| Objective To analyze the non-calcified pulmonary nodules and lung cancers by low-dose computed tomography (LDCT) screening and the relevant epidemiological factors, as well as the other lesions identified simultaneously.Materials and methods We performed LDCT screening without contrast medium for lung cancer on 1,653 eligible participants. They were divided into two age groups:the middle-aged group (between 40 and 64) and the elderly group (between 65 and 79). One thousand five hundred and fifty subjects who had information of smoking or tobacco exposure recorded were divided into ever smokers (smokers or light smokers) and non-smokers (with or without secondhand smoking exposure). The classification for pulmonary nodules and other chest lesions identified by LDCT conformed to the I-ELCAP protocol. At baseline screening, the result was positive when at least one solid or part-solid nodule 5mm or more in diameter or at least one nonsolid nodule 8mm or more in diameter was identified. The result was semi-positive when all of non-calcified nodules identified were too small to imply a positive result. If none of non-calcified nodule was identified the result was negative. Statistical significance was analyzed. The seventh edition of lung cancer staging revised by the IASLC was adopted.Results At baseline screening,1,653 subjects who met the inclusion criteria were recruited (1,293 men and 360 women). Median age was 48 years (range:40 to 79 years). Of these subjects,1,511 were in the middle-aged group and 142 in the elderly group, including 824 ever smokers (803 men,21 women) and 726 non-smokers (429 men,297 women). The difference of population between ever smokers and non-smokers among different gender was statistically significant (P<0.0001). The secondhand smoke exposure rate of non-smokers was high (70.5%) whereas the difference between different gender was not statistically significant (P=0.452). At baseline screening, the results of 244 subjects (14.8%) were positive,717 subjects (43.4%) were semi-positive, and 692 subjects (41.8%) were negative. The positive result rate was higher in the elderly group than in the middle-aged group (30.3% versus 13.3%, P<0.0001). Three hundred and thirty seven positive nodules were identified at baseline screening, in which the malignant rate of part-solid nodules was highest (12.5%). Follow-up CT scans were performed in 190 subjects. Compared with baseline screening,398 nodules had no changes, including 6 positive nodules that maybe malignancies,22 nodules shrank or disappeared,2 nodules increased (1 adenocarcinoma and 1 bronchioalveolar carcinoma, respectively), and 9 nodules newly emerged at follow-up CT. Of the new nodules,8 were considered benign, and 1 undetermined. Fourteen cases of lung cancer confirmed pathologically were identified at baseline and follow-up CT. Of them, the pathological types included adenocarcinoma (10 cases), BAC (3 cases) and carcinoid (1 case). Based on the seventh edition of the TNM stage,7 cases were in stage I A,5 cases in I B, and 2 cases in?A. Positive coronary artery calcification (CAC) were observed in 242 subjects (14.6%), and male, the elderly group, smokers or diabetes were the independent predictors. Emphysema was screened in 58 cases (3.5%), and male, the elderly group and smokers were in the majority. Finally, the median effective dose of 1321 subjects who had recorded radiation dose was 0.89mSv.Conclusion The smoking situation and secondhand smoke exposure rate in our study are consistent with the overall survey trend in China. The malignant rate of positive part-solid pulmonary nodules was highest that should be followed up closely. Early lung adenocarcinoma was the most common pathological type in identified lung cancer and female patients were in the majority. The radiation dose was reduced markedly by LDCT and the CAC and emphysema can be screened simultaneously. Objective To evaluate the detectability and application of computer-aided detection (CAD) system for non-calcified pulmonary nodules in low-dose computed tomography (LDCT) screening.Materials and methods One hundred subjects who underwent LDCT screening were adopted to evaluate. Three analysis methods were designed. Method A:CAD pulmonary nodules detection system; Method B:axial images of 1.25mm thickness with MIP imaging (10mm thickness); Method C:combining A and B. Method B and C were applied independently by one resident and one junior radiologist. The reference standard of true non-calcified nodules was determined in consensus by another two senior radiologists. Nodules'size, location and consistency were recorded. The matching chi-square test and Kappa test were used.Results There were 269 true nodules determined by two senior radiologists. The number of true nodules, sensitivity and false negatives of CAD, method B and C by resident, method B and C by junior radiologist were 181 (67.3%,32.7%),110 (40.9%,59.1%), 206 (76.6%,23.4%),215 (79.9%,20.1%),256 (95.2%,4.8%), respectively. The sensitivity of method C was higher than that of method B by both resident and junior radiologist, and the difference was statistically significant (P<0.0001). The sensitivity of method C by resident was close to that of method B by junior radiologist, and the difference was not statistically significant (P=0.422). The agreement for method C and the reference standard was excellent (Kappa=0.806 by resident, and 0.953 by junior radiologist). There also was an excellent inter-observer agreement (Kappa=0.&35). Missing true nodules by CAD in our study was predominantly due to small size (<4mm), lower attenuation (ground-glass opacity) and the segmentation algorithm (connected or adjacent to blood vessels or pleura). The missing rate of nonsolid nodules by CAD was 89.5%.Conclusion CAD can improve radiologists'performance in detecting pulmonary nodules, particularly for residents or junior radiologists who have limited experience. However, the high false negatives of CAD limit its application as a stand-alone technique, especially to nonsolid pulmonary nodules which are more missed. |
PDF Full Text Request