Artificial Intelligence

Hamburg's Mindpeak using AI to diagnose cancer

8 September 2021
Swifter discovery can improve life expectancy

A longer life expectancy goes hand in hand with the risk of cancer. The Hamburg-based Mindpeak is developing artificial intelligence to help pathologists diagnose cancer swiftly. Until now, specialists have diagnosed cancer mainly by examining tissue samples through a microscope similar to their predecessors 100 years ago. "Artificial intelligence, however, can now help them significantly. A pathologist needs up to 10 minutes for some findings. Our AI only takes only a few seconds," said Felix Faber, who founded Mindpeak with Dr. Tobias Lang in 2018.

Application used in ten laboratories worldwide

Mindpeak has developed the BreastIHC deep learning solution to detect, classify and quantify breast cancer cells. In May, the AI software received its CE-IVD mark. "The CE-IVD mark confirms that BreastIHC meets all essential health and safety requirements outlined in EU directives. That makes us the first company in Germany with such an approval in routine clinical diagnostics in pathology," Faber stressed. The algorithm is now used in ten laboratories including a laboratory chain in the United States, in Poland, and at the Institute of Hematopathology Hamburg.

Felix Faber, founder of Mindpeak

Precise cancer cell analysis improves chances of cure

AI's speed and accuracy can ultimately save lives. The earlier an accurate diagnosis occurs, the more targeted the treatment and the higher the chances of recovery amid increasingly diverse treatment options. Tumors are caused by changing cells. A precise determination of the cancer cell's condition means the most promising therapy can be selected. "BreastIHC recognizes cancer cells and analyses the characteristics of cancer cells," said Faber. Mindpeak has developed an exceptionally robust AI that produces consistently reliable results despite the varying image quality of tissue samples from different laboratories. Crucially, the AI does not have to be adapted to each laboratory.

Other application areas close to approval

Breast cancer is one of the most common cancers in Germany followed by lung cancer. Mindpeak now aims to use its algorithm in that field as well. "Our goal is to obtain approval for two additional products this year. An algorithm that identifies and distinguishes lung cancer cells quickly and reliably, and a solution for lymph node metastases," Faber said. "We are developing a platform that will let us adapt our product even faster to different types of cancer. In five years, we want to be able to help pathologists with 80 per cent of their analyses." This gains even more urgency given that the stagnating number of pathologists worldwide.

Financing international expansion

The Innovationsstarter Fonds Hamburg (IFH) awarded the start-up USD 3 million to develop BreastIHC with APEX Ventures, Nina Capital, Motu Ventures and prominent angel investors. Mindpeak is the first Hamburg-based company to receive combined EIC Accelerator funding from the EU. In August 2020, the start-up obtained an EU grant of EUR 1.4 million and European equity capital of EUR 5 million to expand to other European and U.S. countries.
ys/pb

Sources and further information

Similar articles

Growing number of applications for artificial intelligence

Security crucial as technology undergoes rapid developments - HAW Hamburg researching exciting AI projects

Promising fields of application for AI in Port of Hamburg

Research at Fraunhofer CML focusing on future of the intelligent port - algorithms crucial

Strong AI still a long way off – transparency can boost trust

Can artificial intelligence solve mankind's biggest problems?

How startup cofounder Lena Weirauch is conquering the AI sector

Part 3 of the series “Female AI – four questions for women in AI”. Today: Lena Weirauch, cofounder of the ai-omatic startup
The Consent Management Platform (https://app.usercentrics.eu/) we use could not be loaded. This can happen if AdBlockers incorrectly block this URL. Some features such as maps, proximity search or forms, cannot be used this way. To use these features, please deactivate your AdBlocker or allow access to *.usercentrics.eu.