Background:Carfilzomib is a novel, irreversible inhibitor ofthe proteasome’s chymotrypsin-like activity(Figure 1). It is currently broadly used inrelapsed or refractory multiple myelomapatients. Early, preclinical and clinical datapoint towards a cardiotoxic side effect due tothe fast and highly efficient inhibition of theproteasome in cardiomyocytes. Although thisdrug is known to be cardiotoxic, so far thereare only limited data about assessment ofmyocardial damage using cardiac MRI (CMR)and early identification of patients at risk.

Carfilzomib for Relapsed Multiple Myeloma Patients: Detection of Cardiotoxicity using Cardiac MRI

N.A. Nooman1,2,3, J. Riffel1,2, H. Goldschmidt2,4, F. André1,2, J. Salatzki1,2, M.G. Friedrich1,2,5, H.A. Katus1,2, L.H. Lehmann1,2

1 Klinik für Innere Med. III, Kardiologie, Angiologie und Pneumologie, Universitätsklinikum Heidelberg, Germany; 2 German Center for Cardiovascular Research (DZHK), partner site Heidelberg/Mannheim, Germany; 3 Tanta University Hospital, Diagnostic Radiology Department, Tanta, Egypt; 4 Hematologie, Oncologie und Rheumatologie, Innere Medizin V, Heidelberg, Germany; 5 McGill University Health Centre, Department of Medicine and Diagnostic Radiology, Montreal, Canada

Purpose:This prospective, observational, single centerstudy aimed at identification of patients atrisk based on myocardial deformation (strain)analysis by using state of the art CMR. Ourhypothesis is to detect subclinical cardiotoxicevents and to define parameters which couldpredict late cardiac changes after 3 monthsfollow up (FU) in patients with multiplemyeloma receiving Carfilzomib.

Patients & Methods:Patients (Table 1) diagnosed with multiplemyeloma and scheduled to undergotreatment with Carfilzomib have beenrecruited and performed three CMRexaminations using a clinical 1.5T MRI system(Philips Achieva). Scans were performedbefore onset of therapy (baseline) as well as3-5 days and 3 months thereafter.The CMR parameters included standardmorphological and functional parameters forleft and right ventricles (LV/RV) i.e. ejectionfraction (EF), enddiastolic volume/bodysurface area (EDV/BSA), endsystolicvolume/BSA (ESV/BSA), mitral and tricuspidannular plane systolic excursion (MAPSE,TAPSE) as well as tissue characterizationparameters (LV global T1 & T2 Mapping)(Figure 2) and deformation parameters (LV &RV global longitudinal and circumferentialmyocardial strain (GLS) and (GCS)) from fSENCdata (Myocardial Solutions, Morrisville, US).All data were compared between baseline, 1st

and 2nd follow up.

Results:So far, we have enrolled 14 patients. Twopatients dropped out from the 2nd follow updue to change of chemotherapy regimen ordeath.We found that the mean global LV Strainparameters were within normal range atbaseline and stable during follow-up. Fivepatients however showed an increase in thenumber of abnormal LV segments betweenbaseline and the first follow-up scan (seeFigure 3). Furthermore, at the 2nd follow upthree of those patients showed a decreaseof the LVEF and/or MAPSE which wasassociated with the appearance of clinicalsymptoms such as dyspnea or generalizedweakness in two of them.The mean RV GCS was significantly reducedin follow-up studies (baseline -17.8±1.6%, 1st

-16.5±2.4% & 2nd -14.9±2.7%) indicating apotential effect on cardiac deformation ofthe RV (P = 0.018) (see Figure 4).Other functional and tissue characterizationparameters were within normal range atbaseline and during follow-up.

Conclusion:Our preliminary results indicate thatCarfilzomib may affect ventricular strain inthe absence of detectable tissue pathology.Strain parameters may provide sensitivemarkers to identify an impact ofchemotherapy on the myocardium at an earlystage and thus develop preventive strategies.

Fig 1. Mechanism of CFZ-induced proteasome inhibition.

Figure 2: Exemplary tissue characterization parameters (T1 & T2-Mapping). Images illustrate T1 mapping (upper images) and T2 mapping(lower images) of a patient before carfilzomib therapy. Right imagesillustrate a ‘bullseye’ of the analyzed regions.

References:Fig. 1: Kortuem, M.K. and Stewart, K.A.: Carfilzomib. TheAmerican Society of Hematology. 2013. DOI 10.1182/blood-2012-10-459883.

Conflict of interest:Prof. DR. M. G. Friedrich is a board member inCircle CVI. Other authors have no conflict ofinterest to disclose.

Figure 1: Mechanism of Carfilzomib inducedproteasome inhibition. Carfilzomib binds irreversibly tothe 20s subunit of the proteasome which leads to itscomplete inhibition and subsequently accumulation ofundegraded proteins in the cell. CFZ, Carfilzomib.Modified from Kortuem et al. 2013.

Characteristic ValueMedian age, y (range) 61 (45-74)Male 9Cardiovascular (CV) co-morbidities

Hypertension 3Diabetes mellitus 2Hyperlipidemia 2Obesity 1History of smoking 3Atrial fibrillation 1Coronary artery disease (CAD) 0

Prior therapyBortezomib 9Vincristine 2

Concomitant Immunomodulatory agent (Lenalidomide) 4

Table 1: Baseline Characteristics. (n = 14). All values are expressed asn. y, years; n, number of patients.

Figure 3: Comparison between Myocardial strain analysis at baselineand 3 to 5 days thereafter. Circumferential strain as ‘bull’s eye’ image(upper panel) and 2/3/4 chamber view (lower panel) of a patient atbaseline and 4 days after initiation of Carfilzomib therapy showing theincrease in number of abnormal LV segments (segments with reducedmyocardial strain). The color maps represent continuous strain variables.2Ch, 2 chamber view; 3Ch, 3 chamber view; 4Ch, 4 chamber view; LV, leftventricle; RV, right ventricle.

Figure 5: Biomarkers of included patients. The graphshows the cardiac biomarkers as indicated. Values areexpressed as mean ± SD. Nt-proBNP, n-terminal brainnatriuretic peptide; hsTNT, high sensitive troponin-t.

The decrease in RV GCS was not associatedwith significant increase in cardiacbiomarkers (Nt-proBNP, hsTNT) (Figure 5).Longterm follow up of those patients isneeded to determine whether an increase incardiac biomarkers is associated with highercardiac event rates or increase in mortality.

Figure 4: LV & RV myocardial strain at baseline and follow-up. The graphillustrates the reduction in RV GCS at the 2nd FU (red data). n, number ofpatients; p, measured by paired t-test; symbols indicate mean ± SD.

(ng/l)(pg/ml)