Portable vein checker: AI to help with early detection of thrombosis

Deep vein thrombosis (DVT) poses a considerable health risk. Often only a feeling of heaviness, tingling or pulling in the lower leg, leg swelling or a feeling of heat are signs of such a clot, which can often lead to long-term health problems in the affected veins or even life-threatening complications such as a pulmonary embolism. Up to two thirds of DVT cases even show no symptoms. An international team of experts wants to close this gap with the ThrombUS+ research project. A portable diagnostic device that uses artificial intelligence (AI) to monitor patients and detect blood clots at an early stage is intended to help.

The use of the planned intelligent medical wearable is intended to relieve the burden on doctors and improve prevention and diagnosis. The project was launched in January and is being funded for three and a half years with 9.5 million euros via the EU research framework program Horizon Europe. It involves 18 partners from the EU and the USA and is led by the Greek Athena Research Center. From Germany, the electrical engineering association VDE, the Fraunhofer Institute for Photonic Microsystems (IPMS) and the company Medis, which specializes in medical measurement technology, are involved.

Real-time monitoring of patients with AI

According to the creators, the importance of early detection is statistically proven. According to them, in around half of DVT sufferers, the serious clot that impairs normal blood flow, particularly in the veins of the lower limbs, travels to the lungs and blocks a vessel there. In exceptional cases, such a thrombus - for example through a congenital hole in the heart - can reach the brain, where it can trigger a stroke. Pulmonary embolisms are the third most common cause of cardiovascular death worldwide after strokes and heart attacks.

The ThrombUS+ consortium wants to counteract this with the portable diagnostic technology. In the future, it should enable user-friendly, operator-free and continuous real-time monitoring for patients at high risk of DVT. The approach combines "AI-controlled detection mechanisms based on ultrasound technology, impedance plethysmography (measurement of the electrical alternating current resistance) and light reflection rheography", i.e. data acquisition using infrared light, explains the VDE. The technical development is accompanied by clinical studies to ensure the safety and performance of the medical device.

Data sets should be available to all interested parties

The new method is to be used in particular for patients in the post-operative phase in hospital wards, during long-term surgical procedures, for cancer patients, bedridden people at home or in care facilities and women during and after pregnancy. According to the team, it aims to incorporate "legal, regulatory and safety requirements for complex medical devices as early as possible in the development process" to make findings ready for practical application as quickly as possible. According to the first published project reports, constant monitoring of the relevant state of the art and patent claims is planned.

The participants have also already published a data management plan. In it, they provide information on which data sets the consortium will collect, generate and store and in what form. There are also best practice examples relating to metadata and archiving. These are intended to ensure that the data sets collected or generated are findable, accessible, interoperable and reusable in accordance with the principles for open science prescribed by Horizon Europe. The data plan will be updated regularly, the partners promise. Building on the clinical, technical and regulatory experience gained, they ultimately also want to "sustainably improve the chances of market access for future complex AI-based medical products".

(bme)