应用层析X线摄影合成对膝关节骨赘及软骨下囊的检查

目的评价层析X线摄影组合在描述骨赘及软骨下囊的诊断能力,应用MR成像作为参考,检验伴有疼痛的病变能否被X线摄影及层析X线摄影组合所检测。材料与方法本研究经地方机构审查委员会批准,且所有受试者均填写知情同意书。40名40岁以上的受试者(80个膝关节),     

CT肺动脉成像诊断肺栓塞的一致性研究

目的评价CT肺动脉成像（CTPA）诊断肺栓塞（PE）时,不同经验的读片者间和同一读片者内的一致性。方法 55例临床可疑PE患者行CTPA检查,6位不同经验的放射科医生独立地分析CTPA图像来评价读片者间的一致性。3位放射科医生3个月后第二次分析CTPA图像来评价读片者内的一致性。PE的表现分为阳性、阴性和难以确定。读片者一致性用百分比及Kappa系数表示。结果 6位读片者判定29～31例（平均29.2例）患者CTPA为PE阳性,1～5例（平均3.0例）患者CTPA为难以确定。6位读片者在48例（87.3%）患者CTPA的诊断上取得一致意见,5位读片者在4例患者（7.3%）的诊断上取得一致意见,4位读片者在2例患者（3.6%）的诊断上取得一致意见,3位读片者在1例患者（1.8%）的诊断上取得一致意见。在诊断PE上,如果以每例患者为观察单位,读片者间的一致性＂非常好＂（Kappa值为0.91）。以每个肺动脉为观察单位,读片者间的一致性＂好＂（85%,Kappa值为0.74）;以肺叶动脉为观察单位,读片者间的一致性＂好＂（89%,Kappa值为0.78）;以肺段动脉为观察单位,读片者间的一致性＂中等＂（75%,Kappa值为0.59）。如果以每例患者为观察单位,同一读片者内的平均一致性＂非常好＂（96%,Kappa值为0.93）。结论在CTPA上诊断PE时,经验不同的读片者间和同一读片者内的一致性均较好。     

Application of ultrasonic dual-mode artificially intelligent architecture in assisting radiologists with different diagnostic levels on breast masses classification

PURPOSEWe aimed to compare the diagnostic performance and interobserver variability in breast tumor classification with or without the aid of an innovative dual-mode artificial intelligence (AI) architecture, which can automatically integrate information from ultrasonography (US) and shear-wave elastography (SWE).METHODSDiagnostic performance assessment was performed with a test subset, containing 599 images (from September 2018 to February 2019) from 91 patients including 64 benign and 27 malignant breast tumors. Six radiologists (three inexperienced, three experienced) were assigned to read images independently (independent diagnosis) and then make a secondary diagnosis with the knowledge of AI results. Sensitivity, specificity, accuracy, receiver-operator characteristics (ROC) curve analysis and Cohen’s κ statistics were calculated.RESULTSIn the inexperienced radiologists’ group, the average area under the ROC curve (AUC) for diagnostic performance increased from 0.722 to 0.765 (p = 0.050) with secondary diagnosis using US-mode and from 0.794 to 0.834 (p = 0.019) with secondary diagnosis using dual-mode compared with independent diagnosis. In the experienced radiologists’ group, the average AUC for diagnostic performance was significantly higher with AI system using the US-mode (0.812 vs. 0.833, p = 0.039), but not for dual-mode (0.858 vs. 0.866, p = 0.458). Using the US-mode, interobserver agreement among all radiologists improved from fair to moderate (p = 0.003). Using the dual-mode, substantial agreement was seen among the experienced radiologists (0.65 to 0.74, p = 0.017) and all radiologists (0.62 to 0.73, p = 0.001).CONCLUSIONAI assistance provides a more pronounced improvement in diagnostic performance for the inexperienced radiologists; meanwhile, the experienced radiologists benefit more from AI in reducing interobserver variability.

As a global medical problem, breast cancer is the most commonly diagnosed cancer and the leading cause of cancer death among females (1). The development of imaging technologies have made great contribution to breast cancer diagnosis (2). Ultrasonography (US) is a valuable supplemental scanning modality for women with dense breasts (3–5). Radiologists make final diagnosis based on the morphological characteristics of breast masses according to the interpretation of Breast Imaging Reporting and Data System (BI-RADS) (6). For US BI-RADS classification, interobserver agreement ranging from fair to substantial has been reported among radiologists with different experience levels (7). Therefore, the subjectivity and empirical dependence aroused controversy towards its true diagnostic efficacy. Many studies have shown that the combined use of US-mode and ultrasound elastography has greater diagnostic efficacy than US-mode alone (8–11). However, until now, there was no consistent standard value of shear-wave elastography (SWE) in differentiation between malignant and benign breast masses which might result from the numerous reference parameters including maximum (Emax) and standard deviation (SD) (12). Both US-mode and SWE provide us with abundant uncertain information, which inexperienced radiologists cannot extract from, interpret, and utilize accurately. Therefore, the standardization and precision diagnosis of US and SWE are the challenges and research hotspots of ultrasound technology currently.The field of biomedical image analysis benefited substantially from rapid developments in artificial intelligence (AI) techniques. Previous studies on traditional computer-aided diagnosis (CAD) for breast cancer have shown that the segmentation of US and SWE images and integration of extracted features achieved better classification performance than their individual uses (13–15). With the CAD system, the average diagnostic performance of radiologists was improved or at least comparable to the independent diagnosis based on breast US images (16, 17). An observational study conducted by Van Zelst et al. (18) concluded that CAD software for automated breast US may speed up the screening time without compromising the screening performance of radiologists. As a revolutionary advance, the deep learning algorithm (convolution neural network for image processing) has a strong feature extraction ability, which can extract higher level features rather than superficial features and provide more possibilities for medical image processing. Recently few studies have shown the great potential of deep learning framework in ultrasound image processing of breast lesions (19–22).However, the abovementioned observational studies were focused only on a single mode of ultrasonic breast images. Throughout these papers, there was no deep learning model which is capable of integrating US with SWE images of breast lesions rationally. In clinical practice, comprehensive analysis with dual-mode ultrasound is of great importance for breast lesions. To meet the clinical needs, we designed a novel module named “shared latent subspace learning,” which captures the inter-mode relationship between SWE-mode and US-mode via a shared-parameter Dense Block that is optimized under adversarial loss and orthogonality constraint. Similar to the normal working mode of radiologists and just like the workflow of CAD software (23), radiologists will combine the diagnostic results given by software and achieve the secondary diagnostic results after comprehensive consideration. Therefore, we conducted a dual-mode reader study where we investigated the benefits of incorporating an innovative AI architecture. We compared the diagnostic performance and interobserver agreement of different radiologists in breast mass classification through independent diagnosis and secondary diagnosis.     

Role of different imaging modalities in assessment of temporomandibular joint erosions and osteophytes: a systematic review

OBJECTIVES: To evaluate the ability of different diagnostic imaging techniques for diagnosing the presence of erosions and osteophytes in the temporomandibular joint (TMJ). METHODS: A systematic search of PubMed, Medline, all Evidence Based Medicine (EBM) reviews, Embase, Web of Sciences and Lilacs identified nine articles that met the selection criteria: some type of TMJ diagnostic imaging, data from autopsy or dry skull TMJs as gold standard, absence of diagnosed systemic arthritis and evaluation of the presence of erosions and/or osteophytes. A hand search of the references of the selected articles was also performed. RESULTS: Selected studies evaluated panoramic imaging (unenhanced and colour-enhanced digital subtraction panoramic imaging), axially corrected sagittal tomography, axially corrected frontal tomography, sagittal MRI, CT, high-resolution ultrasound and cone beam CT (CBCT). CONCLUSIONS: Axially corrected sagittal tomography is currently the imaging modality of choice for diagnosing erosions and osteophytes in the TMJ. CT does not seem to add any significant information to what is obtained from axially corrected sagittal tomography. CBCT might prove to be a cost- and radiation dose-effective alternative to axially corrected sagittal tomography. Combining different radiographic techniques is likely to be more accurate in diagnosing erosions and osteophytes in the TMJ than using a single imaging modality. Diagnostic studies that simultaneously evaluate all of the available TMJ imaging technologies are needed.     

基于CT三维重建分析距骨软骨下骨囊肿形态学及分布特点

目的:通过CT三维重建分析距骨软骨下骨囊肿的形态学及分布特点。方法:在天津医院影像报告数据库中检索2015—2020年经踝关节或足部CT扫描后诊断为距骨软骨下骨囊肿患者176例,其中男77例,女99例;年龄14~84岁［（56.1±14.0）岁］。通过Mimics 20.0软件三维重建距骨及囊肿区域后,构建田字解剖网格...     

Diagnostic evaluation of separately acquired PET and CT images by nuclear medicine physicians and radiologists in cancer patients

The aim of our study was to analyze how many oncology patients might benefit from a) integrated positron emission tomography - multidetector computed tomography (PET/MDCT) and additionally b) clinically relevant information provided by either the CT scan or PET scan. A total of 285 consecutive patients 164 male and 121 female, age range 17-84 years, 153 lung cancer, 112 lymphoma, 20 miscellaneous, referred for PET and separate CT scan, were included. The CT scan was performed after the intravenous injection of a soluble contrast media. Patients were retrospectively classified into six Groups: Group I: No pathological uptake on the PET scan, Group II: Suspected lesions were correctly identified by the PET scan alone, Group III: Side-by-side evaluation of PET and CT appeared sufficient to assess the localization of lesions, Group IV: Side-by-side reading was not sufficient and integrated PET/CT was considered beneficial. Additionally all patients with a CT scan with additional clinical relevant information (not visualized by the PET scan) were classified in Group V. Group VI was set for lesions detected by PET alone (not visualized by the CT scan). The CT scan was used as the gold standard to confirm or disprove PET lesion localization. Our results showed: A number of 77 patients, (Group I: 77/285, 27%) had no pathologic fluorine-18-fluorodeoxyglucose (18F-FDG)-uptake. Lesions were correctly localized by either conventional PET alone (Group II: 76/285, 27%) or side-by-side evaluation of PET and CT scans (Group III: 44/285, 15%). Integrated PET/CT or software fusion, was considered beneficial in 31% (88/285) of the patients with pathological 18F-FDG-uptake (Group IV). Additionally to the above, in 15% of all patients clinically relevant information, referring to disseminated small pulmonary lesions, abdominal aortic aneurysms >5 cm, thrombi or pulmonary emboli, was also provided by the CT scan (Group V). Also, in 7% of all patients, unsuspected pathological lesions, mainly bone metastases, were correctly detected by PET alone (Group VI). In conclusion, in 54% of all oncologic patients, PET alone was diagnostic. In 46% of all patients side-by-side reading (15%) or integrated PET/CT images (31%) were considered beneficial for more accurate anatomical localization of the lesions. Additionally, the CT scan added clinically relevant information in 15% of all patients and the PET scan showed unsuspected metastases in 7% of all studied patients. Therefore, integrated reading of PET and MDCT images by nuclear physicians and radiologists may gain quality in the staging of oncology patients.     

Performance evaluation of radiologists with artificial neural network for differential diagnosis of intra-axial cerebral tumors on MR images

BACKGROUND AND PURPOSE: Previous studies have suggested that use of an artificial neural network (ANN) system is beneficial for radiological diagnosis. Our purposes in this study were to construct an ANN for the differential diagnosis of intra-axial cerebral tumors on MR images and to evaluate the effect of ANN outputs on radiologists'' diagnostic performance.MATERIALS AND METHODS: We collected MR images of 126 patients with intra-axial cerebral tumors (58 high-grade gliomas, 37 low-grade gliomas, 19 metastatic tumors, and 12 malignant lymphomas). We constructed a single 3-layer feed-forward ANN with a Levenberg-Marquardt algorithm. The ANN was designed to differentiate among 4 categories of tumors (high-grade gliomas, low-grade gliomas, metastases, and malignant lymphomas) with use of 2 clinical parameters and 13 radiologic findings in MR images. Subjective ratings for the 13 radiologic findings were provided independently by 2 attending radiologists. All 126 cases were used for training and testing of the ANN based on a leave-one-out-by-case method. In the observer test, MR images were viewed by 9 radiologists, first without and then with ANN outputs. Each radiologist''s performance was evaluated through a receiver operating characteristic (ROC) analysis on a continuous rating scale.RESULTS: The averaged area under the ROC curve for ANN alone was 0.949. The diagnostic performance of the 9 radiologists increased from 0.899 to 0.946 (P < .001) when they used ANN outputs.CONCLUSIONS: The ANN can provide useful output as a second opinion to improve radiologists'' diagnostic performance in the differential diagnosis of intra-axial cerebral tumors seen on MR imaging.

Accurate noninvasive radiologic diagnosis is desirable for appropriate treatment planning for brain tumors. MR imaging is an imaging technique of choice for the diagnosis of brain tumors. The MR characteristics of each type of brain tumor have been well documented in the radiologic literature. However, MR diagnosis of brain tumors is usually made subjectively, and its accuracy may be limited by the presence of atypical cases or by a radiologist''s insufficient clinical experience. A computerized scheme that is capable of providing objective information about an image may aid radiologists in the classification of brain tumors. An artificial neural network (ANN), which is a computational model simulating neurons in the human brain, has recently been applied to a variety of pattern recognitions and data classifications in medical imaging. ANNs have been reported to improve the diagnostic performance of radiologists in several fields.^1–9 The objectives of this study were to construct an ANN for the differential diagnosis of intra-axial cerebral tumors on MR images and to evaluate the effect of ANN outputs on radiologists'' diagnostic performance.     

乳腺数字化断层摄影分析乳腺实质结构评估罹患乳腺癌风险:与乳腺数字化摄影比较

目的乳腺密度百分比(PD)是已被确认的罹患乳腺癌的风险因子,本研究目的是在筛查人群中评估数字化乳腺断层摄影(DBT)与数字化乳腺摄影显示的乳腺实质结构特征与PD的相关性。材料与方法本研究经专业委员会核准,受试者均签署书面知情同意书。回顾性分析2007年7月—2008年3月进行的一项经专业委员会批准的DBT筛查     

Lung cancer screening with CT: Evaluation of radiologists and different computer assisted detection software (CAD) as first and second readers for lung nodule detection at different dose levels

Objectives

To find the best pairing of first and second reader at highest sensitivity for detecting lung nodules with CT at various dose levels.

Materials and methods

An anthropomorphic lung phantom and artificial lung nodules were used to simulate screening CT-examination at standard dose (100 mAs, 120 kVp) and 8 different low dose levels, using 120, 100 and 80 kVp combined with 100, 50 and 25 mAs. At each dose level 40 phantoms were randomly filled with 75 solid and 25 ground glass nodules (5–12 mm). Two radiologists and 3 different computer aided detection softwares (CAD) were paired to find the highest sensitivity.

Results

Sensitivities at standard dose were 92%, 90%, 84%, 79% and 73% for reader 1, 2, CAD1, CAD2, CAD3, respectively. Combined sensitivity for human readers 1 and 2 improved to 97%, (p1 = 0.063, p2 = 0.016). Highest sensitivities – between 97% and 99.0% – were achieved by combining any radiologist with any CAD at any dose level. Combining any two CADs, sensitivities between 85% and 88% were significantly lower than for radiologists combined with CAD (p < 0.03).

Conclusions

Combination of a human observer with any of the tested CAD systems provide optimal sensitivity for lung nodule detection even at reduced dose at 25 mAs/80 kVp.     

Delayed regression of huge theca lutein cysts monitored by serial sonograms and beta-HCG levels

The important role of serum beta-human chorionic gonadotropin (beta-HCG) in detecting spread of trophoblastic disease following evacuation of a hydatidiform mole is well established (1,2). Although sonography is accepted as the primary imaging technique in the diagnosis of hydatidiform mole (3), only a few authors have described the post-evacuation appearances of the pelvis, in particular the regression of theca lutein cysts (4,5). We here report a patient in whom there was delayed regression of huge theca lutein cysts compared to the regression of the serum beta-HCG levels after evacuation of a benign non-recurring hydatidiform mole.     

光谱CT头部虚拟平扫图像:不同单能量图像质量的对比

目的:探讨双层探测器光谱CT不同单能量图像在头部影像诊断中的临床应用效能.方法:回顾性分析2019年6-12月因头晕、晕厥和头痛等临床症状而在本院行光谱CT(IQon Spectral CT)头部CT平扫检查的38例患者的影像资料.从光谱CT扫描中可获取常规混合能量图像和40～120 keV的光谱成像数据,以5 keV...