We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Novel semi-automated kidney volume measurements in autosomal dominant polycystic kidney disease.
- Authors
Muto, Satoru; Kawano, Haruna; Isotani, Shuji; Ide, Hisamitsu; Horie, Shigeo
- Abstract
Background: We assessed the effectiveness and convenience of a novel semi-automatic kidney volume (KV) measuring high-speed 3D-image analysis system SYNAPSE VINCENT® (Fuji Medical Systems, Tokyo, Japan) for autosomal dominant polycystic kidney disease (ADPKD) patients.Methods: We developed a novel semi-automated KV measurement software for patients with ADPKD to be included in the imaging analysis software SYNAPSE VINCENT®. The software extracts renal regions using image recognition software and measures KV (VINCENT KV). The algorithm was designed to work with the manual designation of a long axis of a kidney including cysts. After using the software to assess the predictive accuracy of the VINCENT method, we performed an external validation study and compared accurate KV and ellipsoid KV based on geometric modeling by linear regression analysis and Bland-Altman analysis.Results: Median eGFR was 46.9 ml/min/1.73 m2. Median accurate KV, Vincent KV and ellipsoid KV were 627.7, 619.4 ml (IQR 431.5-947.0) and 694.0 ml (IQR 488.1-1107.4), respectively. Compared with ellipsoid KV (r = 0.9504), Vincent KV correlated strongly with accurate KV (r = 0.9968), without systematic underestimation or overestimation (ellipsoid KV; 14.2 ± 22.0%, Vincent KV; − 0.6 ± 6.0%). There were no significant slice thickness-specific differences (p = 0.2980).Conclusions: The VINCENT method is an accurate and convenient semi-automatic method to measure KV in patients with ADPKD compared with the conventional ellipsoid method.
- Subjects
POLYCYSTIC kidney disease; IMAGE recognition (Computer vision); CYSTS (Pathology); ACCURACY; THREE-dimensional imaging; COMPUTER software; PATIENTS
- Publication
Clinical & Experimental Nephrology, 2018, Vol 22, Issue 3, p583
- ISSN
1342-1751
- Publication type
Academic Journal
- DOI
10.1007/s10157-017-1486-6