ESTRO 2025 - Abstract Book

S862

Clinical - Gynaecology

ESTRO 2025

Bologna, Italy. 10 Diagnostic and Interventional Radiology, IRCCS Istituto Ortopedico Rizzoli, Bologna, Italy. 11 Service of Radiology, Imaging Institute of Southern Switzerland, Ente Ospedaliero Cantonale (EOC), Lugano, Swaziland. 12 Medical Physics Unit, Responsible Research Hospital, Gemelli Molise Hospital—Università Cattolica del Sacro Cuore, Campobasso, Italy Purpose/Objective: To stratify treatment outcomes and survival in patients with locally advanced cervical cancer (LACC) undergoing chemoradiotherapy (CRT) using an unsupervised clustering machine learning method. Material/Methods: This retrospective study included 152 patients treated with definitive CRT, which combined external beam radiotherapy to the pelvis with intracavitary brachytherapy, achieving a total equivalent dose of 85–90 Gy at the tumor site. Pre-treatment patient-related data, including age, body mass index, standard blood tests, and complete blood count, were collected. Inflammatory indices such as Neutrophil-Lymphocyte Ratio (NLR), Platelet-Lymphocyte Ratio (PLR), Leukocyte-Lymphocyte Ratio (LLR), Systemic Immune-Inflammation Index (SII), and Aspartate Aminotransferase to Neutrophil Ratio Index (ANRI) were analyzed. An unsupervised clustering approach using the Agglomerative Hierarchical Clustering (AHC) algorithm was applied to stratify patients into clusters based on these parameters. The clusters were compared in terms of local control (LC), disease-free survival (DFS), distant metastasis-free survival (DMFS), and overall survival (OS). Results: Clustering analysis identified two distinct clusters. The analysis of variance highlighted that SII, LLR, ANRI, PLR, NLR, hemoglobin, and white cell count were the most influential variables differentiating the clusters. The centroids for the clusters were as follows: hemoglobin (12.7 vs. 10.9), white cells (7.7 vs. 11.6), NLR (2.8 vs. 10.9), PLR (156 vs. 384), LLR (4.2 vs. 11.2), ANRI (4.2 vs. 2.0), and SII (4.2 vs. 11.2) for clusters 1 and 2, respectively ( Figure 1 ). Significant differences were observed between the clusters for LC (p < 0.001) and DFS (p = 0.017), with 2-year LC and DFS rates of 93.5% and 72.0% in cluster 1, and 92.7% and 72.0% in cluster 2. No significant differences were found for DMFS or OS.( Figure 2 )

Figure 1