PD669 - AI-DERIVED CT BODY COMPOSITION ANALYSIS IDENTIFIES SARCOPENIA AND ITS CLINICAL IMPLICATIONS IN METASTATIC PANCREATIC CANCER: RESULTS FROM THE PANTHEIA-SEOM STUDY

Linked sessions

PD669

AI-DERIVED CT BODY COMPOSITION ANALYSIS IDENTIFIES SARCOPENIA AND ITS CLINICAL IMPLICATIONS IN METASTATIC PANCREATIC CANCER: RESULTS FROM THE PANTHEIA-SEOM STUDY

P. Cerda Serda1,*, A. Mariño Méndez2, F. Palmas Candia3, M. De Toro Carmena4, M. Á. Vicente Conesa5, A. Carmona-Bayonas6, M. Benavente7, P. Jimenez-Fonseca2, A. Muñoz Martin7, V. Pacheco-Barcia8 on behalf of PANTHEIA-SEOM

1Medical Oncology , Hospital Santa Creu i Sant Pau , Barcelona, 2Medical Oncology, Hospital Universitario Central de Asturias, Oviedo, 3Endocrinology, Hospital Universitario Vall d'Hebron, Barcelona, 4Medical Oncology, Hospital Universitario La Princesa, 5Medical Oncology, Hospital Universitario Morales Meseguer, Madrid, 6Medical Oncology , Hospital Universitario Morales Meseguer, Murcia, 7Medical Oncology , Hospital Universitario Gregorio Marañón, 8Medical Oncology , Hospital Universitario La Luz , Madrid, Spain

 

Rationale: Sarcopenia is a frequent but often underdiagnosed condition in metastatic pancreatic cancer. Conventional nutritional markers may fail to detect alterations. 

Methods: This retrospective observational study included patients with metastatic pancreatic cancer. Body composition was assessed from routine abdominal CT scans using the AI-based software FocusedON. Vertebral identification at the L3 level and tissue segmentation were fully automated. Cross-sectional areas (cm²) of skeletal muscle and adipose compartments were extracted from a single axial slice at L3. Statistical analyses were performed using Python (version 3.11.2). Systemic Inflammation Response Index (SIRI) measured by neutrophils x monocytes / lymphocytes.

Results: A total of 71 patients were included, median age of 72.0 years. 59.2% were female. Most patients had good performance status (ECOG 0–1: 85.9%).

Median BMI was 24.7 kg/m², with 80.3% classified as normal weight or overweight and only 4.2% underweight.

Sarcopenia was identified in 32.3% (n=22), with higher prevalence in males (p=0.001). Sarcopenic patients had lower BMI (23.0 vs. 25.6, p=0.001), despite remaining within normal ranges, indicating hidden muscle depletion.

Serum albumin was not useful for identifying sarcopenia and was paradoxically higher in sarcopenic patients (4.0 vs. 3.8 g/dL, p=0.034). Sarcopenia was also associated with a tendency to higher SIRI (median SIRI: 2362 vs. 1955) and a tendency higher incidence of severe treatment-related toxicity (grade ≥3: 40.9% vs. 23.9%; p=0.166).

Conclusion: AI-based analysis of routine CT imaging reveals clinically relevant muscle depletion. In metastatic pancreatic cancer, sarcopenia is associated with increased inflammation and higher treatment toxicity, supporting its integration into clinical risk stratification.

Disclosure of Interest: P. Cerda Serda: None declared, A. Mariño Méndez: None declared, F. Palmas Candia: None declared, M. De Toro Carmena: None declared, M. Á. Vicente Conesa: None declared, A. Carmona-Bayonas: None declared, M. Benavente: None declared, P. Jimenez-Fonseca: None declared, A. Muñoz Martin: None declared, V. Pacheco-Barcia Consultant for: Nutricia Danone, Artis Development, Speakers Bureau of: Nutricia Danone, Other: PANTHEIA-SEOM has received unrestricted grant support from Nutricia Danone, Pfizer, Servier and LEO Pharma. PANTHEIA-CK received a SEOM grant 2024.