Evaluating eligibility criteria of oncology trials using real-world data and AI

Abstract

Line of therapy (LoT); confidence interval (CI); first line of therapy (1L); second line of therapy (2L). Extended Data Table 2 | Validation on progression-free survival hazard ratioThe number of inclusion and exclusion criteria, the number of eligible patients and the hazard ratio of progression-free survival with confidence interval of emulated aNSCLC trials with eligibility criteria under three scenarios: original criteria of the clinical trial, fully relaxed criteria and data-driven criteria learned from results of the hazard ratio of overall survival (same as in Table 1). Article Extended Data Table 3 | Analysis in other cancersEligibility criteria for colorectal cancer (CRC), advanced melanoma and metastatic breast cancer in three scenarios. The number of inclusion and exclusion criteria, the number of eligible patients and the hazard ratio of the overall survival with confidence interval of emulated aNSCLC trials with eligibility criteria under three scenarios: original criteria of the clinical trial, fully relaxed criteria and data-driven criteria.

Machine learning enabled subgroup analysis with real-world data to inform better clinical trial design

“…In a recent study from Liu et al (32) which aims at evaluating ECs for oncology trials using RWD and AI, the authors quantified the representability of each study trait with SHAP (33), and they tried to relax the range of each EC for broadening the participation. Only traits with continuous values are considered in a one-by-one manner.…”

R WE ready for reimbursement? A round up of developments in RWE relating to health technology assessment: part 2

“…This point would require more diverse clinical trials. To this end, a study using RWD from the US Nationwide oncology data source Flatiron Health showed that when using RWD to emulate clinical trials in lung cancer, removing many common trial exclusion criteria had a minimal effect on the trial effect estimate [14]. Relaxing trial entry criteria; however, had an important benefit of making clinical trials more inclusive for women and older patients.…”

A vision of immuno-oncology: the Siena think tank of the Italian network for tumor biotherapy (NIBIT) foundation

“…Moreover, by combining protein modeling and molecular dynamics simulations, nonconservative substitutions in functional regions of the spike glycoprotein of SARS-CoV-2 were demonstrated to contribute to differences in its virulence [109]. In spite of being potentially useful in the present "pandemic contingency", AI will undoubtedly prove useful in the future to design novel clinical trials with more inclusive molecular eligibility criteria in cancer patients [110].…”