High temporal resolution 3D gadolinium-enhanced dynamic MR imaging of renal tumors with pharmacokinetic modeling: Preliminary observations

• Published in: Journal of magnetic resonance imaging 2013-10, Vol.38 (4), p.802-808
• Main Authors: Chandarana, Hersh, Amarosa, Alana, Huang, William C., Kang, Stella K., Taneja, Samir, Melamed, Jonathan, Kim, Sungheon
• Format: Article
• Language: English
• Subjects: Adenoma, Oxyphilic - diagnosis; Adenoma, Oxyphilic - pathology; Adult; Aged; Algorithms; Carcinoma, Renal Cell - diagnosis; Carcinoma, Renal Cell - pathology; Cohort Studies; Contrast Media - pharmacokinetics; Diagnosis, Computer-Assisted; dynamic contrast enhanced MRI; Female; Gadolinium - pharmacokinetics; Humans; Image Processing, Computer-Assisted; Imaging, Three-Dimensional; Kidney Neoplasms - diagnosis; Kidney Neoplasms - pathology; Kinetics; Magnetic Resonance Imaging; Male; Middle Aged; pharmacokinetic modeling; Prognosis; renal cancer; renal neoplasm
• ISSN: 1053-1807; 1522-2586
• DOI: 10.1002/jmri.24035

Purpose To assess dynamic contrast‐enhanced (DCE) magnetic resonance imaging (MRI) tracer pharmacokinetic parameters obtained with Generalized Kinetic Model (GKM) and extended Shutter Speed Model (SSM2) in renal tumors stratified by histologic subtypes. Materials and Methods In all, 24 patients with renal tumors were imaged at 1.5 T utilizing DCE‐MRI with high temporal resolution (1.2 sec/temporal frame) prior to surgery. Tracer kinetic analysis was performed for the entire tumor using individualized aortic input function. GKM and SSM2 were employed to generate transfer constant (Ktrans), plasma volume, and interstitial volume. These parameters, and ΔKtrans (KtransSSM2 − KtransGKM) were compared between tumors stratified by histologic subtype. Results There were 25 renal tumors: 15 clear cell, 4 papillary, 3 chromophobe, and 3 oncocytoma/oncocytic subtype. KtransGKM was significantly higher in chromophobe compared to other subtypes (P < 0.01). Using KtransGKM > 1.0 min−1, chromophobe were diagnosed with 100% sensitivity and 90.9% specificity. KtransSSM2 was higher than KtransGKM for all renal tumors except for all chromophobe and two clear cell subtype. Using KtransGKM > 1.0 min−1 and Δ Ktrans < 0, chromophobe could be discriminated from other lesions with 100% accuracy. Conclusion Ktrans obtained with GKM and SSM2 analysis can potentially discriminate chromophobe from other renal lesions with high accuracy. J. Magn. Reson. Imaging 2013;38:802–808. © 2013 Wiley Periodicals, Inc.