MedElevator
Войти в исследования Journal of gastrointestinal cancer · 2025
MRI and PET-Based Machine Learning Radiomics for Metastasis Prediction in Pancreatic Ductal Adenocarcinoma: A Systematic Review.
СтатьяСистематический обзорНазвание 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.
Elhaie Mohammadreza, Koozari Abolfazl, Koohi Hossein
Аннотация
BACKGROUND: Pancreatic ductal adenocarcinoma (PDAC) is an aggressive malignancy with poor survival, driven in part by early metastatic spread. Conventional imaging lacks sufficient precision to predict metastasis accurately. Machine learning (ML)-based radiomics, integrating quantitative imaging features from modalities such as magnetic resonance imaging (MRI) and positron emission tomography (PET), may enhance prognostic accuracy.
OBJECTIVE: To systematically review the diagnostic accuracy and clinical utility of ML-based radiomics models for predicting metastasis in PDAC.
METHODS: A systematic search of PubMed, Embase, Scopus, Web of Science, and Cochrane Library was conducted according to PRISMA 2020 guidelines (PROSPERO: CRD420251109941). Eligible studies applied ML-based radiomics to MRI, PET, or combined MRI/PET for metastasis prediction in histologically or clinically confirmed PDAC. Data were extracted on study design, patient characteristics, imaging protocols, feature selection, ML algorithms, performance metrics, and validation strategies. Methodological quality was assessed using QUADAS-2. Certainty of evidence was graded using the GRADE framework.
RESULTS: Seven studies met inclusion criteria and were included in the Systematic Review. MRI-based models were the most common, with one multimodal PET/MRI study. Risk of bias was moderate overall, primarily due to retrospective designs and variable reference standards. GRADE certainty was low for pooled diagnostic accuracy and very low for PET/MRI evidence due to imprecision and suspected publication bias.
CONCLUSIONS: ML-based radiomics demonstrates promising accuracy for metastasis prediction in PDAC, particularly with MRI and PET/MRI modalities. Integration with clinical biomarkers further enhances predictive value. However, methodological limitations and low certainty of evidence warrant prospective, multicenter validation with standardized protocols before clinical adoption.
OBJECTIVE: To systematically review the diagnostic accuracy and clinical utility of ML-based radiomics models for predicting metastasis in PDAC.
METHODS: A systematic search of PubMed, Embase, Scopus, Web of Science, and Cochrane Library was conducted according to PRISMA 2020 guidelines (PROSPERO: CRD420251109941). Eligible studies applied ML-based radiomics to MRI, PET, or combined MRI/PET for metastasis prediction in histologically or clinically confirmed PDAC. Data were extracted on study design, patient characteristics, imaging protocols, feature selection, ML algorithms, performance metrics, and validation strategies. Methodological quality was assessed using QUADAS-2. Certainty of evidence was graded using the GRADE framework.
RESULTS: Seven studies met inclusion criteria and were included in the Systematic Review. MRI-based models were the most common, with one multimodal PET/MRI study. Risk of bias was moderate overall, primarily due to retrospective designs and variable reference standards. GRADE certainty was low for pooled diagnostic accuracy and very low for PET/MRI evidence due to imprecision and suspected publication bias.
CONCLUSIONS: ML-based radiomics demonstrates promising accuracy for metastasis prediction in PDAC, particularly with MRI and PET/MRI modalities. Integration with clinical biomarkers further enhances predictive value. However, methodological limitations and low certainty of evidence warrant prospective, multicenter validation with standardized protocols before clinical adoption.