DeePathology: A success story from Batch 0 until today
SagivTech, is a company specializing in artificial intelligence (AI), image processing, and computer vision, founded in 2009 by Chen Sagiv and Nizan Sagiv in Israel.
SagivTech initially worked with Roche Pharma on a project with the mission of creating image analysis solutions to aid pathology work, this collaboration proved to be quite successful for SagivTech as it opened the way to new cooperation opportunities within Roche.
After meeting Jochen Hurlebaus, Head of Digital Health Innovation at Roche, in Penzberg in the autumn of 2017, Chen Sagiv, Nizan Sagiv, and Jacob Gildenblat were invited to join Batch Zero with the initial idea of exploring the needs and pain points of the digital pathology field and how to potentially solve these challenges by using AI technology.
The setup was promising, SagivTech had a good opportunity to talk to different stakeholders inside and outside Roche, bringing different perspectives. At that time one of the outcomes of the pilot project was to get a clear understanding of the impact that AI could have on digital pathology, and on the other hand the huge gap within the existing tools.
Oliver Grimm, Senior Principal Scientist, Clinical Pharmacology Oncology, Pharma Research & Early Development joined as the Roche pilot owner and spotted some features that could distinguish SagivTech, being the training of an algorithm one of them.
“You need human expertise to train an algorithm and that is time-consuming, pathologists don’t want to spend many hours in front of the screen, so a solution that offers this feature making the algorithms’ decisions human-readable will be an advantage”.
Oliver Grimm
As a result, the early exploration gave the concept of creating a solution that could allow pathologists to create powerful AI algorithms. This solution should tackle a few challenges such as the data collection bottleneck, long cycles of data analytics, and algorithm training, in addition, it should be interactive and simple to use.
“I was happy to connect Chen with our German Marketing and Sales team to facilitate direct contact with our Tissue Diagnostics´ customers and collaborating pathologists. Thanks to my colleagues SagivTech could learn more about customer requirements and pain points and were able to shape and augment their offering.”
Christian Korz from the Startup Creasphere team, who was co-leading the pilot project.
At the time, Chen, Nizan, and Jacob didn’t know that this journey would have directly resulted in the creation of DeePathology®, a spin-off from the mother company SagivTech.
The journey still goes on
Following the participation in the Startup Creasphere’s innovation program and the creation of the new company DeePathology® and its main solution DeePathology® Studio, a Do It Yourself platform that allows pathologists and researchers to develop by themselves AI solutions for important problems in Pathology – on a laptop in their lab[1], the collaboration journey between Roche and DeePathology® continued. The company got involved with another division within Roche Pharma where they got a unique opportunity to see how to scale the product, the goal was for DeePathology® to be not just a local solution serving a single pathologist, but rather a solution that would enable collaborative work by multiple pathologists within the same project allowing scalable operations.
“We wanted to bring something that will allow pathologists to create powerful AI algorithms”.
Chen Sagiv, Co-Founder, and Co-CEO, DeePathology Ltd.
The first product idea was to transfer an AI algorithm from the local viewer of DeePathology® into a Roche pathology software solution to allow pathologists to develop their own algorithms using this tandem solution. The solution was intended to provide flexible and specific image recognition solutions for customers and pathologists including automation of routine work, such as rapid annotation, tissue segmentation, or cell detection, and the generation of novel insights, e.g. finding new and distinguishing characteristics to classify certain tissues by using AI.
The great effort of the company to keep improving the technology of its STUDIO AI solutions has paid off, a few months ago they launched DeePathology® STUDIO WEB, this solution has a web interface, runs on the cloud, and allows collaborative work of several pathologists. The input they received from different Roche collaborators was essential to define the need again and to understand how the technology can facilitate faster and more efficient workflows.
Alex Kalogrides is leading the Oncology Research Decision Support Product Management team at Roche, his main role is to set the overall strategy for Roche pathology software solution and grow the overall usage of the product by building the highest priority capabilities that meet their users’ needs. That could mean prioritizing a new feature, building, or seeking an external partner (like DeePathology®) in order to integrate existing functionalities.
“A project that would have taken 10-15 pathologists in 3 or 4 weeks in a research project, was able to be done for 1 pathologist in 3 days thanks to DeePathology®”.
Alex Kalogrides
DeePathology®’s short-term focus will be to keep further scaling their operations and bring their tools as a very important part of large systems, allowing pathologists and researchers to use these sophisticated AI tools, provide explainability, and very clear and easy reporting.
References:
[1] Definition of The DeePathology® Studio. Retrieved November 25, 2022, from https://deepathology.ai/