Current problems in cancer · 2025
Breast cancer neoadjuvant therapy outcome prediction based on clinical patient and tumor features: A cross-sectional study.
СтатьяНазвание и аннотация: Перевод в очереди
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.
Brenner Eva, Bulić Luka, Milković-Periša Marija
Аннотация
INTRODUCTION: Breast cancer is the most common malignant disease in the female population and one of the most common diseases in developed countries. Many factors which may impact the development and outcome of this complex disease have been investigated. The aim of this study was to analyze factors that affect neoadjuvant therapy outcomes and create an outcome prediction model based on these factors.
MATERIALS AND METHODS: Patient data was collected from all patients who underwent breast cancer neoadjuvant therapy at our clinical center from 2018 to 2022. Statistical analysis entailed the identification of patient and tumor features that are significantly associated with RCB index values, using Spearman's correlation coefficient, the Mann-Whitney U-test, and the one-way ANOVA and Kruskal-Wallis test. Significant features were selected and used for the training of a machine-learning model based on the random forest algorithm.
RESULTS: Regarding patient features, age, BMI, and previous history of malignant disease were found significantly associated with the RCB index. Significant tumor features included focality, nuclear grade, immunophenotype, positivity for estrogen receptors, progesterone receptors and HER-2, Ki-67 value, and presence of lymphovascular invasion. Based on these features, a predictive model was created with an accuracy of 80 % and ROC-AUC value of 0.83.
CONCLUSION: The discovered significant features are mostly in line with the published literature. While our predictive model yielded promising results, its training was limited by the number of patients and availability of data. Further research and the creation of more accurate predictive models might facilitate further personalization and improvement of breast cancer neoadjuvant treatment.
MATERIALS AND METHODS: Patient data was collected from all patients who underwent breast cancer neoadjuvant therapy at our clinical center from 2018 to 2022. Statistical analysis entailed the identification of patient and tumor features that are significantly associated with RCB index values, using Spearman's correlation coefficient, the Mann-Whitney U-test, and the one-way ANOVA and Kruskal-Wallis test. Significant features were selected and used for the training of a machine-learning model based on the random forest algorithm.
RESULTS: Regarding patient features, age, BMI, and previous history of malignant disease were found significantly associated with the RCB index. Significant tumor features included focality, nuclear grade, immunophenotype, positivity for estrogen receptors, progesterone receptors and HER-2, Ki-67 value, and presence of lymphovascular invasion. Based on these features, a predictive model was created with an accuracy of 80 % and ROC-AUC value of 0.83.
CONCLUSION: The discovered significant features are mostly in line with the published literature. While our predictive model yielded promising results, its training was limited by the number of patients and availability of data. Further research and the creation of more accurate predictive models might facilitate further personalization and improvement of breast cancer neoadjuvant treatment.