Artificial intelligence generated 3D body composition predicts dose modifications in patients undergoing neoadjuvant chemotherapy for rectal cancer
- PMID: 40379920
- PMCID: PMC12084224
- DOI: 10.1007/s00432-025-06219-5
Artificial intelligence generated 3D body composition predicts dose modifications in patients undergoing neoadjuvant chemotherapy for rectal cancer
Abstract
Purpose: Chemotherapy administration is a balancing act between giving enough to achieve the desired tumour response while limiting adverse effects. Chemotherapy dosing is based on body surface area (BSA). Emerging evidence suggests body composition plays a crucial role in the pharmacokinetic and pharmacodynamic profile of cytotoxic agents and could inform optimal dosing. This study aims to assess how lumbosacral body composition influences adverse events in patients receiving neoadjuvant chemotherapy for rectal cancer.
Methods: A retrospective study (February 2013 to March 2023) examined the impact of body composition on neoadjuvant treatment outcomes for rectal cancer patients. Staging CT scans were analysed using a validated AI model to measure lumbosacral skeletal muscle (SM), intramuscular adipose tissue (IMAT), visceral adipose tissue (VAT), and subcutaneous adipose tissue volume and density. Multivariate analyses explored the relationship between body composition and chemotherapy outcomes.
Results: 242 patients were included (164 males, 78 Females), median age 63.4 years. Chemotherapy dose reductions occurred more frequently in females (26.9% vs. 15.9%, p = 0.042) and in females with greater VAT density (-82.7 vs. -89.1, p = 0.007) and SM: IMAT + VAT volume ratio (1.99 vs. 1.36, p = 0.042). BSA was a poor predictor of dose reduction (AUC 0.397, sensitivity 38%, specificity 60%) for female patients, whereas the SM: IMAT + VAT volume ratio (AUC 0.651, sensitivity 76%, specificity 61%) and VAT density (AUC 0.699, sensitivity 57%, specificity 74%) showed greater predictive ability. Body composition didn't influence dose adjustment of male patients.
Conclusion: Lumbosacral body composition outperformed BSA in predicting adverse events in female patients with rectal cancer undergoing neoadjuvant chemotherapy.
Keywords: Artificial intelligence; Body composition; Chemotherapy; Dose modification; Rectal cancer.
© 2025. The Author(s).
Conflict of interest statement
Declarations. Disclosure: The authorship have no disclosures to declare. Competing interests: The authors declare no competing interests.
Figures
References
-
- Aslani A, Smith RC, Barry J, Allen N, Pavlakis, John AL (2000) The predictive value of body protein for chemotherapy-induced toxicity. Cancer 88:796–803 - PubMed
-
- Bedrikovetski S, Traeger L, Seow W, Dudi-Venkata NN, Selva-Nayagam S, Penniment M, Sammour T (2024) Oncological outcomes and response rate after total neoadjuvant therapy for locally advanced rectal cancer: A network Meta-Analysis comparing induction vs. Consolidation chemotherapy vs. Standard chemoradiation. Clin Colorectal Cancer 23:326–36e9 - PubMed
-
- Besson A, Deftereos I, Gough K, Taylor D, Shannon R, Justin MY (2021) The association between sarcopenia and quality of life in patients undergoing colorectal cancer surgery: an exploratory study. Support Care Cancer 29:3411–3420 - PubMed
MeSH terms
LinkOut - more resources
Full Text Sources