European journal of surgical oncology : the journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology · 2026
The role of artificial intelligence in pre-operative prediction of completeness of cytoreduction for peritoneal surface malignancies: a scoping review.
СтатьяScoping ReviewНазвание 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.
Pau Samuel, Eglinton Timothy, Wang Alan, Fischer Jesse
Аннотация
BACKGROUND: Complete cytoreduction is the most important prognostic factor for patients with peritoneal surface malignancies (PSM) and its prediction remains a clinical challenge. Artificial intelligence (AI) offers a novel opportunity to integrate clinical and imaging features to support surgical decision-making. This scoping review aimed to analyse applications of AI for predicting cytoreduction completeness in PSM.
METHODS: A scoping review was conducted in accordance with PRISMA-ScR guidelines and registered with Open Science Framework. PubMed, Scopus, and Embase were searched from 2000 to 2025. Eligible studies applied AI models to predict cytoreduction completeness in patients with PSM using pre-operative data. Data were extracted on study design, disease type, model architecture, input predictors, performance metrics and explainability strategies.
RESULTS: From 262 records identified from the search strategy, nine studies were included. Seven focused on ovarian cancer, one on synchronous colorectal peritoneal metastases, and one on a mixed PSM. Area under the curve (AUC) values ranged from 0.70 to 0.98. Radiomics-clinical nomograms consistently outperformed single-modality models. The DeAF deep learning framework achieved the strongest multicentre validation (AUC of 0.90), underscoring the potential of deep feature extraction. However, explainability was limited to nomograms, feature importance plots, or calibration analyses; no study adopted modern explainable AI techniques.
CONCLUSION: AI models demonstrate potential for pre-operative prediction of cytoreduction completeness in PSM, particularly when radiomics are combined with clinicopathological factors or when deep learning is applied. Future research should prioritise multicentre external validation, integration of multimodal data and the adoption of explainability tools to enable clinical translation.
METHODS: A scoping review was conducted in accordance with PRISMA-ScR guidelines and registered with Open Science Framework. PubMed, Scopus, and Embase were searched from 2000 to 2025. Eligible studies applied AI models to predict cytoreduction completeness in patients with PSM using pre-operative data. Data were extracted on study design, disease type, model architecture, input predictors, performance metrics and explainability strategies.
RESULTS: From 262 records identified from the search strategy, nine studies were included. Seven focused on ovarian cancer, one on synchronous colorectal peritoneal metastases, and one on a mixed PSM. Area under the curve (AUC) values ranged from 0.70 to 0.98. Radiomics-clinical nomograms consistently outperformed single-modality models. The DeAF deep learning framework achieved the strongest multicentre validation (AUC of 0.90), underscoring the potential of deep feature extraction. However, explainability was limited to nomograms, feature importance plots, or calibration analyses; no study adopted modern explainable AI techniques.
CONCLUSION: AI models demonstrate potential for pre-operative prediction of cytoreduction completeness in PSM, particularly when radiomics are combined with clinicopathological factors or when deep learning is applied. Future research should prioritise multicentre external validation, integration of multimodal data and the adoption of explainability tools to enable clinical translation.