正则化多任务学习在精神分裂症核磁共振成像图像分类中的应用

为实现对精神疾病的计算机辅助诊断与预后,利用机器学习与图像处理技术分析多地区精神疾病的核磁共振成像数据,已成为该领域的必然趋势。本文首先提出切片提取的核磁共振成像图像预处理方法,然后提取图像的纹理特征,最后提出一种lp范数正则化的多任务学习支持向量机精神分裂症分类方法,同时学习3个数据中心精神分裂症图像的共享特征和各自独有的特征,用于分类精神分裂症患者和正常人。实验结果表明,该方法取得了优秀的诊断精度,可为精神分裂症患者的临床诊断与治疗提供生物学依据。     

基于特征选择的阿尔茨海默症辅助诊断

阿尔茨海默症（AD）是一种在老年人中多发的脑部神经疾病,致病原因迄今未明,在疾病发展早期难以诊断。随着 计算机和人工智能技术的大力发展,利用磁共振成像（MRI）技术和机器学习方法辅助医生对AD进行辅助诊断不断取得 新的成果。本研究提出一种基于支持向量机递归特征消除（SVM-RFE）和线性判别分析（LDA）的AD辅助诊断方法。首 先对MRI图像进行预处理,获得90个大脑脑区的灰质体积;然后使用SVM-RFE和LDA相结合的方法,对90个大脑脑区 灰质体积进行特征选择;最后通过SVM进行分类。通过对来自于ADNI数据库中的34名AD、26名主观记忆衰退（SMC） 患者和50名正常被试（NC）的MRI图像分析,得到AD/NC、AD/SMC和NC/SMC的平均分类准确率分别为94.0%、100.0% 和93.6%。实验结果证明,本研究提出的方法可有效提取样本特征,辅助医生诊断AD。     

基于KPCA算法的阿尔茨海默症辅助诊断

阿尔茨海默症（AD）是一种起病隐匿、进行性发展的神经系统退行性疾病,利用磁共振成像和计算机技术对AD患者的辅助诊断是目前不断探索的新课题。本研究先对磁共振图像进行预处理和相关性分析,然后利用核主成分分析法（KPCA）对脑灰质图像进行特征提取,结合Adaboost算法进行分类,并与主成分分析法（PCA）进行对比试验。通过对AD神经影像学计划数据库中的116名AD患者、116名轻度认知障碍患者,以及117名正常对照的脑部功能磁共振成像进行的研究表明,利用机器学习能够很有效地辅助诊断AD脑部疾病,KPCA算法对图像进行特征提取比PCA 算法更加充分完备,分类结果更加精确,能够获得更好的AD辅助诊断结果。     

Diagnosis of cerebrovascular disorders using nuclear magnetic resonance tomography

Regarding cerebrovascular diseases magnetic resonance imaging (MRI) provides informations which could not be obtained before by any other imaging method. MRI permits an early detection of ischemic infarctions, exact delineation of lesions near to skull base, and microangiopathy. Further diagnosis of arteriovenous malformations and venous thromboses can be done without using contrast media. Exclusion of acute hemorrhage or subarachnoidal hemorrhage is still a domain of computed tomography (CT). Other MR methods, like magnetic resonance angiography (MRA) and magnetic resonance spectroscopy (MRS) are expected to replace angiography and positron emission tomography (PET) in the near future, at least for some purposes.     

胎儿胼胝体发育不全的磁共振成像与超声检查对比研究

目的:本研究以磁共振成像(MRI)为参考标准,评价了产前超声检查诊断胎儿胼胝体发育不全(ACC)的能力,并评估了MR图像的神经解剖学特征。方法:超声检查发现胎儿神经发育异常可能的孕妇共计263例,对其胎儿头颅MR图像进行评估。结果:胎儿MRI累计诊断25例ACC,其中24例未被超声检查发现。超声检查诊断ACC的准确性为91%。另外,MRI发现了3例具有相似形态学特征的ACC,即非典型泪滴征。结论:与超声检查相比,MRI检测ACC更加准确。非典型泪滴征是ACC少见且特殊的MRI征象。     

磁共振结肠造影的临床应用研究

目的：探讨磁共振结肠造影（ MRC）检查的临床应用价值。方法：收集167例结肠病变患者的MRC资料,总结其影像表现特点,并将MRC诊断结果与肠镜活检或手术病理结果进行对比分析。所有病例行肠道准备,经肛门注入生理盐水1200～1800 ml后,行多方位、多序列MR扫描及增强扫描。结果：167例患者MRC诊断正确147例,误诊15例,漏诊5例。 MRC诊断的准确率88．02％。结肠恶性肿瘤的MRC主要征象为肠壁局限性增厚或软组织肿块,多伴有肠腔狭窄;结肠良性肿瘤多表现为结节或肿块,边缘较光整;结肠炎性病变MRC表现为不均质肿块或肠壁广泛增厚,境界不清。结论：高质量的MRC图像对结直肠病变的检出及定性诊断有较高的准确性,对临床诊断及治疗具有重要的指导意义。     

Accurate magnetic resonance imaging of atherosclerotic plaques: change future strategies for the diagnosis and therapy of atherosclerotic disease

In recent years, magnetic resonance imaging (MRI) have been developed to image atherosclerosis and is emerging as a useful tool to assess the burden of atherosclerosis, whereas the potential influence on the diagnosis and therapy of atherosclerotic disease have not been fully determined. MRI allows for three-dimensional evaluation of vascular structures and outstanding depiction of various components of the atherosclerotic plaque. The self-contained intravascular MRI probe appears to hold promise in the identification of high-risk coronary and peripheral atherosclerotic lesions. Molecular and targeted contrast MRI can offer exciting possibilities of direct visualization of biologic processes within atherosclerotic tissue. The addition of quantitative hydrogen 1 magnetic resonance spectroscopy and diffusion weighted imaging within atherosclerotic plaques can provide important data on the biological activity of potentially vulnerable lesions. Therefore, we hypothesized that accurate magnetic resonance imaging of atherosclerotic plaques maybe further affect and change future strategies for the diagnosis and therapy of atherosclerotic disease.     

Automatic Cardiac Segmentation Using Semantic Information from Random Forests

We propose a fully automated method for segmenting the cardiac right ventricle (RV) from magnetic resonance (MR) images. Given a MR test image, it is first oversegmented into superpixels and each superpixel is analyzed to detect the presence of RV regions using random forest (RF) classifiers. The superpixels containing RV regions constitute the region of interest (ROI) which is used to segment the actual RV. Probability maps are generated for each ROI pixel using a second set of RF classifiers which give the probabilities of each pixel belonging to RV or background. The negative log-likelihood of these maps are used as penalty costs in a graph cut segmentation framework. Low-level features like intensity statistics, texture anisotropy and curvature asymmetry, and high level context features are used at different stages. Smoothness constraints are imposed based on semantic information (importance of each feature to the classification task) derived from the second set of learned RF classifiers. Experimental results show that compared to conventional method our algorithm achieves superior performance due to the inclusion of semantic knowledge and context information.     

基于随机森林算法的阿尔茨海默症医学影像分类

为实现阿尔茨海默症(AD)的医学影像分类,辅助医生对患者的病情进行准确判断,本研究对采集的34名AD患者、35名轻度认知障碍患者和35名正常对照组成员的功能磁共振影像进行特征提取和分类,具体思路包括:首先利用皮尔逊相关系数计算脑区之间的功能连接,然后采用随机森林算法对被试不同脑区之间的功能连接进行重要性度量及特征选择,最后使用支持向量机分类器进行分类,利用十倍交叉验证估算分类准确率。实验结果显示,随机森林算法可以对功能连接特征进行有效分析,同时得到AD发病过程的异常脑区,基于随机森林和SVM建立的分类模型对AD、轻度认知障碍的识别具有较好的效果,分类准确率可达90.68%,相关结论可以为AD的早期临床诊断提供客观参照。 【关键词】阿尔茨海默症;功能磁共振成像;随机森林;特征选择     

rs-fMRI、DTI及1H-MRS在阿尔茨海默病中的应用进展

随着磁共振成像技术的飞速发展,对阿尔茨海默病的神经影像学研究也越来越深入。其主要目的是通过磁共振影像学特征来识别出早期阿尔茨海默病患者,从而达到早期诊断和治疗的目的。本文就磁共振成像技术包括静息态功能磁共振成像、弥散张量成像及磁共振波谱在阿尔茨海默病的应用进展进行综述。     

Intratumoral hemorrhage, vessel density, and the inflammatory reaction contribute to volume increase of sporadic vestibular schwannomas

Vestibular schwannomas show a large variation in growth rate, making prediction and anticipation of tumor growth difficult. More accurate prediction of clinical behavior requires better understanding of tumor biological factors influencing tumor progression. Biological processes like intratumoral hemorrhage, cell proliferation, microvessel density, and inflammation were analyzed in order to determine their role in vestibular schwannoma development. Tumor specimens of 67 patients surgically treated for a histologically proven unilateral vestibular schwannoma were studied. Preoperative magnetic resonance imaging (MRI) scans were used to determine tumor size and to classify tumors as homogeneous, inhomogeneous, and cystic. Immunohistochemical studies evaluated cell proliferation (histone H3 and Ki-67), microvessel density (CD31), and inflammation (CD45 and CD68). Intratumoral hemorrhage was assessed by hemosiderin deposition. The expression patterns of these markers were compared with tumor size, tumor growth index, MRI appearance, patients' age, and duration of symptoms. No relation between cell proliferation and clinical signs of tumor volume increase or MRI appearance was found. Intratumoral hemosiderin, microvessel density, and inflammation were significantly positively correlated with tumor size and the tumor growth index. Cystic and inhomogeneous tumors showed significantly more hemosiderin deposition than homogeneous tumors. The microvessel density was significantly higher in tumors with a high number of CD68-positive cells. The volume increase of vestibular schwannomas is not based on cell proliferation alone. Factors like intratumoral bleeding, (neo)vascularization, and intensity of the inflammatory reaction also influence tumor volume.     

甲状腺呈胸腺样分化癌病理观察

目的 探讨甲状腺呈胸腺样分化癌（CASTLE）的病理特征、诊断及鉴别诊断。方法 分析2例甲状腺CASTLE的临床表现,对标本进行病理学和免疫组织化学（EnVision^TM法）染色观察。结果 大体检查肿瘤呈质硬灰白色略旱分叶状肿块。光镜下肿瘤组织呈轮廓清楚的巢状,巢间可见丰富的促结缔组织增生性间质。肿瘤细胞呈多边形或梭形,胞质轻度嗜酸,核卵圆形,呈空泡状,有清楚的小核仁,核异型性较轻,核分裂象1～2／10HPF。免疫组织化学肿瘤细胞表达CD5及CD117。结论 CASTLE是一种少见的甲状腺恶性肿瘤,有一定的病理学特征。应与甲状腺未分化癌、甲状腺鳞状细胞癌、转移性淋巴上皮瘤样癌及滤泡树突细胞肉瘤等鉴别。免疫组织化学CD5等标记对鉴别诊断有帮助。     

Synovial Haemangioma — Gadolinium enhanced MRI scanning is the investigation of choice for planned surgical excision

We present three cases of histologically proven synovial haemangiomas of the knee. We discuss the presenting features, differential diagnoses and the best investigative modality of this rare condition, for which historically diagnostic delay has been common. Angiographic and magnetic resonance appearances, including post-gadolinium studies, are demonstrated. Magnetic resonance imaging (MRI) is non-invasive, non-irradiating and accurately demonstrates full extent of the lesions, allowing accurate preoperative planning prior to surgical excision.     

Shape-Adaptive DCT for Denoising of 3D Scalar and Tensor Valued Images

During the last ten years or so, diffusion tensor imaging has been used in both research and clinical medical applications. To construct the diffusion tensor images, a large set of direction sensitive magnetic resonance image (MRI) acquisitions are required. These acquisitions in general have a lower signal-to-noise ratio than conventional MRI acquisitions. In this paper, we discuss computationally effective algorithms for noise removal for diffusion tensor magnetic resonance imaging (DTI) using the framework of 3-dimensional shape-adaptive discrete cosine transform. We use local polynomial approximations for the selection of homogeneous regions in the DTI data. These regions are transformed to the frequency domain by a modified discrete cosine transform. In the frequency domain, the noise is removed by thresholding. We perform numerical experiments on 3D synthetical MRI and DTI data and real 3D DTI brain data from a healthy volunteer. The experiments indicate good performance compared to current state-of-the-art methods. The proposed method is well suited for parallelization and could thus dramatically improve the computation speed of denoising schemes for large scale 3D MRI and DTI.     

基于多参数磁共振影像组学特征的脑垂体瘤质地术前评估方法

为了探究多参数磁共振图像(MP-MRI)特征对脑垂体瘤质地评估的应用价值,提出一种基于影像组学的计算机辅助诊断方法,以期实现术前肿瘤质地的准确判定,从而为手术入路的选择提供影像学依据。对磁共振图像(T1加权、T1加权对比增强、T2加权)的肿瘤区域分别提取6种共296维纹理特征。采用特征选择方法识别重要的影像组学特征,并且使用支持向量机和随机森林两种常用的分类器对垂体瘤质软与质韧进行判别。在84例临床研究样本共计252张MRI图像上,用所述方法进行训练、十折交叉验证及测试。实验结果表明,与单一MRI图像特征相比较,所提出的MP-MRI特征组合能够获得更好的分类效果,分类准确率、敏感性、特异性、AUC分别达到89.80%、90.51%、89.88%、94.08%,表明MP-MRI影像组学特征能够有效准确地识别垂体瘤的软韧质地,有助于垂体瘤疗效和预后的改善。     

Imaging in patients with a first seizure

Neuroimaging has important applications in the diagnosis and treatment of patients with seizures and epilepsy and can contribute to the proper classification of certain epileptic disorders. When a patient presents to the Accident and Emergency Department with a first seizure, the clinician may request brain imaging to determine the cause of the seizure. In new onset seizures or epilepsy with suspected focal pathology, computed tomography (CT) of the brain should be performed if the patient has not fully recovered or if magnetic resonance imaging (MRI) is not available within a reasonable time period. Otherwise, it is reasonable to forego CT and perform MRI. This report emphasises the need for appropriate imaging for accurate diagnosis of suspected new-onset partial epilepsy.     

Contrast agent-free state-of-the-art magnetic resonance imaging on cerebral small vessel disease – Part 2: Diffusion tensor imaging and functional magnetic resonance imaging

Cerebral small vessel disease (cSVD) has been widely studied using conventional magnetic resonance imaging (MRI) methods, although the association between MRI findings and clinical features of cSVD is not always concordant. We assessed the additional contribution of contrast agent-free, state-of-the-art MRI techniques, particularly diffusion tensor imaging (DTI) and functional magnetic resonance imaging (fMRI), to understand brain damage and structural and functional connectivity impairment related to cSVD. We performed a review following the PICOS worksheet and Search Strategy, including 152 original papers in English, published from 2000 to 2022. For each MRI method, we extracted information about their contributions regarding the origins, pathology, markers, and clinical outcomes in cSVD. In general, DTI studies have shown that changes in mean, radial, and axial diffusivity measures are related to the presence of cSVD. In addition to the classical deficit in executive functions and processing speed, fMRI studies indicate connectivity dysfunctions in other domains, such as sensorimotor, memory, and attention. Neuroimaging metrics have been correlated with the diagnosis, prognosis, and rehabilitation of patients with cSVD. In short, the application of contrast agent-free, state-of-the-art MRI techniques has provided a complete picture of cSVD markers and tools to explore questions that have not yet been clarified about this clinical condition. Longitudinal studies are desirable to look for causal relationships between image biomarkers and clinical outcomes.     

微囊型脑膜瘤与脑膜尤文肉瘤/外周原始神经外胚层肿瘤的MRI鉴别诊断

目的:探讨微囊型脑膜瘤与脑膜尤文肉瘤/外周原始神经外胚层肿瘤（pPNET）的MRI征象。方法:回顾性分析经病理证实的14例微囊型脑膜瘤和10例脑膜尤文肉瘤/pPNET患者的临床、影像学及病理资料,结合相关文献分析其MRI差异。结果:微囊型脑膜和脑膜尤文肉瘤/pPNET患者MRI特征在肿瘤形态、T2信号、强化特征、骨侵犯及远处转移方面有统计学意义（P<0.05）;而在T1信号、硬膜尾征和瘤周水肿无统计学意义（P>0.05）。结论:微囊型脑膜瘤和脑膜尤文肉瘤/pPNET影像表现有差异,术前精准鉴别诊断有助于改善临床决策和预测预后。     

CT/MRI双模态造影剂的制备及研究进展

电子计算机断层扫描（CT）用途广泛,常用于解剖学成像,骨成像和肿瘤成像等,但软组织对比度较低。核磁共振成像（MRI）的软组织对比度高,但成像时间长,价格昂贵。将CT成像技术与MRI技术合二为一使其优势互补,可以在短时间内得到更准确的生物信息。如今CT/MRI双模态造影剂相关的综述文献并不是很多。本文在CT/MRI双模态分子影像前驱物的基础上归纳一些制备CT/MRI双模态造影剂的方法,如纳米乳法,表面物理修饰法,表面化学修饰法,氯胺-T碘标记法,氧化还原法,并展望了未来CT/MRI双模态造影剂的发展。     

一种基于光抽运效应的磁共振信号检测方法

与常规磁共振成像相比,超低场磁共振成像技术所需的成像磁体可降至mT级甚至μT级,不仅大大降低了设备的成本和体积,还能够获得普通高场磁共振成像设备无法得到的特殊图像,具有非常广阔的应用前景。本文介绍了一种基于光抽运现象的磁共振信号检测方法,该方法根据外加磁场会影响碱金属原子对特定波长光吸收的原理,通过检测透射光强的变化检测磁共振信号。文章最后利用线圈产生模拟磁共振信号,给出了一维及二维磁共振成像的实验结果,结果表明该方法对于磁共振信号有良好的频率分辨率及相位控制的稳定性。由于该方法具有高灵敏度、低成本以及在信号检测时需要很弱背景场等特点,因此特别适用于超低场磁共振信号的检测。