The abstract is read at the publisher; the summary is JClub's.
Hyperspectral imaging of single immobilized spermatozoa before ICSI enabled predictive models for fertilization, blastulation, and usable blastocyst formation in 93 couples.
- Why it matters: Current ICSI sperm selection relies on morphology and motility, which may not reflect functional competence, and molecular-quality tests are often destructive. A non-invasive way to characterize individual sperm could improve selection.
- What they did: In a prospective two-phase study (March 2020–August 2024) at one IVF center, single immobilized sperm were imaged with a hyperspectral camera on a micromanipulation microscope. XGBoost decision-tree models used hyperspectral spectra and clinical data, including chained models.
- The result: Technical repeatability was high (global intraclass correlation coefficient 0.999), and exploratory classification across repeated acquisitions reached 93.84% sensitivity and 96.76% specificity. In a sibling donor-oocyte safety evaluation, fertilization, blastulation, and usable blastocyst rates were comparable to standard ICSI, supporting feasibility and model development.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.