Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, 75 Francis Street, Boston, MA, 02115, USA.
Department of Imaging, Dana-Farber Cancer Institute, Boston, MA, USA.
Abdom Radiol (NY). 2018 Sep;43(9):2329-2339. doi: 10.1007/s00261-018-1496-1.
To compare hepatocellular phase imaging after intravenous gadoxetate disodium with other MRI pulse sequences and with extracellular agent for assessing hepatic metastases from gastroenteropancreatic neuroendocrine neoplasms (GEP-NEN).
In this IRB-approved, HIPAA-compliant retrospective study, we included 30 patients (15 women, mean age: 58 years, range 44-77 years) with GEP-NEN metastatic to the liver, who underwent MRI with gadoxetate disodium. Six MRI sequences were reviewed by two radiologists to score tumor-liver interface (TLI) on a 5-point scale, to assess lesion detectability in different liver segments (divided into 3 zones/patient), and to measure lesion size. Contrast-to-noise ratio (CNR) was calculated on each sequence. In 19 patients, lesion size and CNR on dynamic imaging with gadopentetate dimeglumine was compared with hepatocellular phase. Wilcoxon signed-rank test was used to compare TLI scores, lesion size, and median CNR, using Bonferroni correction for multiple testing. Interobserver agreement for TLI was analyzed using Krippendorff's alpha, and for lesion size using concordance correlation coefficient (CCC) and mean relative difference.
Hepatocellular phase had the best TLI (mean TLI for reader 1 = 1.2, reader 2 = 1.3) compared to all other sequences (p < 0.0001) with excellent interobserver agreement (Krippendorff's alpha = 1.0), maximum lesion detectability (61/90 zones), highest interobserver agreement for lesion measurement (CCC 0.9875 and smallest mean relative difference - 1.567%), and highest median CNR (31.2, p < 0.008). Hepatocellular phase also had the highest CNR when compared with gadopentetate imaging.
Hepatocellular phase imaging offers significant advantages for assessment of hepatic metastasis in GEP-NEN, and should be routinely considered for follow-up of these patients.
比较静脉注射钆塞酸二钠后肝细胞期成像与其他 MRI 脉冲序列及细胞外造影剂在评估胃肠胰神经内分泌肿瘤（GEP-NEN）肝转移中的应用。
本研究经机构审查委员会批准，符合 HIPAA 规定，回顾性纳入 30 例 GEP-NEN 肝转移患者（15 例女性，平均年龄 58 岁，范围 44-77 岁），均行钆塞酸二钠 MRI。两位放射科医生回顾了 6 种 MRI 序列，对肿瘤-肝脏界面（TLI）进行 5 分制评分，评估不同肝段（每位患者分为 3 个区）的病灶检出情况，并测量病灶大小。计算各序列的对比噪声比（CNR）。在 19 例患者中，对比动态增强 MRI 与肝细胞期成像的病灶大小和 CNR。采用 Wilcoxon 符号秩检验，对 TLI 评分、病灶大小和中位 CNR 进行比较，采用 Bonferroni 校正进行多重比较。采用 Krippendorff 的 α 分析 TLI 观察者间一致性，采用一致性相关系数（CCC）和平均相对差异分析病灶大小的观察者间一致性。
与其他所有序列相比（p<0.0001），肝细胞期 TLI 最佳（读者 1 的平均 TLI=1.2，读者 2=1.3），且观察者间一致性极好（Krippendorff 的 α=1.0），病灶检出率最高（90 个区中的 61 个区），病灶测量的观察者间一致性最高（CCC 0.9875 和最小平均相对差异-1.567%），中位 CNR 最高（31.2，p<0.008）。与钆喷替酸成像相比，肝细胞期也具有最高的 CNR。
肝细胞期成像在评估 GEP-NEN 肝转移方面具有显著优势，应常规用于这些患者的随访。
