Breast (Edinburgh, Scotland) · 2026
Belgian validation of the INFLUENCE 3.0 model predicting 5-year locoregional and contralateral breast cancer recurrence.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Vorst Aline van der, Grimberg Carlijn, van Maaren Marissa C, Hueting Tom A, Weltens Caroline, Siesling Sabine, Groothuis-Oudshoorn Catharina G M
Аннотация
BACKGROUND: Breast cancer (BC) surveillance primarily aims to detect locoregional recurrence (LRR) and contralateral breast cancer (CBC), but current strategies are largely one-size-fits-all and not tailored to individual risks. The Dutch INFLUENCE 3.0 prediction model provides personalised 5-year LRR and CBC risks to support shared decision-making on surveillance. In this study, INFLUENCE 3.0 is externally validated in a Belgian population.
METHODS: All female BC patients curatively treated between 2000 and 2018 at the University Hospitals of Leuven were included. Calibration was assessed using the Integrated Calibration Index (ICI), E50, E90 metrics, and calibration plots. Discrimination was evaluated by calculating the Area Under the Curve (AUC) for each year, based on cumulative events up to that year.
RESULTS: The final Leuven cohort included 8548 patients, of whom 7549 were not treated with neoadjuvant systemic therapy (non-NST) and 999 were (NST). Calibration metrics were generally acceptable, with ICI and E50 values below 0.01 in most subgroups. Calibration plots showed overestimation of predicted risks, particularly at lower predicted risk levels. Discrimination for LRR risk was good to very good (AUC 0.72-0.86 in the non-NST group; >0.82 in the NST group). For CBC risk, discrimination was poor in the non-NST group (AUC ≈ 0.5) and moderate-to-very good in the NST group (AUC 0.68-0.88).
CONCLUSION: INFLUENCE 3.0 showed acceptable calibration and discrimination for LRR risk, and although CBC risk discrimination in the non-NST group was poor. The overall risk estimates - primarily the LRR risks - may still help in guiding tailored surveillance schemes.
METHODS: All female BC patients curatively treated between 2000 and 2018 at the University Hospitals of Leuven were included. Calibration was assessed using the Integrated Calibration Index (ICI), E50, E90 metrics, and calibration plots. Discrimination was evaluated by calculating the Area Under the Curve (AUC) for each year, based on cumulative events up to that year.
RESULTS: The final Leuven cohort included 8548 patients, of whom 7549 were not treated with neoadjuvant systemic therapy (non-NST) and 999 were (NST). Calibration metrics were generally acceptable, with ICI and E50 values below 0.01 in most subgroups. Calibration plots showed overestimation of predicted risks, particularly at lower predicted risk levels. Discrimination for LRR risk was good to very good (AUC 0.72-0.86 in the non-NST group; >0.82 in the NST group). For CBC risk, discrimination was poor in the non-NST group (AUC ≈ 0.5) and moderate-to-very good in the NST group (AUC 0.68-0.88).
CONCLUSION: INFLUENCE 3.0 showed acceptable calibration and discrimination for LRR risk, and although CBC risk discrimination in the non-NST group was poor. The overall risk estimates - primarily the LRR risks - may still help in guiding tailored surveillance schemes.