Fimlab has begun implementing AI-assisted diagnostics in the field of pathology. AI tools targeted at tissue samples support physicians’ work, improve diagnostic accuracy, and open new possibilities for cancer classification.  

Following a competitive tendering process last year, Fimlab selected several AI solution providers under a framework agreement. In the first phase, starting mid-September, AI platforms developed by Finnish company Aiforia and US-based Proscia will be deployed. These platforms use validated AI tools to analyse breast, prostate, and gastrointestinal tissue samples.  

AI supports specialist physicians’ work 

The AI tools have been carefully validated, and their use is closely monitored. AI analyses the digitised tissue sample before the pathologist’s assessment. The physician reviews the AI-generated analysis and makes the final diagnosis with its support.  

According to international studies, AI can accelerate analysis and reporting by 20–40 percent depending on the sample type. In particular, AI offers solutions for breast cancer classification that traditional microscopy cannot provide.  

The adoption of AI is part of Fimlab’s broader development path to address the growing number of samples and the shortage of specialists in pathology.  

– AI does not replace specialist physicians, but it supports their work and enables faster and more accurate diagnostics, says Teemu Tolonen, Chief Physician at Fimlab.  

Fimlab is a pioneer in digital pathology 

Fimlab is a forerunner in digital pathology in Finland. The digitisation of tissue samples began as early as 2021, and now AI takes the development to a new level. In the future, AI algorithms may also offer independent prediction models to support treatment decisions or assist in selecting genetic tests.  

In pathology, AI as a diagnostic aid: 

• helps address increasing sample volumes and specialist shortages  

• improves diagnostic quality and consistency 

• supports physicians’ work without replacing them  

• enables new types of diagnostics and, in the future, the development of prediction models.