Abstract
Texture and shape analysis has been proposed to improve the diagnosis of pla-centa accreta spectrum (PAS) using magnetic resonance (MR) images. To be used clinically, it must be robust to variations in image resolution and slice thickness. Therefore, the aim of this study was to determine whether image dimensions affected texture and shape features extracted from placental MR images. 49 healthy and 35 PAS MR images with a resolution and slice thick-ness range of 0.7-1.8mm and 4-6mm were collected. Images were downsam-pled to a range of image resolutions (0.8-2mm) and slice thicknesses (4.5-7mm) and a standard dimension of 1.8x1.8x6mm³. 2525 texture and shape fea-tures were extracted from the placenta in the original and resampled images us-ing PyRadiomics. 6% (147) of texture and 50% (7) of shape features displayed a significant difference between healthy and PAS, along with low variation (coefficient of variation: CV < 0.2) and strong agreement (intraclass correla-tion coefficient: ICC > 0.8) across all image dimensions investigated. All 10 shape features showed minimal variation (CV < 0.03) and excellent agreement (ICC > 0.99). A higher number of texture features exhibited a significant difference between healthy and PAS when derived from the original MR images with a range of image dimensions (79%, 1990), compared to MR images resampled to a standard dimension (60%, 1500). In conclusion, shape, first or-der and LBP3D-derived texture features exhibited low variation and significant differences between healthy placenta and PAS MR images across a range of image dimensions. Original MR images with a range of image dimensions retained more potentially useful textural information for differentiating healthy placentas and PAS.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/PIPPI_006.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
BibTex
@InProceedings{ColSuz_The_MICCAISAT2026,
author = { Coldman, Suzie AND Zhou, Guanglong AND Wu, Dengjiang AND Guo, Lingzhong AND Whitby, Elspeth AND Li, Xinshan},
title = { { The effect of image resolution on texture and shape features derived from placental magnetic resonanceimages } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17257},
month = {pending},
page = {pending}
}