Cressida Bradley: Navigating AI Challenges in Endometriosis Diagnosis
Cressida Bradley, Program Manager and Research Advisor at APNA, shared a post by Imagendo Aus on LinkedIn about a paper by Alison Deslandes et al. published in Human Reproduction:
“The recent disturbing allegations about endometriosis treatment in Australia have shone a light on the challenges of diagnosis and management and reignited the debate about gold standard approaches. While the path forward is complex, AI is likely to play a crucial role in future endometriosis diagnosis and management approaches. This excellent new Australian research article illustrates the inherent difficulties in finding suitable ‘ground truth’ for training and evaluating ML models when the real-world training data is multimodal, complex and noisy.”
Quoting Imagendo Aus’s post:
“Endometriosis diagnosis has never been simple – and building AI in this space has made that reality impossible to ignore.
Over the past five years, our team has been developing and refining AI models for endometriosis imaging. In doing so, we repeatedly encountered the same challenge: diagnostic ‘ground truth’ is not absolute.
Symptoms vary.
Imaging quality varies.
Surgical and histopathological confirmation have limitations.
When AI is trained on imperfect reference standards, it risks encoding those imperfections into its predictions. Our newly published paper in Human Reproduction is the culmination of five years of technical development, iteration, and critical reflection.
The goal is not to eliminate complexity. It is to harness it.
All diagnostic tests have limitations.
Understanding them is how we build better intelligence.
As AI becomes embedded in clinical imaging workflows, these conversations are no longer theoretical – they are foundational.”
Title: The problem with the ‘truth’: rethinking ground truth for artificial intelligence in endometriosis diagnosis
Authors: Alison Deslandes, Yuan Zhang, Mathew Leonardi, Hsiang-Ting Chen, Gustavo Carneiro, Jodie Avery, George Condous, Steven Knox, M Louise Hull, the IMAGENDO Team
Read the full article.
Stay updated on all scientific advances in the field of fertility with Fertility News.
-
Oct 16, 2026, 19:56Dina Ibrahem: Poor ICM vs. Poor TE
-
Oct 16, 2026, 19:47Sean Lauber: NIPT and Amniocentesis after PGT-A Reveal Few Abnormalities
-
Oct 16, 2026, 19:42Devyanshi Dixit: C-Section vs. Vaginal Birth – 2 Birth Paths, 2 Very Different Recoveries
-
Oct 16, 2026, 19:32Pawel Kordowitzki: Maternal Vitamin D Homeostasis, Pregnancy Outcomes and Offspring Health
-
Oct 16, 2026, 19:13Shizuko Takahashi: Our Study on Women’s Experiences of Assisted Reproductive Care in China Has Been Accepted by RBMO
-
Oct 16, 2026, 19:09ESHRE Announces ECRES-GESEA Reproductive Surgery Training and Certification in Slovenia
-
Oct 16, 2026, 18:20Kabita KB: What Happens if PLCZ1 Is Defective?
-
Oct 16, 2026, 18:12Partha Sarathi Das: From Reproductive Lifespan to Reproductive Healthspan
-
Oct 16, 2026, 18:05Joyce Harper: It Was a Pleasure to Give 2 Talks Celebrating World Menopause Day