Biophotonics discovery · 2025
Identification of colorectal malignancies enabled by phasor-based autofluorescence lifetime macroimaging and ensemble learning.
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.
Lagarto João L, Herrando Alberto I, Rego Rafaela, Fernández Laura, Azevedo José, Domingos Hugo, Vieira Pedro, Parvaiz Amjad, Shcheslavskiy Vladislav I, Silva Pedro G, Castillo-Martin Mireia
Аннотация
SIGNIFICANCE: Colorectal cancer (CRC) remains one of the most frequent cancers and a leading contributor to cancer-associated mortality globally. CRCs are often diagnosed at an advanced stage, which leads to high mortality and morbidity. This outcome is exacerbated by high rates of recurrence and postoperative complications that contribute substantially to poor prognosis. Advancements in endoscopic assessment have improved CRC prevention, early detection, and surveillance over the years. Yet, CRC remains one of the most significant health challenges of the 21st century. Label-free optical spectroscopy methods have long been explored as potential partners to endoscopy, not only to enhance diagnostic accuracy but also to confer predictive capabilities to endoscopic evaluations.
AIM: We investigated the potential of time-resolved autofluorescence measurements excited at 375 and 445 nm to correctly classify benign and malignant tissues in CRC surgical specimens from 117 patients.
APPROACH: Multiparametric autofluorescence lifetime data were collected in two distinct datasets, which were used for training ( n = 73 ) and testing ( n = 44 ) a supervised ensemble learning classification model, with standard histopathology assessment serving as ground truth.
RESULTS: Using 5-fold cross-validation, we achieved 82.6 ± 0.02 % sensitivity, 90.4 ± 0.01 % specificity, 87.4 ± 0.01 % accuracy, and 0.941 ± 0.004 area under the curve (AUC) for training data. Evaluation of unseen test data yielded similar results, with 85.2% sensitivity, 84.5% specificity, 84.8% accuracy, and 0.915 AUC.
CONCLUSIONS: Although preliminary, our findings underscore the potential impact of AI-assisted autofluorescence lifetime measurements in advancing CRC prevention, early detection, and surveillance efforts.
AIM: We investigated the potential of time-resolved autofluorescence measurements excited at 375 and 445 nm to correctly classify benign and malignant tissues in CRC surgical specimens from 117 patients.
APPROACH: Multiparametric autofluorescence lifetime data were collected in two distinct datasets, which were used for training ( n = 73 ) and testing ( n = 44 ) a supervised ensemble learning classification model, with standard histopathology assessment serving as ground truth.
RESULTS: Using 5-fold cross-validation, we achieved 82.6 ± 0.02 % sensitivity, 90.4 ± 0.01 % specificity, 87.4 ± 0.01 % accuracy, and 0.941 ± 0.004 area under the curve (AUC) for training data. Evaluation of unseen test data yielded similar results, with 85.2% sensitivity, 84.5% specificity, 84.8% accuracy, and 0.915 AUC.
CONCLUSIONS: Although preliminary, our findings underscore the potential impact of AI-assisted autofluorescence lifetime measurements in advancing CRC prevention, early detection, and surveillance efforts.