not deemed to be possible. And that was one of the things that really prevented TACE from being used as a

neoadjuvant therapy. And so I think this will be one of the things that will need to be looked at as this continues to be evaluated.

Ongoing Studies in Neoadjuvant Setting

1. https://clinicaltrials.gov/ct2/show/NCT03299946. Accessed January 14, 2020. 2. https://clinicaltrials.gov/ct2/show/NCT03337841. Accessed January 14, 2020. 3. https://clinicaltrials.gov/ct2/show/NCT03916627. Accessed January 14, 2020.

Nivolumab + cabozantinib (CaboNivo)

Phase 1 (NCT03299946)1

Cemiplimab Phase 2 (NCT03916627)3

Pembrolizumab

Phase 2 (NCT03337841)2

Ongoing Studies in

Neoadjuvant Setting

This is only one of many studies that are also going on in the neoadjuvant setting, and you can see the others being listed here. There’s nivolumab plus cabozantinib. There’s pembrolizumab alone, and then there’s cemiplimab. But basically, there are multiple studies that are evaluating this in the neoadjuvant as well as the adjuvant setting.

Patient Case 5: 64-Year-Old Man With NASH

64-year-old man with NASH and compensated cirrhosis Found to have incidental liver mass on imaging MRI shows 6.5-cm LR-5 lesion with two satellite nodules No vascular invasion and no metastatic disease Child–Pugh A; Bili 0.9, Alb 3.7, INR 1.0 ECOG PS 0

So, moving beyond the patients who present with resectable disease, I think that there’s clearly an area of need, as well, in the intermediate stage. And so to discuss the trials that are ongoing in the intermediate stage, I’d like to start with a case of a 64-year-old man with NASH and compensated cirrhosis who was found to have an incidental liver mass.

This patient presents with larger tumor burden. He has a 6.5-cm lesion and two satellite nodules, fortunately, no vascular invasion, and no metastatic disease. He has good liver function, Child–Pugh A, and good performance status.



BARCELONA STAGE

AASLD Guidance: HCC 20181

Level of Evidence STAGE 0 STAGE A STAGE B STAGE C STAGE D

1 Resection TACE Sorafenib (1L); lenvatinib (1L); regorafenib (2L); cabozantinib

(2L); ramucirumab (2L)

2 RFA MWA

Resection OLT, RFA, MWA

TARE, TACE, SBRT

TARE Downsize OLT

Nivolumab (2L); pembrolizumab (2L)

OLT BSC

3 TARE

1. Marrero JA et al. Hepatology. 2018;68:723-750.

So, going back to the BCLC, the Barcelona staging system, you can see that this patient presents beyond an early stage but does present within BCLC stage B. And, as many of you may recall, this is where chemoembolization and radioembolization have really been the main therapies that we’ve used to treat these patients.

And already, as alluded to by Rich, these therapies have prolonged survival, but have not been curative. And this is one of the areas of need in terms of improving both survival as well as the response, as we can observe in these patients.

Lipiodol Transarterial Chemoembolization for Hepatocellular Carcinoma: A Systematic Review of Efficacy and Safety Data

Mortality:

214/34,137

0.6%

Cause of death

Hepatic failure Hepatic encephalopathy Infectious complications Gastrointestinal bleeding Tumor rupture Liver abscess Pulmonary embolization Other/not specified Any cause

No.

59 9

20 17 8 4 4

93 214

Safety Analysis (n = 34,137) Efficacy Analysis (n = 10,108)

1. Lencioni R et al. Hepatology. 2016;64:106-116.

TACE: A Systematic Review of Efficacy and Safety Data1

OS No. of Studies Estimate Lower

95% CI Upper 95% CI

Median, mo ≤2002 >2002

19 44

18.5 19.8

14.6 15.5

22.4 24.1

1-year, % ≤2002 >2002

19 71

70.7 70.4

63.2 65.2

78.3 75.5

2-year, % ≤2002 >2002

21 50

51.1 52.0

37.1 43.9

65.1 60.2

3-year, % ≤2002 >2002

13 53

27.8 43.4

18.3 34.9

37.4 51.8

Analysis of OS After Lipiodol TACE for HCC by Year of Publication

0

20

40

60

80

100

6 mo 1 y 2 y 3 y 5 y

81

70

52

40

32

Time

OS,

%

So, just to briefly review—this was a nice systematic review that was published a couple of years ago that outlines the outcomes that we see with chemoembolization. The systematic review largely focused on lipiodol, traditional chemoembolization. I think there are a couple of things that are worth pointing out.

The first, that chemoembolization has been evaluated in a large number of patients. And I think, going back to Rich’s point, this is largely because of the limitations of systemic therapy in the early days and chemoembolization being the therapy that was used for many, many HCC patients.

But the good thing is that over time we’ve realized that chemoembolization is relatively a safe therapy in most patients. You can see that complications in skilled hands are relatively low, and the mortality also is low, at 0.6%.

But I think that I did say in the setting of skilled hands—and I think that’s important, because there’s a lot of operator dependency with chemoembolization in how this is delivered and the outcomes you will see.

Overall, when you take a look over the entire time period, you can see that chemoembolization is able to provide very high short-term survival. So you see an 81% 6-month survival and a 70% 1-year survival. But as you start going out, you see only a 40% 3-year survival and a 32% 5-year survival. And that’s really where we need to see improvements.

The fortunate thing is, as we learned more about chemoembolization, there has been some improvement. So you can see when you when you stratify by the year of publication, the 3-year survival used to be in the 28% to 30% range. In the more recent publications, we have seen the 3-year survival climb closer to 43%. So we have seen some improvements over time.

1. Lammer J et al. Cardiovasc Intervent Radiol. 2010;33:41-52.

DEB-TACE Versus Conventional TACE: A Randomized Phase 2 Trial1

Prospective Randomized Study of Doxorubicin-Eluting Bead Embolization in the Treatment of Hepatocellular Carcinoma: Results of the PRECISION V Study

Johannes Lammer, Katarina Malagari, Thomas Vogl, Frank Pilleul, Alban Denys, Anthony Watkinson, Michael Pitton, Geraldine Sergent, Thomas Pfammatter, Sylvain Terraz, Yves Benhamou, Tves Avajon, Thomas Gruenberger, Maria Pomoni, Herbert Langenberger, Marcus Schuchmann, Jerome Dumortier, Christian Mueller, Patrick Chevallier, Riccardo Lencioni

One of those improvements has been the transition from conventional TACE to DEB-TACE. So this was evaluated in the PRECISION V study, and I think overall when you take a look at these randomized data, there was no difference in actual outcomes. There was no difference in survival or response rates, but we do see that overall DEB-TACE appears to be better tolerated.

So the AE profile with DEB-TACE is better than that of conventional TACE, and I think many centers have largely transitioned to using DEB-TACE as the conventional way to deliver this therapy.



0

20

40

60

80

100

0 10 20 30 40 50

Surv

ival

Pro

babi

lity,

%

Time Since Randomization, mo

Treatment — cTACE — Y90

Phase 2 PREMIER Trial: TARE Versus TACE1

1. Salem R et al. Gastroenterology. 2016;151:1155-1163.e2

20

30

40

50

60

70

80

90

100

0 5 10 15 20 25 30

Patie

nts

With

out P

rogr

essi

on, %

Time Since Randomization, mo

Y90

P = .0012

cTACE

No. at risk cTACE

Y90

21 8 3 2 0 0 0

24 12 7 3 2 1 0

Overall Survival (Censored)

21 10 2 1 1 0

24 9 2 1 0 0

No. at risk cTACE

Y90

The other advance that we've seen is that many centers, at least in the United States, have also started to increasingly adopt radioembolization. Radioembolization has been primarily evaluated in single-arm studies, and so there's a large number of single-arm studies that show reasonable outcomes with radioembolization, showing good responses and good tolerability. Unfortunately, there have been very little randomized data with radioembolization, and this is one of the things that's precluded it from having a center place in many guidelines for treatment of patients with locoregional disease.

Here you can see data from a small randomized controlled trial that was conducted in a single center, conducted by Riad Salem and colleagues. This trial included, unfortunately, a grand total of 45 patients who were randomized between chemoembolization and radioembolization.

And I think the key things to take away here is that there was a significant difference, even though this was a small trial, in progression-free survival. And you can see that radioembolization had significantly longer progression-free survival than chemoembolization.

However, the overall survival curves nearly overlapped. And using the data that have come out of this center, it was predicted that you would need a trial with over 1,000 patients to actually start to see a survival difference between these two therapies.

Despite that, I will say that even in our center, we have largely shifted to radioembolization for many of these patients, given the tolerability and progression-free survival data. However, although these data are nice, and we’re starting to see improvements, I think that overall I think we can agree we’re not where we want to be.

Moving TKIs to Earlier-Stage Disease: Rationale and Early Evidence1

Angiogenic switch • Proangiogenic >

antiangiogenic factors

• Exposure + shedding of tumor cells into circulation

• ↑ MMP during angiogenesis

• Central hypoxia • Vessel compression by edema • Starvation treatments (eg, TACE)

Leaky vessels • Abnormal flow • Interstitial hypertension • Peritumor edema • Ineffective delivery of O2

and nutrients

• ↑ Growth factors • ↑ Anti-apoptotic proteins • Resistance to chemo/radiotherapy • Impaired tumor immune response

– ↓ CD8+

– ↑ Treg/MDSC – ↑ PD-1/PD-L1

Angiogenesis

Hypoxia Tumor growth (and metastasis)

1. Liu K et al. Clin Transl Gastroenterol. 2017;8:e98.

↑ HIF-1α, VEGF, IGF2

And there has been a desire to try combination therapies to see if we can also improve outcomes for patients with intermediate-stage disease. And there is some scientific rationale for trying to move the TKIs into this space, and so without spending too much time on this slide, given time, I will say that I think all of us are aware that after you do a chemoembolization, you do see surges in some factors, like VEGF, HIF-1α, that can cause angiogenesis, which may explain some of the continued escape after chemoembolization.

And there’s an immunomodulatory change that also can cause continued tumor growth and the presence of metastases. And there’s some rationale for both TKIs as well as IOs being used in combination with many of these locoregional therapies, which may improve outcomes for these patients.

TACE + Systemic Therapy (TKI)1

a PFS was used in the TACE-2 study. 1. Kudo M, Arizumi T. Oncology. 2017;93(suppl 1):127-134.

Post-TACE (N = 458) BRISK-TA (N = 502) SPACE (N = 307) ORIENTAL (N = 888) TACE-2 (N = 313)

Sorafenib (n = 229)

Placebo (n = 227)

Brivanib (n = 249)

Placebo (n = 253)

Sorafenib (n = 154)

Placebo (n = 153)

Orantinib (n = 444)

Placebo (n = 444)

Sorafenib (n = 157)

Placebo (n = 156)

Phase 3 3 (immature/ terminated) 2 3 (terminated due to

interim analysis) 3 (terminated due to

interim analysis)

mOS, mo 29.7 NR 26.4 26.1 NR NR 31.1 32.3 21.1 19.7 HR (95% CI) 1.06 (0.69-1.64) 0.90 (0.66-1.23) 0.898 (0.606-1.330) 1.090 (0.878-1.352) 0.91 (0.67-1.24) P .79 .528 .295 .435 .57

mTTP, mo 5.4 3.7 8.4 4.9 5.6 5.5 ND ND 7.9a 7.8a

HR (95% CI) 0.87 (0.70-1.09) 0.61 (0.48-0.77) 0.797 (0.588-1.080) ND 0.99 (0.77-1.27) P .252 <.0001 .072 ND .94

Primary endpoint TTP OS TTP OS PFS

Definition of progression RECICLE mRECIST mRECIST TACE discontinuation

criteria RECIST 1.1

Median DOR of study drug 17 wk 24 wk 21 wk 43.6 wk 17.1 wk

So, unfortunately, despite the scientific rationale, when you take a look at the data evaluating this, unfortunately, the data have not borne out any beneficial effect for combining TKIs with chemoembolization. So there have been multiple trials that have evaluated this with a grand total of over 2,000 patients.

Those trials have included endpoints of time to progression, and overall survival, as well as progression-free survival. And I think overall, the one thing you can take away here is that all five studies failed to show a benefit of combining TKI therapy with chemoembolization.



These trials include trials that were done in the United States, trials that were done in Europe, as well as trials that were done in Asia. And unfortunately, despite differences in patient populations and small differences in terms of trial design, all of them failed to show a benefit.

Lessons Learned With Locoregional Therapy + TKI Combinations

1. Kudo M et al. ASCO GI 2018. Abstract 206.

No. at Risk TACE + sorafenib 80 56 36 17 12 3 0 TACE alone 76 37 22 8 4 2 0

TACTICS Trial: Phase 21

• Significantly improved PFS with TACE + sorafenib vs TACE alone

• 25.2 vs 13.5 months, respectively; HR = 0.59 (95% CI, 0.41-0.87); P = .006

TACE + sorafenib TACE alone

0.0

0.5

1.0

0 12 24 36 48 60 72

PFS,

%

Time, mo

0

• Different populations, including earlier-stage HCC

• Timing of sorafenib • Dose of sorafenib • Duration of sorafenib • Early termination of study

based on other studies • Study design has

conservative stopping rules

Interestingly enough, there was a study that was presented at ASCO GI 2018, 2 years ago, that did show a potential benefit of combining TKI therapy with locoregional therapy. And I think that some people think that this opens up that box again for us to maybe evaluate this.

So this was the TACTICS trial. It was a phase 2 study that included just around 160 patients who were randomized to TACE alone or TACE plus sorafenib. And I think one of the unique things about this trial is that it did not evaluate typical progression but used the unique endpoint of “un-TACE-able” disease.

So, meaning progression to the point where you can no longer TACE patients, whether that’s progression outside of the liver or significant liver dysfunction. And this resulted in patients staying on the sorafenib significantly longer than the prior studies that evaluated the combination therapy.

And what the authors of this study found was that when you take a look at this progression-free survival, which was the primary endpoint, it was significantly longer with the combination compared to TACE alone, 25.2 months versus 13.5 months, with a 41% reduction in progression-free survival. Unfortunately, they weren't able to evaluate overall survival, both at the time of this presentation as well as the final publication, which is now available (Kudo M et al. Gut. 2019 Dec 4. [Epub ahead of print]).

But I think that overall, you can say that there may be some lessons that you can learn from the difference in terms of the results from this trial versus the prior trials, including, once again, the difference in terms of this outcome, maybe the importance of the duration of the sorafenib.

And then the other thing that was different in terms of this trial was that they started the sorafenib before the chemoembolization. Patients were on the sorafenib for at least

3 weeks going into the chemoembolization, which may have allowed a greater effect and benefit compared with some of the other studies.

Overall, in terms of my opinion, I think that this is an interesting study. I don’t necessarily think that it should open up the box again. I still think that the preponderance of data, in my opinion, still show that there’s no proven benefit of combining the two therapies versus TACE alone.

Immunotherapy

Ablation

TACE Y90

SBRT

1. Duffy AG et al. J Hepatol. 2017;66:545-551.

The Next Wave: Immunotherapy1

Liver-Directed Locoregional Therapy + Immunotherapy

AFP

Afte

r B

efor

e

H&E CD3 CD8 IHC Staining in Response to Treatment

CD3 CD8

Pixe

l Cou

nt, %

8

0

CD3 CD8

Pixe

l Cou

nt, %

15

10

0

5

CD3 CD8 Responders Nonresponders

6

4

2

Before After

Baseline

6 mo

Time, mo Time, mo Time, d

Sum

of L

esio

ns

Rel

. to

Bas

elin

e

AFP,

ng/

mL

Of course, like everything else, the IO agents have created new excitement, and there are a lot of trials that are now evaluating the combination of IO agents in combination with locoregional therapy.

Here you can see a copy from some of the figures from a presentation in Journal of Hepatology, in which they took patients with intermediate- and advanced-stage disease, the majority having advanced-stage disease, and treated them with tremelimumab, and then the patients underwent an ablation.

And what the authors of this study saw and, you can see, was that, compared with before, there was an increase in CD3 and CD8 cells. You can see these patients had a nice decrease in terms of tumor burden and a decrease in terms of AFP.

And so this was a relatively small study, but I think that the authors of this study basically concluded that locoregional therapy can help promote a change in the immune environment that can potentially augment the effect of immunotherapy.



Study Arms Phase Patient Population NCT Identifier

Pembrolizumab + local ablation (IMMULAB) 2 Candidates for local ablation NCT03753659

Nivolumab after SIRT (Y90) 2 Candidates for locoregional therapy NCT03380130

Nivolumab + TACE (IMMUTACE) 2 Intermediate-stage HCC NCT03572582

Durvalumab + tremelimumab + DEB-TACE 2 Intermediate-stage HCC NCT03638141

Pembrolizumab + TACE 1/2 Intermediate-stage HCC NCT03397654

DEB-TACE + nivolumab 1 BCLC B NCT03143270

Durvalumab + tremelimumab + radiation 2 Locally advanced/unresectable or metastatic HCC NCT03482102

Pembrolizumab + SBRT 2 Locally advanced/advanced HCC NCT03316872

Durvalumab + tremelimumab with TACE, RFA, or cryoablation 1/2 Locally advanced/advanced HCC NCT02821754

Pembrolizumab + Y90 1 Locally advanced, high-risk HCC NCT03099564

SBRT then nivolumab ± ipilimumab 1 Unresectable HCC NCT03203304

Ongoing Trials Combining Local and Immune-Based Therapy1

1. http://www.clinicaltrials.gov. Accessed January 14, 2020.

And so, because of the excitement in this area, you can see that there are several trials. I am not going to go through each of these one by one, but you can see that there are trials not only for the intermediate-stage patients but also the locally advanced patients, once again, with the idea that these two can be synergistic in terms of the IO agents potentiating the effect of the locoregional therapy as well as the locoregional therapy maybe potentiating the effect of the systemic therapy for patients with more advanced disease. And I think that these two groups of trials are separate, but I think both are promising and could further advance how we treat these patients in the future.

• Treatment duration = 2.8 months • Of the four radiologically evaluable patients, three had stable disease

on pembrolizumab, and one had progressive disease • All-grade adverse events potentially related to treatment occurred in

50% of patients, including diarrhea (n = 1; grade 3), skin rash (n = 2; grade 2), infusion reaction (n = 1; grade 2), and adrenal insufficiency (n = 1; grade 2)

• Pembrolizumab yielded no synergistic toxicity with TACE, and no DLTs were reported

Phase 1b PETAL Trial: Pembrolizumab Following TACE1,2

1. Pinato DJ et al. ESMO 2019. Abstract 750P. 2. Pinato DJ et al. 2019 American Association for the Study of Liver Diseases Annual Meeting (AASLD 2019). Abstract 327.

Just briefly, I'm going to just highlight a couple to just highlight the potential that we do see in these trials. So this is the phase 1b PETAL trial. So this is pembrolizumab following TACE. Treatment here, you can see 2.8 months. We have only four radically evaluable patients. Three had stable disease on pembrolizumab, and one had progressive disease—very small numbers and very early. This is mainly just looking at safety, and I think that what you can say is that all of the AEs were expected, and I think this really requires further evaluation as we move forward.

Phase 3 EMERALD-1 Trial: TACE + Immunotherapy1

• Primary endpoint: PFS for arm A vs arm C (BICR) • Secondary endpoints: PFS for arm B vs arm C (BICR), OS, and PROs • Other endpoints: safety and PK

Arm C TACE + placebo

Arm A Durvalumab + TACE


Arm B Durvalumab + TACE + bevacizumab

N = 600

R

• Confirmed HCC • Unsuitable for curative therapy

(eg, surgical resection, ablation, transplantation)

• Disease amenable to TACE • No extrahepatic disease • Child–Pugh A to B7 • ECOG PS 0 or 1 • Exclude Vp3 and Vp4

The phase 3 EMERALD-1 study. So this is TACE plus immunotherapy. These patients are unsuitable for curative therapy, either resection, ablation or transplantation, amenable to chemoembolization, and have good liver function, and these patients are being randomized to one of three arms, durvalumab plus TACE, durvalumab plus TACE plus bevacizumab, versus TACE alone.

And the primary outcome is really comparing arm A, which is the durvalumab/TACE arm, to TACE alone, looking at progression-free survival, and then a secondary outcome is looking at arm B progression-free survival versus TACE alone.

• Prognostic model specifically developed for ideal TACE candidates (n = 1,604; treatment naive) – Child–Pugh A-B7 – PS 0 – No VI/mets – No history of tumor rupture – No GIB, ascites, HE, or jaundice

• BCLC B: 74%

6 and 12 Prognostic Score for TACE1

Median OS (mo): 32.9 (95% CI: 30.4-35.4) ≤6: 49.1 (95% CI: 43.7-59.4)

>6 but ≤12: 32.0 (95% CI: 29.9-37.5) >12: 15.8 (95% CI: 14.1-17.7)

1. Wang Q et al. J Hepatology. 2019;70:893-903.

I think the one thing that I would remark on is that these BCLC staging categories, whether this is A, B or C, are nice, because they’re relatively simple. I think many of us in the room, if you asked us, can draw the BCLC staging system, because it is simple. And it gives us a nice tool that we can easily apply to our patients at the bedside.

But I think that it’s clear that these are heterogeneous categories. And so within a B, there are “good Bs” and there are “bad Bs.” And there are Bs that are more like Cs, and there are Bs that are more like As. And so I think that this is a heterogeneous group, and I think there is more and more recognition that I think we shouldn’t be treating all Bs the same.

And so this was, at least in my opinion, a nice prognostic score that broke out Bs and their response to chemoembolization, and it



used essentially the sum of the number of lesions plus the size—so, relatively, once again, simple to do. And when you get this, you have nice cutoffs of 6 and 12.

And so if you have the sum of the lesions plus the size and you’re less than 6, the median survival is 49 months after chemoembolization. The intermediate category is 32 months. And then if your sum of the size plus lesions is greater than 12, you can see the median survival is dramatically lower, at 16 months—actually, much closer to what you see with systemic disease, rather than locoregional disease.

And I bring this up because maybe it shouldn't be one size fits all, and maybe there are some patients with locoregional disease that should be instead treated with systemic therapy. And I think that if there are any interventional radiologists, I expect you to come up here and beat me after.

But I do think that there are some patients with advanced locoregional disease that may be better treated with systemic therapy. Of course, we need further trials to evaluate this.

Lenvatinib Versus TACE1

1. Kudo M et al. AASLD 2019. Abstract 204.

Overall Survival in Patients With Intermediate-Stage HCC Beyond Up-to-7 Criteria and Child–Pugh A Liver Function (Propensity Score Matched)

100

80

60

40

20

0

OS,

%

Time, mo 0 6 12 18 24 30 36 42

Lenvatinib TACE

Median OS, mo (95% CI) 37.9 (23.1-NR) 21.3 (15.7-28.4) HR 0.48 (0.16-0.79); P < .01

So this was an interesting presentation by Kudo and colleagues, taking a look at lenvatinib versus TACE in patients with, once again, larger B tumors. So this is beyond up-to-7—once again, same kind of idea, but using a threshold of 7. Propensity score matched analysis.

And you can see, these patients did better on lenvatinib than they did with TACE. Small numbers, interesting data—I don't think it answers it, and I don't think that this says that we should be doing this routinely in our patients, but I think it is hypothesis generating for how we may move forward in the future.

• Eligibility for downstaging protocol – One lesion >5 cm and ≤8 cm – Two or three lesions each <5 cm and total diameter of all

lesions ≤8 cm – Four or five lesions each <3 cm and total diameter of all lesions ≤8 cm

• Candidates who are eligible and then complete locoregional therapy must be successfully downstaged into T2 (Milan) criteria to receive a MELD exception without need for special case

Downstaging Being Incorporated Into Transplant Criteria1,2

1. https://optn.transplant.hrsa.gov. Accessed January 23, 2020. 2. Abdel-Wahab N et al. J Immunother Cancer. 2019;7:1-10.

As a hepatologist, I want to leave with one last thing. I think that it’s important for us to remember that transplant is the cure for both the cirrhosis as well as the HCC. And I think that we have to recognize that there are some patients with larger tumors who can respond to locoregional therapies and be downstaged into transplant criteria and be transplanted.

And the reason why I bring this up is that, as IO therapies come into this space, I think, at least in my opinion, one of the things that we have to do as we start to think about who should receive IO therapy versus not is to first consider, “Is this patient potentially transplantable?”

And if this patient is potentially transplantable, then I think that it’s important for us to know that and then to discuss that with the transplant team if he or she should be or can be treated with a checkpoint inhibitor.

Downstaging Being Incorporated Into Transplant Criteria1,2

1. https://optn.transplant.hrsa.gov. Accessed January 23, 2020. 2. Abdel-Wahab N et al. J Immunother Cancer. 2019;7:1-10.

Checkpoint Inhibitor

Allograft Rejection, No./Reported

Cases (%)

Median Time to Rejection, d

(range)

Ipilimumab 1/3 (33) 13

Nivolumab 2/4 (50) 12.5 (7-18)

Pembrolizumab 1/3 (33) 7

Ipilimumab followed by pembrolizumab

0/1 (0)

All 4/11 (36) 10 (7-18)

0.0

0.2

0.4

0.6

0.8

1.0

0 5 10 15 20 25 30

OS

Months

P = .03

Median OS, 12 mo (95% Cl, 8-16)

Median OS, 5 mo (95% Cl, 1-9)

Allograft rejection No allograft rejection

And the reason why this is important is because checkpoint inhibitors clearly change your immune environment, and we don’t know how long that lasts. And there have been studies that suggest that you significantly increase the risk of graft rejection, and so you may be withholding a curative treatment from this patient if you put this patient on a checkpoint inhibitor prior to consideration of transplant.

Audience Question 1

Given that IO will likely be moving into the frontline setting, which changes do you foresee happening in the sequencing of treatments?

Dr. Finn: So I think we’ve covered a lot of material tonight, and hopefully you found it useful. But there have been several questions from the audience. Some of them have the same theme, so for the sake of time I’ll try to summarize.

But given the possibility and likelihood that IO comes to the frontline, the IMbrave study, I think, gave us now high-level phase 3 evidence. What will be coming next? Will there be roles for TKIs? Can you use IO after IO? Can you use ramucirumab after bevacizumab?

I think, in general, we do not have high-level evidence to guide us, right? This is progress. Similarly, in the second line, as Katie mentioned, many of the drugs we’ve looked at were done only after prior sorafenib. But we have other drugs now in the front line, lenvatinib. So does that mean we cannot use our clinical judgment?

So personally, I’ll answer this first. If a patient gets perhaps bevacizumab/atezolizumab frontline, I think I would then start to sequence TKIs, first-line TKI, sorafenib or lenvatinib, then second-line, another TKI, regorafenib or cabozantinib. If they have high AFP, then certainly ramucirumab might be an option. Katie, your thoughts? And someone also asked about CTLA-4 after frontline.

Dr. Kelley: Yeah. I think that the best we can do after atezolizumab/bevacizumab is to probably revert to our current TKI sequencing paradigm. However, I think this really shows how incumbent it us upon us as clinical researchers and as a field to start parsing out clinical subanalyses of the data that we are gathering at such a rapid rate, as well as all the biospecimens we’re getting, to really look for clinical and biological biomarkers to identify that subset of patients who tends to be the IO responders.



• Curative therapies such as ablation and resection offerlong-term cure for patients identified at an early stage– High risk of recurrence highlights need for adjuvant

or neoadjuvant therapy• Locoregional therapy options have yielded improved

survival and tolerability over time– Suboptimal long-term survival, particularly

in patients with larger tumor burden, offersopportunity for combinations with systemic therapies

Conclusions

So, in summary, curative therapies such as ablation and resection offer long-term cure but they unfortunately have a high risk of recurrence. There’s a clear need for either adjuvant or neoadjuvant therapy. There are several trials ongoing. If you have one of these high-risk patients in your practice, I highly encourage you to find a clinical trial nearby to enroll those high-risk patients.

Locoregional therapies, likewise, have yielded improved survival and tolerability over time. They’re great therapies. Unfortunately, they’re just not the perfect answer. And I think that those patients who have larger tumor burden offer a perfect opportunity for combinations or maybe even systemic therapy alone for select patients. Thank you.

Dr. Finn: Thank you, Amit, for our ABCs.

Symposium Summary and Audience Q&ARichard S. Finn, MDRobin K. (“Katie”) Kelley, MDAmit Singal, MD, MS


think there’s one answer. I think that the REFLECT trial showed that the overall survival between the two is similar.

Dr. Finn: Yeah, and I guess I would add to that, because I still use a fair amount of sorafenib and because I think OS is the most important endpoint, that there have been data that show that if you do respond to sorafenib or lenvatinib in the frontline setting, perhaps you live longer, and that was from Dr. Kudo’s analysis at the 2019 ASCO GI meeting.

I think tolerability is important, but also response. Some patients have low-burden disease, but they’re advanced. And I think sorafenib is very acceptable.

But then there are patients in whom you want a response, right? They have a large tumor burden. If they don’t respond, they will probably go into liver failure quicker, and therefore maybe lenvatinib might be appropriate there.

Dr. Singal: The one other thing that I want to add is that, you know, you point out the GIDEON data. So we do have a lot of real-world experience with sorafenib. And so for those patients who are extended outside of those trial criteria, particularly those patients with Child B disease, who are often excluded from clinical trials but unfortunately are often the patients who come into our clinical practice, I think those are patients in whom I feel more comfortable using sorafenib while waiting for real-world data in terms of lenvatinib.

Dr. Kelley: I agree.

Audience Question 3

Is there a correlation between HBV status and response to cabozantinib?

Dr. Finn: So, Katie, you alluded to this HBV-cabozantinib correlation. Is there some hypothesis behind that?

Dr. Kelley: Well, I think if we look at the forest plots for the CELESTIAL trial, we do see that the hep B cohort, intriguingly, had a trend towards improved PFS and OS. Whether that’s an artefact or just a function of the population in the prior treatments and other variables, it’s really hard to discern from that level of data.

But MET signaling as a resistance mechanism is one of the

Right now, we’re thinking about 15% of patients will respond to monotherapy based on pembrolizumab and nivolumab data. And with combinations, now with the atezolizumab/bevacizumab, we’re seeing almost 30%. That’s also what we saw with the ipilimumab/nivolumab combination.

We don’t know yet if that’s sort of a ceiling based on the immune profile of the immune-enriched subset of HCC or whether we can do better with different types of combinations. Those are the questions for the next decade of HCC research.

Dr. Finn: Yeah, and the idea that we have these questions reflects the progress, right?

Dr. Kelley: Absolutely.

Dr. Finn: In practice, we need to, I think, still give patients the best chance of getting as much active agents as possible, which goes to Amit’s talk, right? No longer, I think, do you continue to do TACE beyond TACE progression. Patients cannot get TACE until their bilirubin becomes 3 or 4, because we’ll miss the opportunity for the benefit from systemic treatment.

And also, you raise the question of systemic treatment versus TACE. And before, when we did not have so many active drugs, I don’t think anyone would question that, but now that we have drugs with very high response rates, that might be a possibility.

Audience Question 2

How do you choose between lenvatinib and sorafenib in the first-line setting?

Dr. Finn: So there were several questions about choosing a drug in the first-line setting, and given that bevacizuamb and atezolizumab are not yet approved, we have lenvatinib and sorafenib. Amit, how do you determine between the two of them? Is there one best option or optimal option?

Dr. Singal: Yeah. No, I think right now I largely choose on the AE profile. So I think, well, Rich, you did this. You highlighted very nicely the differences between the AE profiles of these two agents in the frontline setting.

We see more hand-foot skin with sorafenib, and I think that you see a little bit more fatigue, anorexia, and then hypertension with lenvatinib. And so I think really what I do is I talk about both options and really choose on AE profile primarily. Overall, I don’t

31Go online to complete the post-test and evaluation for CME/MOC credit




hypotheses. And I think beyond that, we can’t explain it, but it’s a reasonable hypothesis.

Audience Question 4

If a patient gets lenvatinib in the frontline setting, is there an optimal second-line option?

Dr. Finn: So, along those lines, someone gets lenvatinib frontline. Is there an optimal second-line option?

Dr. Kelley: I think right now we don’t really know what the best choice is. And as I mentioned, I think, in patients who have had a long benefit from their first-line TKI, I tend to use a second-line TKI. And if they receive lenvatinib first line, I tend to use cabozantinib second line simply because it was more permissive in its eligibility, allowing one or two prior lines of therapy, and it didn’t have to be simply sorafenib immediately beforehand. So that is the closest proxy to real world in this context of lenvatinib first.

In patients who did very poorly on their first-line TKI, with poor tolerability and rapid progression, now that we have IO therapies, and because I want patients to have a chance to have an IO response before discontinuing all anticancer treatment, I tend to use IO therapies if they’re eligible in patients who did really poorly on a first-line TKI. That’s where I’m curious to hear your practices, as well, because there is no right answer yet.

Dr. Finn: Yeah, there is, and for the sake of time, we’ll go with your answer.

Audience Question 5

Do the results of the phase 3 KEYNOTE-240 study of pembrolizumab change your views on the use of IO agents as monotherapy in the second line and beyond?

Dr. Finn: But I will add, so the negative phase 3 nivolumab frontline data—and I'm biased, but the relatively negative or positive, half-full/half-empty [KEYNOTE-]240 study with

pembrolizumab—does that change your use of these agents in the second line or third line?

Dr. Kelley: For me, no, because I think that we have a preponderance of data reassuring us that we have a subset of patients with a durable, meaningful deep response endpoint that has been proven across drugs and across populations and is reproducible.

That said, where monotherapy fits as a first-, second- or later-line agent really requires more mature data, whether it’s from CheckMate -459 or from further data sets with pembrolizumab.

Dr. Singal: And I think, particularly given the fact that the KEYNOTE-240 is more of a statistical anomaly rather than a lack of effect…

Dr. Kelley: Right.

Dr. Singal: So I think that, if anything, we can enforce the fact that you can continue to use this in the second-line setting, in my opinion.

Audience Question 6

Can you explain the differences in responses to lenvatinib between the REFLECT trial and the combination study with pembrolizumab?

Dr. Finn: So, for the sake of time, I’ll try to wrap up. So, for lenvatinib single agent, the ORR was 40%, and using the two agents, lenvatinib/pembrolizumab, it was 36%. I think it’s comparing mRECIST and RECIST.

So the 40% from REFLECT was modified RECIST by independent review. If we look at RECIST in REFLECT, it was about 18%. If we look at RECIST with the combination of lenvatinib/pembrolizumab, that’s where we get 36%. So it is almost doubled.



Audience Question 7

Did IMbrave150 include patients with macrovascular invasion?

Dr. Finn: Did IMbrave150 include patients with macrovascular invasion? Yes, in fact, they had a very high-risk population of patients who, compared with those in other studies, had a higher incidence of main portal vein invasion or vascular invasion.

And I thank you all for your attendance, and I hope you found it useful.

Narrator: This activity has been jointly provided by Medical Learning Institute, Inc. and PVI, PeerView Institute for Medical Education.

PeerView .com /BXY900The Oncology Topics You Want

From the Experts You Trust

CME/MOC


Sign up for email alerts on new clinical advances and educational activities in your specialty: PeerView.com/signup

Copyright © 2000-2020, PeerView

This CME/MOC activity is jointly provided by Medical Learning Institute, Inc. and PVI, PeerView Institute for Medical Education.

This educational activity is supported by medical education grants from Exelixis, Inc., Genentech, and Merck & Co., Inc.

https://www.PeerView.com/BXY900

cme/moc developing a therapeutic game plan for the

Documents