Tsuda K, Yu K K, Coakley F V, Srivastav S K, Scheidler J E, Hricak H
Department of Radiology, University of California at San Francisco 94143-0628, USA.
J Comput Assist Tomogr. 1999 Jan-Feb;23(1):74-8. doi: 10.1097/00004728-199901000-00016.
The purpose of this work was to compare the efficacy of fat-suppressed and non-fat-suppressed fast spin echo (FSE) endorectal MRI in the detection of extracapsular extension (ECE) of prostate cancer by experienced and inexperienced readers.
Seventy-nine patients with biopsy-proven prostate cancer underwent axial FSE T2-weighted endorectal MRI of the prostate prior to radical prostatectomy. Twenty-one patients were imaged with frequency-selective fat suppression, and 58 were imaged without fat suppression. All images were retrospectively and independently reviewed by two readers of different experience levels who were blinded to clinical and pathological findings. Readers assessed the presence or absence of ECE on a 5 point scale for each side of the prostate, and step-section pathology was used as the standard of reference in all patients. Receiver operating characteristics analysis was used to compare the performance of fat-suppressed and non-fat-suppressed images by both readers.
ECE was present in 33 of 79 (42%) patients. The more experienced reader demonstrated better diagnostic performance (p < 0.05) than the less experienced reader in terms of sensitivity and area under the ROC curve (Az) for MRI without fat suppression. Use of frequency-selective fat suppression did not result in any significant improvement in diagnosis of ECE compared with MRI without fat suppression for either the experienced (Az 0.81 vs. 0.79) or the inexperienced (Az 0.76 vs. 0.68) reader.
Even when reader experience is considered, use of frequency-selective fat suppression did not significantly improve the diagnosis of ECE by MRI. The decision to use fat suppression and the selection of a fat suppression technique can be left to the discretion of the individual reader.
本研究旨在比较脂肪抑制和非脂肪抑制快速自旋回波（FSE）直肠内磁共振成像（MRI）在经验丰富和经验不足的阅片者检测前列腺癌包膜外侵犯（ECE）方面的效能。
79例经活检证实为前列腺癌的患者在根治性前列腺切除术前行前列腺轴向FSE T2加权直肠内MRI检查。21例患者采用频率选择性脂肪抑制成像，58例患者未进行脂肪抑制成像。所有图像均由两名经验水平不同的阅片者进行回顾性独立阅片，阅片者对临床和病理结果不知情。阅片者对前列腺每一侧的ECE存在与否按5分制进行评估，所有患者均以阶梯切片病理结果作为参考标准。采用受试者操作特征分析比较两名阅片者对脂肪抑制和非脂肪抑制图像的诊断性能。
79例患者中有33例（42%）存在ECE。在未进行脂肪抑制的MRI检查中，经验更丰富的阅片者在敏感性和ROC曲线下面积（Az）方面的诊断性能优于经验较少的阅片者（p < 0.05）。对于经验丰富的阅片者（Az 0.81 vs. 0.79）或经验不足的阅片者（Az 0.76 vs. 0.68），与未进行脂肪抑制的MRI相比，使用频率选择性脂肪抑制在ECE诊断方面未带来任何显著改善。
即使考虑阅片者经验，使用频率选择性脂肪抑制也未显著改善MRI对ECE的诊断。是否使用脂肪抑制及脂肪抑制技术的选择可由阅片者自行决定。
