Sequential Therapy Based on Evolvement of Patterns: A New Model for Treatment of Alzheimer’s Disease

In order to solve the problem of long-term (>9 months) efficacy in the treatment of Alzheimer''s disease (AD) by conventional therapy (CT), a staged and multiply-targeted sequential therapy based on the evolvement of patterns (STEP) was developed. Its main innovations include: (1) the time order of evolution of patterns defined by Chinese medicine (CM) in AD was found, that is, "the orderly pattern evolution starting from Shen (Kidney) deficiency, progressing to phlegm, stasis and fire, and worsening to severe toxin as well as functional collapse"; (2) the cascade hypothesis of Shen deficiency in AD and its sequential therapy based on Shen-reinforcing was proposed, that is, "reinforcing Shen in the early stage and throughout the whole process, resolving phlegm, activating blood and purging fire in the middle stage, detoxifying and replenishing vitality to stop the collapse in the advanced stage", and through meta-analysis, clinical drug use was optimized, thus the leap from "inferential selection" to "evidence-based selection" was realized; (3) the STEP regimen combined with CT maintained cognitive and behavioral stability in AD patients for at least 12 months, with cognitive enhancement and behavioral synergy after 9 months, and cognitive benefit was superior to CT at 9, 12, 15, 18, 21, and 24 months, respectively. The 2-year cognitive improvement rate was increased by 25.64% (P=0.020) and the cognitive deterioration rate was decreased by 48.71% (P=0.000). Among them, the cognitive and functional benefits of Shen-reinforcing therapy for very early AD (350 cases) for 1 year were better than the placebo (P<0.001), and the dementia conversion rate was reduced by 8.85% (P=0.002). The behavioral symptomatic relief of patients with vascular dementia received fire-purging therapy (540 cases) was superior to those received CT (P=0.016). These data suggested that the STEP regimen has synergistic effects on CTs at least in terms of cognitive benefit, and the earlier the use, the greater the benefit will have. Therefore, the STEP regimen should be considered as one of the clinical options, particularly for the dearth of effective pharmaceutical or immunological interventions that are currently available for AD.     

改良单图像法手工测量髌骨不稳患者胫骨结节-股骨滑车沟间距的可行性分析

目的 探讨改良单图像法手工测量髌骨不稳患者胫骨结节-股骨滑车沟（tibial tuberosity-trochlear groove，TT-TG）间距的可行性。 方法 选取髌骨不稳患者30例。高年资手术医师A、B，使用改良单图像法手工测量髌骨不稳患者TT-TG间距，对比高资年影像学医师C，使用双图法测量髌骨不稳患者TT-TG间距。应用Cronbach′s alpha系数评价结果的可信度。应用组内相关系数（intraclass correlation coefficient，ICC）评价结果的可重复性。使用Bland-Altman分析图评价结果的一致性。 结果 高年资手术医师A、B使用改良单图像法手工测量髌骨不稳患者TT-TG间距分别和高年资影像学医师C使用双图法测量髌骨不稳患者TT-TG间距比较，Cronbach′s alpha系数分别为0.913和0.959，组内相关系数分别为0.913和0.958。Bland-Altman分析图显示，TT-TG间距差值（mm）的均值为-0.276（95%CI：-3.526~2.974）和0.143（95%CI：-3.110~3.397）。A、B医师之间使用改良单图像法手工测量髌骨不稳患者TT-TG间距比较，Cronbach′s alpha系数为0.891，组内相关系数为0.891。Bland-Altman分析图显示TT-TG间距差值（mm）的均值为-0.276（95%CI：-3.526~2.974）。 结论 改良单图像法手工测量髌骨不稳患者TT-TG间距操作简单，可以准确测量出髌骨不稳患者的TT-TG间距，该方法和双图法可以相互替换使用。     

从温阳透发法论治难治性痤疮

痤疮可辨为阴、阳两证。阳疮者,缘患者劳汗被寒、玄府闭合,汗脂凝聚日久不得外排,蕴热搏结气血而生疮化腐,疮症为热为痛,此为易治;阴疮者,由阳疮转化而来,缘患者外感寒风、内伤生冷,且医者又重用寒凉、失于托透,以至气血冰凝,久不溃脓或溃脓难出,此为难治。根据难治性痤疮临床表现,其多类属于阴疮,以阳虚表郁、邪闭毛窍为致病之关键,其中表郁毛窍、邪闭肌腠为病之标,阳虚阴凝为病之本。治疗当以温阳透发、祛邪外达为法。以麻黄附子细辛汤温阳化阴滞、发阳透毒治其本,待阴邪发越,即用火针强开门户、消疮排脓治其标。标本兼治、针药结合的方法治疗难治性痤疮在临床中能够安全有效地消除痤疮、改善皮损、降低复发率。     

A Multi-Scale DNN Algorithm for Nonlinear Elliptic Equations with Multiple Scales

Algorithms based on deep neural networks (DNNs) have attracted increasing attention from the scientific computing community. DNN based algorithms are easy to implement, natural for nonlinear problems, and have shown great potential to overcome the curse of dimensionality. In this work, we utilize the multi-scale DNN-based algorithm (MscaleDNN) proposed by Liu, Cai and Xu (2020) to solve multi-scale elliptic problems with possible nonlinearity, for example, the p-Laplacian problem. We improve the MscaleDNN algorithm by a smooth and localized activation function. Several numerical examples of multi-scale elliptic problems with separable or non-separable scales in low-dimensional and high-dimensional Euclidean spaces are used to demonstrate the effectiveness and accuracy of the MscaleDNN numerical scheme.