摘要
目的初步探讨基于Bayes理论的计算机辅助诊断(computer-aided diagnosis,CAD)系统在孤立性肺结节(solitary pulmonary nodule,SPN)CT诊断中的价值。方法依据Bayes理论先从352例SPN训练集(恶性135例,良性217例)中求出恶性SPN的验前比及各临床和CT表现的似然比,再运用VC++语言编制基于Bayes理论的CAD系统,用它计算每个SPN的恶性概率,并前瞻性地检验该系统在132例SPN测试集(恶性61例,良性71例)中的诊断效能,与2位高年资和2位低年资放射科医师常规阅片的表现作比较。结果成功构建基于Bayes理论的CAD系统,它诊断训练集SPN的敏感度、特异度和符合率分别为88.9%、93.1%、91.5%,诊断测试集SPN的敏感度、特异度、符合率、阳性预测值及阴性预测值分别为88.5%、85.9%、87.1%、84.4%、89.7%,其诊断符合率与高年资甲、乙医师比较无统计学差异(P>0.05),但高于低年资丙、丁医师(P<0.05)。结论基于Bayes理论的CAD系统可帮助医师尤其是低年资医师提高鉴别SPN良恶性质的能力,并在指导SPN的临床决策中有一定的参考作用。
Objective To preliminarily explore the value of computer-aided diagnosis （CAD） system based on Bayesian theory in the diagnosis of solitary pulmonary nodules （SPNs） with CT. Methods Totally 352 consecutive SPN cases （malignancy n = 135, benignity n =217） were collected retrospectively to form the training set. According to Bayesian theory, the prior odds of malignant SPNs and the likelihood ratios of clinical and CT findings were derived from the training set, then these derived values were used to construct a Bayesian theory based CAD system by using VC^＋＋ language. Utilizing this system, the probability of malignancy in each case with SPN was calculated, and SPNs with ≥50% calculated probability were judged as malignancy and those with 〈 50% calculated probability was judged as benignity. This system was also tested prospectively for its diagnostic validation on the test set （malignancy n =61, benignity n =71 ）, compared with the performance of the two chest radiologists and two radiologic residents using routine diagnostic method. Results The Bayesian theory based CAD system was constructed successfully. The sensitivity, specificity, accuracy of this system for the training set were 88.9% , 93.1%, 91.5%, respectively. On the test set, the sensitivity, specificity, accuracy, positive predictive value, negative predictive value of this system were 88.5% , 85.9% , 87.1% , 84.4% and 89.7%, respectively. The accuracy of this system showed no statistical significance with that made by chest radiologist A （80.3%, X^2 =2.37, P =0. 122） and B （79.5%, X^2=3.12, P =0. 076）, but was higher than that by radiological residents C （74.2%, X^2 =7.05, P =0. 012） and D （74.2%, X^2 =6. 56, P =0. 009）. Conclusion The Bayesian theory based CAD system can help physicians differentiate benign from malignant SPNs, especially for less experienced physicians. This system can provide a reference in deter- mining the management of SPNs.
出处
《第三军医大学学报》
CAS
CSCD
北大核心
2008年第20期1889-1892,共4页
Journal of Third Military Medical University
基金
湖南省自然科学基金(07JJ5010)~~
