Nomogram Based on Body Composition and Prognostic Nutritional Index Predicts Survival After Curative Resection of Gastric Cancer.

RATIONALE AND OBJECTIVES: This study aimed to identify independent prognostic factors for gastric cancer (GC) patients after curative resection using quantitative computed tomography (QCT) combined with prognostic nutritional index (PNI), and to develop a nomogram prediction model for individualized prognosis. MATERIALS AND METHODS: This study retrospectively analyzed 119 patients with GC who underwent curative resection from January 2016 to March 2018. The patients' preoperative clinical pathological data were recorded, and all patients underwent QCT scans before and after curative resection to obtain QCT parameters: bone mineral density (BMD), skeletal muscle area (SMA), visceral fat area (VFA), subcutaneous fat area (SFA) and CT fat fraction (CTFF), then relative rate of change in each parameter (ΔBMD, ΔSMA, ΔVFA, ΔSFA, ΔCTFF) was calculated after time normalization. Multivariate Cox proportional hazards was used to establish a nomogram model that based on independent prognostic factors. The concordance index (C-index), area under the time-dependent receiver operating characteristic (ROC) curve and clinical decision curve were used to evaluate the predictive performance and clinical benefit of the nomogram model． RESULTS: This study found that ΔCTFF, ΔVFA, ΔBMD and PNI are independent prognostic factors for overall survival (OS) (hazard ratio: 1.034, 0.895, 0.976, 2.951, respectively, all p < 0.05). The established nomogram model could predict the area under the ROC curve of OS at 1, 3 and 5 years as 0.816, 0.815 and 0.881, respectively. The C-index was 0.743 (95% CI, 0.684-0.801), and the decision curve analysis showed that this model has good clinical net benefit. CONCLUSION: The nomogram model based on body composition and PNI is reliable in predicting the individualized survival of underwent curative resection for GC patients.

Transfer learning for bearing performance degradation assessment based on deep hierarchical features.

Bearing design, production and complicated operating condition can lead to the scattered life cycle degradation distribution, which will bring a challenge of generalization for performance degradation assessment models. And it is costly and time-consuming to collect a large amount of labeled data for supervised diagnosis, especially when the task comes from a new operating condition. Thus in this paper, a novel bearing degradation assessment model is proposed based on transfer learning and deep hierarchical features extraction. The research of degradation assessment is transformed to the classification task of degradation pattern, which divides degradation process into the normal, slight fault, fault development and damage patterns. The hierarchical network with random weight parameters is introduced to extract the local sub-band characteristics of spectrum, in which the multiple alternately convolution and pooling layers without supervised fine-tuning are employed. Joint Geometrical and Statistical Alignment method is then utilized to obtain projected sharing feature space, and thus the knowledge of bearing degradation process is transferred to accomplish degradation pattern assessment under different operating conditions. Results of the experiments on bearing fault severity and degradation process show that the proposed method reduces the feature distribution divergence between the degradation processes and accomplishes bearing performance degradation assessment in different operating condition.