Modernizing Clinical Trial Eligibility Criteria: Recommendations of the ASCO-Friends of Cancer Research Performance Status Work Group

Abstract

Purpose: Performance status (PS) is one of the most common eligibility criteria. Many trials are limited to patients with high-functioning PS, resulting in important differences between trial participants and patient populations with the disease. In addition, existing PS measures are subjective and susceptible to investigator bias.

Experimental design: A multidisciplinary working group of the American Society of Clinical Oncology and Friends of Cancer Research evaluated how PS eligibility criteria could be more inclusive. The working group recommendations are based on a literature search, review of trials, simulation study, and multistakeholder consensus. The working group prioritized inclusiveness and access to investigational therapies, while balancing patient safety and study integrity.

Results: Broadening PS eligibility criteria may increase the number of potentially eligible patients for a given clinical trial, thus shortening accrual time. It may also result in greater participant diversity, potentially reduce trial participant and patient disparities, and enable clinicians to more readily translate trial results to patients with low-functioning PS. Potential impact on outcomes was explored through a simulation trial demonstrating that when the number of Eastern Cooperative Oncology Group PS2 participants was relatively small, the effect on the estimated HR and power was modest, even when PS2 patients did not derive a treatment benefit.

Conclusions: Expanding PS eligibility criteria to be more inclusive may be justified in many cases and could result in faster accrual rates and more representative trial populations.See related commentary by Giantonio, p. 2369.

Figure 1.

Simulation study results depicting changes in estimated hazard ratio (HR) (Panel A); power (Panel B); and length of trial from first accrual to last required event for analysis (Panel C). Within each panel, six analyses are depicted. In panels A and C, the median from the simulations is plotted as a circle with lines extending vertically to the 5^th and 95^th percentiles. For each analysis, the HR of PS0 and PS1 patients remains constant at 0.7 and the HR of PS2 patients is varied. Red points/lines depict results when PS2 patients are included in the final analysis (N=500); blue points/lines depict results when PS2 patients are excluded (N=450). Regardless of sample size, the trial end is assumed to be when the required number of events (283 events) have been accrued per the power calculation.