CT fat-feature model predicts events within a deliberately enriched cohort
Source check: Source checked. Not independent scientific confirmation.
Finding: DeepFat-assisted segmentation and a 15-feature elastic-net Cox model yielded a test C-index of 0.69 in 80 held-out patients from a 400-person, event-enriched cohort. Test-set risk reclassification was not statistically established. The result was already present in the January 2024 preprint; it is not evidence of general-population risk calibration or clinical benefit.
Technology: DeepFat deep-learning segmentation; elastic-net Cox model
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A previously developed learned segmenter provides manually checked epicardial-fat masks; selected handcrafted spatial/intensity features enter a fitted regularized survival model. AI credit concerns measurement and model fitting, not a new fat-inflammation mechanism.
- Test and result
- The original v1 Methods and Results use 400 archived CT records, with 56% event prevalence and a random 320/80 development/test split. Feature selection is described as training-only. The v1 reports the same 0.69 test C-index as the later journal version.
- Limitation / still unknown
- The cohort was deliberately enriched by event/calcium-score groups, so natural prevalence, absolute risk and transportability are not established. DeepFat masks were manually corrected. The test reclassification statistic had a confidence interval crossing zero and p=0.463 despite a favorable narrative; no significant incremental reclassification claim is adopted. The composite includes revascularization and all-cause death. Later journal Methods/Results were separately read; they are not treated as a new discovery or silently backdated to v1.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://arxiv.org/abs/2401.16190v1 · https://doi.org/10.1016/j.jacadv.2024.101188
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 21 Nov 2026. Review scope: Actual relevant original clinical/computational Methods and Results read;111 original v1 and later journal checked separately; no operational biological protocols.
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.