We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Přesnost a efektivita poloautomatických segmentačních programů pro stanovení objemu jater z MR snímků.
- Authors
Kordač, Petr; Šetinová, Bára; Pajuelo, Dita; Dezortová, Monika; Kovář, Jan; Rossmeislová, Lenka; Šiklová, Michaela; Hájek, Milan; Šedivý, Petr
- Abstract
Aim: To evaluate various segmentation software for liver segmentation from MR images. Method: Three MR examinations were performed on seven healthy volunteers (average age 38.2 ± 5.5 years, BMI = 28.6 ± 8.3 kg/m²) without known liver or cholestatic diseases -- before initiating fasting, after 48 hours of fasting, and after subsequent twoday carbohydrate realimentation. The livers were segmented using seven (semi)automatic methods of software 3D Slicer, LiverLab, ITKSNAP, Myrian and MedSeg and compared to the reference manual segmentation. Results: All methods used for liver volume determination showed good accuracy. Intraclass coefficients of consistency and agreement were above 0.95. The TotalSegmentator module in the 3D Slicer program achieved the best coefficient of variation (CV), and also demonstrated the highest accuracy in the individual assessment of the dietary intervention effect, with an average CV below 10% (other methods ranged from 10-20%). Conclusion: 3D Slicer can be considered the best among all the tested segmentation software for liver segmentation from MR images in terms of program availability, accuracy, and speed. In basic tasks such as organ segmentation, it can compete with commercial software. It can accurately track liver volume changes during short-term dietary interventions.
- Publication
Czech Radiology / Ceska Radiologie, 2023, Vol 77, Issue 3, p156
- ISSN
1210-7883
- Publication type
Article