Скачать или смотреть What others found about EndoClassify

Learn more: https://www.endoclassify.com/
EndoClassify: from routine ultrasound to reliable bedside decision support
Across IVF, ultrasound is routine, yet many decisions still rely on manual measurements and subjective reads. An independent review asked a simple question: can AI turn a standard scan into reliable, real-time decision support?
What EndoClassify does
AI model built for transvaginal ultrasound that segments the endometrium and then classifies images, automating steps that usually depend on operator experience.

Modern deep-learning pipeline with strong performance on retrospective, multicenter data: about 95% accuracy and around 93% sensitivity and specificity.

Clinical value comes from what is measured: endometrial architecture, especially the external hyperechoic layers and their proportion of the whole lining. Thickness alone can miss this.

Translates those features into a simple Good vs Bad classification linked to likelihood of pregnancy after transfer.

Why it matters
Reduced inter-observer variability, faster exams via automated measurement, and improved reproducibility.

Helps teams decide whether to proceed now or optimize the lining first, using evidence distilled from the image rather than relying only on millimeters.

Roadmap and call to action
Integrate AI image analysis with radiomics and multi-omics (proteomics, metabolomics, transcriptomics) for earlier detection, personalized transfer timing, and smarter risk stratification in hyperplasia and endometrial cancer.

Next step: multi-center validation, shared datasets, and standardized workflows.

Vida Shafti1, School of Medicine, Tonekabon Branch, Islamic Azad University, Tonekabon, Iran
Alireza Azarboo, School of Medicine, Tehran University of Medical Sciences, Tehran, Iran
https://saremjrm.com/browse.php?a_id=...

#IVF #Ultrasound #Endometrium #AI #ReproductiveMedicine #Fertility #EndoClassify #IVFInnovation #IVF #fertilitytech #digitalhealth #fertilityhealthcare #healthcareinnovation