Establishment of predictive nomogram and web-based survival risk calculator for desmoplastic small round cell tumor: A propensity score-adjusted, population-based study.

Desmoplastic small round cell tumor (DSRCT) is a rare undifferentiated malignant soft tissue tumor with a poor prognosis and a lack of consensus on treatment. This study's objective was to build a nomogram based on clinicopathologic factors and an online survival risk calculator to predict patient prognosis and support therapeutic decision-making. A retrospective cohort analysis of the Surveillance, Epidemiology and End Results (SEER) database was performed for patients diagnosed with DSRCT between 2000 and 2019. The least absolute shrinkage and selection operator (LASSO) Cox regression analysis was applied to identify the individual variables related to overall survival (OS) and cancer-specific survival (CSS), as well as to construct online survival risk calculators and nomogram survival models. The nomogram was employed to categorize patients into different risk groups, and the Kaplan-Meier method was utilized to determine the survival rate of each risk category. Propensity score matching (PSM) was used to assess survival with different therapeutic approaches. A total of 374 patients were included, and the median OS and CSS were 25 (interquartile range 21.9-28.1) months and 27 (interquartile range 23.6-30.3) months, respectively. The nomogram models demonstrated high predictive accuracy. PSM found that patients with triple-therapy had better CSS and OS than those who received surgery plus chemotherapy (median survival times: 49 vs 34 months and 49 vs 35 months, respectively). The nomogram successfully predicted the DSRCT patients survival rate. This approach could assist doctors in evaluating prognoses, identifying high-risk populations, and implementing personalized therapy.

Structure-Aware Feature Disentanglement With Knowledge Transfer for Appearance-Changing Place Recognition.

Long-term visual place recognition (VPR) is challenging as the environment is subject to drastic appearance changes across different temporal resolutions, such as time of the day, month, and season. A wide variety of existing methods address the problem by means of feature disentangling or image style transfer but ignore the structural information that often remains stable even under environmental condition changes. To overcome this limitation, this article presents a novel structure-aware feature disentanglement network (SFDNet) based on knowledge transfer and adversarial learning. Explicitly, probabilistic knowledge transfer (PKT) is employed to transfer knowledge obtained from the Canny edge detector to the structure encoder. An appearance teacher module is then designed to ensure that the learning of appearance encoder does not only rely on metric learning. The generated content features with structural information are used to measure the similarity of images. We finally evaluate the proposed approach and compare it to state-of-the-art place recognition methods using six datasets with extreme environmental changes. Experimental results demonstrate the effectiveness and improvements achieved using the proposed framework. Source code and some trained models will be available at http://www.tianshu.org.cn.