Utilisation of intramuscular and intermuscular fat to develop a new skeletal muscle grading score which can predict treatment outcomes for locally advanced rectal cancer

Abstract Background Sarcopenia has been widely studied in rectal cancer with increasing evidence to suggest that other body composition parameters, in particular adipose tissue, have an important role. Advances in artificial intelligence (AI) now allow 3D body composition analysis of intermuscular/intramuscular adipose tissue (IMAT) from CT scans. This study aimed to develop and evaluate a skeletal muscle score (SMS), utilising skeletal muscle (SM) and IMAT measurements, to predict treatment response and survival outcomes for rectal cancer patients. Methods A retrospective analysis was performed on 226 patients with localised rectal adenocarcinoma treated at Western Health between 2013 and 2024. Body composition metrics, including SM and IMAT volume and density from the L1-S5 vertebral region, were extracted using validated AI software. A SMS (0–4) was developed to predict overall complete response (oCR). The primary endpoint was oCR, defined as pathological complete response or sustained clinical complete response for at least 3 years. Secondary outcomes included overall, cancer-specific, and disease-free survival. Results An oCR was achieved in 25.7% of patients and was significantly associated with a lower MRI T stage, increased age at diagnosis, and a better SMS, whilst active smoking decreased oCR in a multivariable analysis. Patients with an SMS of zero had a 0% oCR rate, whilst patients with a SMS of four had oCR rate of 60%. A higher SMS correlated with improved overall, cancer-specific, and disease-free survival. Conclusion The SMS is a novel, AI-derived body composition assessment that is strongly correlated with treatment response and survival in rectal cancer patients. This scoring system could provide clinicians with individualised risk stratification to enhance patient counselling.