Oncology

Practical estimands and missing-data methods for PROs truncated by progression

This methods paper proposes a framework to define estimands for longitudinal PRO endpoints when disease progression acts as an intercurrent event. It compares hypothetical and treatment-policy strategies and demonstrates three implementations using simulation. Recommended approaches include implicit multiple imputation in longitudinal models, joint longitudinal–time-to-progression modelling, and control-based imputation post-progression. Authors show supplementary analyses strengthen confidence in PRO interpretation despite progression-related missingness. 1

Nivolumab + ipilimumab shows activity in HTMB breast and pooled histologies but not colorectal cancer

In this phase II basket registry, HTMB (≥10 mut/Mb) patients with breast cancer (n=13) and pooled other histologies (n=26) had disease control rates of 33% and 32%, respectively. The colorectal cohort (n=12) failed to meet the prespecified efficacy threshold and did not expand. Nineteen patients experienced treatment-related grade 3 or serious adverse events, highlighting substantial toxicity risk. Results support histology-specific benefit with N+I in HTMB tumors but are limited by small cohort sizes and nonrandomized design. 2

Most Chinese cancer drug trials use restrictive eligibility beyond NCI recommendations

Review of 2,448 registered anticancer drug trials in China (2013–2021) found 97.0% used at least one restrictive eligibility criterion beyond NCI guidance. Restrictions for performance status, organ function, and comorbidities were highly prevalent. Use of restrictive criteria increased over time, notably for PS, cardiac function, and prior/concurrent malignancies. These patterns likely limit trial generalizability and slow enrollment of patients typical in routine practice. 3