Previous reports have shown favorable performance of artificial intelligence (AI) systems for diagnosing esophageal squamous cell carcinoma (ESCC) compared with endoscopists. However, these findings don’t reflect performance in clinical situations, as endoscopists classify lesions based on both magnified and non-magnified videos, while AI systems often use only a few magnified narrow band imaging (NBI) still images. We evaluated the performance of the AI system in simulated clinical situations. We used 25,048 images from 1433 superficial ESCC and 4746 images from 410 noncancerous esophagi to construct our AI system. For the validation dataset, we took NBI videos of suspected superficial ESCCs. The AI system diagnosis used one magnified still image taken from each video, while 19 endoscopists used whole videos. We used 147 videos and still images including 83 superficial ESCC and 64 non-ESCC lesions. The accuracy, sensitivity and specificity for the classification of ESCC were, respectively, 80.9% [95% CI 73.6–87.0], 85.5% [76.1–92.3], and 75.0% [62.6–85.0] for the AI system and 69.2% [66.4–72.1], 67.5% [61.4–73.6], and 71.5% [61.9–81.0] for the endoscopists. The AI system correctly classified all ESCCs invading the muscularis mucosa or submucosa and 96.8% of lesions ≥ 20 mm, whereas even the experts diagnosed some of them as non-ESCCs. Our AI system showed higher accuracy for classifying ESCC and non-ESCC than endoscopists. It may provide valuable diagnostic support to endoscopists.
31 March 2021 in Journal of Gastroenterology and Hepatology
Endoscopic resection (ER) is the mainstay of treatment for superficial esophageal squamous cell carcinoma (SESCC) instead of esophagectomy because of its minimal invasiveness and favorable clinical outcomes. Developments in endoscopic submucosal dissection have enabled en bloc resection of SESCCs regardless of size, thus reducing the risk of local recurrence. Although ER for SESCC is effective, metastasis may subsequently occur. Additionally, extensive esophageal ER confers a risk of postoperative esophageal stricture. Therefore, accurate assessment of the invasion depth and circumferential extent of SESCCs is important in determining the indications for ER. Diagnostic accuracies for SESCC invasion differ between epithelial (EP)/lamina propria (LPM), muscularis mucosa (MM)/submucosal (SM1), and SM2 cancers. ER is strongly indicated for clinically diagnosed (c)EP/LPM cancers because 90% of these are as pathologically diagnosed (p)EP/LPM, which has a very low risk of metastasis. Remarkably, the diagnostic accuracy for cMM/SM1 differs significantly with lateral spread of cancer. Eighty percent of cMM/SM1 cancers with ≤3/4 circumferential spread prove to be pEP/LPM or pMM/SM1, which have very low or low risk of metastasis. Thus, these are adequate candidates for ER. However, given the relatively low proportion of pEP/LPM or pMM/SM1 and high risk of subsequent stricture, ER is not recommended for whole circumferential cMM/SM1 cancers. For cMM/SM1 cancers that involve >3/4 but not the whole circumference, ER should be considered on a lesion-by-lesion basis because the risk of post-ER stricture is not very high, but the proportion of pEP/LPM or pMM/SM1 is relatively low. ER is contraindicated for cSM2 cancers because 75% of them are pSM2, which has high risk of metastasis.