Ultrasound Texture Analysis for the Diagnosis of Adenomyosis

Florian Delberghe, Zihan Gong, Catarina Dinis Fernandes, C.O. Rees, Huib van Vliet, Massimo Mischi
In: 2025 IEEE Medical Measurements & Applications (MeMeA) · 2025 · pp. 1–6 · doi:10.1109/memea65319.2025.11068108 · W4412171195
article OA: closed CC0
View on OpenAlex View at publisher
⚙ AI-generated summary by claude@2026-06, 2026-06-10 ⓘ

This study quantified uterine texture from ultrasound images using histogram and GLCM features, with a random forest classifier achieving 81.6% accuracy in diagnosing adenomyosis.

One-sentence paraphrase of the abstract; not a substitute for reading it. No clinical advice. How this works

Abstract

Adenomyosis is a common gynecological disease, with transvaginal ultrasonography being the first-line imaging method to diagnose it, based on the qualitative morphological uterus sonographic assessment (MUSA). However, the accuracy of diagnosis depends largely on a clinician's expertise, and scan's quality, often requiring additional investigations by MRI. To enable wider access to accurate and robust diagnostics by ultrasonography, computer-aided diagnosis tools are promising. The use of texture features, serving as a surrogate for the human visual system, can be used to quantitatively analyze ultrasound images. In this study, histogram and gray level co-occurrence matrix features, are used to quantify the texture from 2D grayscale ultrasound images of the uterus. Then several machine learning classifiers are trained to predict the presence or adenomyosis from a dataset of 37 patients with adenomyosis and 45 healthy patients. The best performing classifier, a random forest, achieved an accuracy of 81.6%, an F1 score of 0.747, and an area under the receiver operating curve of 0.804. These metrics show promise for the future use of texture measurements in ultrasound to allow for earlier and more accurate diagnosis of adenomyosis.

My notes (saved in your browser only)

Outcome instruments

MUSA

Condition tags

adenomyosis

Citation neighborhood

Papers in the corpus that this work cites (lower rings, blue) and that cite this one (upper rings, green). Dot size scales with the paper's in-corpus citation count — bigger dot = more influential within the endo/adeno field. Click a dot to open that paper. [ expand to 2 hops ] — adds papers reached through this work's immediate citers/citees. Heavier; up to 60 extra dots.

References (35)

Source provenance

openalex
last seen: 2026-06-10T17:14:06.276822+00:00
unpaywall
last seen: 2026-06-02T02:00:03.124865+00:00
License: CC0 · commercial use OK