3372 - A CD8 T Cell Radiomics Score Correlates with Local Failure in Patients Undergoing Combined SBRT and Immune Checkpoint Inhibition in a Pooled Analysis of Three Phase I Trials

N. Goldrich, A. P. Sivananthan, M. Gutman, M. Arshad, R. R. Katipally, S. P. Pitroda, A. Juloori, S. J. Chmura, and M. C. Korpics; Department of Radiation and Cellular Oncology, University of Chicago, Chicago, IL

Purpose/Objective(s): Combined stereotactic body radiotherapy (SBRT) and immune checkpoint inhibitors (ICI) may work synergistically to improve both locoregional and systemic disease control. However, prognostic markers to identify patients most likely to benefit from combined SBRT and ICI (SBRT+ICI) are lacking. Prior work demonstrated that a validated CD8 T cell radiomics score (RS) is correlated with clinical outcomes in patients undergoing SBRT+ICI. We hypothesize this RS remains associated with clinical outcomes in an expanded cohort of patients from three phase I clinical trials using different ICIs. Materials/

Methods: Patients with metastatic solid tumors treated on three trials of multi-site SBRT with concurrent or sequential ICI were evaluated. Up to 4 metastases per patient were treated with SBRT (30-50 Gy in 3-5 fractions). ICI agents included pembrolizumab or nivolumab combined with either ipilimumab, cabiralizumab, or urelumab. A RS was calculated for each metastasis using 8 variables, including 5 radiomics features extracted from pretreatment CT scans. An average RS was calculated for each patient using all irradiated lesions. A predefined cutoff was used to stratify RS scores and groups of patients as low or high RS. Local failure (LF) of irradiated metastases was analyzed using cumulative incidence with death as a competing risk. Competing risk regression was used to evaluate the association between LF and RS. The Kaplan-Meier method was used to estimate progression free survival (PFS) and overall survival (OS), and Cox proportional hazards models were used to evaluate associations with RS.
Results: We evaluated 183 patients with 427 metastases treated with SBRT+ICI. The majority of patients had NSCLC (57%). The median follow-up was 9 months. The cumulative incidence of LF was 7.6% (95% CI 5.3-10.5%) overall, 4.5% (95% CI 2.5-7.3%) for high-RS metastases, and 16.1% (95% CI 10.0-23.5%) for low-RS metastases. Low-RS metastases had a higher risk of LF compared to high-RS metastases (HR 3.2, 95% CI 1.6 – 6.2, p=0.001). On a continuous scale, there was a decreased risk of LF with increasing RS (HR 0.1, 95% CI 0.01-0.7, p=0.016). For all patients at 12-months, PFS was 19.4% (95% CI 13.6-25.9%) and OS was 49.1% (95% CI 41.2-56.5%). Comparison of low vs. high RS groups was not significant for PFS (HR 1.1, 95% CI 0.8-1.5, p=0.758) or OS (HR 1.2, 95% CI 0.8-1.7, p=0.449). RS on a continuous scale did not correlate significantly with PFS (HR 1.1, 95% CI 0.5-2.3, p=0.895) or OS (HR 1.0, 95% CI 0.4-2.3, p=0.975).
Conclusion: Our results support the prognostic utility of this CD8 T cell RS in patients undergoing SBRT+ICI. Future work will investigate using the RS for treatment selection in patients with advanced solid tumors.