Anatomic resection versus radiofrequency ablation with an ablative margin ≥ 1.0 cm for solitary small hepatocellular carcinoma measuring ≤ 3 cm: Comparison of long-term outcomes using propensity score matching analysis.

PURPOSE: To compare the long-term outcomes of anatomic resection (AR) and radiofrequency ablation (RFA) with an ablative margin (AM) of ≥ 1.0 cm as first-line treatment for solitary hepatocellular carcinoma measuring ≤ 3 cm. METHODS: Two hundred and fifty-one patients who underwent AR (n = 156) or RFA (ablative margin ≥ 1.0 cm, n = 95) at any of 6 tertiary hospitals from 2009 to 2018 were enrolled. Propensity score matched analysis (PSM) were used to compare overall survival (OS), recurrence-free survival (RFS), and perioperative outcomes. Univariate and multivariate analyses were performed to identify prognostic factors associated with RFS and OS. RESULTS: PSM created 67 patient-pairs. After 96 months of follow-up, RFA with an ablative margin ≥ 1.0 cm and AR showed comparable 1-year, 3-year, 5-year, and 8-year OS rates before (P = 0.580) and after (P = 0.640) PSM. However, RFS was better at 1, 3, 5, and 8 years after AR before (P = 0.0036) and after (P = 0.017) PSM. The operation time and postoperative hospital stay were significantly longer in the AR group than in the RFA group before and after PSM (P < 0.05). Multivariate analysis identified age and type of treatment to be independent prognostic factors for RFS and age and hepatitis C to be associated with OS. CONCLUSIONS: Long-term OS was not significantly different between AR and RFA with an AM ≥ 1.0 cm in patients with a solitary hepatocellular carcinoma measuring ≤ 3 cm; but, RFS appeared to be better after AR than after RFA. However, RFA was associated with fewer perioperative complications and a shorter postoperative hospital stay.

A Neural Network with Convolutional Module and Residual Structure for Radar Target Recognition Based on High-Resolution Range Profile.

In the conventional neural network, deep depth is required to achieve high accuracy of recognition. Additionally, the problem of saturation may be caused, wherein the recognition accuracy is down-regulated with the increase in the number of network layers. To tackle the mentioned problem, a neural network model is proposed incorporating a micro convolutional module and residual structure. Such a model exhibits few hyper-parameters, and can extended flexibly. In the meantime, to further enhance the separability of features, a novel loss function is proposed, integrating boundary constraints and center clustering. According to the experimental results with a simulated dataset of HRRP signals obtained from thirteen 3D CAD object models, the presented model is capable of achieving higher recognition accuracy and robustness than other common network structures.