ESTRO 2025 - Abstract Book

S1738

Clinical – Upper GI

ESTRO 2025

68

Digital Poster Skeletal muscle measured on radiotherapy planning CT scans provides predictive utility on early mortality following oesophageal chemoradiotherapy Christopher W Bleaney 1,2 , Donal McSweeney 1 , Ellen Baker 3 , Golnoosh Motamedi 1 , Lubna Bhatt 2 , Hamid Sheikh 2 , Laura Forker 1,2 , Ganesh Radhakrishna 1,2 , Alan McWilliam 1 1 Division of Cancer Sciences, Faculty of Biology Medicine and Health, University of Manchester, Manchester, United Kingdom. 2 Department of Clinical Oncology, The Christie NHS Foundation Trust, Manchester, United Kingdom. 3 School of Medicine, University of Manchester, Manchester, United Kingdom Purpose/Objective: Oesophageal cancer incidence increases with age. Accurate predictive measures of tolerability of radical treatments in the elderly are lacking. Clinical assessment of fitness using performance status and traditional body composition metrics including weight and body mass index (BMI) are not adequately discriminative. Routinely collected radiotherapy planning scans can be opportunistically interrogated to extract quantitative measures of body composition. We postulate that routinely collected clinical and radiographic data can provide predictive value in identifying elderly patients (>70 years) at risk of early mortality following chemoradiotherapy. This will aid stratification of patients to alternative treatments including hypofractionated radiotherapy alone. Material/Methods: Review of electronic health records of patients over 70 years of age treated with chemoradiotherapy between 2018 23 with at least one year of follow up was performed at a tertiary cancer centre. An in-house deep-learning model was applied to radiotherapy planning CT scans to automatically segment and quantify mean cross-sectional skeletal muscle area across three adjacent slices at the T12 vertebral level. Skeletal muscle index (SMI) was determined by normalising skeletal muscle area by patient height squared. A multivariable logistic regression was fitted to predict all-cause mortality within one year of chemoradiotherapy, accounting for clinical factors and SMI. These clinical variables were age, sex, BMI, neutrophil-lymphocyte ratio (NLR), T stage and N stage. Results: Review of 114 patients (37 female) identified 41 (36%) who completed planned chemoradiotherapy. Multivariable analysis found significant associations between SMI (Estimate: -0.1564 [95% CI -0.3113 to -0.002], p=0.048) and age (Estimate: -0.165 [95% CI -0.328 to -0.002], p=0.047) with all cause early mortality following chemoradiotherapy (Table 1; Figure 1). No other significant associations with early mortality were found. Large standard errors for the estimates of T stage are attributed to low patient numbers within these groups. The association between decreased SMI and early mortality is shown in Figure 2. Patients with low SMI at baseline, as measured from RT planning CTs, experience increased risk of early mortality following oesophageal chemoradiotherapy.