Risk factors analysis and prediction model construction of submucosal deep infiltration of early colorectal tumor / 中华肿瘤杂志

Objective: To investigate the risk factors for the development of deep infiltration in early colorectal tumors (ECT) and to construct a prediction model to predict the development of deep infiltration in patients with ECT. Methods: The clinicopathological data of ECT patients who underwent endoscopic treatment or surgical treatment at the Cancer Hospital, Chinese Academy of Medical Sciences from August 2010 to December 2020 were retrospectively analyzed. The independent risk factors were analyzed by multifactorial regression analysis, and the prediction models were constructed and validated by nomogram. Results: Among the 717 ECT patients, 590 patients were divided in the within superficial infiltration 1 (SM1) group (infiltration depth within SM1) and 127 patients in the exceeding SM1 group (infiltration depth more than SM1). There were no statistically significant differences in gender, age, and lesion location between the two groups (P>0.05). The statistically significant differences were observed in tumor morphological staging, preoperative endoscopic assessment performance, vascular tumor emboli and nerve infiltration, and degree of tumor differentiation (P<0.05). Multivariate regression analysis showed that only erosion or rupture (OR=4.028, 95% CI: 1.468, 11.050, P=0.007), localized depression (OR=3.105, 95% CI: 1.584, 6.088, P=0.001), infiltrative JNET staging (OR=5.622, 95% CI: 3.029, 10.434, P<0.001), and infiltrative Pit pattern (OR=2.722, 95% CI: 1.347, 5.702, P=0.006) were independent risk factors for the development of deep submucosal infiltration in ECT. Nomogram was constructed with the included independent risk factors, and the nomogram was well distinguished and calibrated in predicting the occurrence of deep submucosal infiltration in ECT, with a C-index and area under the curve of 0.920 (95% CI: 0.811, 0.929). Conclusion: The nomogram prediction model constructed based on only erosion or rupture, local depression, infiltrative JNET typing, and infiltrative Pit pattern has a good predictive efficacy in the occurrence of deep submucosal infiltration in ECT.

Risk factors for residual cancer or lymph node metastasis after endoscopic noncurable resection of early colorectal cancer / 中华肿瘤杂志

Objective: Risk factors related to residual cancer or lymph node metastasis after endoscopic non-curative resection of early colorectal cancer were analyzed to predict the risk of residual cancer or lymph node metastasis, optimize the indications of radical surgical surgery, and avoid excessive additional surgical operations. Methods: Clinical data of 81 patients who received endoscopic treatment for early colorectal cancer in the Department of Endoscopy, Cancer Hospital, Chinese Academy of Medical Sciences from 2009 to 2019 and received additional radical surgical surgery after endoscopic resection with pathological indication of non-curative resection were collected to analyze the relationship between various factors and the risk of residual cancer or lymph node metastasis after endoscopic resection. Results: Of the 81 patients, 17 (21.0%) were positive for residual cancer or lymph node metastasis, while 64 (79.0%) were negative. Among 17 patients with residual cancer or positive lymph node metastasis, 3 patients had only residual cancer (2 patients with positive vertical cutting edge). 11 patients had only lymph node metastasis, and 3 patients had both residual cancer and lymph node metastasis. Lesion location, poorly differentiated cancer, depth of submucosal invasion ≥2 000 μm, venous invasion were associated with residual cancer or lymph node metastasis after endoscopic (P<0.05). Logistic multivariate regression analysis showed that poorly differentiated cancer (OR=5.513, 95% CI: 1.423, 21.352, P=0.013) was an independent risk factor for residual cancer or lymph node metastasis after endoscopic non-curative resection of early colorectal cancer. Conclusions: For early colorectal cancer after endoscopic non-curable resection, residual cancer or lymph node metastasis is associated with poorly differentiated cancer, depth of submucosal invasion ≥2 000 μm, venous invasion and the lesions are located in the descending colon, transverse colon, ascending colon and cecum with the postoperative mucosal pathology result. For early colorectal cancer, poorly differentiated cancer is an independent risk factor for residual cancer or lymph node metastasis after endoscopic non-curative resection, which is suggested that radical surgery should be added after endoscopic treatment.

Establishment and clinical validation of an artificial intelligence YOLOv51 model for the detection of precancerous lesions and superficial esophageal cancer in endoscopic procedure / 中华肿瘤杂志

Objective: To construct the diagnostic model of superficial esophageal squamous cell carcinoma (ESCC) and precancerous lesions in endoscopic images based on the YOLOv5l model by using deep learning method of artificial intelligence to improve the diagnosis of early ESCC and precancerous lesions under endoscopy. Methods: 13, 009 endoscopic esophageal images of white light imaging (WLI), narrow band imaging (NBI) and lugol chromoendoscopy (LCE) were collected from June 2019 to July 2021 from 1, 126 patients at the Cancer Hospital, Chinese Academy of Medical Sciences, including low-grade intraepithelial neoplasia, high-grade intraepithelial neoplasia, ESCC limited to the mucosal layer, benign esophageal lesions and normal esophagus. By computerized random function method, the images were divided into a training set (11, 547 images from 1, 025 patients) and a validation set (1, 462 images from 101 patients). The YOLOv5l model was trained and constructed with the training set, and the model was validated with the validation set, while the validation set was diagnosed by two senior and two junior endoscopists, respectively, to compare the diagnostic results of YOLOv5l model and those of the endoscopists. Results: In the validation set, the accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the YOLOv5l model in diagnosing early ESCC and precancerous lesions in the WLI, NBI and LCE modes were 96.9%, 87.9%, 98.3%, 88.8%, 98.1%, and 98.6%, 89.3%, 99.5%, 94.4%, 98.2%, and 93.0%, 77.5%, 98.0%, 92.6%, 93.1%, respectively. The accuracy in the NBI model was higher than that in the WLI model (P<0.05) and lower than that in the LCE model (P<0.05). The diagnostic accuracies of YOLOv5l model in the WLI, NBI and LCE modes for the early ESCC and precancerous lesions were similar to those of the 2 senior endoscopists (96.9%, 98.8%, 94.3%, and 97.5%, 99.6%, 91.9%, respectively; P>0.05), but significantly higher than those of the 2 junior endoscopists (84.7%, 92.9%, 81.6% and 88.3%, 91.9%, 81.2%, respectively; P<0.05). Conclusion: The constructed YOLOv5l model has high accuracy in diagnosing early ESCC and precancerous lesions in endoscopic WLI, NBI and LCE modes, which can assist junior endoscopists to improve diagnosis and reduce missed diagnoses.