Berlin’s DeepSpin raises seed funding for its ‘portable, ultra-low-cost’ MRI system

      Comments Off on Berlin’s DeepSpin raises seed funding for its ‘portable, ultra-low-cost’ MRI system

Share This Story

DeepSpin, a Berlin-based startup that is developing what it describes as a “next-generation, AI-powered MRI imaging machine”, has raised €600,000 in seed funding.

Backing the round is APEX Digital Health, with participation from existing investors Entrepreneur First (EF) and SOSV, along with a number of unnamed angel investors. Including grants and earlier investment, it brings the total raised to €1 million pre-launch.

DeepSpin is a graduate of EF’s company builder programme, where its two founders — Clemens Tepel, a former McKinsey consultant, and Pedro Freire Silva, a PhD researcher from KIT — decided to partner in September 2019. Freire Silva drew on his research into small-scale, mass-manufacturable MRI systems and pitched the idea to his future co-founder.

“From the beginning I found the idea very intriguing and so we directly jumped into attempting to prove its feasibility,” says Tepel. “Within 4 weeks we were able to prove it in simulation, get industry-leading advisors on board and get first LOIs [letter of intent] from interested clinicians”.

Yet-to-launch and still in the development phase, DeepSpin aims to build a new type of MRI system at a “fraction of the cost, weight and size” of existing systems. To make this possible, the startup is has developed a new antenna technology combined with AI-controlled operation, which the startup is currently patenting.

“The problem we are solving is that MRI, the most advanced medical imaging method, is currently not easily accessible because it is incredibly expensive, requires specialised operators and needs specifically shielded rooms,” explains Tepel. “We are removing all of these constraints based on our proprietary technology, making MRI universally accessible for any patient, anywhere in the world”.

Adds Freire Silva: “Instead of combining highly expensive hardware with standard software, as it is done on conventional MRI scanners, we will be able to obtain the same clinical information by applying very sophisticated algorithms on simplified hardware, thereby reducing our system’s cost by orders of magnitude”.

Tepel tells me this approach has not been taken before because both key enablers — highly capable AI-algorithms and the specific antenna design – were only available very recently.

Having proven DeepSpin’s methods in simulation, the next step and the team’s current focus is to develop a first fully AI-driven prototype. “Based on that, we will develop an initial product version, aimed at pre-clinical applications, before going into medical certification, which then will allow us to sell our product for clinical use across a range of medical domains and to new geographies that can’t afford conventional systems,” says Tepel.