Predicting Pathologic Response in Locally Advanced Rectal Cancer Using Inflammatory, Nutritional, and...

Area of Science:

• Oncology
• Rectal Cancer Research
• Predictive Modeling in Medicine

Background:

• Total neoadjuvant therapy (TNT) is standard for locally advanced rectal cancer (LARC), but pathological complete response (pCR) rates vary.
• Systemic inflammation, nutritional status, and sarcopenia are known prognostic factors, but integrated predictive models are lacking.
• Predicting treatment response is crucial for optimizing LARC management and patient outcomes.

Purpose of the Study:

• To develop and validate clinical, laboratory, and AI-based models for predicting pathological response in LARC patients undergoing TNT.
• To assess the predictive value of inflammatory markers (CAR, SII), sarcopenia, and clinical factors.
• To establish composite scores (CINR) and machine learning models for risk stratification.

Main Methods:

• Retrospective analysis of 93 LARC patients treated with TNT followed by surgery.
• Assessment of sarcopenia via CT, and inflammatory/nutritional status using C-reactive protein/albumin ratio (CAR) and systemic immune-inflammation index (SII).
• Development of composite CINR scores and Random Forest (RF) models to predict pCR and good pathological response (TRG 0-1).

Main Results:

• Pathological complete response (pCR) was achieved in 21.5% and good pathological response (TRG 0-1) in 46.2% of patients.
• Predictors for pCR included absence of post-TNT sarcopenia, low CAR, low SII, low LDH, and metformin use.
• CINR scores (AUCs 0.846-0.868) and RF models (AUCs 0.910-0.933) demonstrated strong predictive performance for pathological response.

Conclusions:

• Integrated inflammatory, nutritional, and sarcopenia markers, along with CINR scores and AI models, accurately predict pathological response in LARC.
• The developed models and cut-off values can stratify patients into risk groups, aiding clinical decision-making for organ preservation and surgical timing.
• Prospective multicenter validation is recommended to confirm these findings and support broader clinical application.