What we do
Computer Vision
The product roadmap is defined in terms of computer vision specifics.
Computer vision is a subset of machine learning which derives information from (any) images. The challenges that PRMSE defines for itself are to be addressed via computer vision (rather than for example via the science of radiology)
Algorithms
PRMSE develops algorithms by defining target groups for CT image analysis, acquiring relevant data sets, training and testing algorithms, conducting clinical evaluations, and implementing them.
Combined Knowledge
We prioritize patient safety and health, uniting experts from radiology, computer science, regulatory and health economics. By combining all available knowledge, we create a product that ensures patient wellbeing.
Computer Science
PRMSE builds algorithms. We need to be able to explain how they function.
We ask ourselves challenging questions, such as how to build image interpretation tools that function over an ever increasing spectrum of images. From all sorts of population segments. In different countries, and thus genetic differences play a role. The images our algorithms judge are generated on different radiological scanners, built in different time periods.
Our scientists, and our scientific partners are dealing with these and other questions.
Marketing
PRMSE starts with the market.
We learned that lung cancer screening can only happen on a large scale so as to catch early lung cancers at an early stage. When over 90% of the cancers found are well-curable.
We learned that radiologists are overburdened, and report burn-out. Tools are thus needed to take workload off their backs, in a responsible fashion.
PRMSE listens to the market, uncovers the unmet market needs, and translates this into products.
The Company documents the precise needs. And these drive the solutions that our engineers and our technical partners need to go and build.
Our promise to the markets, to the consumers, to the medical staffs, is that the PRMSE: tools are robust. And have been scientifically proven.
We are aware of the fact that autonomous interpretation is (initially) met with reservation. In our approach, in all we do, we deal with this expectation carefully.
Humans remain in the loop of the outcome provided by the PRMSE solutions.
Regulatory
Creating automated image analysis is a novelty in the world of radiology. AI has gained larger acceptance in other industries.
Interpreting images of human beings carries a much higher level of scrutiny. And rightfully so.
We realize that the ultimate approval of US, European and Asian regulatory bodies is a challenge. From day-2 onwards the Company works closely with these regulatory partners, to gain - in a cooperative fashion - a common understanding of the steps needed.