February 22, 2017

Combinostics uses AI to improve early diagnosis

Combinostics can diagnose why the brain isn’t functioning properly. In future, the system could be expanded to cover other neurological diseases as well as cardiology and oncology.
Combinostics can diagnose why the brain isn’t functioning properly. In future, the system could be expanded to cover other neurological diseases as well as cardiology and oncology.
istock.com/triloks

New Finnish technology is tapping into machine learning and artificial intelligence to support healthcare professionals in early diagnosis of neurological disorders.

When a person is suspected of having a memory disorder, they undergo extensive tests ranging from brain scans to neurological examinations. It is then up to their physician to analyse the results. However, even an experienced doctor may struggle to arrive at an accurate diagnosis during the early stages of a disease.

Finnish healthtech company Combinostics has a solution for this in mind: its cloud-based tools for the comprehensive analysis of patient data.

“Today in healthcare, physicians are being overwhelmed with more and more data,” says Lennart Thurfjell, CEO of Combinostics. “Imaging data gets larger and larger in volume and there is new information all the time. We help to turn this data into insights.”

“Our system visualises all the data in a compact way so the physician can understand why the patient’s data is pointing to a certain diagnosis,” explains Lennart Thurfjell.

“Our system visualises all the data in a compact way so the physician can understand why the patient’s data is pointing to a certain diagnosis,” explains Lennart Thurfjell.

Combinostics

The company’s approach is divided into two parts. First, its brain image assessment tool automatically extracts various biomarkers (such as changes in brain atrophy) from brain scans. Then this information is combined with all other available clinical data to create a data profile of the patient, which is compared to a database of previously diagnosed patients. The result? A detailed analysis ascertaining which disease is most likely and how far it has progressed.

Thurfjell emphasises that Combinostics is not trying to replace doctors, but rather offer them intelligent tools to support accurate early diagnosis. All the patient data is available in a visual and compact way for healthcare professionals to examine.

“If a physician suspects the patient has a certain disease and then sees the relevant data points in this direction, they will feel more confident to do the diagnosis at an early stage,” he explains.

Brains of the idea

Combinostics’ technology is built on years of research at the VTT Technical Research Centre of Finland studying AI and machine learning for diagnostics. The company spun off from VTT in 2014 and its first product was launched commercially in December two years later.

Combinostics’ brain image assessment tool is currently used by a handful of university hospitals and radiology service providers in Finland and Sweden, and it will be soon complemented with a clinical decision support tool to provide full patient data analysis. This is where the technology really stands out from the competition.

“The clinical decision support part is a very hot topic,” Thurfjell says. “A lot of people are talking about it, but there are no data-driven clinical decision support tools in neurology out on the market yet. We will launch our clinical decision support product in Europe in the next two months [to be the first].”

Market comes to mind

While Combinostics is currently focused on diagnosing Alzheimer’s disease and other forms of dementia, expansion to other neurological areas could be on the cards in 2018. The Tampere-based company is already laying the groundwork for this in an EU-wide research project studying the prediction of neurodegenerative diseases.

But first Combinostics is looking to expand beyond Finland, Sweden and the UK, where it already has contracts. The company plans to launch in the US during the second half of 2017, subject to regulatory approval. Thurfjell is also eyeing Asian markets, but is waiting for the right local partner.

The healthcare market is moving fast and Combinostics plans to maintain its head start.

“Data-driven diagnostics, big data, machine learning. Everyone realises these technologies have huge potential in healthcare” Thurfjell says. “A lot will happen within the next few years.”

Combinostics’ image analysis tool ensures even the subtlest of changes in the brain will be detected.

Combinostics’ image analysis tool ensures even the subtlest of changes in the brain will be detected.

Combinostics

Text: Eeva Haaramo

Share: