Using a monotone single-index model to stabilize the propensity score in missing data problems and causal inference

The augmented inverse weighting method is one of the most popular methods for estimating the mean of the response in causal inference and missing data problems. An important component of this method is the propensity score. Popular parametric models for the propensity score include the logistic, probit, and complementary log-log models. A common feature of these models is that the propensity score is a monotonic function of a linear combination of the explanatory variables. To avoid the need to choose a model, we model the propensity score via a semiparametric single-index model, in which the score is an unknown monotonic nondecreasing function of the given single index. Under this new model, the augmented inverse weighting estimator (AIWE) of the mean of the response is asymptotically linear, semiparametrically efficient, and more robust than existing estimators. Moreover, we have made a surprising observation. The inverse probability weighting and AIWEs based on a correctly specified parametric model may have worse performance than their counterparts based on a nonparametric model. A heuristic explanation of this phenomenon is provided. A real-data example is used to illustrate the proposed methods.     

Exploratory Study of Apparent Diffusion Coefficient Histogram Metrics in Assessing Pancreatic Malignancy

PurposeTo evaluate whole-lesion 3D-histogram apparent diffusion coefficient (ADC) metrics for assessment of pancreatic malignancy.MethodsForty-two pancreatic malignancies (36 pancreatic adenocarcinoma [PDAC], 6 pancreatic neuroendocrine [PanNET]) underwent abdominal magnetic resonance imaging (MRI) with diffusion-weighted imaging before endoscopic ultrasound biopsy or surgical resection. Two radiologists independently placed 3D volumes of interest to derive whole-lesion histogram ADC metrics. Mann-Whitney tests and receiver operating characteristic analyses were used to assess metrics’ diagnostic performance for lesion histology, T-stage, N-stage, and grade.ResultsWhole-lesion ADC histogram metrics lower in PDACs than PanNETs for both readers (P ≤ .026) were mean ADC (area under the curve [AUC] = 0.787-0.792), mean of the bottom 10th percentile (mean_0-10) (AUC = 0.787-0.880), mean of the 10th-25th percentile (mean_10-25) (AUC = 0.884-0.917) and mean of the 25th-50th percentile (mean_25-50) (AUC = 0.829-0.829). For mean_10-25 (metric with highest AUC for identifying PDAC), for reader 1 a threshold > 0.94 × 10^?3 mm²/s achieved sensitivity 94% and specificity 83%, and for reader 2 a threshold > 0.82 achieved sensitivity 97% and specificity 67%. Metrics lower in nodal status ≥ N₁ than N₀ for both readers (P ≤ .043) were mean_0-10 (AUC = 0.789-0.822) and mean_10-25 (AUC = 0.800-0.822). For mean_10-25 (metric with highest AUC for identifying N₀), for reader 1 a threshold <1.17 achieved sensitivity 87% and specificity 67%, and for reader 2 a threshold <1.04 achieved sensitivity 87% and specificity 83%. No metric was associated with T-stage (P > .195) or grade (P > .215).ConclusionVolumetric ADC histogram metrics may serve as non-invasive biomarkers of pancreatic malignancy. Mean_10-25 outperformed standard mean for lesion histology and nodal status, supporting the role of histogram analysis.     

DVH预测模型在VMAT计划培训中的应用

目的:以宫颈癌为例，探索使用剂量体积直方图（DVH）预测模型培训放疗物理师容积旋转调强（VMAT）计划设计的可行性及其效果。方法:随机选取20例宫颈癌测试病例对3名见习物理师进行两个阶段培训（第一阶段15例，第二阶段5例）。每位物理师分别采用两种方法设计VMAT计划:传统人工计划（MP）和基于DVH预测模型引导的自动计划（KBP）。对比不同阶段两种计划靶区和危及器官的各种剂量学参数，并做相应统计分析。结果:与人工计划相比，自动计划在明显提高PGTV靶区V60覆盖（P<0.001），改善高剂量控制（V66）（P=0.027）的情况下，显著降低膀胱（P<0.001）、直肠（P<0.001）、左右肾（P=0.001和P<0.001）以及左右侧股骨头（P<0.001和P<0.001）等绝大部分正常组织器官的受照剂量。在提高计划合格率的同时，亦减小了计划者间的质量差异。结论:DVH预测模型有助于深化物理师对VMAT初始优化参数设置的理解，加快培训进度，提升VMAT计划设计水平。     

Yttrium-90 Radioembolization to the Prostate Gland: Proof of Concept in a Canine Model and Clinical Translation

PurposeTo investigate the feasibility, safety, and absorbed-dose distribution of prostatic artery radioembolization (RE) in a canine model.Materials and MethodsFourteen male castrated beagles received dihydroandrosterone/estradiol to induce prostatic hyperplasia for the duration of the study. Each dog underwent fluoroscopic prostatic artery catheterization. Yttrium-90 (⁹⁰Y) microspheres (TheraSphere; Boston Scientific, Marlborough, Massachusetts) were delivered to 1 prostatic hemigland (dose escalation from 60 to 200 Gy), with the contralateral side serving as a control. Assessments for adverse events were performed throughout the follow-up (Common Terminology Criteria for Adverse Events v5.0). Positron emission tomography/magnetic resonance (MR) imaging provided a confirmation after the delivery of absorbed-dose distribution. MR imaging was performed before and 3, 20, and 40 days after RE. Tissue harvest of the prostate, rectum, bladder, urethra, penis, and neurovascular bundles was performed 60 days after RE.ResultsAll the animals successfully underwent RE. Positron emission tomography/MR imaging demonstrated localization to and good coverage of only the treated hemigland. No adverse events occurred. The MR imaging showed a significant dose-dependent decrease in the treated hemigland size at 40 days (25%–60%, P < .001). No extraprostatic radiographic changes were observed. Necropsy demonstrated no gross rectal, urethral, penile, or bladder changes. Histology revealed RE-induced changes in the treated prostatic tissues of the highest dose group, with gland atrophy and focal necrosis. No extraprostatic RE-related histologic findings were observed.ConclusionsProstate ⁹⁰Y RE is safe and feasible in a canine model and leads to focal dose-dependent changes in the gland without inducing unwanted extraprostatic effects. These results suggest that an investigation of nonoperative prostate cancer is warranted.     

基于预测剂量引导的宫颈癌自动计划研究

目的:使用机器学习方法建立宫颈癌计划剂量预测回归模型，并将预测剂量引导生成Monaco 计划系统（TPS）可调用 的优化模板文件，实现宫颈癌的自动计划设计。方法:对50例宫颈癌术后调强治疗计划中的危及器官采集基于重叠体积直 方图的几何特征值和基于剂量直方图的剂量目标值，建模后将模型预测剂量结果自动生成Monaco TPS模板文件，进而由TPS 调用优化。使用该方法对另外10例未参与模型训练的测试病例进行自动计划设计，并和人工设计的计划进行对比分析。 结果:自动计划比手动计划的平均设计时间减少了40 min（P<0.05），且平均调优次数降低了3次（P<0.05），剂量学指标和计 划执行效率上两者无明显差异（P>0.05）。结论:基于预测剂量引导的宫颈癌自动计划可以达到临床要求，并且提高了计划 设计效率。     

口腔鳞癌癌旁组织血管的形态定量研究

目的了解口腔鳞癌癌务组织血管的变化及其生物学意义。方法以34例口腔鳞癌瘤旁组织为标本，通过抗第八因子相关抗原免疫组化染色标记血管，作血管的形态定量分析。结果肿瘤结缔组织中血管密度显著高于癌旁组织及正常组织中血管密度，且与肿瘤细胞分化程度相关；紧邻癌旁的组织与远处及正常组织相比也有血管密度的明显增加，证明肿瘤的血管生成作用仅在短距离区域内有效。结论癌旁组织的血管形态学研究对于肿瘤生物学行为的认识、治疗的选择及预后的判断均有重要意义。     

Bp-MRI灰度直方图在鉴别移行带前列腺癌与良性前列腺增生中的应用价值

目的:探讨双参数磁共振成像（Bp-MRI）灰度直方图在鉴别移行带前列腺癌（PCa）与良性前列腺增生（BPH）中的价值。方法:回顾性分析经病理证实的移行带PCa患者43例与BPH患者52例。T2WI和ADC图像上,利用MaZda软件手动勾画感兴趣区（ROI）,进行灰度直方图分析,提取均值、方差、偏度、峰度、百分位数9个直方图纹理特征,对纹理特征进行统计学分析,获得有意义的纹理特征参数,并绘制受试者工作特征（ROC）曲线,以评价其鉴别移行带PCa与BPH的价值。结果:BPH组ADC均值、偏度、各百分位数均高于PCa组,差异均有统计学意义（t=-12.155～5.022,均P<0.05）,方差、峰度差异无统计学意义（均P>0.05）。BPH组T2WI均值、方差、第10百分位数、第50百分位数、第90百分位数、第99百分位数均高于PCa组,差异均有统计学意义（t=-6.618～-2.181,均P<0.05）,偏度、峰度、第1百分位数差异均无统计学意义（均P>0.05）。ADC均值、偏度、第1百分位数、第10百分位数、第50百分位数、第90百分位数、第99百分位数曲线下面积（AUC）差异均有统计学意义（均P<0.05）;T2WI均值、方差、第10百分位数、第50百分位数、第90百分位数、第99百分位数AUC差异均有统计学意义（均P<0.05）。其中,ADC均值的AUC最大,为0.963,以149.29为阈值时,敏感度为90.7%,特异度为92.3%。结论:Bp-MRI灰度直方图分析可以作为鉴别PCa与BPH的重要辅助手段。     

Revision Surgery for Failed Cervical Spine Reconstruction: Review Article

Background

As the number of cervical spine procedures performed continues to increase, the need for revision surgery is also likely to increase. Surgeons need to understand the etiology of post-surgical changes, as well as have a treatment algorithm when evaluating these complex patients.

Questions/Purposes

This study aims to review the rates and etiology of revision cervical spine surgery as well as describe our treatment algorithm.

Methods

We used a narrative and literature review. We performed a MEDLINE (PubMed) search for “cervical” and “spine” and “revision” which returned 353 articles from 1993 through January 22, 2014. Abstracts were analyzed for relevance and 32 articles were reviewed.

Results

The rates of revision surgery on the cervical spine vary by the type and extent of procedure performed. Patient evaluation should include a detailed history and review of the indication for the index procedure, as well as lab work to rule out infection. Imaging studies including flexion/extension radiographs and computed tomography are obtained to evaluate potential pseudarthrosis. Magnetic resonance imaging is helpful to evaluate the disc, neural elements, soft tissue, and to differentiate scar from infection. Sagittal alignment should be corrected if necessary.

Conclusions

Recurrent or new symptoms after cervical spine reconstruction can be effectively treated with revision surgery after identifying the etiology, and completing the appropriate workup.

Electronic supplementary material

The online version of this article (doi:10.1007/s11420-014-9394-8) contains supplementary material, which is available to authorized users.     

斜外侧腰椎体间融合术初次术后邻近节段退变

[目的]评价斜外侧腰椎体间融合术(oblique lumbar interbody fusion,0LIF)治疗腰椎融合术后邻近节段退变的临床效果。[方法]回顾性分析2016年12月一2019年12月本院脊柱外科采用0LIF术治疗腰椎融合术后邻近节段退变50例患者的临床资料。[结果]50例患者均顺利完成手术,均未发生严重并发症。所有患者随访12?16个月,平均(13.74±1.63)个月。术后(12.66±3.64)周患者恢复完全负重活动。随访期间,患者术后疼痛逐步减缓,功能逐步改善。与术前相比较,末次随访时VAS和0DI评分均显著下降(P<0.05)。影像方面,与术前相比,末次随访时患者的腰椎前凸角(LL)显著增加(P<0.05),而侧凸Cobb角显著减少(P<0.05)。至末次随访时,50例患者再次手术椎间隙均达到骨性融合,椎间融合器无移位、下沉。[结论]采用0LIF治疗腰椎融合术后邻近节段退变具有较好的安全性和有效性。