MR renography by semiautomated image analysis: Performance in renal transplant recipients

Abstract

We evaluated a method of semiautomated analysis of dynamic MR image series in renal transplants. Nine patients were studied twice, with an average time interval of 7 days. MR examination consisted of a run of 256 T1-weighted coronal scans (GE; TR/TE/flip: ‫؍‬ 11/3.4/60°; slice thickness ‫؍‬ 6 mm; temporal resolution ‫؍‬ 2 seconds). Gadolinium-DTPA (0.05 mmol/kg) was injected with an injector pump (5 ml/seconds). MR renographs of the cortex and medulla were obtained by segmentation of the renal transplant and placement of two regions of interest (ROIs) MR RENOGRAPHY (MRR) is a dynamic MR examination that provides functional information of the kidney and renal transplant in terms of changes in signal intensity of the renal parenchyma after gadolinium-DTPA injection (1-3). The high spatial resolution of the MR image allows cortical and medullary structures to be separated, a feature that is not possible with scintigraphic techniques and that may be of importance in separating disease processes affecting different parts of the kidney or renal transplant (1-12). Although biopsy is still the standard of reference for diagnosis of diseases affecting the kidney and renal transplant, early reports indicate the possible usefulness of MRR in discriminating disease processes such as acute rejection and acute tubular necrosis (8,10,11).The MR renograph is the time-dependent signal intensity of suitably placed regions of interest (ROIs) over cortical and medullary tissue (1-12). ROIs are usually hand-placed by the operator, who selects cortical and medullary tissue by visual inspection of the early-enhanced images on which contrast between cortical, medullary, and perinephric tissue is optimal.In cases in which contrast between cortex and medulla is suboptimal, however, ROI positioning may be more difficult and time consuming, thereby requiring skill and experience from the operator.In this work we present and evaluate a new method to derive cortical and medullary MR renographs of renal transplants by means of semiautomated ROI placement. The obtained renographs were tested nonparametrically on reproducibility. MATERIALS AND METHODS PatientsNine renal transplant recipients (three men and six women, 46 -66 years old, mean age 53.6 years) were included in this study. These patients were selected from a population of renal transplant recipients who took part in a larger clinical MRI study. Consecutive patients entering this study were asked if they were willing to participate in the reproducibility study and have a second MR session. All patients were in good clinical condition and had a stable renal function for a period of at least 3 months prior to the investigations (creatinine clearances ranging between 22.3 ml/min and 94.2 ml/min, median 62.5 ml/min). Patients with diseased renal transplants, who usually do not have stable renal function, were not included. None of the patients were on diuretics. Immunosuppressive drug regimes consisted of Azathioprine and Prednisone in one patient, Cyclosporine in three patients, and Tac...

Performance of an automated segmentation algorithm for 3D MR renography

“…Partial assessment of kidney function can be done through the analysis of the slopes of transit-time curves (2). Upslope modeling can be done by sampling a small representative cortical and medullary region (26). A more complete quantification of the filtration process, however, requires full segmentation of each kidney compartment because 1) the signal from a small tissue sample may not be representative of the rest of the organ, and 2) clinically, the most significant parameter is not the regional filtration rate (i.e., the rate per unit volume) but its integral over the entire organ.…”

A renal vascular compartment segmentation method based on dynamic contrast-enhanced images

“…Boykov et al [13][14][15] used a semi-automatic algorithm based on interactive graph cuts. Thresholding method was also used to segment the cortex from the kidney [16][17][18].…”

Dynamic MRI contrast enhancement of renal cortex: A functional assessment of renovascular disease in patients with renal artery stenosis

“…With careful application, it is possible to directly visualize contrast uptake in the kidney cortex over time (1,2). CE-MRR methods have been validated on porcine kidney models (3) and subsequently used to examine the renal systems of healthy volunteers (4), patients in need of renal transplants (5), and patients with renovascular disease (RVD) (6). The functional aspect of this method (7) has the potential to supplement parameters such as contrast-enhanced MR angiography (MRA) for characterizing RVD.…”