A disease model for Multiple Myeloma developed using Real World Data
Chanu Pascal 1, Wang Xing (Nina) 2, Li Zao 2, Chen Shang-Chiung 3, Samineni Divya 2, Susilo Monica 2, Ogbu Uzor 4, Williamson Mellissa 5, Marchand Mathilde 6, Jin Yan Jin 2, Li Chunze 2, Bruno René 1
1 Clinical Pharmacology, Genentech/Roche, France; 2 Clinical Pharmacology, Genentech Inc; 3 Travere Therapeutics, Inc; 4 Clinical Science, Genentech Inc; 5 Real World Data Science Genentech Inc; 6 Certara Strategic Consulting, France
Objectives: Multiple myeloma (MM) is a neoplasm characterized by the proliferation and accumulation of malignant plasma cells. End-organ damage resulting from MM includes hypercalcemia, renal insufficiency, anemia, and lytic bone lesions. MM remains incurable despite advances in treatment. Model-based approaches based on tumor dynamics have successfully been used to optimize drug-development in oncology [1]. A drug-independent link between tumor growth inhibition (TGI) metrics and overall survival (OS) was previously established for refractory MM patients where serum M-protein substitutes for tumor size as a marker of tumor burden [2,3]. Indeed M-protein is one of the criteria used to assess clinical response according to the International Myeloma Working Group Uniform Response Criteria. Model-based TGI metrics are biomarkers capturing treatment effect and predict for OS benefit in TGI-OS models, these models have been shown to be drug-independent in most cases [4]. To prevent any confusion, TGI metrics are renamed “M-protein dynamic metrics” in this work. Genentech/Roche has several clinical programs in MM. Real World Data (RWD) were used to establish the link between M-protein dynamic and OS based on recent clinical data, especially with multiple new therapies made available to MM patients during the last decade.
Methods: Data from an oncology-based electronic health record (EHR)-derived de-identified database, Flatiron Health in 1666 patients with relapse or refractory MM were used to develop a model linking M-protein dynamic metrics to OS. Patient data were selected when one the following five treatment combinations was administered: lenalidomide – dexamethasone (N=715), lenalidomide - dexamethasone – daratumumab (N=132), bortezomib – dexamethasone (N=580), and bortezomib - dexamethasone - daratumumab (N=107), pomalidomide – dexamethasone (N=132). Individual OS and M-protein data were used. Majority of the data was collected between 2014 and 2020. Median treatment duration was 17 months with an interquartile interval of 8.5 to 31.7 months. M-protein dynamic metrics were derived using an empirical bi-exponential model [5,6]. M-protein dynamic metrics (M-protein ratio to baseline at different time points, time to growth, growth rate, shrinkage rate), 24 prognostic factors and potential treatment effects were first tested in a univariate analysis using a Cox proportional hazards regression model. Statistically significant covariates in the univariate analysis (p<0.05) were included in a full multivariate parametric distribution survival model. A backward deletion stepwise procedure was performed (p<0.01) to retain covariates in the final OS model. External model validation was performed using daratumumab Phase 3 randomized clinical trials POLLUX and CASTOR (NCT02076009, NCT02136134) obtained in the YODA project (The Yale University Open Data Access, https://yoda.yale.edu/) [7,8].
Results: OS data followed a log-normal distribution. Among all tested M-protein dynamic metrics and prognostic factors, growth rate (G=log(KG)) [6] was found to be the best predictor of OS. Patients with slower growth, better ECOG, higher creatinine clearance and hemoglobin tended to have longer survival, the model was confirmed to be drug-independent.
|
Parameter |
Value |
Std. Error |
p |
|
(Intercept) |
-1.99978 |
0.67836 |
0.00320 |
|
Log(KG (day-1)) |
-0.61366 |
0.07797 |
3.5e-15 |
|
ECOG 0,1,2,>=3 |
-0.35160 |
0.07882 |
8.2e-06 |
|
Creatinine clearance (mL.min-1) |
0.00887 |
0.00229 |
0.00011 |
|
Hemoglobin (g/dL) |
0.12296 |
0.03764 |
0.00109 |
|
Log(scale) |
0.17471 |
0.05274 |
0.00092 |
Model qualification was first performed on the analysis dataset using posterior predictive checks based on Kaplan Meier plots: prediction intervals derived from model-based simulations well included observed OS data for each treatment combination. The model was also validated by reproducing Phase 3 clinical trial outcomes for POLLUX and CASTOR studies: e.g. for CASTOR the simulated OS hazard ratio [95% PI] was 0.51 [0.24;0.91] while the observed value was 0.58.
Conclusions: A model linking M-protein dynamic to OS in MM could be developed based on RWD and qualified in predicting independent trials. The model is consistent with a previous one [2] and can support drug development in MM: e.g., make OS inferences based on early M-protein dynamic data obtained for a new agent, or support design for Phase 3 clinical trials. RWD open a new era with new opportunities for Model-Informed Drug Development.
References:
[1] Bruno R, Bottino D, De Alwis DP, Fojo T, Guedj J, Liu C, et al. Progress and Opportunities to Advance Clinical Cancer Therapeutics Using Tumor Dynamic Models. Clin Cancer Res. 2019 Dec 23. doi: 10.1158/1078-0432.CCR-19-0287.
[2] Bruno R, Jonsson F, Zaki M, Jacques C, Swern A, Richardson PG, et al.. Simulation of Clinical Outcome for Pomalidomide Plus Low-Dose Dexamethasone in Patients with Refractory Multiple Myeloma Based on Week 8 M-Protein Response. Blood 2011;118:1881.
[3] Jonsson F, Ou Y, Claret L, Siegel D, Jagannath S, Vij R et al. A tumor growth inhibition model based on M-protein levels in subjects with relapsed/refractory multiple myeloma following single-agent carfilzomib use. Clin Pharmacol Ther: Pharmacometrics and System Pharmacology, 2015;4:711-9.
[4] Bruno R, Mercier F, Claret L. Evaluation of tumor size response metrics to predict survival in oncology clinical trials. Clin Pharmacol Ther 2014;95:386-93.
[5] Stein WD, Gulley JL, Schlom J, Madan RA, Dahut W, Figg WD, et al. Tumor regression and growth rates determined in five intramural NCI prostate cancer trials: the growth rate constant as an indicator of therapeutic efficacy. Clin Cancer Res 2011;17:907-17.
[6] Claret L, Jin JY, Ferté C, Winter H, Girish S, Stroh M et al. A Model of Overall Survival Predicts Treatment Outcomes with Atezolizumab versus Chemotherapy in Non-Small Cell Lung Cancer Based on Early Tumor Kinetics. Clin Cancer Res. 2018;24:3292-98.
[7] Dimopoulos MA, Oriol A, Nahi H, San-Miguel J, Bahlis NJ, Usmani SZ et al. Daratumumab, lenalidomide, and dexamethasone for multiple myeloma. N Engl J Med 2016;375:1319-31.
[8] Palumbo A, Chanan-Khan A, Weisel K, Nooka AK, Masszi T, Beksac M et al. Daratumumab, bortezomib, and dexamethasone for multiple myeloma. N Engl J Med 2016;375:754-66.