Objective: We aimed to establish an artificial intelligence (AI) model to identify parathyroid glands during endoscopic approaches and compare it with senior and junior surgeons' visual estimation.Methods: A total of 1,700 images of parathyroid glands from 166 endoscopic thyroidectomy videos were labeled. Data from 20 additional full-length videos were used as an independent external cohort. The YOLO V3, Faster R-CNN, and Cascade algorithms were used for deep learning, and the optimal algorithm was selected for independent external cohort analysis. Finally, the identification rate, initial recognition time, and tracking periods of PTAIR (Artificial Intelligence model for Parathyroid gland Recognition), junior surgeons, and senior surgeons were compared.Results: The Faster R-CNN algorithm showed the best balance after optimizing the hyperparameters of each algorithm and was updated as PTAIR. The precision, recall rate, and F1 score of the PTAIR were 88.7%, 92.3%, and 90.5%, respectively. In the independent external cohort, the parathyroid identification rates of PTAIR, senior surgeons, and junior surgeons were 96.9%, 87.5%, and 71.9%, respectively. In addition, PTAIR recognized parathyroid glands 3.83 s ahead of the senior surgeons (p = 0.008), with a tracking period 62.82 s longer than the senior surgeons (p = 0.006).Conclusions: PTAIR can achieve earlier identification and full-time tracing under a particular training strategy. The identification rate of PTAIR is higher than that of junior surgeons and similar to that of senior surgeons. Such systems may have utility in improving surgical outcomes and also in accelerating the education of junior surgeons.

No abstract

Background: The textbook outcome (TO) emerges as a novel prognostic factor in surgical oncology.The present study aimed to evaluate the effect of TO on the risk of death and recurrence in patients with esophageal squamous cell carcinoma (ESCC) after minimally invasive esophagectomy (MIE). Methods:The study involved retrospective analysis of 528 patients with ESCC who were subjected to MIE from January 2011 to December 2017. TO included 8 parameters: complete resection; microscopically tumor-negative resection margins (R0); ≥15 lymph nodes removed and examined; no serious postoperative complications; no postoperative intervention; no re-admission to the intensive care unit (ICU); hospital stay ≤21 days; and no readmission ≤30 days. The Cox and logistic regression model were used to analyze the prognostic factors of survival and risk factors for TO.Results: Among the 528 patients with ESCC who were subjected to MIE, 53.2% reached TO. In the case of patients with locally advanced ESCC, 5-year overall survival (OS) was 51.1% (41.2-61.2%) for the TO group but 33.7% (23.7-43.7%) for the non-TO group (HR =0.644, 95% CI: 0.449-0.924, P=0.015).Similarly, 5-year disease-free survival (DFS) was 47.6% (38.0-57.2%) for the TO group but 29.1% (20.1-38.1%) for the non-TO group (HR =0.671, 95% CI: 0.479-0.940, P=0.018). In addition, 5-year recurrencefree survival (RFS) was 62.9% (53.7-72.1%) for the TO group but 39.8% (29.4-50.2%) for the non-TO group (HR =0.606, 95% CI: 0.407-0.902, P=0.012). Multivariate logistic regression analysis further showed that age, American Society of Anesthesiology (ASA) score, intraoperative blood loss, and smoking status acted as independent risk factors for TO. The results of the multivariate analysis assisted in the establishment of a nomogram for the prediction of TO occurrence. This nomogram exhibited satisfactory consistency and prediction ability [area under the receiving operator characteristic (AUROC) =0.717].

ObjectiveMany surgeons knew the importance of parathyroid gland (PG) in the thyroid surgery, but it was even more difficult to be protected. This study aimed at evaluating the effectiveness of the improved method of searching inferior parathyroid gland (IPG).Methods213 patients were enrolled and divided into test and control groups according to different methods of searching IPG in the surgery. Consequently, we compared the surgical outcome parameters between the two groups, including the operative time, numbers of PG identifying (PG protection in situ, PG auto-transplantation, and PG accidental removal), numbers of the total lymph node (LN) and metastatic LN, parathyroid hormone (PTH), transient hypoparathyroidism, transient recurrent laryngeal nerve palsy, and postoperative bleeding.ResultsWe identified 194 (194/196, 98.98%) and 215 (215/230, 93.48%) PGs in the test group and control group, respectively, and there was a significant difference (P = 0.005), and this result was due to IPG identification differences (96/98, 97.96% vs. 100/115, 86.96%, P = 0.004). Meanwhile, there was a lower ratio of IPG auto-transplantation in the test group compared with that in the control group (46.94% vs. 64.35%, P = 0.013). Serum PTH one day after the operation was 3.65 ± 1.86 vs. 2.96 ± 1.64 (P = 0.043) but with no difference at 6 months. There were no differences in metastatic LN and recurrent laryngeal nerve palsy between two groups.ConclusionThe improved method of searching IPG was simple, efficient, and safe, which was easy to be implemented for searching IPG and protecting it well.