European journal of surgical oncology : the journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology · 2026
Development and external validation of an AI-guided navigation system for the safe dissection plane in robot-assisted nipple sparing mastectomy.
СтатьяВалидационное исследованиеНазвание 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.
Park Woong Ki, Oh Namkee, Choi Gyu-Seong, Kim Bogeun, Kim Ki Jo, Choi Hee Jun, Rhu Jinsoo, Ryu Jai Min
Аннотация
PURPOSE: Robot-assisted nipple-sparing mastectomy (RANSM) has gained acceptance in selected patients; however, identifying the safe dissection plane remains technically challenging due to the absence of tactile feedback. Artificial intelligence (AI)-guided navigation may provide intraoperative assistance, yet no externally validated model has been reported for this procedure.
MATERIALS AND METHODS: This retrospective study developed and validated an AI-guided navigation system to identify the safe dissection plane during RANSM. Surgical video data from 37 procedures performed between January 2021 and December 2024 at two tertiary centers in South Korea were analyzed (internal dataset, n = 29; external dataset, n = 8). The safe dissection plane was annotated as the visual boundary between subcutaneous fat and glandular tissue. An AI segmentation model was trained using 5-fold cross-validation on the internal dataset and tested on the independent external dataset. Model performance was assessed using the Dice Similarity Coefficient (DSC), with intersection over union (IOU), sensitivity, precision, and specificity as secondary metrics.
RESULTS: A total of 1996 internal and 293 external frames were analyzed. The model achieved a mean DSC of 74.0 % (±1.5 %), IOU: 60.0 % (±1.8 %), sensitivity: 79.7 % (±1.9 %), and precision: 71.5 % (±1.6 %) in internal validation. On external validation, the DSC was 70.8 %, IOU: 55.9 %, sensitivity: 73.1 %, precision: 72.2 %, and specificity: 96.8 %.
CONCLUSION: This study is the first to develop and externally validate an AI-guided navigation system for RANSM. The model demonstrated consistent performance across two institutions, suggesting potential to enhance surgical precision and safety. Larger prospective studies are warranted to confirm clinical utility.
MATERIALS AND METHODS: This retrospective study developed and validated an AI-guided navigation system to identify the safe dissection plane during RANSM. Surgical video data from 37 procedures performed between January 2021 and December 2024 at two tertiary centers in South Korea were analyzed (internal dataset, n = 29; external dataset, n = 8). The safe dissection plane was annotated as the visual boundary between subcutaneous fat and glandular tissue. An AI segmentation model was trained using 5-fold cross-validation on the internal dataset and tested on the independent external dataset. Model performance was assessed using the Dice Similarity Coefficient (DSC), with intersection over union (IOU), sensitivity, precision, and specificity as secondary metrics.
RESULTS: A total of 1996 internal and 293 external frames were analyzed. The model achieved a mean DSC of 74.0 % (±1.5 %), IOU: 60.0 % (±1.8 %), sensitivity: 79.7 % (±1.9 %), and precision: 71.5 % (±1.6 %) in internal validation. On external validation, the DSC was 70.8 %, IOU: 55.9 %, sensitivity: 73.1 %, precision: 72.2 %, and specificity: 96.8 %.
CONCLUSION: This study is the first to develop and externally validate an AI-guided navigation system for RANSM. The model demonstrated consistent performance across two institutions, suggesting potential to enhance surgical precision and safety. Larger prospective studies are warranted to confirm clinical utility.