Development and Internal Validation of the AB-IPI using Bootstrapping: A Clinicopathological Prognostic Score Integrating Host Fitness and Tumor Biology in Diffuse Large B-Cell Lymphoma

Abstract Objectives The therapeutic efficacy of rituximab has reduced the discriminatory power of the International Prognostic Index (IPI) in diffuse large B-cell lymphoma (DLBCL), particularly within intermediate-risk categories. To address this “risk dilution,” we aimed to develop and internally validate the AB-IPI (Albumin-BCL2 Refined Prognostic Index) using a hypothesis-driven approach that integrates tumor burden, host fitness, and tumor biology. Methods This multi-center retrospective study analyzed 289 patients with de novo DLBCL treated uniformly with R-CHOP immunochemotherapy. We combined the standard IPI with serum albumin < 3.6 g/dL (representing host fitness/rituximab pharmacokinetics) and BCL2 protein expression > 50% (representing tumor biology). The model was validated internally using bootstrapping with 1,000 resamples in accordance with TRIPOD Type 1b guidelines. This study adhered to the TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis) statement for model development and internal validation (Type 1b). Results During the observation period, 115 death events were recorded. Multivariate Cox regression identified albumin < 3.6 g/dL (Hazard Ratio 2.62), IPI score > 2 (HR 2.13), and BCL2 > 50% (HR 1.72) as independent prognostic factors. The model maintained a robust Events Per Variable (EPV) ratio of 38.3. The AB-IPI stratified patients into four distinct risk groups with 5-year overall survival rates of 88.0% (Low), 76.1% (Intermediate-1), 45.0% (Intermediate-2), and 29.0% (High). The calibration plot demonstrated excellent agreement between predicted and observed probabilities, with a calibration slope of 0.98, indicating minimal optimism and robust risk estimation. Decision Curve Analysis (DCA) demonstrated that the AB-IPI provided a superior Net Benefit across a wide range of clinically relevant threshold probabilities. Conclusions The AB-IPI demonstrates superior clinical utility and calibration compared to the standard IPI. By identifying patients with compounded biological risks who are unlikely to be cured by R-CHOP alone, this score offers a practical framework for optimizing therapeutic strategies, such as the allocation of polatuzumab vedotin.