Centro Hospitalar de São João Porto, Porto, Portugal
Tiago Ribeiro, MD1, Miguel Mascarenhas, MD1, Joao Afonso, MD1, João Ferreira, PhD2, Filipe Vilas Boas, MD1, Marco Parente, PhD2, Renato Natal, PhD2, Pedro Pereira, MD1, Guilherme Macedo, MD, PhD, FACG1 1Centro Hospitalar de São João, Porto, Porto, Portugal; 2University of Porto, Porto, Porto, Portugal
Introduction: The diagnosis and characterization of biliary strictures is challenging. The introduction of digital single-operator cholangioscopy (D-SOC) allowing direct visual inspection of the lesion and targeted biopsies significantly improved the diagnostic yield in patients with indeterminate biliary strictures. However, the diagnostic efficiency of D-SOC remains suboptimal.
Convolutional neural networks (CNNs) are a type of artificial intelligence (AI) architecture resembling the human visual cortex with high performance for automatic image analysis. Studies evaluating the application of these algorithms for the interpretation of endoscopic images has provided promising results. However, to date, the impact of these technological advances on cholangioscopy has not been assessed. We aimed to develop a CNN-based system for automatic detection of malignant biliary strictures in D-SOC images.
Methods: We developed, trained, and validated a CNN based on D-SOC images. Each frame was labeled as normal/benign findings or as a malignant lesion if histopathological evidence of biliary malignancy was available (either in biopsy or surgical specimens). The entire dataset was split for 5-fold cross validation. Also, the image dataset was split for constitution of training and validation datasets. The classification provided by the CNN was compared with the labelling of the lesion (Figura 1a). The performance of the CNN was measured by calculating the area under the receiving operating characteristic curve (AUC), sensitivity, specificity, positive and negative predictive values (PPV and NPV, respectively).
Results: A total of 11855 images from 85 patients were included (9695 of malignant strictures and 2160 of benign findings). The model had an overall accuracy of 94.9% (Figure 1b), a sensitivity of 94.7%, a specificity of 92.1% and an AUC of 1.00 (Figure 1c). The image processing speed of the CNN was 7 ms/frame.
Discussion: The developed deep learning algorithm accurately detected and differentiated malignant strictures from benign biliary conditions. The introduction of artificial intelligence algorithms to D-SOC systems may significantly increase its diagnostic yield for malignant strictures.
Figure: Figure 1: 1a - Output obtained from the application of the CNN. A blue bar represents a correct prediction. A red bar represents an incorrect prediction by the CNN. 1b - Evolution of the accuracy in differentiating malignant from benign biliary strictures by the CNN during training and validation phases, as the training and validation datasets were repeatedly inputted in the neural network. 1c - ROC analyses of the network’s performance. AUC: area under the curve; CNN: convolutional neural network; B: benign finding; M: malignant stricture.
Disclosures: Tiago Ribeiro indicated no relevant financial relationships. Miguel Mascarenhas indicated no relevant financial relationships. Joao Afonso indicated no relevant financial relationships. João Ferreira indicated no relevant financial relationships. Filipe Vilas Boas indicated no relevant financial relationships. Marco Parente indicated no relevant financial relationships. Renato Natal indicated no relevant financial relationships. Pedro Pereira indicated no relevant financial relationships. Guilherme Macedo indicated no relevant financial relationships.
Tiago Ribeiro, MD1, Miguel Mascarenhas, MD1, Joao Afonso, MD1, João Ferreira, PhD2, Filipe Vilas Boas, MD1, Marco Parente, PhD2, Renato Natal, PhD2, Pedro Pereira, MD1, Guilherme Macedo, MD, PhD, FACG1. P1086 - Artificial Intelligence for Automatic Diagnosis of Biliary Strictures Malignancy Status in Single-Operator Cholangioscopy: A Proof-of-Concept Study, ACG 2021 Annual Scientific Meeting Abstracts. Las Vegas, Nevada: American College of Gastroenterology.