Search

Objective By collecting the patient information in the Colon Cancer Database of Peking Union Medical College Hospital since its establishment, we aim to demonstrate the completeness of the clinicopathological characteristics and follow-up data to provide reference for single-center database establishment. Methods We collected 33 surgical patient information entries from the single-center prospective registry database of Colon Cancer at Peking Union Medical College Hospital from January 5, 2016 to May 11, 2022, including demographic characteristics, features of the primary lesion, surgical and pathological information and postoperative complications, and calculated the missing rate of each entry as well as the follow-up data. Results A total of 1682 patients with colon cancer were included in the analysis, in which 981 patients were male and 701 patients were female, with an average age of (62.75±11.97) years. The postoperative complication rate was 15.6% (263/1682). Among the data entries, demographic features such as sex, age, ethnic group as well as the pathologic staging were collected without missing. The relatively high-missing entries were the colonoscopy tumor site with a missing rate of 12.2%, followed by immunohistochemical information of mismatch repair with a missing rate of 8.0%. The missing rate of all 33 entries was distributed in 0-12.2%, of which 81.8% (27/33) were less than 1% and 93.9% (31/33) were less than 5%. As for follow-up data, the percentage of patients with complete missing follow-up was 1.7%, which meant 98.3% patients had available survival data. The 1-year, 2-year, 3-year, 4-year and 5-year follow-up rates of patients enrolled before July 2, 2020 (patients enrolled after this date did not have regular follow-up information in a uniform format due to data migration) were 55.02%, 70.96%, 72.02%, 65.42% and 70.91%, respectively, and the 2-year complete follow-up rate was 19.5%. Conclusion The low rate of missing statistical entries and the high follow-up rate of our center's database has certain reference value for database establishment, but there is still room for improvement in terms of continuous follow-up.

Abstract Turbines are generally oversimplified in conventional algorithms for steady‐state analysis of hydropower plant (HPP) networks. Moreover, conventional algorithms are hardly suitable for serial HPPs. In this paper, a modified global gradient algorithm (MGGA) is developed by extending turbine models and optimizing turbine interpolation. First, the nonlinear model of turbines is established with the proposed variable resistance factor and added to the global matrix. Then turbine unit quantity interpolation is proposed to compute the turbine operating parameters. Finally, turbine links and pipes can be iterated synchronously, and the global matrix is solved using the Newton–Raphson method. Characteristics of MGGA are evaluated by employing two test HPPs, including rapidity of convergence, computational stability of turbine, and applicability for serial HPPs. The results show that: (a) The iterative numbers of the MGGA are 0.72 and 4.98 times on average less than that of CA1 and CA2 for the accuracy of 10−6 in 400 operating conditions, even though MGGA has an obvious hysteretic nature; (b) Only MGGA can successfully calculate the extreme cases that large head losses in HPP networks; (c) Although fixed head nodes are not present between turbine links, such as serial HPPs, MGGA still works.

With the deep integration of "AI + medicine", AI-assisted technology has been of great help to human beings in the medical field, especially in the area of predicting and diagnosing diseases based on big data, because it is faster and more accurate. However, concerns about data security seriously hinder data sharing among medical institutions. To fully exploit the value of medical data and realize data collaborative sharing, we developed a medical data security sharing scheme based on the C/S communication mode and constructed a federated learning architecture that uses homomorphic encryption technology to protect training parameters. Here, we chose the Paillier algorithm to realize the additive homomorphism to protect the training parameters. Clients do not need to share local data, but only upload the trained model parameters to the server. In the process of training, a distributed parameter update mechanism is introduced. The server is mainly responsible for issuing training commands and weights, aggregating the local model parameters from the clients and predicting the joint diagnostic results. The client mainly uses the stochastic gradient descent algorithm for gradient trimming, updating and transmitting the trained model parameters back to the server. In order to test the performance of this scheme, a series of experiments was conducted. From the simulation results, we can know that the model prediction accuracy is related to the global training rounds, learning rate, batch size, privacy budget parameters etc. The results show that this scheme realizes data sharing while protecting data privacy, completes the accurate prediction of diseases and has a good performance.

BackgroundDepressive symptoms are common among patients with lung cancer. We aimed to assess the effects of esketamine on postoperative depressive symptoms after thoracoscopic lung cancer surgery.MethodsIn this randomized, double-blind, placebo-controlled trial, 156 patients undergoing thoracoscopic lung cancer surgery were randomly allocated in a 1:1 ratio to receive intravenous esketamine (intraoperatively and in patient-controlled analgesia until 48 h postoperatively) or normal saline placebo. The primary outcome was the proportion of patients with depressive symptoms at 1 month postoperatively, assessed using the Beck Depression Inventory-II (BDI-II). Secondary outcomes included depressive symptoms at 48 h postoperatively, hospital discharge and 3 months postoperatively, BDI-II scores, anxious symptoms, Beck Anxiety Inventory scores, Quality of Recovery-15 (QoR-15) scores, and 1- and 3-month mortality.Main resultsA total of 151 patients (75 in the esketamine group and 76 in the normal saline group) completed the 1-month follow-up. The esketamine group had a significantly lower incidence of depressive symptoms at 1 month compared to the normal saline group (1.3% vs. 11.8%; risk difference = −10.5, 95%CI = −19.6% to −0.49%; p = 0.018). After excluding patients without lung cancer diagnosis, the incidence of depressive symptoms was also lower in the esketamine group (1.4% vs. 12.2%; risk difference = −10.8, 95%CI = −20.2% to −0.52%; p = 0.018). The secondary outcomes were similar between groups, except that the esketamine group had higher QoR-15 scores at 1 month postoperatively (median difference = 2; 95%CI = 0 to 5; p = 0.048). The independent risk factors for depressive symptoms were hypertension (odds ratio = 6.75, 95%CI = 1.13 to 40.31; p = 0.036) and preoperative anxious symptoms (odds ratio = 23.83, 95%CI = 3.41 to 166.33; p = 0.001).ConclusionPerioperative administration of esketamine reduced the incidence of depressive symptoms at 1 month after thoracoscopic lung cancer surgery. History of hypertension and preoperative anxious symptoms were independent risk factors for depressive symptoms.Clinical trial registration: Chinese Clinical Trial Registry http://www.chictr.org.cn, Identifier (ChiCTR2100046194).

This document is a detailed description of a study on the oxidative chlorination of alkanes using visible light and HCl as the chlorine reagent. The researchers found that simple alkanes cannot be chlorinated effectively under the same reaction conditions, but functionalized alkanes with electron-withdrawing groups can undergo chlorination. The study includes experimental procedures, results, and characterization of the compounds using NMR analysis. The document also acknowledges the funding source and provides additional supporting information. [Extracted from the article]