目的 探索胸部CT值在胸腔积液鉴别诊断的价值。方法 81例胸腔积液患者纳入本研究,胸腔积液分为渗出液、漏出液、恶性胸腔积液及良性胸腔积液。建立平均CT值的ROC曲线,计算曲线下面积。结果 81例胸腔积液患者中59例为渗出液,22例为漏出液;恶性胸腔积液33例,良性胸腔积液48例。渗出液组平均CT值(16.68±6.76)Hu高于漏出液组(5.50±3.42)Hu(P<0.000 1)。ROC曲线分析结果显示,胸腔积液平均CT值对区分渗出液和漏出液具有较高的准确性(曲线下面积为0.944 5)。当最佳界值为≥9.99 Hu时,其敏感度为88.14%,特异度为90.91%;恶性胸腔积液组平均CT值(15.38±7.29)Hu与良性胸腔积液组平均CT值(12.45±8.03)Hu没有差异(P=0.098 1)。结论 在胸腔积液的鉴别诊断过程中,胸部CT的CT值在鉴别漏出液及渗出液中有一定的价值,但尚不能用于鉴别良性及恶性胸腔积液。

Objective To explore the value of chest CT value in the differential diagnosis of pleural effusion. Methods A total of 81 patients with pleural effusion were included in this study, including exudate, transudate, malignant pleural effusion and benign pleural effusion.The ROC curve of average CT value was established and the area under the curve was calculated. Results Among 81 patients with pleural effusion, 59 cases were exudative, 22 cases were transudative, 33 cases were malignant pleural effusion and 48 cases were benign pleural effusion.The mean CT value of the exudate group, (16.68±6.76) Hu, was significantly higher than (5.50±3.42) Hu of the transudate group (P<0.000 1).ROC curve analysis showed that the mean CT value of pleural effusion had high accuracy in distinguishing exudate from transudate (area under the curve was 0.9445).When the cut-off value for exudative effusion was over 9.99 Hu, the sensitivity and specificity were 88.14% and 90.91%, respectively.The mean CT value of malignant pleural effusion group, (15.38±7.29) Hu, was not significantly different from (12.45±8.03) Hu of benign pleural effusion group (P=0.098 1). Conclusions In the differential diagnosis of pleural effusion, the chest CT value can be used to identify transudate and exudate, but not benign and malignant pleural effusion.

目的 分析脑出血(intracerebral hemorrhage,ICH)在SWI和CT中的影像学表现,比较SWI和CT两种检查方式在ICH及脑微小血管出血(cerebral microbleeds,CMBs)中的优势。方法 采用回顾性分析2018年1月—2019年12月在广州市第一人民医院收治的76例疑诊ICH患者行SWI及CT检查,对确诊为ICH及CMBs的病例数据进行统计。结果 SWI与CT在ICH诊断结果的比较,差异无统计学意义(P>0.05);在CMBs诊断结果的比较中,SWI诊断准确率高于CT(P<0.05);另外在SWI与CT的联合检查中,诊断准确率高于SWI与CT独立检查。结论 脑出血患者的SWI和CT的影像表现均能将出血灶显示清楚,除此之外,SWI还能发现CT未能发现的有关血管的微小病灶。SWI作为一种新型技术,在ICH及CMBs方面起着至关重要的作用,与CT相辅相成。两者在医疗诊断中扮演着不可或缺的角色,为医疗的精准实施保驾护航。

目的 通过建立特征参数曲线模型分析在不同CT扫描条件下对肺结节鉴别诊断的量化意义。方法 回顾性分析2018年9月—2019年10月我院收治的肺结节患者的CT胸部平扫图像为研究对象,纳入标准为结节直径≥3 mm,无其他病史。在筛选出的96例样本中,符合条件的样本68例(男性39例,女性29例),按扫描剂量的不同将研究对象分为低剂量观察组(管电压120 kV,管电流20 mA)和常规剂量组(管电压120 kV,管电流150 mA),每组各34例;通过测量并计算扫描长度、有效剂量、样本体型、信噪比相关参数,观察不同管电流与有效辐射剂量之间的相关性以及图像质量;运用PACS人工智能软件以及人工综合提取肺结节特征信息(直径、体积、密度纹理、边缘细节、内部结构)并记录数据,进而通过公式计算出肺结节质量;应用U检验分析比较不同管电流下各参数的组间差异,经过单元逻辑对确定的重要参数体积与质量纳入多元逻辑分析,建立特征参数曲线模型并测量曲线下面积及勾画ROC;使用卡方分析评价不同管电流下建立特征参数曲线模型对肺结节定量诊断分析的差异并同时比较不同管电流下的图像质量。结果 研究中发现,样本接受的有效辐射剂量在管电压一定的条件下,随管电流的增加而线性增加;样本肥胖患者(BMI≥23.9)的CT图像在使用低剂量扫描中呈现出明显噪声,影响组织间观察,而BMI标准(18.5≤BMI≤23.9)的样本的CT图像中,肺结节的信噪比与管电流变化未出现明显趋势阈值,差异无统计学意义(P＜0.05);通过特征参数曲线模型显示,肺结节的体积与质量均质性曲线显示出其变化趋势与管电流变化成相关性,且稳定性和一致性较好,故此二要素为模型主要分析成分,观察组ROC曲线显示其曲线下面积为0.826高于常规组ROC曲线显示其曲线下面积为0.747。结论 与常规剂量相比,低剂量CT扫描模式下建立特征参数曲线模型对肺结节鉴别诊断更有可量化意义。

Objective The quantitative significance of differential diagnosis of pulmonary nodules under different CT scanning conditions was analyzed by establishing characteristic parameter curve model. Methods CT plain chest scan images of patients with pulmonary nodules treated in our hospital from September 2018 to October 2019 were analyzed retrospectively. The inclusion criteria were nodule diameter ≥3 mm and no other medical history. Of the 96 selected samples, 68 met the criteria (39 males and 29 females).According to the different scanning dose, the subjects were divided into low dose observation group (tube voltage 120 kV, tube current 20 mA) and conventional dose group (tube voltage 120 kV, tube current 150 mA).There were 34 cases in each group, and the correlation between different tube current and effective radiation dose and image quality were observed by measuring and calculating the relevant parameters of scanning length, effective dose, sample shape and signal-to-noise ratio. PACS artificial intelligence software and artificial synthesis were used to extract the characteristic information (diameter, volume, density texture, edge details, internal structure) of pulmonary nodules and record the data. Furthermore, the mass of pulmonary nodules was calculated by formula, the differences of parameters under different tube currents were compared by U test, the volume and mass of important parameters determined by unit logic were incorporated into multivariate logic analysis. The curve model of characteristic parameters was established, the area under the curve was measured and ROC was sketched. Chi-square analysis was used to evaluate the difference of characteristic parameter curve model for quantitative diagnosis of pulmonary nodules under different tube currents, and to compare the image quality under different tube currents at the same time. Results In the study, it is found that the effective radiation dose received by the sample increases linearly with the increase of tube current under the condition of constant tube voltage. The CT images of obese patients (BMI≥23.9) showed obvious noise when using low dose scan, which affected the inter-tissue observation. However, in the CT images of the samples with BMI standard (18.5 ≤ BMI ≤ 23.9), there was no obvious trend threshold between the signal-to-noise ratio and tube current of pulmonary nodules, and the difference was not statistically significant (P<0.05).The characteristic parameter curve model showed that the volume and mass homogeneity curve of pulmonary nodules showed the change trend was correlated with the change of tube current, and the stability and consistency were good, so the two elements were the main analytical components of the model. The ROC curve of the observation group showed that the area under the curve was 0.826, which was higher than that of the conventional group, the area under the curve of the ROC curve was 0.747. Conclusion Compared with conventional dose, the establishment of characteristic parameter curve model under low dose CT scan mode is more quantifiable for differential diagnosis of pulmonary nodules.