论文部分内容阅读
摘要目的评价FDG-PET SUV参数的诊断准确性以优化无创性预测软组织肿瘤分级方法。方法回顾性分析123例病人(初诊79例,复发44例)共129个病灶的FDG-PET影像。以病理组织学为肿瘤级别金标准,将一系列SUV参数的绝对值和肿瘤/肝脏比值与肿瘤级别进行相关性分析。结果初诊病人的SUVmax、SUVpeak、SUVmax/SUVliver和SUVpeak/SUVliver与肿瘤级别相关性较好。SUVpeak(AUC-ROC:0.82)和SUVpeak/SUVliver(AUC-ROC:0.82)能够最好地区分低级别(WHO交界
Abstract Objective To evaluate the diagnostic accuracy of FDG-PET SUV parameters in order to optimize non-invasive prediction of soft tissue tumor grade. Methods We retrospectively analyzed FDG-PET images of 129 lesions in 123 patients (79 newly diagnosed and 44 recurrent). With histopathology as the gold standard for tumor grade, a series of correlations between absolute value of SUV parameters and tumor / liver ratio were performed with tumor grade. Results The newly diagnosed SUVmax, SUVpeak, SUVmax / SUVliver and SUVpeak / SUVliver correlated well with tumor grade. SUVpeak (AUC-ROC: 0.82) and SUVpeak / SUVliver (AUC-ROC: 0.82) are best able to distinguish between low-level