A Stacked Generalization of 3D Orthogonal Deep Learning Convolutional Neural Networks for Improved Detection of White Matter Hyperintensities in 3D FLAIR Images

BACKGROUND AND PURPOSE:Accurate and reliable detection of white matter hyperintensities and their volume quantification can provide valuable clinical information to assess neurologic disease progression. In this work, a stacked generalization ensemble of orthogonal 3D convolutional neural networks, StackGen-Net, is explored for improving automated detection of white matter hyperintensities in 3D T2-FLAIR images.MATERIALS AND METHODS:Individual convolutional neural networks in StackGen-Net were trained on 2.5D patches from orthogonal reformatting of 3D-FLAIR (n = 21) to yield white matter hyperintensity posteriors. A meta convolutional neural network was trained to learn the functional mapping from orthogonal white matter hyperintensity posteriors to the final white matter hyperintensity prediction. The impact of training data and architecture choices on white matter hyperintensity segmentation performance was systematically evaluated on a test cohort (n = 9). The segmentation performance of StackGen-Net was compared with state-of-the-art convolutional neural network techniques on an independent test cohort from the Alzheimer’s Disease Neuroimaging Initiative-3 (n = 20).RESULTS:StackGen-Net outperformed individual convolutional neural networks in the ensemble and their combination using averaging or majority voting. In a comparison with state-of-the-art white matter hyperintensity segmentation techniques, StackGen-Net achieved a significantly higher Dice score (0.76 [SD, 0.08], F1-lesion (0.74 [SD, 0.13]), and area under precision-recall curve (0.84 [SD, 0.09]), and the lowest absolute volume difference (13.3% [SD, 9.1%]). StackGen-Net performance in Dice scores (median = 0.74) did not significantly differ (P = .22) from interobserver (median = 0.73) variability between 2 experienced neuroradiologists. We found no significant difference (P = .15) in white matter hyperintensity lesion volumes from StackGen-Net predictions and ground truth annotations.CONCLUSIONS:A stacked generalization of convolutional neural networks, utilizing multiplanar lesion information using 2.5D spatial context, greatly improved the segmentation performance of StackGen-Net compared with traditional ensemble techniques and some state-of-the-art deep learning models for 3D-FLAIR.

White matter hyperintensities (WMHs) correspond to pathologic features of axonal degeneration, demyelination, and gliosis observed within cerebral white matter.¹ Clinically, the extent of WMHs in the brain has been associated with cognitive impairment, Alzheimer’s disease and vascular dementia, and increased risk of stroke.^2,3 The detection and quantification of WMH volumes to monitor lesion burden evolution and its correlation with clinical outcomes have been of interest in clinical research.^4,5 Although the extent of WMHs can be visually scored,⁶ the categoric nature of such scoring systems makes quantitative evaluation of disease progression difficult. Manually segmenting WMHs is tedious, prone to inter- and intraobserver variability, and is, in most cases, impractical. Thus, there is an increased interest in developing fast, accurate, and reliable computer-aided automated techniques for WMH segmentation.Convolutional neural network (CNN)-based approaches have been successful in several semantic segmentation tasks in medical imaging.⁷ Recent works have proposed using deep learning–based methods for segmenting WMHs using 2D-FLAIR images.^8-11 More recently, a WMH segmentation challenge¹² was also organized (http://wmh.isi.uu.nl/) to facilitate comparison of automated segmentation of WMHs of presumed vascular origin in 2D multislice T2-FLAIR images. Architectures that used an ensemble of separately trained CNNs showed promising results in this challenge, with 3 of the top 5 winners using ensemble-based techniques.¹²Conventional 2D-FLAIR images are typically acquired with thick slices (3–4 mm) and possible slice gaps. Partial volume effects from a thick slice are likely to affect the detection of smaller lesions, both in-plane and out-of-plane. 3D-FLAIR images, with isotropic resolution, have been shown to achieve higher resolution and contrast-to-noise ratio¹³ and have shown promising results in MS lesion detection using 3D CNNs.¹⁴ Additionally, the isotropic resolution enables viewing and evaluation of the images in multiple planes. This multiplanar reformatting of 3D-FLAIR without the use of interpolating kernels is only possible due to the isotropic nature of the acquisition. Network architectures that use information from the 3 orthogonal views have been explored in recent works for CNN-based segmentation of 3D MR imaging data.¹⁵ The use of data from multiple planes allows more spatial context during training without the computational burden associated with full 3D training.¹⁶ The use of 3 orthogonal views simultaneously mirrors how humans approach this segmentation task.Ensembles of CNNs have been shown to average away the variances in the solution and the choice of model- and configuration-specific behaviors of CNNs.¹⁷ Traditionally, the solutions from these separately trained CNNs are combined by averaging or using a majority consensus. In this work, we propose the use of a stacked generalization framework (StackGen-Net) for combining multiplanar lesion information from 3D CNN ensembles to improve the detection of WMH lesions in 3D-FLAIR. A stacked generalization¹⁸ framework learns to combine solutions from individual CNNs in the ensemble. We systematically evaluated the performance of this framework and compared it with traditional ensemble techniques, such as averaging or majority voting, and state-of-the-art deep learning techniques.     

普萘洛尔联合其他治疗措施治疗婴幼儿血管瘤有效率的网状Meta分析

目的:采用网状Meta分析方法,将普萘洛尔联合其他治疗与各对照治疗措施进行对比。方法:以“普萘洛尔”,“血管瘤”为关键词检索CNKI、VIP、万方数据;以“propranolol”“hemangioma”为关键词检索Cochrane Library、Embase、pubmed。检索期限为建库至2019年6月1日。采用STATA14.0软件Network程序包进行数据分析。结果:共纳入18篇文献,涉及8项治疗措施,共纳入1469例血管瘤患者。网状Meta分析结果显示:4种联合治疗方式与单纯口服普萘洛尔相比疗效均优于单纯口服普萘洛尔治疗。普萘洛尔联合注射平阳霉素、普萘洛尔联合外用噻吗洛尔、普萘洛尔联合敷贴器的有效率均优于对应的注射平阳霉素、外用噻吗洛尔及敷贴器。各治疗措施的有效性排序为:普萘洛尔联合敷贴器>普萘洛尔联合外用噻吗洛尔>普萘洛尔联合口服糖皮质激素>普萘洛尔联合注射平阳霉素>敷贴器照射>外用噻吗洛尔>口服普萘洛尔>注射平阳霉素。结论:对于婴幼儿血管瘤的治疗,普萘洛尔联合治疗措施疗效均优于单纯的口服普萘洛尔治疗,其中普萘洛尔联合敷贴器的疗效最佳。     

12-year survival analysis of 322 Hintegra total ankle arthroplasties from an independent center

Background and purpose — Total ankle arthroplasties (TAAs) have larger revision rates than hip and knee implants. We examined the survival rates of our primary TAAs, and what different factors, including the cause of arthritis, affect the success and/or revision rate.Patients and methods — From 2004 to 2016, 322 primary Hintegra TAAs were implanted: the 2nd generation implant from 2004 until mid-2007 and the 3rd generation from late 2007 to 2016. A Cox proportional hazards model evaluated sex, age, primary diagnosis, and implant generation, pre- and postoperative angles and implant position as risk factors for revision.Results — 60 implants (19%) were revised, the majority (n = 34) due to loosening. The 5-year survival rate (95% CI) was 75% (69–82) and the 10-year survival rate was 68% (60–77). There was a reduced risk of revision, per degree of increased postoperative medial distal tibial angle at 0.84 (0.72–0.98) and preoperative talus angle at 0.95 (0.90–1.00), indicating that varus ankles may have a larger revision rate. Generation of implant, sex, primary diagnosis, and most pre- and postoperative radiological angles did not statistically affect revision risk.Interpretation — Our revision rates are slightly above registry rates and well above those of the developer. Most were revised due to loosening; no difference was demonstrated with the 2 generations of implant used. Learning curve and a low threshold for revision could explain the high revision rate.

Arthritis in the ankle often develops earlier than in the hip or knee, and 70% have a traumatic etiology (Saltzman et al. 2005, Brown et al. 2006). Total ankle arthroplasty (TAA) can be indicated for severe arthritis in the ankle joint, but the anatomical preconditions, like a small surface area and high stress from compression and torque (Bouguecha et al. 2011, Kakkar and Siddique 2011), makes it less durable than hip and knee prosthetics. The Hintegra TAA, a 3-component mobile bearing, uncemented implant (Hintermann et al. 2004) is widely used and results from the development center demonstrate survival rates of 94% and 84% after 5 and 10 years’ follow-up (Barg et al. 2013). This is considerably more than the survival rates from national registries. Labek et al. (2011) demonstrated that development centers report only half of the revision rate that can be found in the few existing national registers. In a systematic review of primary Agility total ankle arthroplasty (DePuy Synthes Orthopedics, Warsaw, IN, USA), the author (Roukis 2012) found that the incidence of complications increased from 7% to 12%, in studies where the inventor was excluded. Similar results were found by Prissel and Roukis (2013), who found an increased incidence of complications from 6% to 13% in studies where the inventor or faculty consultants were excluded. These studies indicated the risk of selection (inventor) and publication (conflict of interest) bias.Planning and surgical technique, including significant experience, are mandatory for a successful outcome. The better result from development centers may reflect, besides the above-mentioned bias, that there is a long learning curve and that the indication for revision surgery varies.We examined the survival rates of primary Hintegra TAAs performed at Hvidovre Hospital, with revision rate as outcome. We report primary diagnosis for primary TAA and examine whether sex, generation of the implant, preoperative angles and implant position affect the revision rate.     

Skin closure with 2-octyl cyanoacrylate and polyester mesh after primary total knee arthroplasty offers superior cosmetic outcomes and patient satisfaction compared to staples: a prospective trial

European Journal of Orthopaedic Surgery & Traumatology - The goals of this study were to compare patient satisfaction and wound-related complications in patients receiving 2-octyl cyanoacrylate...     

足趾移植长手指全形再造术的临床应用

目的探讨足趾移植长手指全形再造手术的临床疗效。方法自2015年6月至2019年6月,对16例因外伤致手指缺损患者采用足趾移植长手指全形再造术,术后评估供区及受区的感觉功能、运动功能及外观形态,分析指甲畸形及增生性瘢痕的发生情况,并记录术后发生感染、血肿、皮片坏死及供区愈合不良等情况;通过调查问卷的方式分析患者的满意度。结果所有患者术后获随访1~12个月,其中2例受区发生感染,1例受区皮片边缘发生坏死,经换药后予以缓解;其余患者的供、受区均未出现长时间的痛疼感觉,受区感觉功能恢复达87.50%,受区运动功能恢复均较满意,手指外形基本满意。供区感觉受影响者2例,运动功能受限者1例,外形一般者2例。所有患者无指甲畸形及增生性瘢痕发生;满意者1例,基本满意者14例,不满意者1例。结论采用足趾移植长手指全形再造手术,基本可以满足患者及医师对于缺损手指进行完美修复的目标。     

手术治疗结节性甲状腺肿合并甲状腺癌的临床效果

目的:分析手术治疗结节性甲状腺肿合并甲状腺癌的临床效果。方法:从2017年3月~2019年3月期间到我院接受治疗的结节性甲状腺肿合并甲状腺癌患者中选出60例完成此次研究,60例患者全部接受手术治疗,经1年的随访后对比两组患者接受治疗前后的生活质量变化情况及治疗效果。结果:接受治疗后患者的心理、身体、物质生活、社会关系等多项生活质量指标均优于治疗前,治疗前后的数据对比存在明显差异(P<0.05)。所有患者接受治疗后均为表现出较为明显的不良反应,其治疗存在较高安全性,总治疗有效率为87.50%(55/60)。结论:帮助结节性甲状腺肿合并甲状腺癌患者选择合理的手术方式,能够提高临床治疗效果,减少并发症。     

不同健康状况的中老年人社会资本量表的信度及效度评价

目的 评价中老年人社会资本量表的信效度，探索其在不同健康状况的中老年人群间的区分度。方法 针对已编制的中老年人社会资本量表，采用多阶段分层随机抽样方法，在四川省成都市及泸州市共抽取1 367名50岁及以上中老年居民进行问卷调查。运用Cronbach’s α系数、Pearson相关系数、t检验、探索性因子分析等方法评价量表的信度和效度。结果 1 367名中老年调查对象的社会资本量表总Cronbach’s α系数为0.67，各维度条目与总分的相关系数在0.30～0.79之间，具有良好的信度；探索性因子分析表明，14个条目提取出5个特征根大于1的因子，累积方差贡献率为68.84%；现患慢性疾病的中老年病例组人群的个人社会资本得分、家庭社会资本得分、社团社会资本得分、社区社会资本得分及社会资本总得分均低于未患病的对照人群组，宏观社会资本得分高于对照人群组，且差异均有统计学意义 (均 P＜0.01)，提示该量表具有较好的区分度。结论 中老年人社会资本量表具有较好的信度和效度，且对不同健康状况的50岁及以上中老年人中有较好的区分度。     

应用国产封堵器在治疗先天性心脏病的临床疗效分析

目的 总结国产封堵器经皮和经胸途径治疗先天性心脏病的临床疗效。方法 回顾性分析我院2013年1月至2017年12月在X线透视下或单纯超声心动图引导下采用经皮穿刺股静脉或股动静脉法，或者食管超声心动图监测下经胸小切口行先天性心脏病封堵1186例，其中经皮X线下封堵1081例、经皮单纯超声引导下封堵42例、经胸封堵63例；其中动脉导管未闭（ patent ductus arteriole，PDA）426例、房间隔缺损（atrial septal defect，ASD）363例、室间隔缺损（ventricular septal defect，VSD）348例、卵圆孔未闭11例、房间隔缺损合并室间隔缺损9例、房间隔缺损合并动脉导管未闭6例、房间隔缺损合并肺动脉瓣狭窄（pulmonary stenosis，PS）12例、动脉导管未闭合并肺动脉瓣狭窄8例、主肺动脉侧支封堵3例[经胸封堵66例改为63例，PDA443例改为426例]。结果 全组病例成功率98.2％（1165/1186），无死亡病例。随访1~36个月，术后第1、3、6、12个月及术后每年常规行超声心动图及心电图检查。术后第1、6、12个月的随访率分别为 92.9%（1102/1186）、84.1%（998/1186）、70.5%（836/1186）。超声心动图提示少量残余分流（<3 mm）18例；三尖瓣少量反流33例，中量反流5例；主动脉瓣轻度反流5例，中度反流1例；心律失常Ⅲ°房室传导阻滞（Avionics Bulletin，AVB[房室传导阻滞（Avionics Bulletin，AVB）]）1 例，Ⅱ°AVB 3例，完全性左束支3例，交界性心动过速3例，交界性逸搏2例。结论 国产封堵器在先天性心脏病封堵治疗中具有成功率高、创伤小、并发症低、操作容易、疗效确切、恢复快等特点，是治疗先天性心脏病的理想方法。