Artificial Intelligence-Based Machine Learning to Diagnose and Classify Adenomyosis from Ultrasound Scans: a Multicentre Model Development Study

NCT06765512 | Active Not Recruiting

• 1 other identifier: AIMLadeno
• Study Type: observational
• Target: 10,000
• Locations: 1 country
• Sites: 1

Brief Summary

The aim of this study is to use the vast dataset of annotated ultrasound images of normal uterus and of adenomyosis of varying severity to train a neural network using deep learning framework (Pytorch) and automated machine learning tool (Vertex AI). The main question it aims to answer are:

1. 1.Diagnostic performance of automated (Google Vertex AI (Artificial intelligence) vision) and deep learning (Pytorch) machine learning model
2. 2.Time saved in assessment of adenomyosis per healthcare professional

Conditions

Adenomyosis of Uterus

Keywords

diagnosis; deep learning; machine learning; classification; artificial learning; automated machine learning

Outcome Measures

Primary Outcomes (4)

• Precision

  Precision will indicate how well the model is capturing information and how much it is leaving out. From all the test set of images that were assigned a category, precision indicates how many actually were supposed to be categorised with that label.

  through study completion, an average of 1 year

• Recall

  From all the test set of images that were assigned a category, recall indicates indicates how many were actually assigned the label.

  through study completion, an average of 1 year

• Average precision

  The model accuracy will be determined by the area under the precision-recall curve and in Vertex AI, this metric is called average precision. It measures how well the model performs across all score threshold. The closer the score is to 1, the better is the model's performance on the test set.

  through study completion, an average of 1 year

• Number of images correctly identified as per their original classification

  Images correctly identified as per their original classification label e.g. Mild labelled image of adenomyosis identified as mild. This will be true positive. Images incorrectly identified and not in line with original classification system e.g. moderate labelled image of adenomyosis classified as mild. This will be false positive

  through study completion, an average of 1 year

Secondary Outcomes (2)

• Interrater agreement

  through study completion, an average of 1 year

• Time taken to classify all the set of images

  through study completion, an average of 1 year

Study Arms (1)

None (Non-adenomyosis) and adenomyotic uterus ultrasound images

None group will be set of images with homogenous myometrium. The Adenomyosis group will be a set of images with ultrasound features of adenomyosis as per MUSA diagnostic criteria. These will be divided into mild, moderate and severe.

Other: use of deep learning and automated machine learning to diagnose and classify adenomyosis

Interventions

use of deep learning and automated machine learning to diagnose and classify adenomyosis OTHER

Vertex AI Vision V1 software will be used as an automated machine learning tool and Pytorch 2.5 as deep learning framework. The complete set of reviewed, formatted and labelled images will be uploaded and split manually into two different datasets in 9:1 ratio; 90% of the selected images will be used as training dataset (training + validation) and 10% as test dataset.

None (Non-adenomyosis) and adenomyotic uterus ultrasound images

Eligibility Criteria

• Sex: female
• Healthy Volunteers: No
• Age Groups: Child (0-17), Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

The study cohort will comprise of ultrasound images of patients who attended CARE Fertility centres for Ultrasound between February 2022 to February 2024 and were diagnosed with normal uterus and adenomyosis using Morphological Uterus Sonographic Assessment (MUSA) criteria on screening of images. The MUSA criteria outlines the ultrasound features of myometrium and myometrial lesions using standardised terms, definitions and measurements. Previously published schematic mapping system of adenomyosis severity will be used for determining the severity of uterine adenomyosis on review of images. A score ranging from to 1 to 4 is attributed to each grade and the sum of the score numbers is used to calculate the extension of the disease: mild (1-3), moderate (4-6), and severe (\>7)

You may qualify if:

• Participants: Women attending CARE Fertility centre for ultrasound between February 2022 to February 2024 for any indication and are diagnosed with normal uterus and adenomyosis on ultrasound on screening of images.
• Input data: Good quality and conclusive 2D and/ or 3D images of normal uterus and adenomyotic uterus where the ultrasound characteristics of adenomyosis are clearly visible.

You may not qualify if:

• Participants: Women with co-existing single or multiple intramural fibroids and endometrial cavity abnormalities.
• Input data: Inconclusive ultrasound on assessment by the second reviewer, poor-quality images where the ultrasound characteristics of adenomyosis are unclear and images which cannot be classified into one of the four categories will be excluded.

Sponsors & Collaborators

• CARE Fertility UK: lead
• University of Birmingham: collaborator

Study Sites (1)

CARE Fertility

Birmingham, England, United Kingdom

Location

Related Publications (19)

• Bergholt T, Eriksen L, Berendt N, Jacobsen M, Hertz JB. Prevalence and risk factors of adenomyosis at hysterectomy. Hum Reprod. 2001 Nov;16(11):2418-21. doi: 10.1093/humrep/16.11.2418.

  PMID: 11679531 BACKGROUND

• Munro MG. Uterine polyps, adenomyosis, leiomyomas, and endometrial receptivity. Fertil Steril. 2019 Apr;111(4):629-640. doi: 10.1016/j.fertnstert.2019.02.008.

  PMID: 30929720 BACKGROUND

• Drukker L, Noble JA, Papageorghiou AT. Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology. Ultrasound Obstet Gynecol. 2020 Oct;56(4):498-505. doi: 10.1002/uog.22122.

  PMID: 32530098 BACKGROUND

• Sim J, Wright CC. The kappa statistic in reliability studies: use, interpretation, and sample size requirements. Phys Ther. 2005 Mar;85(3):257-68.

  PMID: 15733050 BACKGROUND

• Zapf A, Castell S, Morawietz L, Karch A. Measuring inter-rater reliability for nominal data - which coefficients and confidence intervals are appropriate? BMC Med Res Methodol. 2016 Aug 5;16:93. doi: 10.1186/s12874-016-0200-9.

  PMID: 27495131 BACKGROUND

• Moons KGM, Wolff RF, Riley RD, Whiting PF, Westwood M, Collins GS, Reitsma JB, Kleijnen J, Mallett S. PROBAST: A Tool to Assess Risk of Bias and Applicability of Prediction Model Studies: Explanation and Elaboration. Ann Intern Med. 2019 Jan 1;170(1):W1-W33. doi: 10.7326/M18-1377.

  PMID: 30596876 BACKGROUND

• Harmsen MJ, Van den Bosch T, de Leeuw RA, Dueholm M, Exacoustos C, Valentin L, Hehenkamp WJK, Groenman F, De Bruyn C, Rasmussen C, Lazzeri L, Jokubkiene L, Jurkovic D, Naftalin J, Tellum T, Bourne T, Timmerman D, Huirne JAF. Consensus on revised definitions of Morphological Uterus Sonographic Assessment (MUSA) features of adenomyosis: results of modified Delphi procedure. Ultrasound Obstet Gynecol. 2022 Jul;60(1):118-131. doi: 10.1002/uog.24786.

  PMID: 34587658 BACKGROUND

• Van den Bosch T, Dueholm M, Leone FP, Valentin L, Rasmussen CK, Votino A, Van Schoubroeck D, Landolfo C, Installe AJ, Guerriero S, Exacoustos C, Gordts S, Benacerraf B, D'Hooghe T, De Moor B, Brolmann H, Goldstein S, Epstein E, Bourne T, Timmerman D. Terms, definitions and measurements to describe sonographic features of myometrium and uterine masses: a consensus opinion from the Morphological Uterus Sonographic Assessment (MUSA) group. Ultrasound Obstet Gynecol. 2015 Sep;46(3):284-98. doi: 10.1002/uog.14806. Epub 2015 Aug 10.

  PMID: 25652685 BACKGROUND

• Ito Y, Unagami M, Yamabe F, Mitsui Y, Nakajima K, Nagao K, Kobayashi H. A method for utilizing automated machine learning for histopathological classification of testis based on Johnsen scores. Sci Rep. 2021 May 10;11(1):9962. doi: 10.1038/s41598-021-89369-z.

  PMID: 33967273 BACKGROUND

• LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015 May 28;521(7553):436-44. doi: 10.1038/nature14539.

  PMID: 26017442 BACKGROUND

• Collins GS, Dhiman P, Andaur Navarro CL, Ma J, Hooft L, Reitsma JB, Logullo P, Beam AL, Peng L, Van Calster B, van Smeden M, Riley RD, Moons KG. Protocol for development of a reporting guideline (TRIPOD-AI) and risk of bias tool (PROBAST-AI) for diagnostic and prognostic prediction model studies based on artificial intelligence. BMJ Open. 2021 Jul 9;11(7):e048008. doi: 10.1136/bmjopen-2020-048008.

  PMID: 34244270 BACKGROUND

• Kwong JCC, Khondker A, Lajkosz K, McDermott MBA, Frigola XB, McCradden MD, Mamdani M, Kulkarni GS, Johnson AEW. APPRAISE-AI Tool for Quantitative Evaluation of AI Studies for Clinical Decision Support. JAMA Netw Open. 2023 Sep 5;6(9):e2335377. doi: 10.1001/jamanetworkopen.2023.35377.

  PMID: 37747733 BACKGROUND

• Guerriero S, Pascual M, Ajossa S, Neri M, Musa E, Graupera B, Rodriguez I, Alcazar JL. Artificial intelligence (AI) in the detection of rectosigmoid deep endometriosis. Eur J Obstet Gynecol Reprod Biol. 2021 Jun;261:29-33. doi: 10.1016/j.ejogrb.2021.04.012. Epub 2021 Apr 14.

  PMID: 33873085 BACKGROUND

• Singhal N, Kudavelly S, Ramaraju GA. Deep Learning Based Junctional Zone Quantification using 3D Transvaginal Ultrasound in Assisted Reproduction. Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:2133-2136. doi: 10.1109/EMBC44109.2020.9175358.

  PMID: 33018428 BACKGROUND

• Lazzeri L, Morosetti G, Centini G, Monti G, Zupi E, Piccione E, Exacoustos C. A sonographic classification of adenomyosis: interobserver reproducibility in the evaluation of type and degree of the myometrial involvement. Fertil Steril. 2018 Nov;110(6):1154-1161.e3. doi: 10.1016/j.fertnstert.2018.06.031.

  PMID: 30396560 BACKGROUND

• Exacoustos C, Morosetti G, Conway F, Camilli S, Martire FG, Lazzeri L, Piccione E, Zupi E. New Sonographic Classification of Adenomyosis: Do Type and Degree of Adenomyosis Correlate to Severity of Symptoms? J Minim Invasive Gynecol. 2020 Sep-Oct;27(6):1308-1315. doi: 10.1016/j.jmig.2019.09.788. Epub 2019 Oct 7.

  PMID: 31600574 BACKGROUND

• Naftalin J, Hoo W, Pateman K, Mavrelos D, Holland T, Jurkovic D. How common is adenomyosis? A prospective study of prevalence using transvaginal ultrasound in a gynaecology clinic. Hum Reprod. 2012 Dec;27(12):3432-9. doi: 10.1093/humrep/des332. Epub 2012 Sep 20.

  PMID: 23001775 BACKGROUND

• Timmerman D, Verrelst H, Bourne TH, De Moor B, Collins WP, Vergote I, Vandewalle J. Artificial neural network models for the preoperative discrimination between malignant and benign adnexal masses. Ultrasound Obstet Gynecol. 1999 Jan;13(1):17-25. doi: 10.1046/j.1469-0705.1999.13010017.x.

  PMID: 10201082 BACKGROUND

• Mishra I, Melo P, Easter C, Sephton V, Dhillon-Smith R, Coomarasamy A. Prevalence of adenomyosis in women with subfertility: systematic review and meta-analysis. Ultrasound Obstet Gynecol. 2023 Jul;62(1):23-41. doi: 10.1002/uog.26159. Epub 2023 Apr 28.

  PMID: 36647238 BACKGROUND

MeSH Terms

Conditions

Adenomyosis; Disease

Condition Hierarchy (Ancestors)

Uterine Diseases; Genital Diseases, Female; Female Urogenital Diseases; Female Urogenital Diseases and Pregnancy Complications; Urogenital Diseases; Genital Diseases; Pathologic Processes; Pathological Conditions, Signs and Symptoms

Study Officials

• Mohammed Khairy

  CARE Fertility, Birmingham

  PRINCIPAL INVESTIGATOR

Study Design

• Study Type: observational
• Observational Model: COHORT
• Time Perspective: RETROSPECTIVE
• Sponsor Type: OTHER
• Responsible Party: SPONSOR

Study Record Dates

• First Submitted: January 2, 2025
• First Posted: January 9, 2025
• Study Start: June 4, 2024
• Primary Completion: December 4, 2025
• Study Completion: February 6, 2026
• Last Updated: January 9, 2025

Record last verified: 2025-01

Data Sharing

• IPD Sharing: Will not share

Locations

GB (1)