aleqabiayat-del/NHS-PREDICT-Breast-v3-Egypt-Calibration

Validation and calibration of the NHS PREDICT Breast v3.0 model using data from an Egyptian patient cohort to improve breast cancer risk prediction accuracy # NHS-PREDICT-Breast-v3-Egypt-Calibration Validation and calibration of the NHS PREDICT Breast v3.0 model using data from an Egyptian patient cohort to improve breast cancer risk prediction accuracy ## Executive Summary **Project Goal:** This study aims to address this gap by rigorously validating and calibrating the NHS PREDICT Breast v3.0 model in a contemporary cohort of Egyptian patients. We hypothesize that while the model may show some discriminatory power, its calibration may be suboptimal due to differences in tumour biology, prevalence of risk factors, and treatment patterns unique to the Egyptian population. A successful validation will provide a robust, evidence-based tool for Egyptian oncologists, potentially improving clinical decision-making and patient counselling. Conversely, if the model proves to be poorly calibrated, it will highlight the need for either recalibration or the development of a more locally-relevant prognostic model. **Key Finding:** The model showed **moderate discriminatory power** (C-index = 0.684) and exhibited **suboptimal calibration**, particularly by overestimating mortality (being too pessimistic) in moderate-risk groups. This highlights the critical need for recalibration or development of locally relevant models. --- ## Methodology and Cohort (Methods) ### Research Question and Hypothesis * **Research Question:** How well does the NHS PREDICT Breast v3.0 model maintain its predictive performance when applied to an Egyptian breast cancer cohort? * **Hypothesis:** We hypothesize that while the model may show some discriminatory power, its calibration may be suboptimal due to differences in tumour biology, prevalence of risk factors, and treatment patterns unique to the Egyptian population. ### Study Design and Analysis * **Study Type:** Single-center, retrospective cohort study. * **Cohort:** 964 patients with non-metastatic invasive breast cancer treated at Cairo University Hospitals (2000–2018). * **Primary Outcom …