Predictors of Early Polymetastatic Relapse After SABR for up to 5 Oligometastases: A Secondary Analysis of the Phase II SABR-5 Trial.

Abstract

A subset of patients with oligometastatic cancer experience early widespread cancer dissemination and do not benefit from metastasis-directed therapy such as SABR. This study aimed to identify factors associated with early polymetastatic relapse (PMR). The SABR-5 trial was a single arm phase 2 study conducted at all 6 regional cancer centers across British Columbia (BC), Canada. SABR for oligometastases was only offered on trial. Patients with up to 5 oligometastatic lesions (total, progressing, or induced) received SABR to all lesions. Patients were 18 years of age or older, Eastern Cooperative Oncology Group 0 to 2 and life expectancy ≥6 months. This secondary analysis evaluated factors associated with early PMR, defined as disease recurrence within 6 months of SABR, which is not amenable to further local treatment. Univariable and multivariable analyses were performed using binary logistic regression. The Kaplan-Meier method and log-rank tests assessed PMR-free survival and differences between risk groups, respectively. Between November 2016 and July 2020, 381 patients underwent treatment on SABR-5. A total of 16% of patients experienced PMR. Worse performance status (Eastern Cooperative Oncology Group 1-2 vs 0; hazard ratio [HR] = 2.01, P = .018), nonprostate/breast histology (HR = 3.64, P <.001), and oligoprogression (HR = 3.84, P <.001) were independent predictors for early PMR. Risk groups were identified with median PMR-free survival ranging from 5 months to not yet reached at the time of analysis. Rates of 3-year overall survival were 0%, 53% (95% confidence interval [CI], 48-58), 77% (95% CI, 73-81), and 93% (95% CI, 90-96) in groups 1 to 4, respectively (P <.001). Four distinct risk groups for early PMR are identified, which differ significantly in PMR-free survival and overall survival. The group with all 3 risk factors had a median PMR-free survival of 5 months and may not benefit from local ablative therapy alone. This model should be externally validated with data from other prospective trials.