Artificial intelligence-assisted staging in Barrett’s carcinoma

Abstract

Background and study aims: Artificial intelligence (AI) is increasingly being used to detect neoplasia and interpret endoscopic images. The T stage of Barrett’s carcinoma is a major criterion for deciding on subsequent treatment. Although endoscopic ultrasound is still the standard for preoperative staging, its value is debatable. Novel tools are required to assist with staging, to optimize results. This study aimed to investigate correctly identified T stages of Barrett’s carcinoma using an artificial intelligence system, on the basis of endoscopic images. Patients and methods: 1020 images (minimum one per patient, maximum three) from 577 patients with Barrett’s adenocarcinoma were used for training and internal validation of a convolutional neuronal network. In all, 821 images were selected to train the model and 199 images were ultimately used for validation. Results: AI recognized Barrett’s mucosa without neoplasia with an accuracy of 85% (95% CI, 82.7–87.1). Mucosal cancer was identified with a sensitivity of 72% (95% CI, 67.5–76.4), specificity of 64% (95% CI, 60.0–68.4), and accuracy of 68% (95% CI, 64.6–70.7). The sensitivity, specificity, and accuracy for early Barrett’s neoplasia < T1b (sm2) were 57% (95% CI, 51.8–61.0), 77% (95% CI, 72.3–80.2), and 67% (95% CI, 63.4–69.5), respectively. More advanced stages (T3/T4) were diagnosed correctly with a sensitivity of 71% (95% CI, 65.1–76.7) and specificity of 73% (95% CI, 69.7–76.5). The overall accuracy was 73% (95% CI, 69.6–75.5). Conclusions: Clinically important decision-making on endoscopic therapy can be achieved with a high degree of certainty using artificial intelligence.

The role of artificial intelligence in the endoscopic diagnosis of esophageal cancer: a systematic review and meta-analysis

“… 34 In 2022, Knabe et al . 35 used AI to assess the tumor (T) staging of adenocarcinoma in BE in 1020 images. They were able to identify mucosal cancer with accuracy of 68% and larger T3/T4 lesions with accuracy of 73%.…”

“…They were able to identify mucosal cancer with accuracy of 68% and larger T3/T4 lesions with accuracy of 73%. 35 …”

“…Sixteen out of 24 ESCC studies and 10 out of 15 EAC studies had independent or external validation datasets to assess their AI algorithms. 11 , 12 , 14 , 16–23 , 27 , 28 , 30 , 31 , 35 , 36 , 38–40 , 42–46 Validation dataset size ranged from 52 to 2123 cases for ESCC and 9 to 199 cases for EAC. 11 , 35 , 38 , 44 …”

“… 11 , 12 , 14 , 16–23 , 27 , 28 , 30 , 31 , 35 , 36 , 38–40 , 42–46 Validation dataset size ranged from 52 to 2123 cases for ESCC and 9 to 199 cases for EAC. 11 , 35 , 38 , 44 …”

Care Delivery Interventions for Individuals with Cancer: A Literature Review and Focus on Gastrointestinal Malignancies

“…Interestingly, previous randomized controlled trials using AI computer-aided diagnosis alongside colonoscopies suggested a significantly higher detection rate for colorectal neoplasms than traditional colonoscopy [ 1 , 10 ]. Multiple studies also found that the training of a convolutional neural network to detect Barrett’s esophagus produced the same or higher accuracy rates than its human comparison [ 17 , 18 , 19 , 20 ]. With further trials to expand AI developments in endoscopy, the field of gastroenterology’s advances suggest that AI diagnostic models could be used as a supplemental tools alongside physicians’ judgments [ 21 , 22 , 23 ].…”