[Establishment and validation of a predictive nomogram model for advanced gastric cancer with perineural invasion]

Objective: Peripheral nerve invasion (PNI) is associated with local recurrence and poor prognosis in patients with advanced gastric cancer. A risk-assessment model based on preoperative indicators for predicting PNI of gastric cancer may help to formulate a more reasonable and accurate individualized diagnosis and treatment plan. Methods: Inclusion criteria: (1) electronic gastroscopy and enhanced CT examination of the upper abdomen were performed before surgery; (2) radical gastric cancer surgery (D2 lymph node dissection, R0 resection) was performed; (3) no distant metastasis was confirmed before and during operation; (4) postoperative pathology showed an advanced gastric cancer (T2-4aN0-3M0), and the clinical data was complete. Those who had other malignant tumors at the same time or in the past, and received neoadjuvant radiochemotherapy or immunotherapy before surgery were excluded. In this retrospective case-control study, 550 patients with advanced gastric cancer who underwent curative gastrectomy between September 2017 and June 2019 were selected from the Affiliated Hospital of Qingdao University for modeling and internal verification, including 262 (47.6%) PNI positive and 288 (52.4%) PNI negative patients. According to the same standard, clinical data of 50 patients with advanced gastric cancer who underwent radical surgery from July to November 2019 in Qingdao Municipal Hospital were selected for external verification of the model. There were no statistically significant differences between the clinical data of internal verification and external verification (all P>0.05). Univariate analysis and multivariate logistic regression analysis were used to determine the independent risk factors for PNI in advanced gastric cancer, and the clinical indicators with statistically significant difference were used to establish a preoperative nomogram model through R software. The Bootstrap method was applied as internal verification to show the robustness of the model. The discrimination of the nomogram was determined by calculating the average consistency index (C-index). The calibration curve was used to evaluate the consistency of the predicted results with the actual results. The Hosmer-Lemeshow test was used to examine the goodness of fit of the discriminant model. During external verification, the corresponding C-index index was also calculated. The area under ROC curve (AUC) was used to evaluate the predictive ability of the nomogram in the internal verification and external verification groups. Results: A total of 550 patients were identified in this study, 262 (47.6%) of which had PNI. Multivariate logistic regression analysis revealed that carcinoembryonic antigen level ≥ 5 μg/L (OR=5.870, 95% CI: 3.281-10.502, P<0.001), tumor length ≥5 cm (OR=5.539,95% CI: 3.165-9.694, P<0.001), mixed Lauren classification (OR=2.611, 95%CI: 1.272-5.360, P=0.009), cT3 stage (OR=13.053, 95% CI: 5.612-30.361, P<0.001) and the presence of lymph node metastasis (OR=4.826, 95% CI: 2.729-8.533, P<0.001) were significant independent risk factors of PNI in advanced gastric cancer (all P<0.05). Based on these results, diffused Lauren classification and cT4 stage were included to establish a predictive nomogram model. CEA ≥ 5 μg/L was for 68 points, tumor length ≥ 5 cm was for 67 points, mixed Lauren classification was for 21 points, diffused Lauren classification was for 38 points, cT3 stage was for 75 points, cT4 stage was for 100 points, and lymph node metastasis was for 62 points. Adding the scores of all risk factors was total score, and the probability corresponding to the total score was the probability that the model predicted PNI in advanced gastric cancer before surgery. The internal verification result revealed that the AUC of nomogram was 0.935, which was superior than that of any single variable, such as CEA, Lauren classification, cT stage, tumor length and lymph node metastasis (AUC: 0.731, 0.595, 0.838, 0.757 and 0.802, respectively). The external verification result revealed the AUC of nomogram was 0.828. The C-ndex was 0.931 after internal verification. External verification showed a C-index of 0.828 from the model. The calibration curve showed that the predictive results were good in accordance with the actual results (P=0.415). Conclusion: A nomogram model constructed by CEA, tumor length, Lauren classification (mixed, diffuse), cT stage, and lymph node metastasis can predict the PNI of advanced gastric cancer before surgery.

[Establishment and clinical testing of pancreatic cancer Faster R-CNN AI system based on fast regional convolutional neural network]

Objective: To investigate the effectiveness of an enhanced CT automatic recognition system based on Faster R-CNN for pancreatic cancer and its clinical value. Methods: In this study, 4 024 enhanced CT imaging sequences of 315 patients with pancreatic cancer from January 2013 to May 2016 at the Affiliated Hospital of Qingdao University were collected retrospectively, and 2 614 imaging sequences were input into the faster R-CNN system as training dataset to create an automatic image recognition model, which was then validated by reading 1 410 enhanced CT images of 135 cases of pancreatic cancer.In order to identify its effectiveness, 3 750 CT images of 150 patients with pancreatic lesions were read and a followed-up was carried out.The accuracy and recall rate in detecting nodules were recorded and regression curves were generated.In addition, the accuracy, sensitivity and specificity of Faster R-CNN diagnosis were analyzed, the ROC curves were generated and the area under the curves were calculated. Results: Based on the enhanced CT images of 135 cases, the area under the ROC curve was 0.927 calculated by Faster R-CNN. The accuracy, specificity and sensitivity were 0.902, 0.913 and 0.801 respectively.After the data of 150 patients with pancreatic cancer were verified, 893 CT images showed positive and 2 857 negative.Ninety-eight patients with pancreatic cancer were diagnosed by Faster R-CNN.After the follow-up, it was found that 53 cases were post-operatively proved to be pancreatic ductal carcinoma, 21 cases of pancreatic cystadenocarcinoma, 12 cases of pancreatic cystadenoma, 5 cases of pancreatic cyst, and 7 cases were untreated.During 5 to 17 months after operation, 6 patients died of abdominal tumor infiltration, liver and lung metastasis.Of the 52 patients who were diagnosed negative by Faster R-CNN, 9 were post-operatively proved to be pancreatic ductal carcinoma. Conclusion: Faster R-CNN system has clinical value in helping imaging physicians to diagnose pancreatic cancer.

[Clinicopathological features and gene phenotypes of benign metastasizing leiomyoma]

Objective: To study the clinicopathological features, immunophenotypes and MED12 gene status in benign metastasizing leiomyoma (BML). Methods: Nine cases of BML diagnosed at the Affiliated Hospital of Qingdao University from 2012 to 2018 were collected, and the radiologic and histologic features were analyzed. The protein expression of leiomyosarcoma-related driver genes, including RB1, PTEN,ATRX,p16,p53, as well as ER,PR,CD34,FH, and Ki-67 were detected using immunohistochemistry, and the mutation status of MED12 gene exon 2 was detected by Sanger sequencing. Results: All the nine patients with BML were female, and the age range was 48 to 64 years (median 55 years). All patients had history of uterine fibroids. The morphologic features of BML were similar to a benign uterine leiomyoma and did not exhibit malignant characteristics. All cases were positive for ER and PR, and negative for CD34. In addition, RB1, PTEN, ATRX, and FH were positive in all cases (wild type), while p16 showed a focally positive pattern. P53 positive index was less than 5% (wild type), and Ki-67 positive index was less than 1%. Sanger sequencing was done in six BML samples; one sample harbored a nonsense mutation c. 142_144delinsTAA (p.Glu48Ter), and another exhibited a synonymy mutation (c.192C>T, p.Phe64=)and one missense mutation c.196C>T (p.Pro66Ser). Conclusions: The present study suggests that BML is a unique leiomyoma entity that is pathologically and genetically different from leiomyosarcomas and conventional uterine leiomyomas. Evaluating the genetic phenotype of BML, especially the expression of leiomyosarcoma-related driver genes protein and MED12 gene status, may be helpful in understanding the pathogenesis of BML and in its differentiation from leiomyosarcoma.