King's College London

Research portal

Characterization of Small Renal Tumors with MR Elastography: a Feasibility Study

Research output: Contribution to journalArticle

Davide Prezzi, Radhouene Neji, Christian Kelly-Morland, Tim O'Brien, Ben Challacombe, Ashish Chandra, Ralph Sinkus, Vicky Goh

Original languageEnglish
Pages (from-to)344-351
JournalInvestigative Radiology
Volume53
Issue number6
Early online date1 Jun 2018
DOIs
StateE-pub ahead of print - 1 Jun 2018

King's Authors

Abstract

Objectives: To explore the feasibility of magnetic resonance elastography (MRE) for characterizing indeterminate small renal tumors (SRT) as part of a multiparametric MR imaging protocol.
Materials and Methods: Following institutional review board approval and informed consent, 21 prospective adults (15 men, median age 55 [range 25-72] years) with SRT were enrolled. Tumors (2 - 5 cm Ø) were imaged using three-directional, gradient echo MRE. Viscoelastic parametric maps (shear wave velocity [c] and attenuation [α]) were analyzed by two independent radiologists. Interobserver agreement (Bland-Altman statistics and intraclass correlation coefficients [ICC]) was assessed. Anatomical T2-weighted, dynamic contrast enhanced (DCE) and diffusion sequences completed the acquisition protocol. Imaging parameters were compared between groups (Mann-Whitney U test).
Results: MRE quality was good in 18 cases. A cohort of 5 oncocytomas and 11 clearcell renal cell carcinomas (ccRCC) was analyzed. MRE viscoelastic parameters were the strongest imaging discriminators: oncocytomas displayed significantly lower shear velocity c (median, 0.77 m/s; interquartile range [IQR], 0.76-0.79) (P = 0.007) and higher shear attenuation α (median, 0.087 mm-1; IQR, 0.082-0.087) (P = 0.008) than ccRCC (medians, 0.92 m/s and 0.066 mm-1; IQR, 0.84-0.97 and 0.054-0.074 respectively). T2 signal intensity ratio (tumor/renal cortex) was lower in oncocytomas (P = 0.02). DCE and diffusion MR parameters overlapped substantially (P ≥ 0.1). Oncocytomas displayed a consistent MRE viscoelastic profile, corresponding to data point clustering in a bidimensional scatter plot. MRE ICC were 0.982 for c and 0.984 for α, indicating excellent interobserver agreement.
Conclusions: MRE is feasible for SRT characterization; MRE viscoelastic parameters stronger discriminators between oncocytoma and ccRCC than anatomical, DCE and diffusion MR imaging parameters.

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454