论文部分内容阅读
目的评价计算机辅助诊断在MDCT肝脏多期增强扫描识别小肝细胞癌(hepatocellular carcinoma,HCC)及对所识别的HCC强化特点定量的可行性。方法2名有经验的放射科诊断医师共同阅读已经确诊的49例HCC患者的CT图像,用常规方法标记可疑病变区域并测量病变大小,达成一致的诊断结果作为金标准。这2名医师使用计算机辅助工具IQQA-Liver再次共同阅读所选的图像。分析IQQA-Liver识别病灶结果与金标准的相关性,用IQQA-Liver自动定量病灶强化特点。结果金标准标记了86个HCC。IQQA-Liver识别了327个可疑病灶,其中63个与金标准标记的一致,占金标准的73.3%(63/86),伴有平均每例患者5.4个假阳性。金标准标记的86个HCC中,65.1%显示为等/低-高-低-低强化类型,24.3%显示为等/低-低-低-低强化类型,其他为不典型强化类型;从平扫期至动脉期、从动脉期至门静脉期和从门静脉期至延迟期的病灶强化变量分别为(24.2±10.3)HU、(-24.3±11.1)HU和(-15±5.1)HU。结论IQQA-Liver有助于识别、诊断小HCC。
Objective To evaluate the feasibility of computer-aided diagnosis in the identification of hepatocellular carcinoma (HCC) by multi-phase enhanced scan in the liver of MDCT and the feasibility of quantifying the enhancement characteristics of HCC identified. Methods Two experienced diagnostic radiologists read the CT images of 49 confirmed cases of HCC, and used conventional methods to mark the suspicious lesion area and measure the size of the lesion, and reached a consensus diagnosis result as the gold standard. The two physicians again used the computer aided tool IQQA-Liver to read the selected images together. Analysis IQQA-Liver identification lesion results associated with the gold standard, with IQQA-Liver auto-quantitative lesion enhancement features. Results The gold standard has 86 HCCs tagged. IQQA-Liver identified 327 suspicious lesions, 63 of which were consistent with the gold standard, accounting for 73.3% (63/86) of the gold standard, with an average of 5.4 false positives per patient. Among the 86 HCCs, 65.1% showed low / high-low-low enhancement type, 24.3% showed low / low-low enhancement type and the rest were atypical types Swelling to arterial phase showed a significant enhancement of (24.2 ± 10.3) HU, (-24.3 ± 11.1) HU, and (-15 ± 5.1) HU, respectively, from arterial phase to portal venous phase and from portal phase to delayed phase. Conclusion IQQA-Liver helps to identify and diagnose small HCC.