ESTRO 2024 - Abstract Book

S440

Clinical - Biomarkers

ESTRO 2024

clinical outcomes of patients treated with immunotherapy (5). In a cohort of 97 patients treated with immunotherapy and radiotherapy (RTIO cohort), we have shown that the CD8 radiomics score may help to predict and identify lesions which are at higher risk of progression (”cold tumors”) (6), which has been confirmed in a second independent radiomic study in 136 melanoma patients (Melanoma cohort) (7).

The objective of this study is to assess whether the CD8 radiomics score applied on liver metastasis may help to assess patient prognosis in a post-hoc analysis of these two studies (RTIO and Melanoma cohorts).

Material/Methods:

Data from the two previously published studies were retrospectively reviewed.

Melanoma cohort consisted of 136 advanced melanoma patients included in our center treated with anti-PD1 based immunotherapy. The RTIO cohort consisted of 94 patients with advanced solid tumors treated in six independent clinical studies of immuno-radiotherapy combinations, most patients receiving stereotactic RT to one lesion (3 x 8 Gy). CD8 Radiomics score of delineated liver metastases were retrieved. This score was used to dichotomized patient according to the level of CD8 T cell estimated in the liver metastases. Cold liver metastases referred to liver metastases with low CD8 radiomic score. For patients having several analyzed liver metastases, patients with no cold liver metastasis were in the “hot liver metastases” group. The cutoff to define hot vs cold was the one refined and published in the Melanoma cohort. Since no cutoff were predefined in the RTIO cohort, median value was used to dichotomize patients. In the Melanoma cohort, 46 patients (33.8%) presented with liver metastases, for whom 186 liver metastases were delineated (Median per patient= 3.00, IQR[1.00-4.75]). The CD8 radiomic score of liver metastases was not extracted for 12 patients whose liver metastases were smaller than 1 mL. In the RTIO cohort, 32 patients (34%) presented with liver metastases. The CD8 radiomic score of liver metastases was available for 30 patients, accounting for 112 liver metastases (Median: 2.00, IQR[2.00-2.00]). Radiotherapy was delivered to a liver metastasis for 14 of them. Overall survival (OS) of patients without liver lesion was better than the OS of patients with cold liver lesions in both cohorts (HR=0.07, 95%CI [0.02-0.21], p=0.0003 in the melanoma cohort, and HR=0.42, 95%CI [0.19-0.94], p=0.0498, in the RTIO cohort, respectively). However, it was not significantly different than the OS of patients with hot liver lesions in both cohorts (HR=0.99, 95%CI [0.53-1.8], p=0.97 in the melanoma cohort, and HR=0.76, 95%CI [0.31-1.9], p=0.56, in the RTIO cohort, respectively) (FIGURE 1). Results:

Figure 1. Overall survival of patients in the Melanoma cohort (A,B) and RTIO cohort (C,D) according to the presence of cold liver lesions (A,C) or hot liver lesion (B,D) compared to the absence of liver lesions.