01/21/2022 Daniela Kuklova
Breast cancer affects 6,000 women in the Czech Republic annually. Breast cancer is the most prevalent form of cancer in women. And the numbers have been steadily rising. One way to early detect abnormalities and save their lives is prevention – full-scale and regular mammography screening. It has been proven that regular mammography scans can detect the majority of tumors at an early stage. It means that up to 90% of breast cancer patients could be cured. However, most women wait too long for a preventive examination, some of them up to 6 months.
The rapid development of machine learning, especially deep learning, is fueling radiology's interest in using the technology to improve the process of breast cancer screening. Therefore, the goal of introducing Artificial Intelligence (AI) into practice is that, in combination with humans, it might lead to a higher level of detection of early-stage malignant tumors and to a reduction in false negative or positive mammographic findings.
As a result, Trask – a leading Czech technological company, validated a feasibility of the product at the market and established an organisation in the medical device industry and named the product - Pledio. It is a fully integrated AI platform with its first solution that will identify and localize risk factors for early detection of breast cancer.
During the pandemic crisis, we were wondering at Trask how we could help the healthcare system. First, we were thinking to utilize our artificial intelligence competencies to support the COVID team at Bulovka University Hospital. However, after the discussion with leading doctors, the deployment of AI was more suitable and sustainable for the mammography screening.
Following the European juridical system, mammography images are independently assessed by two radiologists working under certain pressure and increasingly with limited capacity due to the overall stagnation of number of radiologists with respect to the number of performed mammography screenings. This sparked us to work together with the Bulovka University Hospital on a project that could speed up, refine, and therefore scale to find the “red flags”. The solution combined Artificial Intelligence, optimization of processes, and integration into the existing hospital IT environment and doctors’ workflow taking into consideration the sensitivity of processed data.
We developed an AI solution for mammography that has been trained on millions of mammography images and validated on the local data. The solution version 1.0 is fundamentally analyzing input mammography screening and predict while flagging risk areas to be further examined by the radiologist. It is fully embedded solution in the hospital systems (PACS and visualization terminals) that does not affect the workflow of doctors. Furthermore, it shall lead to a fully automated product that can be used as an AI-driven diagnostic tool to decrease the pressure implied for radiologists assessing the screenings.
Our ambition is to create a medical AI lab that on top of the breast cancer diagnosis aggregates AI models for other diseases in the form of a platform. Therefore, Pledio offers an opportunity for AI scientists from with the academic background or feasible commercial AI solutions to roll-out their models directly to hospitals and make significant impact on the healthcare sector. On the other hand, thanks to Trask’s proven ability of large-scale integration projects, it offers quick and easy ability to extend the AI-supported solutions in medical imaging into hospitals. The entire team of Pledio aims to build a collaborative ecosystem in healthcare while making the AI-driven science accessible and affordable to its users.
+420 722 985 264