Department of Diagnostic Imaging, Careggi University Hospital, Largo Brambilla 3, 50134, Firenze (FI), Italy.
Department of Experimental and Clinical Biomedical Sciences, School of Human Health Sciences, University of Florence, Largo Brambilla 3, 50134, Firenze (FI), Italy.
Radiol Med. 2021 Oct;126(10):1345-1355. doi: 10.1007/s11547-021-01376-2. Epub 2021 Jun 28.
Morphologic magnetic resonance imaging (MRI) for characterization of salivary gland tumors has limited utility, and the use of perfusion MRI data in the clinical setting is controversial. We examined the potential of tissue-normalized dynamic contrast-enhanced (DCE) MRI pharmacokinetic parameters of salivary gland tumors as imaging biomarkers for characterization and differentiation between benign and malignant lesions.
DCE-MR images acquired from 60 patients with parotid and submandibular gland tumors were retrospectively reviewed. Pharmacokinetic parameters as transfer constant (Ktrans), rate constant (Kep), extracellular space volume (Ve), fractional plasma volume (Vp), and AEC (area of all times enhancement curve) were measured on both the lesion and the normal contralateral salivary gland parenchyma. Lesion/parenchyma ratio (L/P) for each parameter was calculated.
Five groups of lesions were identified (reference: histopathology): pleomorphic adenomas(n = 20), Warthin tumors(n = 16), other benign entities(n = 4), non-Hodgkin lymphomas(n = 4), and malignancies(n = 16). Significant differences were seen for mean values of L/PKtrans (higher in malignancies), L/PKep (lower in adenomas than Warthin tumors), L/PVe (lower in Warthin tumors and lymphomas), L/PVp (higher in Warthin tumors and malignancies than adenomas), and L/PAEC (higher in malignancies). Significant differences were found between benign and malignant (non-lymphoproliferative) lesions in mean value of L/PKtrans (0.485 and 1.581), L/PVp (1.288 and 2.834), and L/PAEC (0.682 and 1.910). ROC analysis demonstrated the highest AUC (0.96) for L/PAEC, with sensitivity and specificity for malignancy of 93.8% and 97.5% (cutoff value = 1.038).
Lesion/parenchyma ratio of DCE-MRI pharmacokinetic data could be helpful for recognizing the principal types of salivary gland tumors; L/PAEC seems a valuable biomarker for differentiating benign from malignant tumors.
形态磁共振成像(MRI)在唾液腺肿瘤的特征描述中应用有限,且灌注 MRI 数据在临床中的应用仍存在争议。我们旨在探讨唾液腺肿瘤组织归一化动态对比增强(DCE)MRI 药代动力学参数作为成像生物标志物的潜力,以对良恶性病变进行特征描述和鉴别。
回顾性分析 60 例腮腺和颌下腺肿瘤患者的 DCE-MRI 图像。在病变和对侧正常唾液腺实质上测量药代动力学参数,包括转移常数(Ktrans)、速率常数(Kep)、细胞外容积(Ve)、血浆容积分数(Vp)和增强曲线下面积(AEC)。计算每个参数的病变/实质比值(L/P)。
根据参考标准(病理)将病变分为五组:多形性腺瘤(n=20)、Warthin 瘤(n=16)、其他良性实体瘤(n=4)、非霍奇金淋巴瘤(n=4)和恶性肿瘤(n=16)。恶性肿瘤的 L/PKtrans 平均值较高(P<.001),多形性腺瘤的 L/PKep 平均值较低(P<.001),Warthin 瘤和淋巴瘤的 L/PVe 平均值较低(P<.001),Warthin 瘤和恶性肿瘤的 L/PVp 平均值较高(P<.001),恶性肿瘤的 L/PAEC 平均值较高(P<.001)。良性和恶性(非淋巴增生性)病变之间 L/PKtrans(0.485 和 1.581)、L/PVp(1.288 和 2.834)和 L/PAEC(0.682 和 1.910)的平均值差异有统计学意义。ROC 分析显示 L/PAEC 的 AUC 值最高(0.96),用于恶性肿瘤的敏感性和特异性分别为 93.8%和 97.5%(截断值=1.038)。
DCE-MRI 药代动力学数据的病变/实质比值有助于识别唾液腺肿瘤的主要类型;L/PAEC 似乎是鉴别良恶性肿瘤的有价值的生物标志物。