论文部分内容阅读
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 investigate the value of computer-aided diagnosis (CAD) system based on Bayes theory in CT diagnosis of solitary pulmonary nodules (SPN). Methods According to the Bayes theory, the pre-test ratio of malignant SPN and the likelihood ratios of clinical manifestations and CT findings were obtained from 352 SPN training sets (135 cases of malignant and 217 benign cases), then the CAD software based on Bayes theory System that uses it to calculate the malignancy probability of each SPN and prospectively verifies the diagnostic efficacy of the system in 132 SPN test sets (malignant 61, benign 71), with 2 senior and 2 lower-year Radiologist routine reading performance comparison. Results The CAD system based on Bayes theory was successfully constructed. The sensitivity, specificity and coincidence rate of the training set SPN were 88.9%, 93.1% and 91.5%, respectively. The sensitivity, specificity, coincidence rate and positive rate of diagnostic test set SPN The predictive value and negative predictive value were 88.5%, 85.9%, 87.1%, 84.4%, 89.7% respectively. The diagnostic coincidence rate was not significantly different from that of senior A and B doctors (P> 0.05) Years of C, Ding physician (P <0.05). Conclusion The CAD system based on Bayes theory can help physicians, especially junior physicians, improve the ability to identify the benign and malignant features of SPN and has certain reference value in guiding the clinical decision-making of SPN.