Authors
Mona Dabiri, Henry Stewart, Shivani Ahlawat, Laura Fayad
Published in
Journal of computer assisted tomography. Nov 18, 2026. Epub Nov 18, 2026.
Abstract
To determine whether CT-like grayscale-inverted images generated from 2-dimensional gradient-recalled echo (2D GRE) and 3-dimensional gradient-recalled echo (3D GRE) magnetic resonance imaging approximate conventional computed tomography (CT) for evaluating pelvic bone tumors.
In this retrospective study, 36 patients with pelvic bone tumors who underwent CT and MRI, including 2D GRE and 3D GRE sequences, were reviewed by 2 fellowship-trained musculoskeletal radiologists. Tumor location, internal lesion appearance, cortical erosion, periosteal reaction, calcification, fracture, image quality, artifacts, and additional abnormalities were assessed. Agreement between CT-like grayscale-inverted GRE images and CT was evaluated using the Cohen κ statistic, and diagnostic performance of 2D GRE and 3D GRE was directly compared using the McNemar test.
Tumors were most commonly located in the sacrum (16/36) or ilium (14/36). Agreement with CT was highest for cortical erosion (2D GRE, κ=1.000 for both readers; 3D GRE, κ=0.899 to 0.922) and periosteal reaction (2D GRE, κ=0.756 to 0.879; 3D GRE, κ=0.746 to 0.760). The McNemar test revealed no significant differences in diagnostic performance between 2D GRE and 3D GRE for any evaluated feature. Fractures were underdetected on CT-like GRE images relative to CT. MRI identified additional abnormalities not seen on CT, including suspected osseous lesions, soft-tissue edema, and soft-tissue or adnexal masses. MRI was considered sufficient for diagnostic assessment in 91.7% (33/36) of cases.
CT-like grayscale-inverted images from routine 2D GRE and 3D GRE sequences provided information comparable to CT for most pelvic bone tumor features, although CT remained more sensitive for fractures. These preliminary data support CT-like GRE imaging as a potential radiation-free adjunct in selected pelvic tumor assessments.
PMID:
43049304
Bibliographic data and abstract were imported from PubMed on 18 Nov 2026.
Read full publication at:
Please sign in
to see all details.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 8
- Comments 0