Overestimation of survival in individuals with cancer treated in the real world with new cancer drugs: A population-based cohort study.

Saroj Niraula BC Cancer/ University of British Columbia, Kelowna, BC, Canada info_outline Saroj Niraula

Authors Saroj Niraula BC Cancer/ University of British Columbia, Kelowna, BC, Canada info_outline Saroj Niraula Organizations BC Cancer/ University of British Columbia, Kelowna, BC, Canada Abstract Disclosures Research Funding No funding sources reported Background: Oncologists and individuals with cancer often rely on outcomes from pivotal randomized controlled trials (RCTs) for real-world treatment decisions. However, stringent inclusions and standardized protocols limit their applicability. This study quantifies overall survival (OS) disparities for novel cancer drugs between real-world and pivotal RCTs. Methods: Ethics approvals were secured, and data on frequently used targeted cancer drugs for solid cancer treatment in all individuals with cancer in Manitoba, Canada from inception to June 30, 2023, collected using pharmacy database and the Manitoba Cancer Registry. Evidence consistently indicates similar cancer outcomes in Manitoba to most high-resource countries. This analysis includes six common immunotherapies and monoclonal antibodies. Kaplan-Meier Method was used to plot survival probabilities, and descriptive statistics to elucidate variable differences. Comparison between survival outcomes in pivotal RCTs and real-world scenarios was expressed as "Overestimation Quotient or OQ" (OS in RCT over OS in the real world for same indication). Results: A total of 941 individuals diagnosed with cancer were evaluated across eight indications. The observed median OS in the real world for included indications was 11.9 months, ranging from 6.6 to 37.6 months— a duration significantly shorter than the median OS reported in pivotal Randomized Controlled Trials (RCTs) (31.35 months, range 9.2 to 72.1) (P < 0.001). This translated to a median difference of 15.6 months (range 2.6 to 59.9) per indication and an "Overestimation Quotient" ranging from 1.5 to 5.9. Notably, the observed OS was even inferior in the real-world setting compared to the control group used in pivotal RCTs in half of the indications (Table 1). Conclusions: In routine practice, new cancer drugs offer starkly inferior survival compared to that reported in pivotal RCTs used for approval of those drugs, with the difference in survival ranging from a few months to several years, translating up to six-fold difference. Our results highlight potential for alarming misinformation influencing treatment decisions in clinics. We trust that oncologists will consider our findings during treatment discussions in clinics. Drug Indication Median OS(95% CI) Real World (Months) Sample Size Median OS Pivotal RCT (Months) Median OS (Control Arm) RCT (Months) Ref Author Overestimation Quotient Bevacizumab CNS 3.7(2.9-5.1) 53 9.8 N/A Vredenburgh 2.6 Durvalumab NSCLC 29.6(24.4-34.9) 102 47.5 29.1 Spigel 1.6 Ipilimumab+Nivolumab Melanoma 12.2(8.4-27.6) 57 72.1 19.9 Wolchok 5.9 Nivolumab Kidney 11.7(7.2-16.7) 61 25 19.6 Motzer 2.1 Nivolumab NSCLC 6.6(5.3-7.8) 135 9.2 6 Brahmer 1.4 Pertuzumab Breast 37.6(29.8-64.6) 134 56.5 40.8 Swain 1.5 Pembrolizumab Melanoma 21.2(12.12-38.4) 81 32.7 15.9 Robert 1.5 Pembrolizumab NSCLC 11.1(9.4-13.1) 318 30 14.2 Reck 2.7