Nivolumab in patients with advanced solid tumors with ctDNA-defined high blood tumor mutational burden: a phase II trial

Background: Tissue-based tumor mutational burden (TMB) is a biomarker for immune checkpoint inhibitors; however, data on circulating tumor DNA (ctDNA)-based blood TMB (bTMB) are limited. Patients and methods: We conducted an amended phase II basket study enrolling patients with gastrointestinal (GI) cancers with high bTMB identified using Guardant360® CDx (cohorts 1 and 2) and patients with GI cancers with ultrahigh bTMB identified using Guardant360® CDx (Ultrahigh TMB GI cohort; top 2% within each GI tumor type). After early discontinuation of the Ultrahigh TMB GI cohort, the bTMB-High cohort enrolled patients with solid tumors across cancer types with bTMB score ≥14, as measured by FoundationOne® Liquid CDx. Nivolumab 360 mg was administered intravenously every 3 weeks until progression or unacceptable toxicity. The primary endpoint was objective response rate (ORR) and hypothesis testing was prespecified for the bTMB-High cohort. Results: Seventy patients were treated (cohort 1, n = 28; cohort 2, n = 4; Ultrahigh TMB GI cohort, n = 19; bTMB-High cohort, n = 19). The ORR was 14.3%, 0%, 21.1%, and 21.1%, respectively. In the bTMB-High cohort, the lower bound of the 90% confidence interval (CI) exceeded the prespecified 5% threshold, meeting the primary endpoint. Treatment-related grade ≥3 adverse events occurred in 8.6% of patients. Early ctDNA decline at week 3 was associated with longer progression-free survival (hazard ratio 0.51, 95% CI 0.29-0.87). Baseline bTMB did not clearly distinguish responders from nonresponders, whereas higher baseline ctDNA concentrations were associated with a lower ORR. Conclusions: Nivolumab showed activity in patients with ctDNA-defined bTMB-High cancers. Absolute bTMB alone did not reliably identify benefits, whereas baseline ctDNA concentrations and early ctDNA decline provided additional insight into response. These findings provide a preliminary signal supporting the prospective evaluation of ctDNA-based bTMB selection in a tumor-agnostic setting and warrant further validation in larger, randomized trials. © 2026 The Author(s)