New Study Aims to Revolutionize the Diagnosis of Endometriosis

One in ten women of reproductive age suffer from endometriosis[1]. Out of these women, 30-50% will experience fertility problems because of anatomical distortion of crucial reproductive structures and hampering of ovarian reserves[2]. Despite such life-altering symptoms, it takes a woman in the UK on average eight years to receive a diagnosis by surgical laparoscopy[3], by which point, the condition will have developed significantly.

“Diagnosis through surgery is outdated, invasive and comes with its own risks. More often than not, patients are also referred for laparoscopy after the disease has progressed visibly and significantly, and at a point where the chances of preserving fertility are much lower,” says Chief Investigator Dr Ippokratis Sarris, Consultant in Reproductive Medicine and Director of King’s Fertility.

“In addition, the cost of surgery when you factor in the operating theatre time, the surgeon, the anaesthetist and the wider team – not to mention the cost of disrupting the patient’s life – is substantial, and we should be moving away from surgery towards less invasive diagnostic tools,” Dr Sarris continues.

A new world of AI

Joined by NCIMI (National Consortium of Intelligent Medical Imaging) and industry partners, GE Healthcare and Perspectum, Dr Sarris began a study this year to investigate whether less invasive methods such as ultrasound and MRI may be used, supported by AI.

“AI opens up a brave new world of possibilities and the field of fertility has been one of the first to adopt it. In a medical sense, it can be used to substitute two critical limitations: the intensity of expert human resources and the subjectivity of imaging,” says Dr Sarris.  

One of the barriers to effectively diagnosing endometriosis is a lack of awareness in the public and in the medical field. “Endometriosis can be difficult to diagnose,” says Dr Sarris. “It requires an expert with extensive experience and a keen eye for identifying the complexities of the condition”.

The diagnostic pathway for endometriosis has not changed in over ten years, giving women with symptoms very little confidence about their chances of a rapid diagnosis and prompt treatment. An inquiry published by the All-Party Parliamentary Group on endometriosis (APPG) reported that 58% of patients visited their GP more than 10 times, 43% visited the hospital more than five times, and 53% visited A&E, all before receiving a diagnosis[4].

This is where AI will play an important part. “Even experts will not always agree on a diagnosis. AI takes away the reliance on manual and subjective processes and therefore lends itself nicely to the area of endometriosis,” says Dr Sarris.

Using less invasive methods

Led by Dr Sarris, the study named DEFEND (Developing an Ultrasound-MRI-biomarker fusion model for Endometriosis), will explore the effectiveness of using 2D and 3D ultrasound and MRI scanning in the diagnosis of endometriosis.

“Ultrasound has always been a strong tool in imaging for gynaecology and endometriosis, but it is very specialised. Our goal for this project is to optimise the benefits of established ultrasound imaging technology in real time; to see motion within the tissue, look for areas of scarring and identify the source of pain,” says Barbara Del Prince, Global Product & Clinical Management, GE Healthcare.

“Ultimately, ultrasound will give us information that, in combination with other exams, will provide a more objective evaluation of the tissue itself, giving a complete overall picture of endometriosis. The addition of the AI component will allow us to move from an extremely specialist examination to a more routine type of assessment to enhance accuracy and standardise results,” Del Prince adds.

In tandem, Perspectum will be using multiple parametric MRI sequences with various different imaging angles to map the pelvis. “What Perspectum is bringing to the table is a paradigm shift in how we interpret MRI. Instead of using pattern recognition, we will be able to read and interpret the actual makeup of the tissue on a point-by-point basis,” says Professor Sally Collins, Clinical Lead for Women’s Health at Perspectum Ltd.

“This is important because endometriosis manifests itself in a number of ways and with various subtypes so that no two cases will be the same. It is valuable to be able to distinguish between the different subtypes – ovarian, deep and common superficial peritoneal endometriosis – to tailor treatment plans accordingly. However, this often goes beyond the capabilities of standard imaging or even laparoscopy,” Professor Collins continues.

The DEFEND study

The project began to recruit patients in January 2022 and is set to run until March 2023 but has already sparked lots of interest in the field. “The study is already encouraging a lot of passion and self-reflection in clinical settings, asking questions such as ‘how can we train experts to identify hidden signs of endometriosis?’” says Dr Sarris.

Women between the ages of 18 and 40 with suspected endometriosis, and who are already scheduled for laparoscopy, are being recruited for 2D and 3D ultrasound and MRI scans with top of the range imaging systems. These machines will produce raw data from specific sequences and captures to give a detailed visual mapping of the affected pelvic region.

“It is still early days, so we haven’t been able to determine anything about the data just yet, but what we have seen is a great pickup from patients who want to take part in the project, showing that this is truly an unmet need,” Dr Sarris concludes.

 

REFERENCES