Construction of a Standardized Time-Lapse Imaging Database and a Gradient Boosting Ensemble Framework for Integrating Zygote Morphokinetic Parameters with Conventional Embryo Assessment
- 1Guangdong Provincial Fertility Hospital, Guangzhou, China
Abstract
In vitro fertilization (IVF) laboratories equipped with time-lapse incubators generate vast quantities of sequential embryo images, yet the absence of standardized, annotated databases impedes the development of reproducible computational tools for embryo assessment. Here we describe the standardized time-lapse imaging database built upon prior research groundwork comprising 631 two-pronuclear (2PN) zygotes from 218 treatment cycles performed at Guangdong Provincial People’s Hospital (2020–2023), together with a gradient boosting decision tree (GBDT) ensemble framework designed to fuse heterogeneous data types for blastocyst outcome prediction. Each embryo record integrates 84 zygote-stage morphokinetic parameters extracted from EmbryoScope time-lapse sequences via a previously validated convolutional neural network segmentation pipeline with 8 conventional embryo assessment features recorded at cleavage and blastocyst stages according to the Istanbul consensus. The fusion framework employs LightGBM with equal-weight initialization and iterative residual-decreasing training, augmented by recursive feature elimination and nested five-fold cross-validation. Ablation experiments demonstrate that the full model (AUC = 0.78) outperforms morphokinetics-only (AUC = 0.71) and conventional-only (AUC = 0.65) configurations, confirming that zygote-stage temporal dynamics carry complementary information beyond standard morphological grading. SHAP analysis identifies cytoplasmic area slope, zona pellucida grayscale trend, and pronuclear fading time as the three most influential predictors. The database and fusion methodology provide a reproducible framework for integrating time-series imaging features with categorical clinical assessments in reproductive medicine.
Related
- A retrospective comparative study of a full-course precision MDT nursing studio model integrating rational emotive behavior therapy for standardized health management during pregnancy
- Quantified standardized individualized small craniotomy vs. conventional large craniotomy for cerebral herniation secondary to acute supratentorial epidural hematoma: a propensity score-matched study for emergency and battlefield trauma care
- Integrating technology in leech therapy: Toward standardized evidence-based Ayurvedic practice.
- Benefits of Integrating Standardized Patient Encounters into Pre-clerkship Neurology Curriculum.
- A standardized Orthosiphon aristatus extract improves forced-swimming performance and modulates hematological and hepatic biochemical parameters in paclitaxel-treated mice
BibTeX
@misc{zhao2026construction,
title = {{Construction of a Standardized Time-Lapse Imaging Database and a Gradient Boosting Ensemble Framework for Integrating Zygote Morphokinetic Parameters with Conventional Embryo Assessment}},
author = {M. ZHAO and J. LIU and D. HAN and C. ZHANG and Y. ZHOU and S. CHEN and K. Appiah and C. LIU},
howpublished = {medRxiv},
year = {2026},
month = aug,
doi = {10.64898/2026.08.20.26359523},
url = {https://www.medrxiv.org/content/10.64898/2026.08.20.26359523v2},
note = {Public domain},
}
RIS
TY - UNPB
TI - Construction of a Standardized Time-Lapse Imaging Database and a Gradient Boosting Ensemble Framework for Integrating Zygote Morphokinetic Parameters with Conventional Embryo Assessment
AU - M. ZHAO
AU - J. LIU
AU - D. HAN
AU - C. ZHANG
AU - Y. ZHOU
AU - S. CHEN
AU - K. Appiah
AU - C. LIU
PY - 2026
DA - 2026/08/20
PB - medRxiv
DO - 10.64898/2026.08.20.26359523
UR - https://www.medrxiv.org/content/10.64898/2026.08.20.26359523v2
AB - In vitro fertilization (IVF) laboratories equipped with time-lapse incubators generate vast quantities of sequential embryo images, yet the absence of standardized, annotated databases impedes the development of reproducible computational tools for embryo assessment. Here we describe the standardized time-lapse imaging database built upon prior research groundwork comprising 631 two-pronuclear (2PN) zygotes from 218 treatment cycles performed at Guangdong Provincial People’s Hospital (2020–2023), together with a gradient boosting decision tree (GBDT) ensemble framework designed to fuse heterogeneous data types for blastocyst outcome prediction. Each embryo record integrates 84 zygote-stage morphokinetic parameters extracted from EmbryoScope time-lapse sequences via a previously validated convolutional neural network segmentation pipeline with 8 conventional embryo assessment features recorded at cleavage and blastocyst stages according to the Istanbul consensus. The fusion framework employs LightGBM with equal-weight initialization and iterative residual-decreasing training, augmented by recursive feature elimination and nested five-fold cross-validation. Ablation experiments demonstrate that the full model (AUC = 0.78) outperforms morphokinetics-only (AUC = 0.71) and conventional-only (AUC = 0.65) configurations, confirming that zygote-stage temporal dynamics carry complementary information beyond standard morphological grading. SHAP analysis identifies cytoplasmic area slope, zona pellucida grayscale trend, and pronuclear fading time as the three most influential predictors. The database and fusion methodology provide a reproducible framework for integrating time-series imaging features with categorical clinical assessments in reproductive medicine.
ER -