In the AI Startup of the Week, the editorial staff of ai.nl is featuring promising AI startups, their innovations, solutions and challenges. In this fourteenth episode, we are taking a look at WSK Medical, a Dutch AI startup based in Amsterdam building AI solutions for early cancer detection.
Artificial Intelligence‘s biggest impact has been felt in the field of diagnostics, pharmaceutical, and healthcare industry. The field has seen an influx of new medtech startups that are completely changing the face of the industry. Most of these startups have focussed their efforts on early diagnosis and Amsterdam-based WSK Medical wants to do the same for early detection of cancer.
According to Crunchbase, WSK Medical has raised a total of €750,000 so far and is trying to bridge the gap between the medical field and AI. It aims to do that by developing easy-to-use, workflow integrated, clinically valuable AI systems for early cancer detection.
AI-supported clinical systems
One of the important things to note is the fact that AI is increasingly becoming sophisticated enough to do the things that humans are good at. AI has also been found to do the work of humans quickly, more efficiently, and often at lower cost. The healthcare industry is one where AI and robotics are playing an outsized role.
From training, research, and end of life care to diagnosis, decision making, and early detection, AI and robotics have become part of our healthcare ecosystem. WSK Medical wants to play a key role in the early detection by implementing AI to detect cancer in their early stages.
It does this by building solutions that can assist doctors and clinicians in their daily work. “We believe in assisting clinical specialists with early cancer detection using AI solutions,” the Dutch startup says on its website.
With a vision to use data to create value for assisting in cancer prevention, treatment, and care, WSK Medical is a team of people that puts compliance, standards at the front and counts the likes of The Netherlands Cancer Institute, Slide Score, Founda Health, and others as partners.
WSK Medical: Team
- Derick Montaque (CEO, CTO & Co-founder)
- Marius Wellenstein (COO & Co-founder)
- Jonathan Woodburn (Software Developer and Data Scientist)
- Tubay Yüceyalçın (Global Sales and Business Development)
WSK Medical and Zeno Pathology
The product that WSK Medical makes for early diagnosis of cancer is called Zeno Pathology. It is an automated AI analysis pipeline for digital pathology and it performs slide image analysis as a background task. Where it stands out is its ability to be integrated into existing workflow for a seamless integration to all relevant clinical or administration systems.
The company says the Zeno Pathology can also be configured with following modules:
- Nottingham Grading/Mitosis counting (H&E)
- Nuclei detection/lymph node metastasis & heatmap of tumour probability (H&E)
- TIL detection (H&E)
- ER/PR detection (IHC)
- Her2 detection (IHC)
- Ki-67 detection (IHC)
With Zeno Pathology, WSK Medical says it is essentially solving the problem of increased workload that leads to decrease in available manpower. It is also trying to reduce the error and eliminate observer variability in the analysis of cancerous tissue.
Zeno Pathology starts by analysing from your existing slide viewer or workflow. It then extracts the desired slide or slides automatically from the database. Once the desired slide is extracted, Zeno Pathology performs an automated whole slide image analysis. The results are then produced and downloaded to the customer’s desired format or location or system.
Medical software powered by AI and ML
WSK Medical is an innovative software company that is looking to bring digital transformation to the healthcare industry using AI and machine learning. This Dutch startup does this using knowledge acquisition, reasoning, problem solving, perception, learning, planning, and robotics.
The core of WSK Medical’s AI powered technology seems to work on the same fundamentals as a traditional AI, which is also sometimes referred to as knowledge engineering.
The traditional AI basically deals with the concept of machines being able to act and react like humans only if they have abundant information related to the world. WSK Medical thus makes sure that its software and AI systems have abundant information including “access to objects, categories, properties, and relations between all of them to implement knowledge engineering.”
Once this AI part is built, WSK Medical says it uses machine learning to deal with programming knowledge, common sense, and reasoning, which have proven to be difficult and tedious for a software or a machine to learn and implement.
The ingenuity of WSK Medical can be seen from the way it implements machine learning. It does start with supervised learning, of course, which leads to a form of learning called classification. This helps WSK Medical to determine the category of an object, a key component in its effort to build a program that is able to facilitate early diagnosis of cancer.
WSK Medical says it uses the aforementioned AI algorithms and supervised machine learning paired with high performance computing to deliver tailored AI-based solutions to its healthcare clients.
How does WSK Medical work with clients?
As mentioned above, WSK Medical’s key product is a tailored AI-based solution for its clients. For its data modelling, mining and software development, WSK Medical relies on two end-to-end processes called Agile and CRISP-DM. Agile is used by the company for rapid software prototyping while CRISP-DM (CRoss-Industry Standard Process for Data Mining) allows it to deliver a deep learning model.
In order to deliver tailored AI-based solutions, it starts the process by understanding the problem, assessing the success criteria of its AI application and the clinical environment.
Once it sees the ability and possibility to implement its AI solution, it starts the process of data quality assessment. This process involves accessing the initial data and then assessing the structure and quality of the data. The goal here is to understand whether any data is missing to reach the result.
These steps are followed by data cleansing and preparation of the dataset. WSK Medical will then run the data mining tools and select a suitable model for supplying trained data to the learning algorithm. As you would have understood by now, the next step is evaluation to identify if the result meets the project objectives and finally, the model is deployed.
What to read next?
- 🤗Hugging Face wants to be GitHub of machine learning: a look at its history, key members, and major achievements
- 🕶️ Darktrace wants to be the torchbearer for AI-powered cyber defence: a look at its history, key members, and major achievements
- 🧠 DeepMind wants to build a general-purpose AI: a look at its history, key members, and major achievements like AlphaGo