基于域对齐的深度学习方法在急性缺血性卒中的CT/MRI影像分割上的应用

目的  基于平扫CT提出一种域对齐方法来显著提高急性缺血性卒中(AIS)的早期快速诊断能力。方法  回顾性分析南方医科大学第三附属医院神经内科和神经外科2020年1月~2022年12月收治的入院后3 d内同时接受平扫头颅CT和MRI/DWI、ADC以及T2-Flair序列扫描的AIS患者,构建了一个由318例AIS病例组成的成对CT/MRI影像数据集,分别对每一组配对的教师-学生影像特征进行归一化;再以8∶2的比例随机分为训练集和验证集。设计一种新的生成性对抗性网络来对齐特征层上的跨模式输入,将细节丰富的MRI图像中的语义知识传递到CT图像中进行AIS分割,开发了一种新的域适应算法（Our DA）。结果  与目前性能表现较优异的医学影像分割模型nnUNet相比,Our DA明显优于nnU-Net,每一层验证集之间的分割精度提升约15%。结论  本研究构建的Our DA模型基于MRI/DWI序列的影像特征并迁移到平扫头颅CT上,对平扫头颅CT上的AIS病灶具有较高的自动分割性能,有助于早期自动识别AIS病灶。     

Rare bioparticle detection via deep metric learning

Recent deep neural networks have shown superb performance in analyzing bioimages for disease diagnosis and bioparticle classification. Conventional deep neural networks use simple classifiers such as SoftMax to obtain highly accurate results. However, they have limitations in many practical applications that require both low false alarm rate and high recovery rate, e.g., rare bioparticle detection, in which the representative image data is hard to collect, the training data is imbalanced, and the input images in inference time could be different from the training images. Deep metric learning offers a better generatability by using distance information to model the similarity of the images and learning function maps from image pixels to a latent space, playing a vital role in rare object detection. In this paper, we propose a robust model based on a deep metric neural network for rare bioparticle (Cryptosporidium or Giardia) detection in drinking water. Experimental results showed that the deep metric neural network achieved a high accuracy of 99.86% in classification, 98.89% in precision rate, 99.16% in recall rate and zero false alarm rate. The reported model empowers imaging flow cytometry with capabilities of biomedical diagnosis, environmental monitoring, and other biosensing applications.

Conventional deep neural networks use simple classifiers to obtain highly accurate results. However, they have limitations in practical applications. This study demonstrates a robust deep metric neural network model for rare bioparticle detection.     

早期功能训练对脑卒中急性下肢深静脉血栓形成的影响

目的：比较脑卒中后急性下肢深静脉血栓形成（DVT）早期康复治疗与卧床制动对下肢功能影响及并发肺梗死风险的差异。方法：脑卒中并急性下肢深静脉血栓形成患者60例,随机分为观察组和对照组各30例,2组均予常规抗凝及消肿治疗,对照组予卧床、弹力袜支持及患肢抬高治疗7～14d后开始行肢体功能康复,观察组予早期（下肢血栓形成后1～2d）弹力袜支持下行肢体功能康复治疗。比较2组患者3个月内肺梗死的累计发生率和功能恢复情况。结果：2组患者3个月内均无症状性肺梗死病例发生。治疗3个月后,2组患者下肢FMA评分均较治疗前明显升高（P〈0．05,0．01）,且观察组更高于对照组（P〈0．01）;治疗后,2组VAS评分及双侧小腿内踝上缘治疗前后周径差值均较治疗前明显下降（P〈0．05,0．01）,且观察组更低于对照组（P〈0．01）。结论：脑卒中患者急性下肢深静脉血栓形成早期开始康复治疗较卧床治疗患肢功能改善更明显,且不增加肺梗死的发生率。     

Simulation of hyperelastic materials in real-time using deep learning

The finite element method (FEM) is among the most commonly used numerical methods for solving engineering problems. Due to its computational cost, various ideas have been introduced to reduce computation times, such as domain decomposition, parallel computing, adaptive meshing, and model order reduction. In this paper we present U-Mesh: A data-driven method based on a U-Net architecture that approximates the non-linear relation between a contact force and the displacement field computed by a FEM algorithm. We show that deep learning, one of the latest machine learning methods based on artificial neural networks, can enhance computational mechanics through its ability to encode highly non-linear models in a compact form. Our method is applied to three benchmark examples: a cantilever beam, an L-shape and a liver model subject to moving punctual loads. A comparison between our method and proper orthogonal decomposition (POD) is done through the paper. The results show that U-Mesh can perform very fast simulations on various geometries and topologies, mesh resolutions and number of input forces with very small errors.     

动态血压监测应用于脑卒中急性期患者的研究

目的:对脑卒中急性期患者进行24 h动态血压监测(Ambulatory Blood Pressure Monitor,ABPM),分析不同血压类型患者的血压特点和预后,并制定个体化血压护理干预措施。方法:采用MOBIL-O-GRATH动态血压监护仪,对265例脑卒中急性期患者入院48 h内进行24 h动态血压监测,并对不同血压类型患者ABPM监测结果和发病3个月预后进行比较分析。结果:动态血压监测结果为勺型、非勺型和反勺型血压患者分别为80例、90例和95例;3种血压类型患者的晨起收缩压、夜间脉压、夜间收缩压及舒张压、夜间舒张压变异性、24小时平均动脉压及发病后3个月的预后比较差异有统计学意义(P0.05)。结论:根据脑卒中急性期患者动态血压监测结果,不仅可以分析不同血压类型患者的血压特点和预后,还可为制定个体化血压护理干预措施提供依据。     

Early mobilization of acute stroke patients

Objective. To examine the early mobilization of acute stroke patients. Design. Postal survey. Setting. Thirteen health boards in Scotland. Participants. Ninety‐nine health professionals of whom 39 were doctors, 39 nurses and 21 physiotherapists. Results. There was a lack of understanding and agreement across the three professions in terms of what was meant by ‘early mobilization’. Further, the duration, frequency, intensity, risk/benefits and activities associated with early mobilization are undescribed despite clinical guidelines urging its use. Multi‐disciplinary decision making regarding early mobilization was not self‐evident. Conclusions. (i) An evidence‐base for early mobilization is required along with agreement on what physiological monitoring should be undertaken while early mobilization is on going; (ii) Health professionals need a greater awareness of the evidence linking stroke complications with patient immobilization and in particular in relation to pressure sores, painful shoulder and falls; (iii) The clinical decision to mobilize an acute stroke patient early should be made explicitly within a multi‐disciplinary acute stroke team; (iv) There is an absolute need for further research into early mobilization in terms of intensity, duration, frequency, risks and benefits in relations to types of stroke of early mobilization. Relevance to clinical practice. Early mobilization in acute stroke care is recommended in a range of European, American and UK policy guidelines as a strategy to minimize or prevent complications. However the evidence‐base to support early mobilization in acute stroke is missing. Health professionals require a research‐based approach in order to deliver safe and effective early mobilization to acute stroke patients.     

Predicting outcomes of patients in Japan after first acute stroke using a simple model

OBJECTIVE: Prediction of patient outcome can be useful as an aid to clinical decision making. Many studies, including my own, have constructed predictive multivariate models for outcome following stroke rehabilitation therapy, but these have often required several minutes work with a pocket calculator. The aim is to develop a simple, easy-to-use model that has strong predictive power. METHODS: Four hundred sixty-four consecutive patients with first stroke who were admitted to a rehabilitation hospital during a period of 19 mo were enrolled in the study. Sex, age, the stroke type, Functional Independence Measure total score on admission (X), onset to admission interval (number of days from stroke onset to rehabilitation admission), and length of rehabilitation hospital stay (number of days from hospital admission to discharge) were the independent variables. Functional Independence Measure total score at discharge (Y) was the dependent variable. RESULTS: Stepwise multiple regression analysis resulted in the model containing age (P < 0.0001), X (P < 0.0001), and onset to admission interval (P < 0.0001). The equation was: Y = 68.6 - 0.32 (age) + 0.80X - 0.13 (onset to admission interval), a multiple correlation coefficient (R) = 0.82, and a multiple correlation coefficient squared (R2) = 0.68. Simple regression analysis revealed the relation between Xand Y: Y = 0.85X + 37.36, and R = 0.80 R2 = 0.64. In fact, plots of X vs. Ywere nonlinear, but seemed to be able to be linearized by some form of equation. It was found that there is a linear relation between logX and Y. The equation is Y = 106.88x - 95.35, where x = logX, R = 0.84, and R2 = 0.70. The correlation is improved by a regression analysis of a natural logarithmic transformation of X (R = 0.84 vs. R = 0.82). CONCLUSION: The results in this study confirm that the simple regression model using a logarithmic transformation of X (R = 0.84) has predictive power over the simple regression model (R = 0.80). This model is well validated and clinically useful.     

急性期脑卒中患者下肢深静脉血栓风险预测模型的构建及应用

目的 构建并应用急性期脑卒中患者下肢深静脉血栓（deep vein thrombosis,DVT）风险列线图预测模型。方法 采用前瞻性研究设计,便利选取2020年1月—2021年4月在南宁市某三级甲等综合医院住院的602例急性期脑卒中患者作为研究对象。其中2020年1月—12月的415例作为建模组,2021年1月—4月的187例作为验证组对模型进行外部验证。采用单因素和多因素Logistic回归分析急性期脑卒中患者下肢DVT危险因素,建立风险预测模型并绘制列线图。采用受试者操作特征曲线（receiver operating characteristic,ROC）和Hosmer-Lemeshow检验验证模型预测效果。结果 建模组415例中有35例发生DVT,发生率为8.4%;验证组187例中有19例发生DVT,发生率为10.2%。建模组中单因素分析结果显示,年龄、诊断、卧床时间、意识状态、偏瘫程度,是否有吸烟史、房颤史、血栓史,是否使用脱水药物、是否留置中心静脉导管、血浆纤维蛋白原、D-二聚体定量是急性期脑卒中患者发生DVT的影响因素。多因素Logistic回归分析结果显示,年龄、意识状态、偏瘫程度、是否使用脱水药物是急性期脑卒中患者发生DVT的独立影响因素（OR值分别为1.901、1.702、1.940、3.231,均P<0.05）,以上述4个因素为自变量构建列线图,模型ROC曲线下面积为0.850,约登指数最大值为0.758时,灵敏度为83%,特异度为82%,最佳临界值为0.071。Hosmer-Lemeshow拟合优度检验 χ²=2.143,P=0.951;外部验证组ROC曲线下面积为0.893,约登指数最大值为0.746时,灵敏度为90%,特异度为85%,最佳临界值为0.084。结论 构建的列线图可个性化预测急性期脑卒中患者DVT发生风险,有助于护理人员制订相应的干预措施。     

Nurse-led dysphagia screening in acute stroke patients

This article discusses the findings of an audit to assess the improved outcomes of a systematic approach to training nurses working in an emergency assessment area (EAA) to conduct dysphagia screening for patients who have had a stroke. The investment in training has reduced the time patients wait for dysphagia screening from 35 hours to less than one hour. As a result of this audit dysphagia screening competencies have been established.     

急性脑卒中后下肢深静脉血栓形成的研究

目的观察急性脑卒中后下肢深静脉血栓形成(DVT)的发病情况。方法对215例发病72h内的脑卒中住院患者于发病后4d及14d进行双下肢深静脉超声检测，确定下肢DVT的发生率。结果急性脑卒中下肢DVT的发病率为7．9％，卒中后4d即可出现，多数在发病14d后被检出。DVT主要影响患肢，也有部分病例累及双下肢。结论急性脑卒中患者早期即可出现下肢DVT，应尽早采取预防措施以降低其危害性。     

尿激酶超早期动脉溶栓并支架成形术治疗急性缺血性脑梗死患者的护理

目的 总结急性缺血性脑梗死超早期动脉溶栓并支架成形术的护理.方法 回顾性分析9例急性缺血性脑梗死患者动脉溶栓并支架成形术的术后临床资料.应用美国国立卫生院卒中评分(即NIHSS评分)以判断患者的意识和神经功能恢复情况.结果 9例患者中7例患者缺血症状均缓解,2例因术后颅内出血死亡.神经功能评分明显改善.结论 溶栓并支架成形术能够提高急性缺血性卒中患者的治愈率,但也具有一定危险性.做好术后监测对挽救患者生命,有着重要意义.     

Levodopa improves procedural motor learning in chronic stroke patients

Rösser N, Heuschmann P, Wersching H, Breitenstein C, Knecht S, Flöel A. Levodopa improves procedural motor learning in chronic stroke patients.

Objective

To test the hypothesis that administration of dopamine precursor levodopa improves procedural motor learning (defined as the ability to acquire novel movement patterns gradually through practice) in patients with residual motor deficits in the chronic phase after stroke (≥1y after stroke).

Design

A double-blind, placebo-controlled, randomized crossover design.

Setting

Neurology department in a German university.

Participants

Eighteen patients with chronic motor dysfunction because of stroke (13 men, 5 women; age range, 53-78y; mean time poststroke ± SD, 3.3±2.1y).

Intervention

Patients received 3 doses of levodopa (100mg of levodopa plus 25mg of carbidopa) or placebo before 1 session of procedural motor learning.

Main Outcome Measures

Procedural motor learning performed by using the paretic hand assessed by using a modified version of the serial reaction time task with a probabilistic sequence. The primary outcome measure was the difference in reaction times between random and sequential elements.

Results

Levodopa significantly improved our primary outcome measure, procedural motor learning, compared with placebo (P<.05). Reaction times to random elements, analysis of error rates, psychophysical assessments, and performance in a simple motor task were comparable between conditions, indicating that better learning under levodopa was not caused by differences in response styles, arousal, mood, or motor reaction times but that levodopa modulated learning.

Conclusions

Our results show that levodopa may improve procedural motor learning in patients with chronic stroke, in line with our hypothesis. These findings suggest that this interventional strategy in combination with customary rehabilitative treatments could significantly improve the outcome of neurorehabilitation in the chronic stage after stroke. (Clinicaltrials.gov identifier NCT00126087.)     

急性脑卒中患者发生误吸的危险因素分析

摘要 目的：探讨急性脑卒中患者误吸的发生率及其危险因素。 方法：71例急性脑卒中患者根据纤维鼻咽喉镜吞咽功能检查（FEES）的结果分为误吸组和无误吸组。对可能影响卒中后误吸发生的危险因素进行多因素Logistic回归分析。 结果：急性脑卒中患者误吸的发生率为50.7%,36.1%的误吸患者为无症状性误吸。单因素分析显示年龄、脑卒中史和卒中严重程度是卒中后误吸的危险因素(P<0.05)。多因素Logistic回归分析表明严重卒中（OR＝5.778,95%CI 1.123—29.737）和既往有脑卒中史（OR=3.302,95%CI 1.174—9.293）是卒中后误吸的独立危险因素。 结论：误吸是脑卒中急性期的常见问题,严重卒中（NIHSS>10分）和脑卒中史是卒中后误吸的独立危险因素。 关键词 脑血管意外;误吸;危险因素 中图分类号：R743.3,R493 文献标识码：A 文章编号：1001-1242（2010）-02-0131-04     

Large scale deep learning for computer aided detection of mammographic lesions

Recent advances in machine learning yielded new techniques to train deep neural networks, which resulted in highly successful applications in many pattern recognition tasks such as object detection and speech recognition. In this paper we provide a head-to-head comparison between a state-of-the art in mammography CAD system, relying on a manually designed feature set and a Convolutional Neural Network (CNN), aiming for a system that can ultimately read mammograms independently. Both systems are trained on a large data set of around 45,000 images and results show the CNN outperforms the traditional CAD system at low sensitivity and performs comparable at high sensitivity. We subsequently investigate to what extent features such as location and patient information and commonly used manual features can still complement the network and see improvements at high specificity over the CNN especially with location and context features, which contain information not available to the CNN. Additionally, a reader study was performed, where the network was compared to certified screening radiologists on a patch level and we found no significant difference between the network and the readers.