Detecting blood clots with AI — Solving A Major Global Burden Of Disease
ThinkSono has now completed our CrowdCube funding campaign and reached 120% of our initial target!
Because of this, I wanted to dig deeper into why we are specifically working on DVT detection, and the huge potential of our technology in transforming the management of this condition.
First, some more context about DVT, which I will borrow from our Nature publication.
Venous thromboembolism — A major global burden of disease
Venous thromboembolism (VTE) is a major global burden of disease. Worldwide, the incidence of VTE is 1–3 per 1000 individuals.
VTE, deep vein thrombosis (DVT) and pulmonary embolus (PE) are the leading cause of hospital-related disability-adjusted life years lost. At least 7.7 million people will require investigation for VTE every year. An ageing population across many countries will lead to a greater health burden, particularly in middle- and low-income countries where early death from infection is decreasing.
Mortality from VTE is common, a European study estimated 534,000 deaths per year and a similar study in the US reported 60,000-300,000 deaths per year. As many as 900,000 people are affected by VTE in the US alone.
DVT has a high level of morbidity. 30–50% of the surviving patients develop long-term symptoms in their affected leg (post-thrombotic syndrome).
In high-income countries, the routine practice to diagnose patients after a positive D-dimer blood test and an indicative evaluation using the Wells score (a risk assessment) is to confirm or rule out a suspected DVT with a two- or three-point ultrasound scan.
Ultrasound scans are most commonly performed in a radiology or cardiovascular department of a hospital by a highly trained radiographer/radiologist.
The Inefficient DVT patient pathway
Between 85 and 90% of patients presenting to their GP in high-income countries with a suspected DVT will be investigated only to find no evidence of a thrombus.
Many patients will receive unnecessary anticoagulants with numerous potential side-effects through an often-painful subcutaneous injection whilst waiting more than the recommended four hours for their scan.
Safely negating this wait would improve patient satisfaction, reduce the burden of high-risk treatment (anticoagulants confer haemorrhagic complication risks) and discount healthcare costs.
The problem is clear — What is the solution?
The first step in treating a DVT is to diagnose it. Therefore, making the diagnosis as accessible and affordable as possible is the key to managing this condition better for millions of patients worldwide.
This is exactly the solution that we’re building at ThinkSono. In collaboration with our hospital partners such as Oxford University Hospital and NHS Blood Transplant in the UK, as well as other hospitals across the UK, Germany and Greece we have been building and developing our AutoDVT software.
AutoDVT is a simple App that you can download on your smartphone, coupled with a handheld ultrasound device a nurse can scan suspected DVT patients in as little as 5 minutes.
This enables all healthcare staff to scan suspected DVT patients at the point of care (e.g at their GP practice, at home or in the hospital) and avoids up to 90% of the referrals going to the radiology department.
The clinical pathway can then be simplified as shown in the image below.
What is the economic and clinical evidence for this?
We’re now building up this clinical evidence as we speak. We have published several peer reviewed papers that can be found on our website here.
However, a basic summary is given below:
- We have published our Nature paper on accuracy and health economics here, showing our technology can potentially generate a positive net monetary benefit up to £72 to £175 per software-supported examination.
- We published some of our clinical work at ECR2022 showing a sensitivity of 100% and specificity of 65%. We also obtained diagnostic quality data in 93% of our scans.
- Our technology has shown initial DVT diagnostic accuracy of 100% and a specificity of 77% by recruiting patients with a suspected DVT at two tertiary centres: Magdeburg, Germany and Athens, Greece (see here).
- We’re running a double blinded multi-centered clinical trial across 8 Hospitals in the UK and have recruited more than 400 patients. The study information is found here. As far as we are aware, this is the largest prospective AI Ultrasound study for DVT in the world!
What will happen next?
We’re gathering data from across our studies and plan on obtaining full regulatory clearance in the EU and USA.
The timing of this will depend on when our clinical data will be fully analysed, but we expect this will happen within the next 18–24 months.
We’re hugely excited about the progress to date and hope to continue to develop this hugely important technology for the patients, staff and healthcare providers worldwide.
If you wish to connect, email us at hello@thinksono.com
