To assess whether MRI-derived abscess volume, sphericity, and compactness predict ICU admission and prolonged hospital length of stay better than manually measured maximum diameter.
Approach:
Design and cohort: Retrospective, single-center study of 227 patients with surgically confirmed neck abscesses who underwent contrast-enhanced MRI within 48 hours of presentation between April 2013 and August 2021.
Segmentation: Abscesses were manually segmented on MRI using post-contrast T1-weighted images, with T2-weighted and diffusion-weighted images as guidance. A neuroradiologist reviewed all masks; a random 30-case sample was used to assess interobserver reliability.
Morphometry: Volume and sphericity were extracted using PyRadiomics and IBSI recommendations. The supplied context ends before the compactness definition and outcome analyses.
Key Findings:
The provided context reports the study rationale, cohort, and segmentation methods but does not provide study results.
Interpretation:
The study was designed to evaluate whether quantitative MRI morphology measures add predictive value beyond maximum abscess diameter; the supplied material does not establish whether they did.
Limitations:
The study was retrospective and conducted at a single tertiary referral center.
Of 535 screened patients, 227 were included; exclusions included absent abscess, unavailable surgical confirmation, non-standard MRI protocols, insufficient imaging data, or lack of co-registered imaging.
The provided context does not include the full methods or results.
