Multicenter study of the impact of trial eligibility criteria on enrollment to KRAS G12C inhibitor trials in patients with non-small cell lung cancer.

11171 Background: Eligibility criteria (EC) are the primary method to assess patient appropriateness for clinical trials. There is a tradeoff between narrowing EC for patient safety and matching trial populations to the real-world population likely to receive the study agents. There is little evidence to guide optimal EC design. In 2017, ASCO proposed modifications to EC to increase the generalizability of trial findings. We previously reported a single-center experience and now report a multi-center cohort of non-small cell lung cancer (NSCLC) patients (pts) with KRAS G12C mutations to determine whether EC for trials of KRAS G12C inhibitors allowed enrollment of pts seen as part of routine care at three academic medical centers. Methods: We extracted EC for Phase I-III trials of six KRAS G12C inhibitors (sotorasib, adagrasib, olomorasib, divarasib, JDQ443 and RMC-6291) that were published or made available by sponsors. We defined a consensus set of eligibility criteria. We retrospectively reviewed pts with NSCLC and KRAS G12C mutations detected on universal testing of NSCLCs at Columbia University Irving Medical Center, Memorial Sloan Kettering and Weill Cornell Medicine from 2018 to 2023. Pts were re-evaluated at times of progression and last follow up. Pts were deemed trial-eligible if they met all EC, borderline if they had one laboratory value <20% from cutoff, or otherwise ineligible. Associations between demographic factors with odds of meeting eligibility criteria were determined using a multivariate logistic regression. Results: We identified 185 pts with KRAS G12C mutant advanced NSCLC who received treatment. Only 64 (35%) of these pts would have qualified for a second-line (2L) study of a KRAS G12C inhibitor. 15 (8%) had borderline eligibility and 106 (57%) were ineligible. 33/56 (60%) pts who received 2L KRAS G12C inhibitors would not have met consensus EC. Common reasons for 2L ineligibility included poor performance status (59, 49%), renal dysfunction (45, 37%), active brain metastases (33, 18%) and cytopenias (18, 15%). Age was associated with ineligibility (OR 1.07 per year, p = 0.006). Medicaid insurance was associated with a four-fold higher rate of ineligibility compared to Medicare but was not statistically significant (OR 4.86, p = 0.079). Liberalizing criteria for renal dysfunction and brain metastases would increase enrollment potential by 25% without decreasing the median overall survival of the broadened eligible cohort, whereas allowing worse performance status would decrease survival and effect sizes (1L HR 0.86 versus 0.74, p = 0.04; 2L HR 0.530 versus 0.423, p < 0.001). Conclusions: Our data indicate substantial differences between the real-world population of patients treated with KRAS G12C inhibitors and those who were trial eligible. Efforts should focus on improving clinical trial generalizability without compromising safety.