Automated identification of blastocyst regions at different development stages

The selection of the best single blastocyst for transfer is typically based on the assessment of the morphological characteristics of the zona pellucida (ZP), trophectoderm (TE), blastocoel (BC), and inner cell-mass (ICM), using subjective and observer-dependent grading protocols. We propose the fir...

Full description

Saved in:
Bibliographic Details
Published in:Scientific reports 2023-01, Vol.13 (1), p.15-15, Article 15
Main Authors: Farias, Adolfo Flores-Saiffe, Chavez-Badiola, Alejandro, Mendizabal-Ruiz, Gerardo, Valencia-Murillo, Roberto, Drakeley, Andrew, Cohen, Jacques, Cardenas-Esparza, Elizabeth
Format: Article
Language:English
Subjects:
631/114/1305
631/114/1564
631/114/2397
631/136/1455
631/136/756
639/705/117
Blastocyst
Blastocysts
Developmental stages
Embryo, Mammalian
Hatching
Humanities and Social Sciences
Morphology
multidisciplinary
Physical characteristics
Science
Science (multidisciplinary)
Sensitivity analysis
Trophectoderm
Zona Pellucida
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The selection of the best single blastocyst for transfer is typically based on the assessment of the morphological characteristics of the zona pellucida (ZP), trophectoderm (TE), blastocoel (BC), and inner cell-mass (ICM), using subjective and observer-dependent grading protocols. We propose the first automatic method for segmenting all morphological structures during the different developmental stages of the blastocyst (i.e., expansion, hatching, and hatched). Our database contains 592 original raw images that were augmented to 2132 for training and 55 for validation. The mean Dice similarity coefficient (DSC) was 0.87 for all pixels, and for the BC, BG (background), ICM, TE, and ZP was 0.85, 0.96, 0.54, 0.63, and 0.71, respectively. Additionally, we tested our method against a public repository of 249 images resulting in accuracies of 0.96 and 0.93 and DSC of 0.67 and 0.67 for ICM and TE, respectively. A sensitivity analysis demonstrated that our method is robust, especially for the BC, BG, TE, and ZP. It is concluded that our approach can automatically segment blastocysts from different laboratory settings and developmental phases of the blastocysts, all within a single pipeline. This approach could increase the knowledge base for embryo selection.
ISSN:2045-2322
2045-2322
DOI:10.1038/s41598-022-26386-6