Unraveling the role of tumor sidedness in prognosis of stage II colon cancer.

Background: Stage II colon cancer has varying risks for metastasis, and treatment strategies depend on molecular and clinicopathological features. While tumor-sidedness is a well-accepted prognostic factor for stage III/IV colon cancer, its role in stage II is controversial. Understanding its effect in stage II is crucial for improving treatment strategies. Methods: We analyzed clinical and follow-up data of colon cancer from the Surveillance, Epidemiology, and End Results database (2004-2017). Patients were divided into a primary study cohort (2010-2017) and a validation cohort (2004-2009). The baseline characteristics between right-sided colon cancer (RCC) and left-sided colon cancer (LCC) groups were compared. Moreover, the effect of tumor size on cancer-specific survival (CSS) was evaluated using Kaplan-Meier analysis. Results: The study involved 87,355 patients in the study cohort and 65,858 in the validation cohort. Of the study cohort, 52.3% were diagnosed with RCC. The median age was 64 years old, with 48.5% females and 76.8% of white people. In addition, stage II RCC showed better CSS compared with LCC (5-year CSS 88.0% vs 85.5%, P < 0.001), while stage III/IV RCC demonstrated poorer outcomes. Multivariate Cox regression analysis identified that the right-sidedness was a positive prognostic factor in stages I/II but negative in stages III (HR 1.10, P < 0.001) and IV (HR 1.26, P < 0.001). Chemotherapy rates decreased in RCC, particularly in stage II (RCC vs LCC: 16.2% vs 28.5%, P < 0.001). Subgroup analysis, stratified by T3/T4 stages and chemotherapy status, further highlighted better survival outcomes in RCC. Conclusions: RCC is associated with a significantly better prognosis in stage II. The importance of considering tumor-sidedness in clinical decision-making and the design of future clinical trials should be emphasized.

Diagnostic Value of Inflammation-Related Indicators in Distinguishing Early Colon Cancer and Adenomatous Polyps.

BACKGROUND: There are few clinical symptoms in early colorectal cancer, so it is necessary to find a simple and economical tumor detection index for auxiliary diagnosis. This study aims to explore the diagnostic value of preoperative inflammation-related indicators, such as neutrophil, lymphocyte, platelet count, platelet to lymphocyte ratio (PLA), neutrophil to lymphocyte ratio (NLR), and systemic immune-inflammation index (SII), for early colorectal cancer, and determine whether inflammation-related indicators can provide more accurate diagnostic judgment for patients. METHODS: This study was a retrospective study. Patients who were first diagnosed with colorectal cancer or colorectal adenomatous polyp at Beijing Friendship Hospital from October 2016 to October 2017 were retrospectively collected. According to inclusion and exclusion criteria, a total of 342 patients were included, including 216 patients with colorectal cancer and 126 patients with colorectal adenomatous polyp. Fasting venous blood and other clinical features were collected to compare the differences between colorectal cancer and colorectal adenoma. RESULTS: There were statistically significant differences in age, carcinoembryonic antigen, albumin, hemoglobin, mean platelet volume, lymphocyte, monocyte, NLR, PLA, SII, and mean platelet volume to platelet count ratio between colorectal cancer group and colorectal adenoma group (P < .05), and a Nomogram model was established. Using inflammatory markers to differentiate colorectal and colorectal polyps produced greater AUC than using tumor markers alone (.846 vs .695). CONCLUSION: Inflammation-related indicators, such as lymphocyte, monocyte, and mean platelet volume, may serve as potential indicators to assist in the diagnosis of early colorectal cancer.

Sarcobesity, but not visceral fat, is an independent risk factor for complications after radical resection of colorectal cancer.

Background: The influence of body composition on the outcome of colorectal cancer surgery is controversial. The aim of this study was to evaluate the effects of visceral obesity and sarcobesity on the incidence of total and surgical complications after radical resection of colorectal cancer. Methods: We collected a total of 426 patients who underwent elective radical resection of colorectal cancer at Beijing Friendship Hospital, Capital Medical University from January 2017 to May 2018. According to the inclusion and exclusion criteria, 387 patients were finally included. A CT scan at the level of the L3-L4 intervertebral disk was selected to measure the values of visceral fat area and skeletal muscle area. Multivariate analysis was used to explore the independent risk/protective factors affecting postoperative complications. Results: 128 (33.1%) patients developed complications, and 44 (11.4%) patients developed major complications. Among them, 111 patients developed surgical complications and 21 developed medical complications. Visceral fat area (Z = -3.271, p = 0.001), total fat area (Z = -2.613, p = 0.009), visceral fat area to subcutaneous fat area ratio (V/S, Z = -2.633, p = 0.008), and sarcobesity index (Z = -2.282, p = 0.023) were significantly associated with total complications. Visceral fat area (Z = -2.119, p = 0.034) and V/S (Z = -2.010, p = 0.044) were significantly associated with total surgical complications. Sarcobesity index, smoking, stoma, blood loss, surgery time, and American Society of Anesthesiology (ASA) score were selected as risk factors for total postoperative complications according to LASSO regression. Multivariate logistic regression analysis suggested that sarcobesity index was an independent risk factor for postoperative total complications and surgical complications. Subgroup analysis suggested that albumin level was an independent protective factor for postoperative total complications in male patients. Smoking, operative time, and sarcobesity index were independent risk factors, and cholesterol was an independent protective factor for total postoperative complications in female patients. Conclusion: Increased sarcobesity index is an independent risk factor for postoperative complications in patients with colorectal cancer, while visceral fat area is not. For female patients, smoking, operation time, and obesity index are independent risk factors for postoperative complications, while cholesterol is an independent protective factor. For male patients, serum albumin is an independent protective factor for postoperative complications.

Hierarchical Weight Averaging for Deep Neural Networks.

Despite simplicity, stochastic gradient descent (SGD)-like algorithms are successful in training deep neural networks (DNNs). Among various attempts to improve SGD, weight averaging (WA), which averages the weights of multiple models, has recently received much attention in the literature. Broadly, WA falls into two categories: 1) online WA, which averages the weights of multiple models trained in parallel, is designed for reducing the gradient communication overhead of parallel mini-batch SGD and 2) offline WA, which averages the weights of one model at different checkpoints, is typically used to improve the generalization ability of DNNs. Though online and offline WA are similar in form, they are seldom associated with each other. Besides, these methods typically perform either offline parameter averaging or online parameter averaging, but not both. In this work, we first attempt to incorporate online and offline WA into a general training framework termed hierarchical WA (HWA). By leveraging both the online and offline averaging manners, HWA is able to achieve both faster convergence speed and superior generalization performance without any fancy learning rate adjustment. Besides, we also analyze the issues faced by the existing WA methods, and how our HWA addresses them, empirically. Finally, extensive experiments verify that HWA outperforms the state-of-the-art methods significantly.

Impact of preoperative biliary drainage on postoperative complications and prognosis after pancreaticoduodenectomy: A single-center retrospective cohort study.

Background and objectives: Obstructive jaundice is common in patients with pancreaticobiliary malignancies. Preoperative biliary drainage (PBD) can alleviate cholestasis; however, no consensus has been reached on the impact of PBD on the incidence of surgery-related complications and patient survival. This study aimed to evaluate the effect among patients treated with PBD. Methods: This retrospective study examined the clinical and follow-up prognostic data of 160 patients with pancreaticobiliary malignancies who underwent pancreaticoduodenectomy (PD) at Beijing Friendship Hospital, Capital Medical University, from January 2016 to July 2020. Outcomes were compared between patients who underwent PBD (PBD group) and those who did not (control group). Changes in biochemical indicators were evaluated before and after drainage in the PBD group. Between-group differences in inflammatory indicators after PD were assessed using the Wilcoxon signed-rank test. Postoperative complications were classified according to the Clavien-Dindo classification system. The effects of PBD and biliary drainage efficiency on postoperative complications were evaluated using the chi-square test and binary logistics regression. The Kaplan-Meier analysis was used for between-group comparison of survival analysis. Univariate and multivariate regression analyses were performed to identify prognostic factors of survival. Results: Total 160 patients were enrolled,the mean age of the study sample was 62.75 ± 6.75 years. The distribution of pancreaticobiliary malignancies was as follows: 34 cases of pancreatic head cancer, 61 cases of distal bile duct cancer, 20 cases of duodenal papilla cancer, 39 cases of duodenal ampullary cancer, and 6 cases of malignant intraductal papillary mucinous neoplasm (IPMN). PBD was performed in 90 of the 160 patients, with PBD performed using an endoscopic retrograde cholangiopancreatography (ERCP) approach in 55 patients and with percutaneous transhepatic cholangiography (PTC) used in the remaining 35 cases. The mean duration of drainage in the PBD group was 12.8 ± 8.8 days. The overall rate of complications was 48.05% (37/77) in the control group and 65.55% (59/90) in the PBD group with non-significant difference (χ2 = 3.527, p=0.473). In logsitics regression analysis, PBD was also not a risk factor for postoperative complications OR=1.77, p=0.709). The overall rate of postoperative complications was significantly higher among patients who underwent PBD for >2 weeks (χ2 = 6.102, p=0.013), with the rate of severe complications also being higher for this subgroup of PBD patients (χ2 = 4.673, p=0.03). The overall survival time was 47.9 ± 2.45 months, with survival being slightly lower in the PBD group (43.61 ± 3.26 months) than in the control group (52.24 ± 3.54 months), although this difference was not significant (hazard ratio (HR)=0.65, p=0.104). Conclusion: In patients with malignant biliary obstruction, PBD does not affect the incidence of postoperative complications after pancreaticoduodenectomy nor does it affect patient survival. Prolonged biliary drainage (>2 weeks) may increase the incidence of overall postoperative complications and severe complications.

Efficient Spiking Neural Networks With Radix Encoding.

Spiking neural networks (SNNs) have advantages in latency and energy efficiency over traditional artificial neural networks (ANNs) due to their event-driven computation mechanism and the replacement of energy-consuming weight multiplication with addition. However, to achieve high accuracy, it usually requires long spike trains to ensure accuracy, usually more than 1000 time steps. This offsets the computation efficiency brought by SNNs because a longer spike train means a larger number of operations and larger latency. In this article, we propose a radix-encoded SNN, which has ultrashort spike trains. Specifically, it is able to use less than six time steps to achieve even higher accuracy than its traditional counterpart. We also develop a method to fit our radix encoding technique into the ANN-to-SNN conversion approach so that we can train radix-encoded SNNs more efficiently on mature platforms and hardware. Experiments show that our radix encoding can achieve 25 × improvement in latency and 1.7% improvement in accuracy compared to the state-of-the-art method using the VGG-16 network on the CIFAR-10 dataset.

Smart and Rapid Design of Nanophotonic Structures by an Adaptive and Regularized Deep Neural Network.

The design of nanophotonic structures based on deep learning is emerging rapidly in the research community. Design methods using Deep Neural Networks (DNN) are outperforming conventional physics-based simulations performed iteratively by human experts. Here, a self-adaptive and regularized DNN based on Convolutional Neural Networks (CNNs) for the smart and fast characterization of nanophotonic structures in high-dimensional design parameter space is presented. This proposed CNN model, named LRS-RCNN, utilizes dynamic learning rate scheduling and L2 regularization techniques to overcome overfitting and speed up training convergence and is shown to surpass the performance of all previous algorithms, with the exception of two metrics where it achieves a comparable level relative to prior works. We applied the model to two challenging types of photonic structures: 2D photonic crystals (e.g., L3 nanocavity) and 1D photonic crystals (e.g., nanobeam) and results show that LRS-RCNN achieves record-high prediction accuracies, strong generalizibility, and substantially faster convergence speed compared to prior works. Although still a proof-of-concept model, the proposed smart LRS-RCNN has been proven to greatly accelerate the design of photonic crystal structures as a state-of-the-art predictor for both Q-factor and V. It can also be modified and generalized to predict any type of optical properties for designing a wide range of different nanophotonic structures. The complete dataset and code will be released to aid the development of related research endeavors.