计算机辅助诊断在肺结节诊断中的应用价值

目的:评价计算机辅助检测系统对肺结节检测的价值。材料和方法:从受检病人中随机抽取120例有肺部单发结节,直径9～30mm的病例,所有肺结节病灶均经病理证实。另从健康体检人群中随机抽取120个年龄、性别构成与受检组相仿的作为对照组。两组胸片均经CT证实,并由多位放射学专家共同诊断。所有胸片均通过数字化放射成像获得。由5位高年资影像医生和5位低年资影像医生作为观测者分别对不用和用CAD输出图像的胸片进行诊断。观察者积分用受试者操作特性曲线分析来衡量。结果:平均ROC曲线下面积从不用CAD输出图像的0.762增加到用CAD输出图像的0.825(P<0.05)。在使用CAD输出图像时,低年资影像医生比高年资影像医生的平均ROC曲线下面积增加的更多。结论:运用数字化胸片所配置的CAD系统能帮助影像诊断医生提高对肺结节性病灶的检出率。     

Evaluation of a real-time interactive pulmonary nodule analysis system on chest digital radiographic images: a prospective study

RATIONALE AND OBJECTIVES: We sought to assess the performance of a real-time interactive pulmonary nodule analysis system for evaluation of chest digital radiographic (DR) images in a routine clinical environment. MATERIALS AND METHODS: A real-time interactive pulmonary nodule analysis system for chest DR image softcopy reading (IQQA-Chest; EDDA Technology, Princeton Junction, NJ) was used in daily practice with a Picture Archiving and Communication System in a National Cancer Institute-designated cancer teaching hospital. Patients referred for follow-up of known cancer underwent digital chest radiography. Posteroanterior and lateral DR images were first read by resident radiologists along with experienced chest radiologists using a Picture Archiving and Communication System work station. The computer-assisted detection (CAD) program was subsequently applied to the posteroanterior DR images, and changes (if any) in diagnosis were recorded. For reference standard, a follow-up chest radiograph at least 6 months following the initial examination or a follow-up computed tomographic scan of the chest within 3 months was used to establish diagnostic accuracy. RESULTS: Of 324 DR examinations, follow-up imaging according to our parameters was available for 214 patients (67%). Lung nodules were found and subsequently confirmed in 35 patients (10%) without CAD. Using CAD, nodules were found and subsequently confirmed in 51 patients (15%), improving sensitivity from 63.8% (95% confidence interval [CI], 0.49%-0.76%) to 92.7% (95% CI, 0.82%-0.98%) (P < .0001, McNemar). Nodules were subsequently proved to be malignant in five of the 16 additional cases (31%). False-positive readings increased from three to six cases; specificity decreased from 98.1% (95% CI, 0.95%-0.99%) to 96.2% (95% CI, 0.92%-0.98%) (not significant). There were 153 true-negative cases (71.4%). CONCLUSIONS: This study suggests that the interpretation of chest radiographs for lung nodules can be improved using an automated CAD nodule detection system. This improvement in reader performance comes with a minimal number of false-positive interpretations.     

CsI-detector-based dual-exposure dual energy in chest radiography for lung nodule detection: results of an international multicenter trial

To assess both sensitivity and specificity of digital chest radiography alone and in conjunction with dual-exposure dual-energy chest radiography for the detection and classification of pulmonary nodules. One hundred patients with a total of 149 lung nodules (3-45 mm; median, 11 mm) confirmed by CT were included in this study. Dual-exposure dual-energy chest radiographies of each patient were obtained using a CsI detector system. Experienced board-certified chest radiologists from four different medical centers in Europe reviewed standard chest radiographs alone and in conjunction with dual-energy images blinded and in random order. The reviewers rated the probability of presence, calcification and malignancy of all lung nodules on a five-point rating scale. Lesions detected were identified by applying a specific coordinate system to enable precise verification by the study leader. A receiver-operating characteristic (ROC) analysis was performed. In addition to the 149 true-positive CT proven lesions, 236 false-positive lung nodules were described in digital chest radiographies in conjunction with dual-energy chest radiographies. The cumulative sensitivity of chest radiography in conjunction with dual energy was 43%, specificity was 55%. For digital radiography alone, sensitivity was 35% and specifity was 83%. For the dual energy system, positive predictive value was 58%, and negative predictive value was 66% compared to the digital radiography with a positive predictive value of 59% and a negative predictive value of 65%. Areas under the curve in a ROC analysis resulted in 0.631 (95% confidence interval =0.61 to 0.65) for radiography with dual energy and 0.602 (95% confidence interval =0.58 to 0.63) for digital radiography alone. This difference was not statistically significant. For the detection of lesion calcification or the determination of malignancy, ROC analysis also failed to show significant differences. CsI-based flat-panel dual-exposure dual-energy imaging added to standard chest radiography did not show statistically significant improvement for the detection of pulmonary nodules, nor the identification of calcifications, nor the determination of malignancy.     

Bayesian regression methodology for estimating a receiver operating characteristic curve with two radiologic applications: prostate biopsy and spiral CT of ureteral stones

RATIONALE AND OBJECTIVES: The authors evaluated two Bayesian regression models for receiver operating characteristic (ROC) curve analysis of continuous diagnostic outcome data with covariates. MATERIALS AND METHODS: Full and partial Bayesian regression models were applied to data from two studies (n = 180 and 100, respectively): (a) The diagnostic value of prostate-specific antigen (PSA) levels (outcome variable) for predicting disease after radical prostatectomy (gold standard) was evaluated for three risk groups (covariates) based on Gleason scores. (b) Spiral computed tomography was performed on patients with proved obstructing ureteral stones. The predictive value of stone size (outcome) was evaluated along with two treatment options (gold standard), as well as stone location (in or not in the ureterovesical junction [UVJ]) and patient age (covariates). Summary ROC measures were reported, and various prior distributions of the regression coefficients were investigated. RESULTS: (a) In the PSA example, the ROC areas under the full model were 0.667, 0.769, and 0.703, respectively, for the low-, intermediate-, and high-risk groups. Under the partial model, the area beneath the ROC curve was 0.706. (b) The ROC areas for patients with ureteral stones in the UVJ decreased dramatically with age but otherwise were close to that under the partial model (ie, 0.774). The prior distribution had greater influence in the second example. CONCLUSION: The diagnostic tests were accurate in both examples. PSA levels were most accurate for staging prostate cancer among intermediate-risk patients. Stone size was predictive of treatment option for all patients other than those 40 years or older and with a stone in the UVJ.     

应用统计学体模评价CR乳腺摄影影像的探讨

目的应用ALVIM统计学体模TRM,探讨乳腺计算机X线摄影(CR)获得尺寸最小的Al2O3斑点(钙化灶)和尼龙纤维(肿块灶)信号的能力。方法将具有各5种大小不同Al2O3斑点和尼龙纤维组成的ALVIM统计学体模TRM置于成像板(IP)上,用钼靶X线机26kV,调节毫安量,选择适当图像处理参数,获得一张密度值D=0·70±0·05的实验照片,用5值判断法取得每一行由5个信号和噪声所组成的10行信号的记分总值,用ALVIM统计学体模的计算公式计算出一组真阳性概率P(S/s)和假阳性概率P(S/n)值,绘制ROC曲线,并计算出每种信号的判断概率值Pdet,再用SPSS10·0统计学单因素方差分析软件处理数据,获得尺寸最小的钙化灶和软组织肿块灶。结果用概率统计方法获得制作ROC曲线的数据和判断概率平均值Pdet,5种大小不同的Al2O3斑点(钙化灶)中,0·20mm的Pdet=0·6250最小,0·55mm的Pdet=0·9000最大,而0·20mm与0·25mm的Pdet差异无统计学意义,与其他的差异有统计学意义;5种大小不同的尼龙纤维(肿块灶)中,0·45mm的Pdet=0·5313最小,1·00mm的Pdet=0·8813最大,而0·45mm的与0·60mm的差异无统计学意义,0·45mm与其他的差异有统计学意义。结论应用ALVIM统计学体模TRM制作ROC曲线和获得判断概率值Pdet的计算简单、快捷,宜于日常临床工作开展的评价影像质量控制,值得推广。     

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.     

计算机辅助检测肺结节在数字化X线胸片肺癌筛查中的应用价值

目的 评价计算机辅助检测(CAD)肺结节系统在数字化X线胸片上肺癌筛查中的应用价值.方法 由1名放射科医师和CAD肺结节检测系统独立阅读100例连续的数字摄影(DR)X线胸片,CAD系统可以检出最长直径在5～15 mm的肺结节.由2名放射科专家(有15年胸部影像诊断经验)进行回顾性阅读,参照相应的CT图像,两人意见达成一致后标记真结节的个数和位置并保存标记结果,将标记结果作为金标准来比较放射科医师和CAD系统的肺结节检测敏感性和假阳性率.结果 放射科医师共检测到95个结节,CAD系统共检测到304个结节.在回顾性检查中2名放射科专家共标记134个真结节,其中放射科医师检测到82个(61.2%),CAD检测到105个(78.4%),CAD系统检测到而被放射科医师漏诊的结节35个,放射科医师检测到而CAD系统漏诊的结节10个.放射科医师应用CAD系统后检测到112个真结节,检测率提高到83.6%.放射科专家意见一致后认为CAD系统检出199个假阳性结节,平均每张胸片约2.0个.结论 在肺癌筛查中放射科医师和CAD系统必须联合应用才可以识别X线胸片中所有的结节.     

Monte Carlo validation of the Dorfman-Berbaum-Metz method using normalized pseudovalues and less data-based model simplification

RATIONALE AND OBJECTIVES: Two problems of the Dorfman-Berbaum-Metz (DBM) method for analyzing multireader receiver operating characteristic (ROC) studies are that it tends to be conservative and that it can produce AUC estimates outside the parameter space--ie, greater than one or less than zero. Recently it has been shown that the problem of AUC (or other accuracy) estimates outside the parameter space can be eliminated by using normalized pseudovalues, and it has been suggested that less data-based model simplification be used. Our purpose is to empirically investigate if these two modifications--normalized pseudovalues and less data-based model simplification--result in improved performance. MATERIALS AND METHODS: We examine the performance of the DBM procedure using the two proposed modifications for discrete and continuous ratings in a null simulation study comparing modalities with respect to the ROC area. The simulation study includes 144 different combinations of reader and case sample sizes, normal/abnormal case sample ratios, and variance components. The ROC area is estimated using parametric and nonparametric estimation. RESULTS: The DBM procedure with both modifications performs better than either the original DBM procedure or the DBM procedure with only one of the modifications. For parametric estimation with discrete rating data, use of both modifications resulted in the mean type I error (0.043) closest to the nominal .05 level and the smallest range (0.050) and standard deviation (0.0108) across the 144 type I error rates. CONCLUSIONS: We recommend that normalized pseudovalues and less data-based model simplification be used with the DBM procedure.     

Digital mammography: observer performance study of the effects of pixel size on the characterization of malignant and benign microcalcifications

RATIONALE AND OBJECTIVES: The authors performed this study to evaluate the effects of pixel size on the characterization of mammographic microcalcifications by radiologists. MATERIALS AND METHODS: Two-view mammograms of 112 microcalcification clusters were digitized with a laser scanner at a pixel size of 35 microm. Images with pixel sizes of 70, 105, and 140 microm were derived from the 35-microm-pixel size images by averaging neighboring pixels. The malignancy or benignity of the microcalcifications had been determined with findings at biopsy or 2-year follow-up. Region-of-interest images containing the microcalcifications were printed with a laser imager. Seven radiologists participated in a receiver operating characteristic (ROC) study to estimate the likelihood of malignancy. The classification accuracy was quantified with the area under the ROC curve (Az). The statistical significance of the differences in the Az values for different pixel sizes was estimated with the Dorfman-Berbaum-Metz method and the Student paired t test. The variance components were analyzed with a bootstrap method. RESULTS: The higher-resolution images did not result in better classification; the average Az with a pixel size of 35 microm was lower than that with pixel sizes of 70 and 105 microm. The differences in Az between different pixel sizes did not achieve statistical significance. CONCLUSION: Pixel sizes in the range studied do not have a strong effect on radiologists' accuracy in the characterization of microcalcifications. The low specificity of the image features of microcalcifications and the large interobserver and intraobserver variabilities may have prevented small advantages in image resolution from being observed.     

用血清和呼气样品的拉曼指纹谱方法筛查早期肺癌与乳腺癌

我国每年大约有310万新发生癌症病例,约200万人死亡,目前还没有找到可普及的早期癌症筛查方法。我们在前期研究中,采用血清拉曼指纹谱筛查方法,联合筛查肺癌（43例）和早期乳腺疾病（58例）,对照健康人（35例）队列,用10条特征谱线（636,805,945,1 017,1 135,1 330,1 399,1 453,1 620,1 688 cm-1）处理。联合主成份分析结果：真阳性（+）准确率分别为100%,86%,91.4%;假阴性（-）误判率分别为接近0,14%,8.6%;联合聚类分析结果：真阳性（+）准确率分别为81.4%,86%,100%;假阴性（-）误判率分别为18.6%,14%,0。我们建议,早期肺癌拉曼筛查的研究重点：①为了增強拉曼指纹谱筛查早期癌症方法的敏感性和特异性,需要采集血清和呼气两种样本,在该两者样品中,将包含患者的与全部癌变细胞的代谢有关的全部特征小分子的信息;②以大样本队列的Ⅰ期肺癌病例创建早期肺癌代谢特征拉曼指纹谱数据库,然后,再创建早期乳腺癌、胃癌和其他癌症的数据库;③推进普及的早期癌症筛查国家系统工程的研究,包括硬件、软件和管理等建设。     

Low-dose computed tomography screening for lung cancer in a general population: characteristics of cancer in non-smokers versus smokers

RATIONALE AND OBJECTIVES: To report the detection rate for lung cancers in computed tomography (CT) screening in Japanese adults, and to analyze differences in the appearance of the cancers in non-smokers versus smokers. MATERIALS AND METHODS: Subjects consisted of 7,847 Japanese adults who received low-dose CT screening at least once in a 3-year period. The detection rate of lung cancers and the correlation of imaging, clinical, and pathologic findings of cancers in non-smokers versus smokers were examined. RESULTS: The detection rate for lung cancer was 1.1% for both non-smokers (45 of 4,251) and smokers (39 of 3,596). The prevalence of well-differentiated adenocarcinomas was greater in non-smokers (88%; 22 of 25) than in smokers (29%; 4 of 14) (P < .001). The prevalence and incidence of pathologic stage IA disease were greater in non-smokers than in smokers (92%; [22 of 24] vs 58% [7 of 12], and 100% [19 of 19] vs 70% [14 of 20]) (both P < .05). The mean size of the tumors in the non-smokers (12.4 mm) was smaller than that in smokers (18.2 mm) (P < .001). The percentage of cancers categorized as pure or mixed ground-glass opacity (86%; 38 of 44) on CT was greater in non-smokers than in smokers (46%; 16 of 35) (P < .001). CONCLUSION: Most of the lung cancers in non-smokers were slow-growing adenocarcinomas appearing as faint ground-glass opacities on CT, whereas rapidly growing cancers appearing as solid nodules were more commonly seen in smokers.     

A meta-analysis of the accuracy of prostate cancer studies which use magnetic resonance spectroscopy as a diagnostic tool

OBJECTIVE: We aimed to do a meta-analysis of the existing literature to assess the accuracy of prostate cancer studies which use magnetic resonance spectroscopy (MRS) as a diagnostic tool. MATERIALS AND METHODS: Prospectively, independent, blind studies were selected from the Cochrane library, Pubmed, and other network databases. The criteria for inclusion and exclusion in this study referenced the criteria of diagnostic research published by the Cochrane center. The statistical analysis was adopted by using Meta-Test version 6.0. Using the homogeneity test, a statistical effect model was chosen to calculate different pooled weighted values of sensitivity, specificity, and the corresponding 95% confidence intervals (95% CI). The summary receiver operating characteristic (SROC) curves method was used to assess the results. RESULTS: We chose two cut-off values (0.75 and 0.86) as the diagnostic criteria for discriminating between benign and malignant. In the first diagnostic criterion, the pooled weighted sensitivity, specificity, and corresponding 95% CI (expressed as area under curve [AUC]) were 0.82 (0.73, 0.89), 0.68 (0.58, 0.76), and 83.4% (74.97, 91.83). In the second criterion, the pooled weighted sensitivity, specificity, and corresponding 95% CI were 0.64 (0.55, 0.72), 0.86 (0.79, 0.91) and 82.7% (68.73, 96.68). CONCLUSION: As a new method in the diagnostic of prostate cancer, MRS has a better applied value compared to other common modalities. Ultimately, large scale RCT (randomized controlled trial) randomized controlled trial studies are necessary to assess its clinical value.     

Angle between 1 and 4 min gives the most significant difference in time–intensity curves between benign disease and breast cancer: analysis of dynamic magnetic resonance imaging in 103 patients with breast lesions

The aim of the present study was to clarify the morphologic characteristics of time–intensity curves (TICs) that are useful for distinguishing benign from malignant lesions. One hundred three patients with breast lesions underwent dynamic breast MRI. The areas under the receiver operating characteristic (ROC) curves (AUC) from every component of TIC were compared between benign and malignant disease. As a result, angle of cross line between 1 and 4 min is more useful than rapid enhancement for distinguishing benign from malignant lesions.     

3D CT versus axial helical CT versus conventional tomography in the classification of acetabular fractures: a ROC analysis

AIM: To assess the diagnostic power of three-dimensional computed tomography (3D CT), axial helical computed tomography (CT) and conventional tomography in the classification of acetabular fractures by interdisciplinary review. MATERIALS AND METHODS: Receiver operating characteristics (ROCs) were assessed for two radiologists and two surgeons blinded to the presence of acetabular fractures in an animal model (a total of 62 porcine hips, 40 of them with artificial acetabular fractures). Main target parameter was the diagnostic accuracy in the classification of the artificial fractures following Judet et al. RESULTS: ROC analysis for radiologists showed A(z) values of 0.83 for 3D CT, 0.81 for axial helical CT, and 0.78 for conventional tomography; differences between the three techniques were not significant (P = 0.46-0.73). A(z) values for the surgeons were 0.87 for 3D CT, 0.68 for axial helical CT, and 0.60 for conventional tomography; 3D CT was significantly better than axial helical CT (P = 0.01) and conventional tomography (P = 0.001). The differences between axial helical CT and conventional tomography were not significant (P = 0.37). CONCLUSION: Acetabular fractures are best classified by 3D CT, followed by axial helical CT and conventional tomography when assessed by surgeons. 3D CT did not provide any additional significant benefit in the classification performed by radiologists.     

Comparison of independent double readings and computer-aided diagnosis (CAD) for the diagnosis of breast calcifications

RATIONALE AND OBJECTIVES: The aim of the study is to compare independent double readings by radiologists and computer-aided diagnosis (CAD) in diagnostic interpretation of mammographic calcifications. MATERIALS AND METHODS: Ten radiologists independently interpreted 104 mammograms containing clustered microcalcifications. Forty-six of these were malignant and 58 were benign at biopsy. Radiologists read the images with and without a computer aid by using a counterbalanced study design. Sensitivity and specificity were calculated from observer biopsy recommendations, and receiver operating characteristic (ROC) curves were computed from their diagnostic confidence ratings. Unaided double-reading sensitivity and specificity values were derived post hoc by using three different objective rules and an additional rule of simulated-optimal double reading that assumed that consultations for resolving two radiologists' different independent diagnoses always produce the correct clinical recommendation. ROC curves of unaided double readings were obtained according to the literature. RESULTS: Single reading without computer aid yielded 74% sensitivity and 32% specificity, whereas CAD reading yielded 87% sensitivity and 42% specificity and appeared on a higher ROC curve (P < .0001). Three methods of formulating independent double readings generated sensitivities between 59% and 89%, specificities between 50% and 13%, and operating points that moved essentially along the average unaided single-reading ROC curve. ROC curves of unaided independent double readings showed small, statistically insignificant improvement over those of unaided single readings. Results of the simulated-optimal double reading were similar to CAD: 89% sensitivity and 50% specificity. CONCLUSION: Independent double readings of mammographic calcifications may not improve diagnostic performance. CAD reading improves diagnostic performance to an extent approaching the maximum possible performance.     

Unilateral lower lung field opacities on chest radiography: a comparison of the clinical manifestations of tuberculosis and pneumonia

Background

The radiological manifestations of lower lung field (LLF) tuberculosis (LLFTB) are similar to those of LLF pneumonia (LLFP), making diagnosis challenging. The aim of this study was to determine if there are differences in the clinical manifestations of LLFTB and LLFP in patients with unilateral LLF opacities.

Methods

We performed a retrospective review of patient records to identify those with unilateral LLF opacities who were subsequently diagnosed with LLFTB or LLFP. We compared demographics, clinical manifestations, hematological data, and radiographic findings between the groups of patients.

Results

We identified 22 and 72 patients diagnosed with LLFTB and LLFP, respectively. Multivariate analysis revealed that age (odds ratio [OR] = 1.05, 95% confidence interval [CI] = 0.99–1.11, P = 0.072), lack of fever > 38 °C (OR = 9.04, 95% CI = 1.69–48.40, P = 0.001), duration of symptoms ≥ 7 days (OR = 4.57, 95% CI = 1.09–19.26, P = 0.038), and the lack of air bronchograms upon radiography (OR = 12.08, 95% CI = 1.98–73.64, P = 0.007) were significant predictors of LLFTB in patients with LLF opacities. We used these predictors to construct a mathematical model for predicting LLFTB in patients with LLF opacities.

Conclusions

Our findings suggest that older age, prolonged duration of symptoms, lack of fever > 38°C, and the absence of air bronchograms are more common in patients with LLFTB than patients with LLFP. These findings may help clinicians differentiate between LLFTB and LLFP and thus initiate timely and appropriate treatment.