Altay Canan, Başara Akın Işıl, Özgül Hakan Abdullah, Şen Volkan, Bozkurt Ozan, Tuna Emine Burçin, Yörükoğlu Kutsal, Seçil Mustafa
Dokuz Eylül University, Izmir, Turkey.
Izmir Atatürk Eğitim Ve Araştırma Hastanesi, Izmir, Turkey.
Abdom Radiol (NY). 2025 Mar;50(3):1254-1265. doi: 10.1007/s00261-024-04596-y. Epub 2024 Sep 27.
This study retrospectively assessed the diagnostic accuracy of fat quantification based on proton density fat fraction (PDFF) for differentiating renal tumors.
In this retrospective study, 98 histologically confirmed clear cell renal cell carcinomas (ccRCCs), 35 papillary renal cell carcinomas (pRCCs), 14 renal oncocytomas, 16 chromophobe renal cell carcinomas (chRCCs), 10 lymphomas, 19 uroepithelial tumors, 10 lipid-poor angiomyolipomas (AMLs), and 25 lipid-rich AMLs were identified in 226 patients (127 males and 99 females) over 5 years. All patients underwent multiparametric kidney MRI. The MRI protocol included an axial plane and a volumetric 3D fat fraction sequence known as mDIXON-Quant for PDFF measurement. Demographic data were recorded, and PDFF values were independently reviewed by two radiologists blinded to pathologic results. MRI examinations were performed using a 1.5 T system. MRI-PDFF measurements were obtained from the solid parts of all renal tumors. Fat quantification was performed using a standard region of interest for each tumor, compared to histopathological diagnoses. Sensitivity and specificity analyses were performed to calculate the diagnostic accuracy for each histopathological tumor type. Nonparametric variables were compared among the subgroups using the Kruskal-Wallis H test and Mann Whitney U test. P-values < 0.05 were considered statistically significant.
In all, 102 patients underwent partial nephrectomy, 70 patients underwent radical nephrectomy, and the remaining 54 had biopsies. Patient age (mean: 58.11 years; range: 18-87 years) and tumor size (mean: 29.5 mm; range: 14-147 mm) did not significantly differ across groups. All measurements exhibited good interobserver agreement. The mean ccRCC MRI-PDFF was 12.6 ± 5.06% (range: 11.58-13.61%), the mean pRCC MRI-PDFF was 2.72 ± 2.42% (range: 2.12-3.32%), and the mean chRCC MRI-PDFF was 1.8 ± 1.4% (range: 1.09-2.5%). Clear cell RCCs presented a significantly higher fat ratio than other RCC types, uroepithelial tumors, lymphomas, and lipid-poor AMLs (p < 0.05). Lipid-rich AMLs demonstrated a very high fat ratio.
MRI-PDFF facilitated accurate differentiation of ccRCCs from other renal tumors with high sensitivity and specificity.
本研究回顾性评估基于质子密度脂肪分数（PDFF）的脂肪定量在鉴别肾肿瘤中的诊断准确性。
在这项回顾性研究中，在5年期间对226例患者（127例男性和99例女性）进行了分析，共识别出98例经组织学证实的透明细胞肾细胞癌（ccRCC）、35例乳头状肾细胞癌（pRCC）、14例肾嗜酸细胞瘤、16例嫌色肾细胞癌（chRCC）、10例淋巴瘤、19例尿路上皮肿瘤、10例少脂血管平滑肌脂肪瘤（AML）和25例富脂AML。所有患者均接受了多参数肾脏MRI检查。MRI检查方案包括轴位平面和用于PDFF测量的容积3D脂肪分数序列，即mDIXON-Quant。记录人口统计学数据，由两名对病理结果不知情的放射科医生独立审查PDFF值。使用1.5T系统进行MRI检查。从所有肾肿瘤的实性部分获取MRI-PDFF测量值。对每个肿瘤使用标准感兴趣区域进行脂肪定量，并与组织病理学诊断结果进行比较。进行敏感性和特异性分析以计算每种组织病理学肿瘤类型的诊断准确性。使用Kruskal-Wallis H检验和Mann-Whitney U检验在亚组之间比较非参数变量。P值<0.05被认为具有统计学意义。
共有102例患者接受了部分肾切除术，70例患者接受了根治性肾切除术，其余54例进行了活检。患者年龄（平均：58.11岁；范围：18 - 87岁）和肿瘤大小（平均：29.5mm；范围：14 - 147mm）在各组之间无显著差异。所有测量结果均显示出良好的观察者间一致性。ccRCC的平均MRI-PDFF为12.6±5.06%（范围：11.58 - 13.61%），pRCC的平均MRI-PDFF为2.72±2.42%（范围：2.12 - 3.32%），chRCC的平均MRI-PDFF为1.8±1.4%（范围：1.09 - 2.5%）。透明细胞RCC的脂肪比率显著高于其他RCC类型、尿路上皮肿瘤、淋巴瘤和少脂AML（p<0.05）。富脂AML显示出非常高的脂肪比率。
MRI-PDFF有助于以高敏感性和特异性准确区分ccRCC与其他肾肿瘤。
