Popovic M, Talarico O, van den Hoff J, Kunin H, Zhang Z, Lafontaine D, Dogan S, Leung J, Kaye E, Czmielewski C, Mayerhoefer M E, Zanzonico P, Yaeger R, Schöder H, Humm J L, Solomon S B, Sofocleous C T, Kirov A S
Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, NY, 10065, USA.
Cornell University, Ithaca, NY, 14850, USA.
EJNMMI Res. 2020 Nov 23;10(1):142. doi: 10.1186/s13550-020-00707-0.
Deriving individual tumor genomic characteristics from patient imaging analysis is desirable. We explore the predictive value of 2-[18F]FDG uptake with regard to the KRAS mutational status of colorectal adenocarcinoma liver metastases (CLM).
2-[18F]FDG PET/CT images, surgical pathology and molecular diagnostic reports of 37 patients who underwent PET/CT-guided biopsy of CLM were reviewed under an IRB-approved retrospective research protocol. Sixty CLM in 39 interventional PET scans of the 37 patients were segmented using two different auto-segmentation tools implemented in different commercially available software packages. PET standard uptake values (SUV) were corrected for: (1) partial volume effect (PVE) using cold wall-corrected contrast recovery coefficients derived from phantom spheres with variable diameter and (2) variability of arterial tracer supply and variability of uptake time after injection until start of PET scan derived from the tumor-to-blood standard uptake ratio (SUR) approach. The correlations between the KRAS mutational status and the mean, peak and maximum SUV were investigated using Student's t test, Wilcoxon rank sum test with continuity correction, logistic regression and receiver operation characteristic (ROC) analysis. These correlation analyses were also performed for the ratios of the mean, peak and maximum tumor uptake to the mean blood activity concentration at the time of scan: SUR, SUR and SUR, respectively.
Fifteen patients harbored KRAS missense mutations (KRAS+), while another 3 harbored KRAS gene amplification. For 31 lesions, the mutational status was derived from the PET/CT-guided biopsy. The Student's t test p values for separating KRAS mutant cases decreased after applying PVE correction to all uptake metrics of each lesion and when applying correction for uptake time variability to the SUR metrics. The observed correlations were strongest when both corrections were applied to SUR and when the patients harboring gene amplification were grouped with the wild type: p ≤ 0.001; ROC area under the curve = 0.77 and 0.75 for the two different segmentations, respectively, with a mean specificity of 0.69 and sensitivity of 0.85.
The correlations observed after applying the described corrections show potential for assigning probabilities for the KRAS missense mutation status in CLM using 2-[18F]FDG PET images.
从患者影像分析中得出个体肿瘤基因组特征是很有必要的。我们探讨了2-[18F]FDG摄取对于结直肠癌肝转移（CLM）KRAS突变状态的预测价值。
在一项经机构审查委员会（IRB）批准的回顾性研究方案下，对37例行CLM的PET/CT引导下活检患者的2-[18F]FDG PET/CT图像、手术病理及分子诊断报告进行了回顾。使用不同商业软件包中实现的两种不同自动分割工具，对37例患者39次介入性PET扫描中的60个CLM进行了分割。PET标准摄取值（SUV）针对以下情况进行了校正：（1）使用从不同直径的体模球体得出的冷壁校正对比恢复系数对部分容积效应（PVE）进行校正；（2）根据肿瘤与血液标准摄取比（SUR）方法，对动脉示踪剂供应的变异性以及注射后直至PET扫描开始的摄取时间变异性进行校正。使用学生t检验、带连续性校正的Wilcoxon秩和检验、逻辑回归及受试者工作特征（ROC）分析，研究KRAS突变状态与平均、峰值及最大SUV之间的相关性。还分别针对扫描时平均、峰值及最大肿瘤摄取与平均血液活性浓度的比值：SUR、SUR及SUR进行了这些相关性分析。
15例患者存在KRAS错义突变（KRAS+），另有3例存在KRAS基因扩增。对于31个病灶，其突变状态源自PET/CT引导下的活检。在对每个病灶的所有摄取指标应用PVE校正后，以及对SUR指标应用摄取时间变异性校正后，用于区分KRAS突变病例的学生t检验p值降低。当对SUR同时应用两种校正，且将存在基因扩增的患者与野生型归为一组时，观察到的相关性最强：p≤0.001；两种不同分割的曲线下ROC面积分别为0.77和0.75，平均特异性为0.69，敏感性为0.85。
应用所述校正后观察到的相关性表明，使用2-[18F]FDG PET图像为CLM中KRAS错义突变状态分配概率具有潜力。
