Cardiac MRI radiomics and texture analysis for differentiating the causes of left ventricular hypertrophy: a systematic review.
Isaac Eli Martínez Safar, Juan Pablo De La Ossa Díaz, Daniel A Martínez D et al.
✦ AI-curated · Sources linked
In 30 seconds
This PRISMA-DTA systematic review assessed whether cardiac MRI radiomics and texture analysis can credibly distinguish causes of left ventricular hypertrophy (LVH). It included 11 studies published from 2019 to 2025, mostly retrospective and single-centre, covering contrasts such as hypertrophic cardiomyopathy (HCM) versus hypertensive heart disease and cardiac amyloidosis versus HCM. Evidence was synthesised without meta-analysis. A validation 2 × 2 table could be reconstructed for only 4 of 11 studies, and certainty was very low throughout. The authors concluded that current evidence does not support clinically credible differentiation.
Key findings
- A validation 2 × 2 table was reconstructable in 4 of 11 studies; only 3 used a held-out or cross-validated set.
- For cardiac amyloidosis versus HCM, none of the 4 studies reported per-class counts.
- No study reported external geographic validation, and overfitting or high risk of it was evident in 8 of 11 studies.
- Certainty was very low for every etiological contrast, with no contrast supporting valid pooling.
Why it matters
LVH has multiple causes with different management implications, so a reliable imaging-based distinction could be clinically useful. This review indicates that reported radiomics accuracy is not yet sufficiently verifiable or robust to establish that role. The findings highlight the need to interpret apparent performance cautiously and to distinguish promising model development from evidence of clinically credible diagnostic differentiation.
What to keep in mind
The evidence base comprised 11 studies, mostly retrospective and single-centre; no study reported external geographic validation, and the review could not validly pool results.