0000-0001-7413-5002...2020/06/23 · 2015 to 15april 2019. 60 patients received cicb consisting of...

1

Checkpoint-blocker induced autoimmunity is associated with pretreatment T cell expression profiles and

favourable outcome in melanoma

Short title: Relationship between immune-related adverse events and clinical outcomes

W. Ye1*, A Olsson-Brown2,3*, R. A. Watson4,8, V. T. F. Cheung5,9, R. D. Morgan6, I. Nassiri4,8, R. Cooper4,8,

C.A. Taylor4,8, O. Brain5, R. N. Matin4,7, N. Coupe4, M. R. Middleton8,9, M. Coles9,10, J.J. Sacco2,3, M. J. Payne4,

B. P. Fairfax4,8,9

* Equal Contributions

Affiliations:

1. Oxford University Clinical Academic Graduate School, University of Oxford, Oxford, UK

2. The Clatterbridge Cancer Centre, Wirral, UK

3. University of Liverpool, Liverpool, UK

4. Department of Oncology, Churchill Hospital, Oxford, UK

5. Translational Gastroenterology Unit, John Radcliffe Hospital, University of Oxford, Oxford, UK

6. Department of Oncology, The Christie NHS Foundation Trust, Manchester, UK

7. Department of Dermatology, Churchill Hospital, Oxford, UK

8. The MRC Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, UK

9. NIHR Oxford Biomedical Research Centre, Oxford University Hospitals NHS Foundation Trust, John

Radcliffe Hospital, Oxford, UK

10. Kennedy Institute of Rheumatology, NDORMS, University of Oxford, Oxford, UK

Corresponding author:

Dr Benjamin P Fairfax

The MRC Weatherall Institute of Molecular Medicine,

University of Oxford,

John Radcliffe Hospital,

Headley Way, Oxford

OX3 9DS, UK

Tel: 01865 222310

Email: [email protected]

ORCID: 0000-0001-7413-5002

Article type: Original article

Key words: melanoma, pembrolizumab, ipilimumab, nivolumab, immune-related adverse events, survival

All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprintthis version posted June 24, 2020. ; https://doi.org/10.1101/2020.06.23.20138594doi: medRxiv preprint

NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.

https://doi.org/10.1101/2020.06.23.20138594

2

Key message:

Immune-related adverse events (irAEs) commonly occur in patients with metastatic melanoma treated with

immune checkpoint blockade (ICB) therapy. In real world setting we find development of early irAEs post-ICB

treatment is associated with survival benefit, indicative of a shared mechanism with anti-tumour efficacy. CD8+

T cells from patients who develop irAE show increased receptor diversity, and pre-treatment samples from

patients who develop irAE post single-agent anti-PD1 show over-expression of inflammatory pathways,

indicating baseline immune state can determine irAE development.

1 Abstract

1.1 Background

Immune checkpoint blockers (ICBs) activate CD8+ T cells to elicit anti-cancer activity but frequently lead to

immune-related adverse events (irAEs). The relationship of irAE with baseline parameters and clinical outcome

is unclear. We investigated associations between irAE development, CD8+ T cell receptor diversity and

expression and clinical outcome in a non-trial setting.

1.2 Methods

Patients ≥18 years old with metastatic melanoma (MM) receiving combination ICB (ipilimumab plus nivolumab

– cICB, n=60) or single-agent ICB (nivolumab/pembrolizumab – sICB, n=78) were prospectively recruited. We

retrospectively evaluated the impact of irAEs on survival. This analysis was repeated in an independent cohort

of MM patients treated at a separate institution (n=210, cICB:74, sICB:136). We performed RNA sequencing of

CD8+ T cells isolated from patients prior to treatment, analysing T cell receptor clonality differential transcript

expression according to irAE development.

1.3 Results

48.6% of patients experienced treatment-related irAEs within the first 5 cycles of treatment. Development of

irAE prior to the 5th cycle of ICB was associated with longer progression-free and overall survival (PFS, OS) in

the primary cohort (log-rank test, PFS: P=0.00034; OS: P<0.0001), replicated in the secondary cohort (OS:

P=0.00064). Across cohorts median survival for those patients not experiencing irAE was 14.4 (95% CI:9.6-

19.5) months vs not-reached (95% CI:28.9 - Inf), P=3.0x10-7. Pre-treatment performance status and neutrophil

count, but not BMI, were additional predictors of clinical outcome. Analysis of CD8+ T cells from 128 patients



https://doi.org/10.1101/2020.06.23.20138594

3

demonstrated irAE development was associated with increased T cell receptor diversity post-treatment

(P=4.3x10-5). Development of irAE in sICB recipients was additionally associated with baseline differential

expression of 224 transcripts (FDR<0.1), enriched in pro-inflammatory pathway genes including CYP4F3 and

PTGS2.

1.4 Conclusions

Early irAE development post-ICB is strongly associated with favourable survival in MM and increased diversity

of peripheral CD8+ T cell receptors after treatment. irAE post-sICB is associated with pre-treatment

upregulation of inflammatory pathways, indicating irAE development may reflect baseline immune activation

states.

2 Introduction

The introduction of immune checkpoint blockade (ICB) therapy into clinical practice has transformed the

outlook for metastatic melanoma (MM) patients. Current ICB standard of care consists of either monotherapy

with anti-PD1 agents (pembrolizumab or nivolumab, sICB), or combined anti-CTLA-4/ anti-PD1 (ipilimumab

and nivolumab, cICB). Nivolumab and pembrolizumab block T cell programmed death 1 (PD1) receptor,

preventing ligation by the ligands PDL1 and PDL2 which are frequently over-expressed in tumour and stroma

[1]. Ipilimumab is an anti-CTLA-4 monoclonal antibody that enhances effector T cell activity by antagonising

the homeostatic function of CTLA-4 which is induced upon T-cell antigen presentation [2]. Whereas untreated

MM is associated with a 5-year overall survival (OS) of approximately 10% [3], the 5-year OS following anti-

PD1 monotherapy in treatment-naive patients is 30 – 40% [1] and combination ICB was associated with a

median OS exceeding 5 years in the Checkmate 067 study, with a subset of patients potentially cured [4].

A key concern with ICB is the high incidence of immune-related adverse events (irAEs). Notably, compared to

patients receiving sICB, patients receiving cICB show increased incidence and severity of irAEs [5]. IrAEs can

be challenging to manage, requiring treatment interruption or discontinuation and systemic immunosuppression.

At present, it remains controversial as to whether the development of irAEs is associated with survival. Several

retrospective studies have observed an association between the development of irAEs and improved treatment

response and/or survival in MM patients treated with ICB, suggesting that reduced tolerance to self-antigens and

reduced tolerance to tumour antigens are closely linked [5–11]. This is not consistently observed however

[12,13], which may be due to confounders including differences between clinical trial and standard clinical



https://doi.org/10.1101/2020.06.23.20138594

4

populations, divergences from standard dosing and frequent concurrent exposure to other non-standard of care

agents such as vaccines.

Data from the clinical setting regarding the prognostic implications of irAE development remain limited. The

treatment pathway within the UK National Health Service is standardised, with patients stratified to cICB or

sICB depending on clinical features and patient preferences. Patients whose disease progresses on anti-PD1

sICB have the option of targeted therapy if they have activating BRAF mutations, or alternatively second-line

ipilimumab; whilst patients with disease progression on cICB have no-further standard of care options available.

With this in mind, we have assayed the incidence and severity of irAEs in MM patients treated with either cICB,

or anti-PD1 sICB across two prospectively recruited cohorts from tertiary UK centres, to explore the question:

does the development of irAEs early in ICB therapy impact clinical outcome? We have subsequently explored

CD8+ T cell RNA sequencing from a subset of the cohort to investigate the relationship between gene

expression and irAE development.

3 Methods

3.1 Patients

Eligible patients were ≥ 18 years with a confirmed diagnosis of MM, and had received at least one cycle of ICB

therapy. 138 Patients were recruited prospectively through the Oxford Radcliffe Biobank, from 23 November

2015 to 15 April 2019. 60 Patients received cICB consisting of ipilimumab 3 mg/kg plus nivolumab 1 mg/kg

every 3 weeks for a maximum of 4 treatment cycles, followed by maintenance nivolumab 240 mg fortnightly or

480 mg monthly. 78 Patients receiving sICB therapy included either nivolumab 480 mg monthly, or

pembrolizumab 2 mg/kg every 3 weeks. In the replication cohort 74 patients were treated at the Clatterbridge

Cancer Centre in Liverpool with cICB and 136 with sICB between 1 January 2016 and 7 January 2019. Patients

receiving ICB therapy were treated until unacceptable irAE, progressive disease, death, or patient withdrawal.

3.2 Study design

Patients provided written consent to participation within the Oxford Radcliffe Biobank (09/H0606/5+5) and this

research was approved via applications: OCHRe 16/A019, 18/A064. For the Clatterbridge dataset, research was

approved via the HYST study (12/NW/0525) and local audit approval (17-18/40). Data on patient demographic

and clinical characteristics, in addition to the type, severity, date of onset and management of irAEs, were



https://doi.org/10.1101/2020.06.23.20138594

5

collected retrospectively using electronic medical records. Treatment decisions were determined by the treating

clinician. IrAEs were reported according to the National Cancer Institute Common Terminology Criteria for

Adverse Events (CTCAE) version 4.03 with pneumonitis being diagnosed via CT. We collected efficacy data

consisting of radiological response as defined by the Response Evaluation Criteria in Solid Tumours (RECIST)

version 1.1 [15], OS and PFS.

3.3 Outcomes

We evaluated irAE characteristics, predictors of irAE development, and the OS and PFS in patients who

developed early irAEs compared to those who did not. Early irAEs were defined as those from treatment

initiation to completion of the 4th cycle of treatment, i.e. occurring prior to cycle 5. We chose this time point as

cICB therapy consists of a maximum of 4 cycles of ipilimumab plus nivolumab, prior to commencement of

maintenance nivolumab monotherapy. OS was defined as the time from first ICB dose to death from any cause.

PFS was defined as the time from first ICB dose to disease progression, as determined by serial cross-sectional

imaging, or death.

3.4 Statistical analysis

Baseline characteristics were analysed using descriptive statistics. Categorical variables were summarised using

frequencies and percentages, and continuous variables using medians and ranges. Given the time-dependent

nature of developing irAEs and the susceptibility to guarantee-time bias [15], we also performed a 12 week

landmark analysis. Only patients who are alive or have not progressed at 12 weeks are included in the OS and

PFS landmark analysis respectively, and patients were grouped according to whether they experienced irAEs

prior to 12 weeks. OS and PFS was estimated using Kaplan-Meier analysis, and the log-rank test was used to

determine the statistical significance between the curves. Logistic regression was used to determine predictors

of developing irAEs. Univariable and multivariable Cox proportional hazards models were used to investigate

the association between prognostic factors and survival. P < 0.05 was considered statistically significant. All

analyses were performed using the survminer [17] and survival [18] packages in R, version 3.5.1.

3.5 Sample collection

Patients provided written informed consent to donate samples for analysis to the Oxford Radcliffe Biobank

(Oxford Centre for Histopathology Research ethical approval nos. 16/A019 and 18/A064); 30–50�ml blood



https://doi.org/10.1101/2020.06.23.20138594

6

was collected into EDTA tubes (BD vacutainer system) taken immediately pre-treatment. Peripheral blood

mononuclear cells were obtained by density centrifugation (Ficoll Paque). CD8+ cell isolation was carried out

by positive selection (Miltenyi) according to the manufacturer’s instructions, with all steps performed either at

4°C or on ice. Post-selection cells were spun down and resuspended in 350μl of RLTplus buffer with 1% beta-

mercaptoethanol or DTT, and transferred to 2-ml tubes. Samples were stored at −80°C for batched RNA

extraction. Homogenization of the sample was carried out using the QIAshredder (Qiagen). The AllPrep

DNA/RNA/miRNA kit (Qiagen) was used for RNA extraction. DNase�I was used during the extraction

protocol to minimize DNA contamination. RNA was eluted into 35�μl of RNase- free water. The amount of

RNA present was quantified by Qubit analysis, and RNA samples stored at − 80°C until ready for sequencing.

3.6 Expression analysis

Poly(A) RNA was sequenced on Illumina HiSeq-4000 (75bp paired-end reads) and Illumina Novaseq machines

(150bp paired-end reads) both at the Oxford Genome Centre, Wellcome Centre for Human Genetics. Reads

were aligned to GCRh38/hg38 using HISAT2, and High-mapping quality reads were selected based on MAPQ

score using bamtools. Marking and removal of duplicate reads were performed using picard (v.1.105), and

samtools was used to pass through the mapped reads and calculate statistics. 128 high-quality transcriptomes

were used for expression analysis of pre-treatment samples. Read count information was generated using HTSeq

and DESeq2 [19][20]was used for differential expression analysis, comparing individuals who developed irAE

within the initial 5 cycles of treatment versus unaffected. We controlled for sequencer, age and sex in the

analysis. Only transcripts with mean of >10 reads were analyzed, using the binomial Wald test with

750�iterations after correcting for size factors and dispersion.

3.7 T cell receptor diversity

MiXCR[21] was used to map reads on reference sequences of V, D and J genes, and to quantitate TCR

clonotypes from mapped reads using complementarity-determining region 3 (CDR3) gene regions for RNA

samples from pre and post first cycle ICB (n=106 paired samples). The non-default partial alignments option

(OallowPartialAlignments = true) was applied to preserve partial alignments for the assembly step. Three

iterations of read assembly were performed using the assemblePartial setting. Shannon diversity was calculated

within the Vegan R package[22].



https://doi.org/10.1101/2020.06.23.20138594

7

4 Results

4.1 Patient characteristics

138 MM patients treated with ICB therapy were prospectively recruited (Table 1). Of 78 patients receiving sICB

therapy, 69 received pembrolizumab and 9 received nivolumab. 6 patients had a history of prior autoimmune

disease (1 type 1 diabetes, 2 inflammatory bowel disease, 2 thyroid abnormalities, 1 Sjogren’s syndrome). The

median number of cycles received per patient was 4 for cICB therapy, and 8 for sICB therapy. The median

follow up duration was 12.1 (0.3 – 42.7) months.

4.2 Immune related adverse events

Any grade treatment-related irAEs were reported in 89 (64.5%) patients, of which 74 (53.6%) experienced an

irAE prior to cycle 5. Grade 3 or 4 irAEs were reported in 45 (32.6%) patients (Table 2). There were no drug-

related deaths. Cutaneous irAEs, colitis and hepatitis occurred early post ICB initiation (median 37, 34 and 49

days respectively). In contrast, gastritis and pneumonitis are late complications, with median time to onset of

378 and 386 days (Supplementary Figure 1a, Table 2).

Compared to patients treated with sICB, patients receiving cICB had over two-fold increase in the frequency of

any grade irAEs, and more than a four-fold increase in the frequency of Grade 3 or 4 irAEs (Table 2). Multi-

organ system irAEs were more common with cICB therapy, with 24 (40%) patients experiencing irAEs

affecting 3 or more organs, compared to 3 (4%) patients treated with sICB (Supplementary Figure 1b). Systemic

steroids were required in 35/40 (88%) cICB and 10/29 (34%) sICB patients who experienced irAEs. 11 patients

had steroid-refractory disease, requiring mycophenolate (4 hepatitis, 1 bullous pemphigoid), infliximab (3

colitis, 1 gastritis), methotrexate (1 arthritis) and hydroxychloroquine (1 arthritis). ICB discontinuation occurred

in 30/69 (43%) patients who experienced irAEs (23/40 cICB, 7/29 sICB). Events included colitis (N = 7),

hepatitis (N = 4), cutaneous manifestations (N = 3) and hypophysitis (N = 3). Neurological irAEs, pneumonitis,

autoimmune haemolytic anaemia, diabetes, nephritis and myositis led to ICB discontinuation in 1-2 patients

each. In a further 14 patients, ICB therapy was interrupted due to irAEs (11 cICB, 3 sICB). Prior to treatment

interruption, the median number of cycles received was 2 and 10 for patients treated with cICB and sICB,

respectively.



https://doi.org/10.1101/2020.06.23.20138594

8

We used logistic regression to test putative risk factors for irAE development prior to cycle 5, evaluating age,

sex, baseline BMI, medical history of autoimmune disease, and type of treatment. We found the only significant

clinical predictor of increased likelihood of developing early irAEs being treatment with cICB (OR=19.1, 95%

CI 7.1-59.1, P =3.8x10-8, Supplementary Figure 1c).

4.3 Oncological outcomes

Among 138 patients, 58 (42%) experienced a complete or partial response to ICB therapy at the first

radiological assessment (3 month CT), whereas 26 patients (18.8%) had stable disease, and 43 (31.2%) had

progressive disease. For the remaining 11 (8%) patients, no sequential cross-sectional imaging was available,

however 9 (6.5%) patients had clear clinical progression.

The median OS across the cohort was 28.9 months (95% CI 18.7 months – Inf) and the median PFS was 9.0

months (95% CI 5.8- 20.5 months ). At 1 year, the OS and PFS rates were 72% (95% CI 65 – 81) and 48%

(95% CI 40 – 58), respectively. At 2 years, the OS and PFS rates were 56% (95% CI 47 – 67) and 37% (95% CI

28.9 – 48.2), respectively.

Comparing patients who develop an early irAE prior to the 5th cycle of treatment to those who do not, early

irAE development was associated with significantly longer OS and PFS (Figure 1a,b, OS log-rank P < 0.0001,

PFS log-rank P = 0.00034). This observation remained significant when stratifying patients according to

treatment received (Figure 1c,d, Supplementary Figure 2a,b). Similarly this remained significant when analysis

was confined to patients suffering Grade 3 or 4 irAE only (Supplementary Figure 2c,d). To adjust for guarantee-

time bias, we performed a landmark analysis including only patients who are alive (N = 127) or who had not

progressed (N = 98) at 12 weeks. We found that the development of an irAE prior to week 12 was associated

with significantly improved OS (Supplementary Figure 2e, log-rank P =0.0013), however this did not reach

significance for PFS (Supplementary Figure 2f, log-rank P = 0.099)

4.4 Independent replication

To explore whether our findings were reproducible in a different institution we repeated the analysis in a

separately collected cohort of MM patients receiving ICB based at The Clatterbridge Cancer Centre, Liverpool,

consisting of 136 recipients of sICB and 74 recipients of cICB. The cohort had a similar demographic make-up



https://doi.org/10.1101/2020.06.23.20138594

9

to the Oxford cohort, although the patients who received cICB tended to be older (median age 65 vs. 59, Table

3), whereas the recipients of sICB were younger (median age 68 vs. 74, Table 3). Across this cohort we

replicated the observation that irAE during the first 4 cycles of immunotherapy was significantly associated with

prolonged overall survival (median 13 (95% CI:8-23) months vs. not reached, P =0.00064, Figure 2a). When we

assessed each treatment type independently we observed a suggestive but non-significant benefit of irAE within

the sICB cohort (median 15 (9-23) vs. 26 (16-Inf) months, P=0.1). Conversely, this observation remained

robust within the cICB cohort (median 5 (3-Inf) months vs not reached (20-Inf), P=0.00023). When we

combined the data from both cohorts early irAE were found to be highly significantly associated with overall

survival time (median 14.4 (95% CI:9.6-19.5) months vs not-reached (95% CI:28.9 - Inf), P=3.0x10-7), Figure

2b), and this remained the case for sICB (median 15.2 (12-23) months vs not-reached (18-Inf), P=0.0028) or

cICB treatments (median 5 (3-Inf) months vs 31 (28.9 – Inf) months, P=3.7x10-7).

4.5 Other variables associated with outcome

We used univariable Cox proportional hazard models to explore the effect of different parameters on OS (Table

4). Development of an irAE prior to cycle 5 and baseline albumin levels were positively associated with

significantly improved OS. In contrast, non-cutaneous melanoma subtype, raised performance status, neutrophil

count, monocyte count and lactate dehydrogenase levels at baseline were negatively associated with OS (Table

4). Using retrospective trial data from immunotherapy and targeted agents it has recently been shown that a

raised BMI in females is associated with a superior clinical outcome[23]. We explored the association between

BMI and outcome in both the Oxford and Liverpool cohorts but notably did not see an effect in either

univariable or multivariable analyses with clinical outcome, with a trend in both datasets towards raised BMI at

treatment start being negatively associated with outcome.

Using multivariable models to explore the interaction between these different factors, we found that

development of an irAE prior to cycle 5, monocyte count, performance status and neutrophil count were

nominally associated with OS (Table 4), although irAE remained the only significant predictor after correcting

for multiple testing (Figure 3a, HR 0.09, 95% CI: 0.03 – 0.27, (Benjamini Hochberg corrected) FDR=0.0002).

We had access to more limited clinical data from the Liverpool cohort, but in the subset of individuals with

neutrophil counts available we replicated the negative association with baseline neutrophil count and clinical

outcome (Table 4).



https://doi.org/10.1101/2020.06.23.20138594

10

Performing multivariable Cox proportional hazard models on the combined datasets from both institutions

(n=348) we again observed a protective effect of irAE (HR 0.35, 95% CI:0.24-0.53, P=6.8 x10-7) and a weaker

effect of age (reduction in HR 0.98, 95% CI:0.97-0.99, P=0.02, Figure 3b). Across both sites we had data

including baseline lymphocyte and neutrophil counts for 272 individuals. Analysis of these combined data

demonstrated that baseline neutrophil count (HR 1.24 per unit increase, 95% CI:1.16-1.32, FDR=1.3x10-9) and

development of irAE prior to cycle 5 (HR 0.36, 95% CI:0.23-0.57, FDR=5.4x10-5) were associated with clinical

outcome (Figure 3c).

4.6 Association of irAE development with CD8+ T Cell Receptor diversity and baseline gene expression

Identification of markers predictive of irAE is of high interest to immuno-oncology. There is growing evidence

that variation in propensity to irAE might be ascertainable by analysis of blood samples[24]. We have

previously shown that analysis of peripheral CD8+ T cell transcriptomics can provide insights into treatment

outcome[25]. Given the association between clinical outcome and irAE development we sought to further

investigate the relationship between irAE development and features of peripheral CD8+ T cells. We analysed T

Cell Receptor (TCR) diversity from CD8+ T cells from pre (day 0) and post-treatment (day 21) samples from

128 of the patients in the Oxford cohort. Analysis of the Shannon Diversity index, a marker of abundance and

evenness, of TCR demonstrated that patients who proceeded to develop irAE within the first five cycles of

treatment had significantly greater CD8+ T cell TCR diversity on day 21 post treatment (Figure 4a). Notably we

found a strong inverse-relationship between age and diversity, and this was reflected in a difference between the

pre-treatment diversity in recipients of sICB and cICB where recipients in the Oxford cohort differed in age. To

further explore the relationship between irAE development and CD8+ T cell TCR diversity we fitted a linear

model, taking into account treatment type, patient age, day 0 and day 21 CD8+ TCR diversity, sex and baseline

monocyte counts. This demonstrated that, in addition to treatment type, TCR diversity on day 21 was a key

predictor of irAE (P=0.038, Figure 4b).

4.7 Association with CD8+ T cell baseline gene expression

To explore the relevance of baseline transcriptomics to irAE development we analysed gene expression in CD8+

T cells isolated from ICB recipients immediately prior to treatment. Whilst we saw limited significant variation

in pre-treatment samples according to development of irAE post cICB (data not shown), analysis of CD8+ T cell



https://doi.org/10.1101/2020.06.23.20138594

11

expression data of sICB recipients (n=71), of whom 24% (17/71) developed irAE prior to the fifth cycle of

treatment, identified 224 transcripts (FDR<0.1, Supplementary Data 1) differentially expressed according to

subsequent development of an irAE (Figure 4c). Notable upregulated genes included several involved in

arachidonic acid metabolism and eicosanoid synthesis including CYP4F3, encoding Leukotriene-B(4) omega-

hydroxylase 2 (FDR 0.014), and PTGS2, encoding prostaglandin-endoperoxide synthase 2 (FDR 0.021).

Pathway analysis of transcripts upregulated according to subsequent irAE development showed 58 pathways

where there was enrichment of genes upregulated in patients who developed irAE (FDR<0.05, Supplementary

Data 2). Notably the greatest fold enrichment was observed in genes involved in TNF synthesis (GO:0042535,

24.7 fold-change, P=2.5x10-7), a key cytokine putatively involved in irAE development (Figure 4d,

Supplementary Data 2). These data demonstrate a baseline tendency to inflammation before treatment in

individuals who develop irAE post sICB, suggesting that the pre-treatment immune state sensitizes to irAE

development.

5 Discussion

Whilst ICB therapy constitutes the main treatment for MM, it frequently results in irAE development.

Consistent with the literature [5], we find a markedly higher incidence of irAEs in patients treated with cICB

compared to sICB therapy. Moreover, irAEs secondary to cICB are typically more severe, with a greater

proportion of patients experiencing multi-organ involvement. Neither sex nor prior autoimmune history

influenced the development of early irAEs in our cohort, whereas in the general population, autoimmune

diseases have a higher prevalence in females and in those with a history of autoimmunity [26]. Although we

were underpowered to exclude absence of effect of prior autoimmune history, the lack of sex effect suggests

classical risk factors for autoimmune disease may be of less relevance to the development of irAEs. Further

studies are important to ascertain whether this observation remains robust. We could not replicate the reported

association between pre-treatment BMI and clinical outcome [23] in either the Oxford or Liverpool datasets in

either sex and there was no association between BMI and tendency to irAEs. Our data is notable in that the BMI

was collected prospectively, and it did not incorporate targeted agents or immunotherapy and chemotherapy

regimens together. Whilst further prospective data is required, these results suggest the putative link between

BMI and favourable outcome in ICB is not clear-cut and prospective studies analysing this would be helpful.



https://doi.org/10.1101/2020.06.23.20138594

12

Awareness of the different irAEs and their clinical characteristics is paramount to their successful management

in the clinical setting. As previously described, we observe that distinct irAEs have a different median time-to-

onset post ICB therapy initiation, which may reflect diverging mechanisms of development. Cutaneous irAEs,

colitis and hepatitis typically occur early, whilst late complications include gastritis and pneumonitis. The

timescales within our cohort broadly corroborate those reported in previous studies [9,19], except pneumonitis

which occurs at a median time of 378 days within our cohort compared to 62 days in a previous study [10].

Despite the diversity of irAEs that occur with ICB therapy, most can be managed symptomatically or with

corticosteroids, and only a small number require other systemic therapies.

Whilst association between development of irAEs and oncological response to ICBs has been described

previously, it has been predominantly in the trial setting [6,7]. Previous studies have demonstrated vitiligo

development to be linked to an objective response to ICB therapy [11,12], although the association between

vitiligo and good clinical outcome in melanoma is known outside ICB treatment. A link between irAE and

survival benefit in ICB therapy has also been reported in non-small cell lung cancer and therefore may not be

specific to melanoma [28]. However, not all available evidence support the link between irAE and improved

survival with retrospective studies of single agent Ipilimumab response and nivolumab responses failing to

observe an association [9,12,13].

With data from a standard-of-care non-trial dataset, we clearly demonstrate that development of irAE of any

grade is associated with substantially improved OS, across univariable and multivariable Cox models. This

observation remains significant in the 12-week landmark analysis, ruling out a guarantee-time bias where

patients died before being able to develop irAE. In contrast, the association between early irAE and improved

PFS was significant across univariable and multivariable analyses, but not in the 12-week landmark analysis.

Importantly, we were able to replicate these observations in an independent cohort, although here, upon analysis

of individual treatments the effect only remained significant in recipients of cICB, although a strong trend to this

effect in the sICB group existed with more marked variation in response. Notably, there were significant

differences between the Oxford and Liverpool cohorts, with Liverpool sICB recipients tending to be younger

and Liverpool cICB recipients significantly older on average than those in Oxford. When we combined the

datasets in multivariable analyses however, development of irAE was highly associated with outcome, even

when taking treatment type into account. Similarly, neutrophil count at start of treatment was strongly



https://doi.org/10.1101/2020.06.23.20138594

13

negatively associated with outcomes, indicating baseline myeloid proliferation leading to neutrophilia is a clear

poor prognostic marker of response.

The incorporation of the TCR diversity data and transcriptomic analysis from a reasonably powered dataset

(n=128 individuals) demonstrated that higher diversity of TCR post treatment was found amongst those who

developed irAE. Notably, TCR diversity is associated with age, as was the type of treatment patients were

placed on in Oxford. However, when accounting for treatment type, age, sex, monocyte count, and baseline

diversity we still found that development of irAE remained associated with day 21 TCR diversity. A greater

TCR diversity equates to increased ability to recognise diverse antigens, and thus an association with propensity

to irAEs may be anticipated.

Whereas we did not note a strong association between baseline CD8+ cell gene expression and irAE

development in recipients of cICB, across sICB recipients who developed irAEs, increased expression of

inflammatory pathway mediators was noted in the pre-treated state. cICB leads to markedly greater changes in

gene expression in CD8+ T cells than sICB[25] and we postulate that with such broad activation, baseline

variation in expression is of less importance compared to the larger effect of treatment. Within the sICB

recipients who develop irAE we suspect that the enrichment in neutrophil and platelet activation pathways in

CD8+ T cells at baseline may reflect increased intercellular adherence in these samples with platelet adherence

to immunoprecipitated T cells, a common finding in activated T cells from chronic viral infection[29], which is

postulated to aid CD8+ recruitment to sites of injury. Increased cellular adherence indicates predisposal to

activation, and this is supported by enrichment of numerous inflammatory mediator pathways including

hydrogen peroxide catabolic process (GO:0042744), positive regulation of tumour necrosis factor biosynthetic

process (GO:0042535) and positive regulation of nitric oxide biosynthetic process (GO:0045429). Thus our data

suggests the propensity to develop irAE to sICB is in part due to the baseline CD8+ T cell activation, which can

be predicted and may reflect pre-treatment anti-tumour responses.

It can be inferred that because irAE were managed as per standard, frequently with steroids or other

immunosuppressants, these treatments are unlikely to adversely affect prognosis. Conversely, although the

divergence in likelihood of irAE induction between anti-CTLA-4 and anti-PD1 agents shows that the

relationship between anti-tumour effects and irAE development is not linear, our results also suggest that it may



https://doi.org/10.1101/2020.06.23.20138594

14

be difficult to completely separate efficacy from irAE propensity in novel agents. The comparative strengths of

our study are that, unlike data from trials, the results reflect real-world scenarios, with particular relevance to the

UK healthcare setting and our observations independently replicate in a separate tertiary centre. Our ability to

combine clinical observations with prospectively collected transcriptomic data identified differences in the TCR

diversity after treatment according to irAE development, as well as baseline predisposition to activation in sICB

recipients who develop irAEs. The limitations of this study include the retrospective collection of clinical data

and the relatively small sample size in the primary cohort. Larger transcriptomic series involving other cell

subsets will be vital for future studies in teasing out the relationship between clinical response and irAE

development. A positive association between increased tumour mutational burden (TMB) and response to ICB

is well recognised[30]. Here we show that irAE similarly are positively associated with outcome in MM.

Although this likely represents a separate association to that of response and TMB; given the increased baseline

inflammatory response in sICB irAE sufferers, it is tempting to speculate that TMB may relate to irAE, with

higher TMB eliciting more neo-antigens and potentially off target effects – a hypothesis that warrants further

exploration.

In conclusion, in our clinical practice, the development of early irAEs in MM patients treated with ICBs is

associated with a survival benefit and, in sICB recipients, development of irAE is associated with baseline

immune state. Whether specific irAEs have more favourable prognostic associations than others remains

unknown and will require larger datasets to ascertain. These findings highlight the crucial importance of

investigating the biological mechanisms underlying irAE development, to determine whether their presence

could be a robust indicator of response to treatment. Collectively, this could aid early identification of

individuals likely to benefit from ICB therapy, and enable the application of stratified medicine.

Data Availability Statement

All sequencing data will be made freely available to organizations and researchers to conduct research in

accordance with the UK Policy Framework for Health and Social Care Research via a data access agreement.

Sequence data will be deposited at the European Genome–phenome Archive, which is hosted by the European

Bioinformatics Institute and the Centre for Genomic Regulation under accession no. EGAS00001004081.

Patient anonymized irAE data will be shared on reasonable request.



https://doi.org/10.1101/2020.06.23.20138594

15

Acknowledgements

With thanks to all patients who have generously contributed samples and clinical data to the Oxford Radcliffe

Biobank for this study. We are also very grateful to the staff of the Day Treatment Unit at the Oxford Cancer

Centre and Brodey Cancer Centre at The Horton Hospital for their invaluable assistance.

Funding

The work was supported by the National Institute for Health Research (NIHR) Oxford Biomedical Research

Centre (BRC). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR

or the Department of Health. WY is an Academic Foundation Programme Trainee, RAW is an NIHR Academic

Clinical Fellow and recipient of a CRUK predoctoral award (C64881/A26189); BPF is funded by a Wellcome

Intermediate Clinical Fellowship (201488/Z/16/Z). ACOB is a MRC Academic Clinical Fellow (Award Ref.

MR/N025989/1). VTFC is funded by grants from the Norman Collisson Foundation and Oxfordshire Health

Services Research Committee (OHSRC) part of Oxford Hospitals Charity

Disclosures

All authors have declared no conflicts of interest.

References

[1] O. Hamid et al., ‘Five-year survival outcomes for patients with advanced melanoma treated with

pembrolizumab in KEYNOTE-001’, Ann. Oncol., doi: 10.1093/annonc/mdz011.

[2] D. Schadendorf et al., ‘Pooled Analysis of Long-Term Survival Data From Phase II and Phase III Trials

of Ipilimumab in Unresectable or Metastatic Melanoma’, J. Clin. Oncol., vol. 33, no. 17, pp. 1889–1894,

Feb. 2015, doi: 10.1200/JCO.2014.56.2736.

[3] C. Garbe, T. K. Eigentler, U. Keilholz, A. Hauschild, and J. M. Kirkwood, ‘Systematic review of medical

treatment in melanoma: current status and future prospects’, The Oncologist, vol. 16, no. 1, pp. 5–24,

2011, doi: 10.1634/theoncologist.2010-0190.

[4] J. Larkin et al., ‘Five-Year Survival with Combined Nivolumab and Ipilimumab in Advanced Melanoma’,

N. Engl. J. Med., vol. 381, no. 16, pp. 1535–1546, Oct. 2019, doi: 10.1056/NEJMoa1910836.



https://doi.org/10.1101/2020.06.23.20138594

16

[5] F. S. Hodi et al., ‘Nivolumab plus ipilimumab or nivolumab alone versus ipilimumab alone in advanced

melanoma (CheckMate 067): 4-year outcomes of a multicentre, randomised, phase 3 trial’, Lancet Oncol.,

vol. 19, no. 11, pp. 1480–1492, Nov. 2018, doi: 10.1016/S1470-2045(18)30700-9.

[6] M. Freeman-Keller, Y. Kim, H. Cronin, A. Richards, G. Gibney, and J. S. Weber, ‘Nivolumab in

Resected and Unresectable Metastatic Melanoma: Characteristics of Immune-Related Adverse Events and

Association with Outcomes’, Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res., vol. 22, no. 4, pp. 886–

894, Feb. 2016, doi: 10.1158/1078-0432.CCR-15-1136.

[7] P. Attia et al., ‘Autoimmunity correlates with tumor regression in patients with metastatic melanoma

treated with anti-cytotoxic T-lymphocyte antigen-4’, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol., vol. 23,

no. 25, pp. 6043–6053, Sep. 2005, doi: 10.1200/JCO.2005.06.205.

[8] S. G. Downey et al., ‘Prognostic Factors Related to Clinical Response in Patients with Metastatic

MelanomaTreated by CTL-Associated Antigen-4 Blockade’, Clin. Cancer Res. Off. J. Am. Assoc. Cancer

Res., vol. 13, no. 22 Pt 1, pp. 6681–6688, Nov. 2007, doi: 10.1158/1078-0432.CCR-07-0187.

[9] A. A. Sarnaik et al., ‘Extended dose ipilimumab with a peptide vaccine: immune correlates associated

with clinical benefit in patients with resected high-risk stage IIIc/IV melanoma’, Clin. Cancer Res. Off. J.

Am. Assoc. Cancer Res., vol. 17, no. 4, pp. 896–906, Feb. 2011, doi: 10.1158/1078-0432.CCR-10-2463.

[10] J. S. Weber et al., ‘Safety Profile of Nivolumab Monotherapy: A Pooled Analysis of Patients With

Advanced Melanoma’, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol., vol. 35, no. 7, pp. 785–792, Mar.

2017, doi: 10.1200/JCO.2015.66.1389.

[11] C. Hua et al., ‘Association of Vitiligo With Tumor Response in Patients With Metastatic Melanoma

Treated With Pembrolizumab’, JAMA Dermatol., vol. 152, no. 1, pp. 45–51, Jan. 2016, doi:

10.1001/jamadermatol.2015.2707.

[12] Y. Nakamura et al., ‘Correlation between vitiligo occurrence and clinical benefit in advanced melanoma

patients treated with nivolumab: A multi-institutional retrospective study’, J. Dermatol., vol. 44, no. 2, pp.

117–122, Feb. 2017, doi: 10.1111/1346-8138.13520.

[13] T. Z. Horvat et al., ‘Immune-Related Adverse Events, Need for Systemic Immunosuppression, and

Effects on Survival and Time to Treatment Failure in Patients With Melanoma Treated With Ipilimumab

at Memorial Sloan Kettering Cancer Center’, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol., vol. 33, no. 28,

pp. 3193–3198, Oct. 2015, doi: 10.1200/JCO.2015.60.8448.



https://doi.org/10.1101/2020.06.23.20138594

17

[14] P. A. Ascierto et al., ‘Clinical experience with ipilimumab 3 mg/kg: real-world efficacy and safety data

from an expanded access programme cohort’, J. Transl. Med., vol. 12, p. 116, May 2014, doi:

10.1186/1479-5876-12-116.

[15] E. A. Eisenhauer et al., ‘New response evaluation criteria in solid tumours: revised RECIST guideline

(version 1.1)’, Eur. J. Cancer Oxf. Engl. 1990, vol. 45, no. 2, pp. 228–247, Jan. 2009, doi:

10.1016/j.ejca.2008.10.026.

[16] A. Giobbie-Hurder, R. D. Gelber, and M. M. Regan, ‘Challenges of Guarantee-Time Bias’, J. Clin.

Oncol., vol. 31, no. 23, pp. 2963–2969, Jul. 2013, doi: 10.1200/JCO.2013.49.5283.

[17] A. Kassambara, M. Kosinski, P. Biecek, and S. Fabian, survminer: Drawing Survival Curves using

‘ggplot2’. 2018.

[18] Therneau T, ‘A Package for Survival Analysis in S. version 2.38, https://CRAN.R-

project.org/package=survival.’ 2015.

[19] M. I. Love, W. Huber, and S. Anders, ‘Moderated estimation of fold change and dispersion for RNA-seq

data with DESeq2’, Genome Biol., vol. 15, no. 12, Dec. 2014, doi: 10.1186/s13059-014-0550-8.

[20] S. Anders, P. T. Pyl, and W. Huber, ‘HTSeq--a Python framework to work with high-throughput

sequencing data’, Bioinformatics, vol. 31, no. 2, pp. 166–169, Jan. 2015, doi:

10.1093/bioinformatics/btu638.

[21] D. A. Bolotin et al., ‘MiXCR: software for comprehensive adaptive immunity profiling’, Nat. Methods,

vol. 12, no. 5, pp. 380–381, May 2015, doi: 10.1038/nmeth.3364.

[22] P. Dixon, ‘VEGAN, a package of R functions for community ecology’, J. Veg. Sci., vol. 14, no. 6, pp.

927–930, Dec. 2003, doi: 10.1111/j.1654-1103.2003.tb02228.x.

[23] J. L. McQuade et al., ‘Association of body-mass index and outcomes in patients with metastatic

melanoma treated with targeted therapy, immunotherapy, or chemotherapy: a retrospective, multicohort

analysis’, Lancet Oncol., vol. 19, no. 3, pp. 310–322, Mar. 2018, doi: 10.1016/S1470-2045(18)30078-0.

[24] R. Das et al., ‘Early B cell changes predict autoimmunity following combination immune checkpoint

blockade’, J. Clin. Invest., vol. 128, no. 2, pp. 715–720, Jan. 2018, doi: 10.1172/JCI96798.

[25] B. P. Fairfax et al., ‘Peripheral CD8+ T cell characteristics associated with durable responses to immune

checkpoint blockade in patients with metastatic melanoma’, Nat. Med., vol. 26, no. 2, pp. 193–199, Feb.

2020, doi: 10.1038/s41591-019-0734-6.



https://doi.org/10.1101/2020.06.23.20138594

18

[26] W. W. Eaton, N. R. Rose, A. Kalaydjian, M. G. Pedersen, and P. B. Mortensen, ‘Epidemiology of

Autoimmune Diseases in Denmark’, J. Autoimmun., vol. 29, no. 1, pp. 1–9, Aug. 2007, doi:

10.1016/j.jaut.2007.05.002.

[27] J. S. Weber, K. C. Kähler, and A. Hauschild, ‘Management of Immune-Related Adverse Events and

Kinetics of Response With Ipilimumab’, J. Clin. Oncol., May 2012, doi: 10.1200/JCO.2012.41.6750.

[28] K. Haratani et al., ‘Association of Immune-Related Adverse Events With Nivolumab Efficacy in Non-

Small-Cell Lung Cancer’, JAMA Oncol., vol. 4, no. 3, pp. 374–378, Mar. 2018, doi:

10.1001/jamaoncol.2017.2925.

[29] S. A. Green et al., ‘Activated platelet–T-cell conjugates in peripheral blood of patients with HIV

infection: coupling coagulation/inflammation and T cells’, AIDS, vol. 29, no. 11, pp. 1297–1308, Jul.

2015, doi: 10.1097/QAD.0000000000000701.

[30] T. A. Chan et al., ‘Development of tumor mutation burden as an immunotherapy biomarker: utility for the

oncology clinic’, Ann. Oncol., vol. 30, no. 1, pp. 44–56, Jan. 2019, doi: 10.1093/annonc/mdy495.



https://doi.org/10.1101/2020.06.23.20138594

19

Figures and Tables

Figure 1. Kaplan Meier curves of overall survival (OS) and progression free survival (PFS) according to

development of any grade of irAEs prior to cycle 5. a) OS whole cohort, shaded areas showing 95% CI (n=138),

b) PFS whole cohort, shaded areas showing 95% CI, c) Kaplan Meier of OS specific to recipients of cICB

therapy (n=60), d) Kaplan Meier of OS specific to recipients of sICB therapy (n=78). All P-values refer to log-

rank test.

Figure 2. a) Kaplan Meier curves of overall survival (OS) for Liverpool Replication dataset stratified according

to development of irAE prior to the fifth cycle of treatment, shaded areas showing 95% CI, (n=218), b) Kaplan

Meier curves for combined Oxford and Liverpool datasets (n=346). All P-values refer to log-rank test.

Figure 3. a) Results from multivariable Cox Proportional Hazard analysis of factors associated with OS in

Oxford cohort (n=138), P values corrected for multiple testing cohort, b) Factors associated with OS across

combined Oxford and Liverpool datasets where data available for all individuals (n=348), c) Factors associated

with OS across combined Oxford and Liverpool datasets where data available including cell counts (n=272).

Figure 4. a) The diversity of T Cell receptor beta chain, calculated from CD8+ TCR data mapped from RNA on

day 21 of ICB treatment was significant greater in those who developed irAE (student T test, n=106), b) Results

from multivariable logistic regression analysis demonstrating significant association with both ICB type

(reference cICB) and day 21 TCR diversity with the development of irAE, c) Volcano plot of differentially

expressed transcripts from pre-treatment CD8+ T cells from sICB recipients comparing those who did and did

not proceed to develop irAE within the first 5 cycles of treatment (n=71, 116 transcripts at FDR=0.05 and 224

transcripts at FDR=0.1) d) Go Ontology Biological Process (GOBP) pathway analysis of genes differentially

associated with development from 4c. Here the most significant 12 pathways are shown (y axis: GOBP code, x

axis: fold change enrichment).

Supplementary Figure 1. Profile of immune-related adverse events (irAEs) within the cohort. a) Density plot

of the development of different irAEs over time. b) Distribution of the number of different irAEs per patient. c)

Logistic regression model of the risk factors for development of irAEs prior to cycle 5



https://doi.org/10.1101/2020.06.23.20138594

20

Supplementary Figure 2. a) Kaplan Meier of Oxford PFS for cICB recipients, b) Kaplan Meier of Oxford PFS

for sICB recipients, c) Kaplan Meier of Oxford OS for cICB recipients based on grade 3 or grade 4 toxicity

prior to cycle 5, d) as per c, but for PFS, e) Landmark analysis for Oxford OS according to irAE prior to cycle 5,

f) Landmark analysis for Oxford PFS according to irAE prior to cycle 5.

Table 1. Baseline demographic and clinical characteristics of patients

Characteristic Single ICB therapy

(n=78)

Combination ICB

therapy (n=60)

Complete cohort

(n=138)

Median age (range), years 74 (29 – 95) 59 (21 – 74) 68 (21 – 95)

Males, n (%) 42 (53.8) 34 (56.7) 76 (55.1)

Median BMI (range) 28 (21 – 59) 27 (19 – 43) 28 (19 – 59)

Prior autoimmune disease, n (%) 4 (5.1) 2 (3.3) 6 (4.3)

ECOG performance status, n (%)

0 29 (37.2) 47 (78.3) 76 (55.1)

1 35 (44.9) 10 (16.7) 45 (32.6)

2 12 (15.4) 3 (5.0) 15 (10.9)

3 1 (1.3) 0 (0.0) 1 (0.7)

Not reported 1 (1.3) 0 (0.0) 1 (0.7)

BRAF status, n (%)

Mutation 19 (24.3) 23 (38.3) 42 (30.4)

Wild type 51 (65.4) 29 (48.3) 80 (58.0)

Not reported 8 (10.2) 8 (13.3) 16 (9.0)

Elevated LDH, n (%) 22 (28.2) 19 (31.7) 41 (29.7)

Prior systemic treatment, n (%) 27 (34.6) 8 (13.3) 35 (25.4)

BRAF inhibitor 10 (12.8) 5 (8.3) 15 (8.7)

Ipilimumab 17 (21.8) 0 (0.0) 17 (12.3)

Other 0 (0.0) 3 (5.0) 3 (2.2)



https://doi.org/10.1101/2020.06.23.20138594

21

Table 2. Immune related adverse events (irAEs) across the Oxford cohort

irAE Median

time to

irAE

onset

(days)

Single ICB therapy

(n=78)

Combination ICB therapy

(n=60)

Any grade irAE,

n (%)

Grade 3/4 irAE,

n (%)

Any grade irAE,

n (%)

Grade 3/4 irAE,

n (%)

Total - 35 (44.9) 7 (9.0) 54 (90.0) 38 (63.3)

Cutaneous

manifestation

37 26 (33.3) 0 (0.0) 33 (55.0) 3 (5.0)

Colitis 34 5 (6.4) 1 (1.3) 25 (41.7) 11 (18.3)

Hepatitis 49 2 (2.6) 1 (1.3) 22 (36.7) 13 (21.7)

Adrenal axis* 133 5 (6.4) 5 (6.4) 15 (25.0) 14 (23.3)

Thyroiditis 70 2 (2.6) 0 (0.0) 13 (21.7) 1 (1.7)

Arthritis 78 4 (5.1) 1 (1.3) 5 (8.3) 1 (1.7)

Pneumonitis 378 3 (3.8) 1 (1.3) 2 (3.3) 2 (3.3)

Gastritis 386 3 (3.8) 2 (2.6) 4 (6.7) 2 (3.3)

Neurological 63 0 (0.0) 0 (0.0) 5 (8.3) 2 (3.3)

Nephritis 87 1 (1.3) 1 (1.3) 2 (3.3) 1 (1.7)

Diabetes 65 0 (0.0) 0 (0.0) 2 (3.3) 2 (3.3)

Myositis 43 0 (0.0) 0 (0.0) 2 (3.3) 2 (3.3)

Haematological 136 0 (0.0) 0 (0.0) 1 (1.7) 1 (1.7)

*adrenal axis encompasses hypophysitis and adrenal insufficiency



https://doi.org/10.1101/2020.06.23.20138594

22

Table 3. Comparison of Oxford and Liverpool cohorts

Characteristic sICB

(n=136)

P vs.

Oxford

cICB (n=74) P vs.

Oxford

All treatments

Liverpool

cohort

(n=210)

Oxford

cohort

(n=138)

P

Median age

(range), years

68 (28 –

89)

P=0.02 65 (29 – 87) P=0.002 66 (28 –

89)

68 (21-95 0.51

Males, n (%) 77 (56.6) P=0.80 42 (56.7) P=1 119 (56.7) 76 (55.1) 0.77

Median BMI

(range)

28 (21 –

59)

P=0.31 27 (19 – 43) P=0.21 27 (17-49) 28 (19 –

59)

0.12



https://doi.org/10.1101/2020.06.23.20138594

23

Table 4. Predictors of oncological outcomes using univariable and multivariable Cox proportional hazards

models

Characteristic Univariable Multivariable

HR (95% CI) p value HR (95% CI) p value

OS Oxford

Age (mean=64.9) 1.01 (0.99 – 1.03) 0.41 0.99 (0.97 – 1.02) 0.47

Male sex 1.55 (0.88 – 2.72) 0.11 1.48 (0.71 – 3.11) 0.30

performance status >0 2.39 (1.35 – 4.23) 0.0028 2.83 (1.05 – 2.86) 0.03

Baseline BMI 1.01 (0.97-1.05) 0.59 1.01 (0.95-1.07) 0.79

Baseline BMI >25 1.29 (0.66-2.51) 0.46 - -

Development of irAE prior to

cycle 5

0.31 (0.17 – 0.57) 0.00015

0.09 (0.03 – 0.27) 0.000016

Anti-PD1 treatment 1.39 (0.77 – 2.52) 0.28 0.43 (0.13 – 1.43) 0.17

Prior systemic therapy 1.42 (0.81 – 2.53) 0.22 - -

Raised LDH (>255 iu/L) 2.22 (1.10 – 4.51) 0.02 - -

Baseline neutrophil(1x106/L) 1.25 (1.15 – 1.35) 7x10-8 1.22 (1.05 – 1.42) 0.009

Baseline lymphocyte(1x106/L) 1.13 (0.70 – 1.81) 0.62 1.39 (0.71 – 2.74) 0.33

Baseline monocyte(1x106/L) 16.7 (6.28 – 42.5) 1x10-8 7.47 (1.34 – 41.73) 0.022

Albumin (g/dl) 1.16 (1.10-1.24) 7x10-7 1.05 (0.87-1.05) 0.31

BRAF mutation 0.75 (0.40 – 1.41) 0.37 0.95 (0.41-2.18) 0.90

Non-cutaneous melanoma 2.71 (1.46 – 5.06) 0.002 - -

OS Liverpool

Age (mean=64.3) 0.98 (0.96 – 0.99) 0.004 0.98 (0.96 – 0.99) 0.009

Male sex 1.09 (0.74 – 1.61) 0.65 1.22 (0.82 – 1.82) 0.32

Baseline BMI 1.00 (0.97 – 1.04) 0.83 - -

Baseline BMI >25 1.09 (0.72-1.64) 0.69 1.00 (0.97 – 1.04) 0.68

irAE prior to cycle 5 0.50 (0.34 – 0.75) 0.0008 0.43 (0.27 – 0.69) 0.00051


Prior systemic therapy 2.09 (1.05-4.16) 0.04 2.33 (1.09 – 4.95) 0.03

Baseline neutrophil (1x106/L) 1.21 (1.11 – 1.32) 8.6x10-6 - -

Baseline lymphocyte (1x106/L) 0.76 (0.50 – 1.15) 0.20 - -

Low Albumin (<36g/dl) 7.42 (3.54-15.58) 1.2x10-7 - -

Multivariate analysis only performed on variables where 95% or more of values were available for analysis Albumin threshold as determined for normal range by institution



https://doi.org/10.1101/2020.06.23.20138594

24

Table 5. Univariable and multivariable analyses of factors associated with OS - combined datasets

Characteristic Univariable Multivariable

HR (95% CI) p value HR (95% CI) p value

OS - baseline indices (n=348)

Age 0.99 (0.98 – 1.00) 0.11 0.98 (0.97 – 1.00) 0.02

Male sex 1.23 (0.90 – 1.69) 0.2 1.29 (0.93 – 1.79) 0.13

Baseline BMI 1.01 (0.98-1.03) 0.67 - -

Baseline BMI >25 1.13 (0.79-1.60) 0.51 1.10 (0.77 – 1.57) 0.60

irAE prior to cycle 5 0.43 (0.31 – 0.60) 6.5x10-7 0.38 (0.25 – 0.57) 2.5x10-6


Prior systemic therapy 1.10 (0.68 – 1.79) 0.69 1.14 (0.64 – 2.05) 0.65

Region (vs. Oxford) 1.32 (0.95-1.85) 0.10 1.31 (0.88-1.95) 0.19

OS - including biochemical and haematological indices (n=281)

Age - - 0.99 (0.97 – 1.00) 0.14

Male sex - - 1.35 (0.91 – 1.99) 0.14

Baseline BMI >25 - - 0.96 (0.61 – 1.50) 0.85

irAE prior to cycle 5 - - 0.36 (0.23 – 0.58) 1.6x10-5

Anti-PD1 treatment - - 1.25 (0.74 – 2.10) 0.41

Prior systemic therapy - - 1.22 (0.66 – 2.25) 0.52

Baseline neutrophil 1.22 (1.15 – 1.29) 1.2x10-11 1.17 (1.09-1.27) 3.4x10-5

Baseline lymphocyte 1.11 (0.66 – 1.23) 0.53 0.96 (0.69-1.33) 0.80

Low Albumin (per centre) 4.7 (2.87-7.70) 8.1x10-10 3.68 (1.96-6.91) 5.3x10-5

Region (vs. Oxford) - - 1.55 (0.41-1.01) 0.06

Multivariate analysis only performed on variables where 95% or more of values were available for analysis Albumin threshold as determined for normal range by institution



https://doi.org/10.1101/2020.06.23.20138594

25

Supplementary Data 1. List of genes differentially associated with irAE development in pretreatment samples

from sICB recipients

Supplementary Data 2. List of GOBP pathways enriched for members of genes associated with irAE

development



https://doi.org/10.1101/2020.06.23.20138594

Figure 1 a b

c d

+++

+ ++ ++

+ ++++++++ +

+++++++++++

++++ ++++++ ++ ++++ +++ ++ ++ +++ + + +++ +

p = 0.00034

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Prog

ress

ion

surv

ival p

roba

bilit

yfre

e

irAE prior to cycle 5

++

NoYes

64 29 20 17 12 4 1 1 0

74 46 31 24 15 11 6 5 1Yes

No

Number at risk

+++++

++ +++

+ +++ ++ ++ ++ ++

+++ + +

+++++++++++++++++ ++++ +++++++++++++ ++++ ++ + +++ + ++++++

+ + +++ +

p < 0.0001

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Ove

rall

surv

ival p

roba

bilit

y

irAE prior to cycle 5

++

NoYes

64 45 35 25 17 12 4 2 1

74 58 43 28 20 18 9 5 1Yes

No

Number at risk

+

+ +

+ ++++++++++++ ++++ +++++++++

++ ++ + + ++ + +

+++

p = 0.0013

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30Months

Ove

rall

surv

ival p

roba

bilit

y

irAE prior to cycle 5++

NoYes

7 3 1 1 1 1 0

53 42 29 15 10 9 3Yes

NoNumber at risk

++++

+ +++

+ +++ ++ ++ + ++

+++ + +

++++ + + ++ + + ++ + + +++ +

p = 0.0023

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Ove

rall

surv

ival p

roba

bilit

y


NoYes

57 42 34 24 16 11 4 2 1

21 16 14 13 10 9 6 5 1Yes

NoNumber at risk



https://doi.org/10.1101/2020.06.23.20138594

++ +

+

++ + + + + + + + + + +

+ + + ++ + + + + +

+ + ++ + + +

+ + + + + + + + + ++ + + + + + + + + +

p = 0.00064

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Ove

rall

surv

ival p

roba

bilit

y


No

Yes

114 82 52 40 24 11 9 6 1

94 87 68 49 29 18 9 4 0Yes

No

Number at risk

++++++++

+++

++ ++++++ + ++ + ++++++ ++++ +++ ++ +++

++ +++ ++ + + + + +

++++++++++++++++++ ++++++ ++++++++++++++++

+++++++++++++ + +++ ++++++++ +++++++++ + ++ +++ +++ +

p < 0.0001

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Ove

rall

surv

ival p

roba

bilit

y


No

Yes

178 127 87 65 41 23 13 8 2

168 145 111 77 49 36 18 9 1Yes

No

Number at risk

Figure 2

a

b



https://doi.org/10.1101/2020.06.23.20138594

0.05 0.25 0.50 1.00 2.00 5.00 10.00 50.00Hazard Ratio (95% Confidence Interval)

Cov

aria

te

−log10 PValue1234

corrected PValue>0.05<0.05

age

irAE pre C5

albumin

baseline BMI

BRAF mutated

lymphocyte

type (non cut)

monocytes

neutrophils

PS

sex (male)

treatment (sICB)

0.25 0.50 1.00 1.50Hazard Ratio (95% Confidence Interval)

corrected PValue>0.05<0.05

−log10PValue2.55.07.5C

ovar

iate

age

irAE pre C5

centre (Oxford)

lymphocyte

neutrophils

sex (male)

baseline BMI

treatment (sICB)

0.25 0.50 1.00 1.50Hazard Ratio (95% Confidence Interval)

corrected PValue > 0.05< 0.05

−log10 PValue246

age

irAE pre C5

baseline BMI

centre (Oxford)

treatment (sICB)

sex (male)

Cov

aria

te

Figure 3

a

b

c



https://doi.org/10.1101/2020.06.23.20138594

Figure 4 a b

c d

neutrophil degranulation

antimicrobial humoral response

positive regulation of tumour necrosis factor biosynthetic process

platelet degranulation

response to lipopolysaccharide

antibacterial humoral response

hydrogen peroxide catabolic process

leukocyte migration

inflammatory response

positive regulation of nitric oxide biosynthetic process

actin filament polymerization

response to tumour necrosis factor

GO:0006954

GO:0050900

GO:0043312

GO:0032496

GO:0002576

GO:0045429

GO:0034612

GO:0030041

GO:0042744

GO:0019731

GO:0019730

GO:0042535

5 10 15 20 25fc

GO

BP

nOverlap10

20

30

40

20

30

40

−log(pvalue)

n=71 COL17A1

ESAMCEACAM1

FAR2 ABCA13

FAXDC2PTTG1IP

AFF2

CHIT1CTDSPL

VSIG2

H1−0

PTMAP2AZU1

TACSTD2VIL1

MGLL

SIGLEC17P

SEC22B4P WDR54

DRAXIN

NEO1 PCNX2

0.0

2.5

5.0

7.5

−2 0 2lo g 2 (fold change)

−lo

g 10(p−

valu

e)

association with irAE

suppressednoneupregulated

age

day 0 diversity

day 21 diversity

ICB

male sex

monocyte

0.01 0.10 1.00 10.00Odds Ratio (95% Confidence Interval)

varia

ble

−log10(PVal)1234

P=4.3-5

no yes

6

7

8

9TRB diversity day 21

Autoimmune IRAE pre-cycle 5

Dive

rsity



https://doi.org/10.1101/2020.06.23.20138594

0

10

20

30

40

0 1 2 3 4 5 6Number of irAEs per patient

Num

ber o

f pat

ient

s

Type of ICB therapycombinationsingle

Neurological Pneumonitis Thyroiditis

Cutaneous Gastritis Hepatitis

Adrenal axis Arthritis Colitis

0 200 400 600 0 200 400 600 0 200 400 600

0.000

0.005

0.010

0.015

0.000

0.005

0.010

0.015

0.000

0.005

0.010

0.015

Days post initiation of ICB therapy

Den

sity

−4 −2 0 2

Odds Ratio (95% Confidence Interval)

Cha

ract

eris

tic

−log10(PVal)2

4

6

single versus combination

immunotherapy

prior autoimmunity

male sex

baselineBMI

age

Supplementary Figure 1

a b

c



https://doi.org/10.1101/2020.06.23.20138594

+

++

+++++++

+++ + ++ ++ + ++

+ ++ + + +

p = 0.0019

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30Months

Prog

ress

ion

free

surv

ival p

roba

bilit

y


NoYes

7 1 1 1 1 1 0

53 31 17 11 5 2 0Yes

No

Number at risk

+++

++ ++ +

++ +++++++ +

++ + + + ++ + + ++ + + +++ +p = 0.00047

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Prog

ress

ion

free

surv

ival p

roba

bilit

y

irAE prior to cycle 5NoYes

57 28 19 16 11 3 1 1 0

21 15 14 13 10 9 6 5 1Yes

NoNumber at risk

++

+++++++++++++ +++ +++++

++ ++ ++ ++ ++ +++ ++++++ + + + + ++ + +

++++++++ +++ +++ ++++++++++ +++ ++ + ++ + +++++ +

p = 0.0014

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Ove

rall

surv

ival p

roba

bilit

y


NoYes

93 65 50 36 26 21 10 6 2

45 38 28 17 11 9 3 1 0Yes

No

Number at risk

+++++

++++ ++ + ++ +

+ + +++++++++ + + + + ++ +

++++++++

+++ +++++ + +++

+++ ++ + ++ +

p = 0.0047

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Ove

rall

surv

ival p

roba

bilit

y


NoYes

93 44 31 27 19 11 6 5 1

45 31 20 14 8 4 1 1 0Yes

No

Number at risk

+++++

+++++

+ +++ ++ ++ ++

++++

+ + +

+++++++++++++++++ ++++ +++++++++++++ ++++ ++ + +++ + ++++++

+ + +++ +

p = 0.0013

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Ove

rall

surv

ival p

roba

bilit

y


NoYes

55 45 35 25 17 12 4 2 1

72 58 43 28 20 18 9 5 1Yes

No

Number at risk

++

+

+ ++ +

++ ++++++++ +

++++++++++

++++ ++++++ ++ ++++ +++ ++ ++ +++ + + +++ +

p = 0.099

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25 30 35 40Months

Prog

ress

ion

Free

sur

viva

l pro

babi

lity


NoYes

37 29 20 17 12 4 1 1 0

61 46 31 24 15 11 6 5 1Yes

No

Number at risk

Supplementary Figure 2

a b

c d

e f



https://doi.org/10.1101/2020.06.23.20138594

0000-0001-7413-5002...2020/06/23 · 2015 to 15april 2019. 60 patients received cicb consisting of...

Documents