Department of Surgery, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands.
J Surg Oncol. 2011 Jul 1;104(1):10-6. doi: 10.1002/jso.21874. Epub 2011 Mar 4.
Preoperative radiological assessment of hepatic steatosis is recommended in patients undergoing a liver resection, but few studies investigated the diagnostic accuracy after neoadjuvant chemotherapy. The aim of this study was to compare diagnostic accuracy of preoperative CT or MRI measurements of steatosis in patients with colorectal liver metastases after induction chemotherapy.
MRI measurements (relative signal intensity decrease; RSID), N = 36, and CT scan measurements (Hounsfield units; HU), N = 32, were compared with histological steatosis assessment. Diagnostic accuracy was determined for detecting any (>5%) or marked macrovesicular steatosis (>33%).
MRI showed the highest correlation with histology (r = 0.82, P < 0.001), compared to CT measurements (r = -0.65, P < 0.001). Based on linear regression analysis, radiological cut-off values for 5% and 33% macrovesicular steatosis, corresponded to 0.7% and 19.2% RSID in the MRI-group, and 60.4 and 54.2 HU in the CT-group, respectively. Sensitivity and specificity for the detection of any and marked macrovesicular steatosis using MRI was 87% and 69%, and 78% and 100%, respectively, and for CT, 83% and 64%, and 70% and 87%, respectively.
In patients treated with neoadjuvant chemotherapy MRI measurements of steatosis showed the highest correlation coefficient and the best diagnostic accuracy, as compared to CT measurements.
建议在接受肝切除术的患者中进行术前影像学评估肝脂肪变性,但很少有研究调查新辅助化疗后肝脂肪变性的诊断准确性。本研究旨在比较结直肠癌肝转移患者诱导化疗后术前 CT 或 MRI 测量的脂肪变性的诊断准确性。
比较了 MRI 测量(相对信号强度降低;RSID),N=36,和 CT 扫描测量(Hounsfield 单位;HU),N=32,与组织学脂肪变性评估。确定了检测任何(>5%)或明显大泡性脂肪变性(>33%)的诊断准确性。
与 CT 测量(r=-0.65,P<0.001)相比,MRI 与组织学显示出最高的相关性(r=0.82,P<0.001)。基于线性回归分析,MRI 组中 5%和 33%大泡性脂肪变性的放射学临界值分别对应 0.7%和 19.2%的 RSID,而 CT 组分别对应 60.4 和 54.2 HU。使用 MRI 检测任何和明显大泡性脂肪变性的敏感性和特异性分别为 87%和 69%,78%和 100%,而 CT 分别为 83%和 64%,70%和 87%。
与 CT 测量相比,在接受新辅助化疗的患者中,MRI 脂肪变性测量显示出最高的相关系数和最佳的诊断准确性。