A single-center prospective study evaluating the usefulness of artificial intelligence for the diagnosis of esophageal squamous cell carcinoma in a real-time setting

Abstract

Background Several pre-clinical studies have reported the usefulness of artificial intelligence (AI) systems in the diagnosis of esophageal squamous cell carcinoma (ESCC). We conducted this study to evaluate the usefulness of an AI system for real-time diagnosis of ESCC in a clinical setting. Methods This study followed a single-center prospective single-arm non-inferiority design. Patients at high risk for ESCC were recruited and real-time diagnosis by the AI system was compared with that of endoscopists for lesions suspected to be ESCC. The primary outcomes were the diagnostic accuracy of the AI system and endoscopists. The secondary outcomes were sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and adverse events. Results A total of 237 lesions were evaluated. The accuracy, sensitivity, and specificity of the AI system were 80.6%, 68.2%, and 83.4%, respectively. The accuracy, sensitivity, and specificity of endoscopists were 85.7%, 61.4%, and 91.2%, respectively. The difference between the accuracy of the AI system and that of the endoscopists was − 5.1%, and the lower limit of the 90% confidence interval was less than the non-inferiority margin. Conclusions The non-inferiority of the AI system in comparison with endoscopists in the real-time diagnosis of ESCC in a clinical setting was not proven. Trial registration Japan Registry of Clinical Trials (jRCTs052200015, 18/05/2020).

Efficiency of endoscopic artificial intelligence in the diagnosis of early esophageal cancer

“…Finally, 19 papers relating to AI‐assisted diagnosis of early EC and its invasion depth were included in the study. 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 By analyzing endoscopic WLI/NBI, the sensitivity and specificity of AI‐assisted systems, novices, and experts in the diagnosis of early EC and its infiltration depth were compared in these studies, respectively. The literature search process is shown in Figure 1 , and the basic information of each study was extracted (Table 1 ).…”

“…This enables the AI system to identify ECs of different types and degrees of condition, including early lesions, thus improving the comprehensiveness of diagnosis. 17 In addition, the AI system has a unique advantage in image analysis, which allows it to quickly and accurately analyze a large number of endoscopic images and effectively detect tiny lesions or abnormalities, thus helping in the timely detection of early EC. Yang et al 21 found that the AI system, trained through simulation, diagnosed early EC with an accuracy of 88.1% in the evaluation of 1097 WLI, which was higher than that of experienced endoscopists (84.5%) and low‐ranking physicians (68.5%), and there was an obvious increase in sensitivity (90.1% vs. 86.4% vs. 72.7%) and specificity (85% vs. 82.5% vs. 63.7%), showing that the AI system can reach the level of endoscopists after training.…”

Novel milestones for early esophageal carcinoma: From bench to bed

“…The accuracy, sensitivity and specificity of the AI systems were 80.6%, 68.2%, and 83.4%, respectively. In contrast, endoscopic physicians exhibited increased accuracy (85.7%) but decreased sensitivity (61.4%) and specificity (91.2%)[ 115 ]. Although there are still several limitations in the early detection of ESCC using AI, such as issues related to data quality, unreliable results, and the need for verification by professional doctors, significant improvements can be achieved by establishing comprehensive databases, enhancing model interpretability, integrating multimodal information, and strengthening self-learning capabilities.…”