Ultrasound Assessment in Polycystic Ovary Syndrome Diagnosis: From Origins to Future Perspectives-A Comprehensive Review.
Level 5 - mechanism / opinion, no new human data
Narrative review synthesizing published literature without systematic search protocol or quantitative data aggregation.
PubMed 40002866 · doi:10.3390/biomedicines13020453
What was done
The authors conducted a narrative review of English-language articles retrieved from PubMed and Embase focused on ultrasound assessment in polycystic ovary syndrome (PCOS). The review evaluated 2D, 3D, and Doppler imaging modalities, diagnostic criteria, ovarian stroma metrics, and emerging machine-learning algorithms.
What was found
The abstract reports no pooled quantitative data or exact numerical estimates of diagnostic accuracy. It notes that 2D ultrasound follicle number per ovary (FNPO) and ovarian volume (OV) sensitivity and specificity vary substantially depending on probe frequency, patient factors (such as adolescence or obesity), and cut-off thresholds (≥12, ≥20, or ≥25 follicles). Three-dimensional and Doppler ultrasound enable automated measurements, stromal-to-ovarian area assessments, and vascular indices correlated with hyperandrogenism, while artificial intelligence models reduce observer variability.
Why it matters
It summarizes how technological advancements in imaging and machine learning are addressing longstanding limitations in standardized diagnostic criteria and operator subjectivity for polycystic ovarian morphology.
Limits
As a narrative review, it lacks a formal systematic search strategy, quality appraisal, and meta-analytic pooling. The abstract provides no specific numerical diagnostic metrics, and practical challenges including operator dependency, lack of age- and BMI-specific cutoffs, and limited multicenter validation of AI tools persist.
Cited by
- supports In PCOS, a polycystic ovary appearance on ultrasound is characterized by seeing 20 or more follicles arranged in a 'string of pearls' pattern.