Serum Bile Acid Metabolomics-Based Machine Learning Model for Non-Invasive Diagnosis of Gastric Cancer.
Аннотация
BACKGROUND
Gastric cancer (GC) has posed a severe threat to human health. Metabolomics analysis has been applied in the context of multiple diseases, offering new perspectives for the development of non-invasive detections.
OBJECTIVE
We aimed to develop a diagnostic model for GC based on a serum bile acid (BA) panel.
METHODS
A total of 632 participants were recruited for serum samples collection, and their clinical characteristics were recorded. Targeted BA profiling was conducted using ultra-performance liquid chromatography-mass spectrometry (UPLC-MS). Participants from Xi'an were randomly allocated into a training cohort and an internal tuning cohort at a 7:3 ratio. The key BAs were identified through the two-step selection method, and nine machine learning (ML) models were constructed and optimized. The optimal model was further evaluated externally using an independent validation cohort from Lanzhou.
RESULTS
A panel of 10 serum BAs was confirmed for the construction of a diagnostic model. The XGBoost model exhibited the best performance on the tuning cohort, with an area under the receiver operating characteristic curve (AUROC) of 0.858. It also exhibited robust performance on the external validation cohort (AUROC = 0.816), significantly outperforming tumor markers (with the highest AUROC of CA72-4 = 0.68). The developed model also demonstrated satisfactory predictive performance for patients with resectable stage GC (Stage I-II) (average precision [AP] = 0.75).
CONCLUSION
GC patients exhibit BA metabolic disorders. The XGBoost model based on a 10-BA panel exhibited superior efficacy in GC detection, indicating its promising potential for clinical application.
Предметные рубрики MeSH
Подготовка и перевод публикации
Обработка статьи · этапы, ошибки и управление
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.
Тип публикации указан по данным источника. Уровень доказательности автоматически не присваивается.