In part 2 of our “neat start-ups” series, we look a simple application of AI and cloud-based technology that might transform how we assess and treat skin cancer
Shazam is almost so old some younger folk may not even recognise it, although it’s strictly in the domain of very successful early consumer apps.
It is a machine learning, AI-type app where you can point your mobile phone towards some music you love, but can’t recognise the artist and/or song title, and it will immediately tell you what you are listening to.
So would “Shazam for moles” be an app that you can point your mobile phone camera at and it will tell you what mole you are looking at?
Very nearly.
DermEngine is a workflow management and a AI-driven knowledge base of skin cancers for doctors, from a Canadian-founded, Australian-funded start-up called MetaOptima.
It’s foundation idea is simple.
Build an ever-increasing database of diagnosed skin cancers globally in one place with a highly accessible and integratable system, driven largely by mobile phone app and camera utility, to make the decisions around diagnosing skin cancers, particularly in a general practice setting, much faster, easier and cheaper.
The DermEngine database is cloud-based and quickly accessible on any system, mobile or desktop.
Using a dermatoscope mobile phone attachment, (either one manufactured by MetaOptima called MoleScope, or a plethora of others on the market) patients’ moles can be photographed, stored and compared with the world’s largest database of images to help make an informed decision.
When one doctor told local MetaOptima general manager, Peter Birch: “Oh, I get it, it’s Shazam for moles”, the phrase immediately stuck, for obvious reasons.
The system also integrates with major pathology systems, and with most of the patient-management systems used by doctors.
As a cloud-based system, DermEngine has been available world-wide for more than four years, but in the past 12 months it has established a quickly expanding local office to serve the Australian marketplace.
You quickly get the sense that it’s coming to a practice near you in the not too distant future. There are strong indications that some of the bigger corporates might be picking up on the technology and rolling it out en masse to many of their practices as one of the many newly emerging upstream services that they will be able to offer to add to their base revenues.
And it’s already being picked up by some of the major skin clinics, as they migrate away from their more expensive older technology. Some of the more advanced older systems can cost from $150,000 to $300,000, whereas the accuracy provided by DermEngine is now available to most practices for a fraction of this cost.
Birch emphasises in his pitch that DermEngine is not for diagnosis, it is for clinical decision support.
“It’s a sort of fast-evolving online encyclopaedia of skins cancers at your finger tips, or at the end of your mobile,” he told the Wild Health Summit, which was held in Sydney recently.
There are a couple of other reasons to think this business might be going places.
If you think about its centralised database and AI model – MetaOptima’s IP resides in a few complex algorthms that are constantly being tweaked by the company’s scientists – then you have a classic case of a “network effect” type business. That is, the more data that gets sent to the middle, the smarter the whole ecosystem becomes, and the more valuable it becomes to every user, until it’s the biggest system and the only one you can really use.
The business is backed in Australia by one of our smartest local VCs, Airtree Ventures, the principal of which is Daniel Petre, formerly of Microsoft fame. Airtree Ventures other investments include the highly rated healthcare appointment engine business Hotdoc, and one of only two Australia software unicorns by valuation, the Adobe Design-busting, self-serve cloud-design business, Canva.
And if you need any more intrigue to prompt you to check out this business, DermEngine is promising early next year the ability to generate a total-body 3D map within two minutes, using a drone driven by the power of AI and the push of a button.
Can’t wait for that one.
