“We make a big effort to really understand our customer’s project and their precise needs.”
“When it comes to AI, first and foremost, we focus on the robustness of our AI models… Our products also need to work at scale, they need to be intuitive to use for a pathologist and must not slow him or her down in any way. They have to make life easier or in some way improve productivity or diagnostic accuracy.”
Conversation with Steven Hashagen
CEO at Indica Labs, Albuquerque, New Mexico, United States
BIOSKETCH: Since founding Indica Labs in 2011 with only a small amount of seed funding, Steven Hashagen has become one of the most successful entrepreneurs in our industry. He now sits at the helm of a highly successful organization which specializes in image management, image analysis and artificial intelligence for the field of tissue pathology. Today, Indica Labs operates globally and employs over 110 people in a dozen countries around the world. The business continues to grow and to innovate with a focus on creating the best possible tools for their customers’ needs, a heavy R&D spend and the continual recruitment of best-in-class engineers and scientists.
Interview by Jonathon Tunstall – 25 Apr 2023
Published – 13 Jun 2022
JT – Welcome to Pathology News. My name is Jonathan Tunstall and today I’m interviewing Steven Hashagen, CEO of Indica Labs. Indica Labs is a well-known company in our industry which specializes in image management, image analysis and artificial intelligence applications for tissue pathology.
SH – Hi Jonathon. Good to speak with you again.
JT – Now Steven, you and I obviously have some history together because we both worked at Aperio at the same time, around 2008 to 2011. After that, we went our separate ways, and you went off to found Indica Labs. Now we’ll talk about that decision in a moment, but perhaps you could start by telling us something about your early career and the events that took you to that point.
SH – Well, my background was in computer science. I studied at the University of New Mexico and then San Diego State University. Aperio Technologies was probably my second real job out of college, that is, my second job in the field of computer science. Aperio was later acquired by Leica Biosystems, which most people know today.
At Aperio, I learned a lot about digital pathology and a lot about software development, but it was the early days (I think I was around employee number eight), where I also learned a lot about startup culture, team building and leadership.
When I left Aperio, I think the company was close to 200 employees, so obviously a lot changed in those years. I would say there were some good changes and some not so good ones, but overall, it was an incredible learning experience and a very positive overall experience.
JT – So there you are, with all this new experience, in the early days of image analysis in digital pathology, in a rapidly growing organisation and in a rapidly growing field. Then in 2011 you made this very bold decision to start your own business. I wonder if there was a catalyst behind that decision. Was it perhaps the right time for you personally, or did you perhaps see a gap in the market for a higher quality image analysis product?
SH – Well maybe it was kind of a perfect storm of both of those things. I was young enough to take some risk, and that’s an important thing. When you’re older, it’s perhaps harder to bounce back financially from your dumb mistakes, but when you’re young enough, you can take on more risk. So, to start Indica, I cleaned out my 401K and that was the basis for starting the company. That, plus some money that we raised from a few seed investors.
I probably couldn’t have taken that risk much later on in life, especially as my family grew and I had more responsibility. I was also perhaps mature enough at that time to have had enough experience to maybe not make too many foolish decisions as well.
So, looking back, it was the perfect time for me personally, but also a time when, from a business perspective, there was a clear opportunity in the market. There were a lot of customers in the field who had been asking for certain image analysis modules that didn’t exist yet and Aperio was not interested in building them. In fact, Aperio was moving in the other direction where they were looking to sunset their image analysis product line and send all those customers off to a company called Definiens.
So, looking back, it was a perfect storm of personal factors and business factors that allowed me to make this decision and to achieve some early successes.
JT – Well I remember that at this period of time, many companies were moving their focus toward the clinical side of the market, and yet you established yourself very much in the research, biopharma and drug discovery sector. Was that a strategic decision at the time?
SH – Yeah, it was actually and that’s kind of a funny one, because I made a distinct promise to myself when I formed Indica Labs that we’d be a research-focused company and our products would be intended for research use only. I also promised myself that we’d never build a clinical or a diagnostic product. Now, fast forward 12 years, and I’m breaking my own promises. Indica is now heavily invested in our diagnostic product line.
So, you are going to ask me ‘why the change’, and I think that it’s just a different time. Back in those early days, scanners were not as fast as they are today, and the image quality was not as good. AI or deep learning didn’t even exist, at least in the way we know it now. Also, clinical labs had a really difficult time justifying an investment in digital pathology. Now, that’s all changing. We now see that clinicians are ready for digital pathology and many of them are very far along the adoption path.
So, we’ve broken our own promise, but we are living in different times and any successful company needs to evolve both its strategy and its product line.
JT – So, tell me how you have managed that evolution at Indica. I am thinking specifically that you now have the HALO Link product which is used for image management and somewhat of a departure from your main focus on image analysis.
SH – Well, we’ve been very disciplined about adding new products and new services in making sure that they have really strong synergy with our other existing products and services. I consider that each new product needs to have a kind of flywheel effect, it needs to add to the momentum as a whole, that is, the growth of the company and the adoption of our platform as a whole.
We were already creating tools for the image analyst who’s sitting at the workbench, head down, doing image analysis all the time, but those results need to be shared and they need to be reviewed by pathologists or by other researchers. HALO Link was a natural progression, and it was an easy sell to our existing HALO® customers, because it added a nice benefit to the existing platform. So, that’s how I view it, each new product and service complements the others and adds to the overall momentum of the entire platform.
JT – Well, I guess HALO Link is also an important product in the context of scanner integration, because I believe you are scanner agnostic, is that right?
SH – That’s correct. Our entire platform supports all of the major open scanner formats as well as the proprietary ones.
JT – I think any software player today that isn’t providing their own instrument, their own hardware, simply has to be software agnostic, instrument agnostic in fact.
SH – Yes, and we do go out of our way to try to encourage other vendors and our partners down the road of open file formats, open APIs etc. Not all have played nicely on that front, but I think it’s moving in the right direction.
JT – Well, we’ve talked about products, but obviously over the years, the company has changed a lot as well. Tell me something about what Indica looks like today. I’m thinking in terms of structure and personnel and the territories where you operate.
SH – Roughly speaking, we’re about 40% engineering, 40% scientific staff and 20% services and support, and you can probably glean from those figures that we’re investing very heavily in research and development. Our engineering team is primarily based at our headquarters in New Mexico and then our field applications people, that is the scientific team, are global. They’re out in the field supporting our customers, providing training etc. Their job is to help our customers complete their projects and achieve success through the use of our products.
I believe that one of the unique things about the way we have built our culture at Indica Labs, is that we don’t offer sales commissions. There are no commissioned sales reps at Indica. Instead, we offer a generous profit-sharing plan that everyone participates in. I participate in exactly the same plan as the rest of the staff and that plan includes every single person, whatever their position in the company. I believe this profit-sharing scheme has uniquely shaped Indica’s culture and has helped the entire team to align and focus on the same mission.
We also believe in creating long lasting, fruitful partnerships with our customers. Our interactions with the customer should never be about closing a deal or making a number for the next quarter. We focus on trying to do the right thing for the customer, helping them to find the right product and providing advice which can help that customer to be successful.
JT – So to sum up what you have just told me, Indica Labs has developed a culture which focuses on the science and uses scientists to sell to other scientists. Is that correct?
SH – That’s correct. I think if we are confident that we’ve built the right products and that we’ve built them well, then our job is simply to make sure that the customer is aware of our products and understands what they can do for them. They should know how Indica’s tools can help them to achieve their own goals. Once we have achieved that, then, hopefully, the product sells itself. Of course, this is an oversimplification, but that’s the kind of high-level sales strategy that we operate under.
JT – Okay great. I wanted to circle back a little to explore a little further the change in strategy at Indica Labs around clinical products. I notice that Indica now has a fairly new product called HALO AP® that I believe is designed purely for use in clinical laboratories, is that correct?
SH – Yes, correct. HALO AP is designed for clinical use and by that, I mean a variety of clinical workflows and diagnostic image analysis applications. We’re excited about HALO AP as a major growth area for the company. It’s still a young product, but the early traction we’ve experienced has been really incredible. We have quite a few customers now, including major hospitals, reference labs, CROs and even big pharma companies who are doing clinical studies. Also, major healthcare networks such the British National Health Service.
The feedback from these groups has been overwhelmingly positive and we’ve even been told by some customers that the user experience for pathologists is the best they’ve seen in digital pathology software. We’re really proud of that. So many pathologists still prefer to work with a microscope, but that shouldn’t be the case. To me, that’s just a sign that the software hasn’t been designed well. So, I do really believe that HALO AP has taken some major steps towards changing that.
JT – Well, HALO AP is also a major strategic step for Indica Labs, so I take it that you personally believe we are now seeing the long-awaited expansion of adoption in the clinical market. A lot of careers have been damaged over that prediction in the last ten years. Are things really different this time?
SH – I think COVID has had a massive impact. The effect of the pandemic has been to bring about perhaps ten years of evolution for digital pathology condensed into one or two years. That is both from a technical standpoint and in terms of acceptance of the technology by the pathologist. It has also helped from a regulatory perspective as well, as regulators have had to make concessions to allow the use of digital pathology during the COVID emergency.
JT – And you are not alone in seeing this of course. The ‘Clinical AI’ market, (let’s call it that) is becoming extremely crowded, isn’t it? I was at a conference recently, and I think there were perhaps a dozen companies in the room, maybe more, offering image analysis applications for clinical and diagnostic use. In this crowded market, how does Indica Labs continue to stand out from the crowd?
SH – You’re absolutely right, this is a crowded space. Personally, I think that stems from the fact that there’s currently so much investment going into AI in pathology. Those investments inevitably lead to new companies, new startups, new products. Let’s face it, there is a lot of excitement and buzz around this space.
And as you say though, in this type of market it can be a real challenge to stand out. I believe it’s about really making sure that you know your customers and that you have real products, that is, products that work extremely well, and by themselves can stand out above the noise.
When it comes to AI, first and foremost, we focus on the robustness of our AI models and making sure we have high quality results coming out of those models. Our products also need to work at scale, they need to be intuitive to use for a pathologist and must not slow him or her down in any way. They have to make life easier, or in some way improve productivity or diagnostic accuracy. They also have to cleanly and seamlessly integrate with the other laboratory systems.
There are other factors to consider as well. We already discussed that at Indica Labs we use scientists to sell to scientists and we make a big effort to really understand our customer’s project and their precise needs. We also have a lot of references that we can call on from successful labs, who have integrated our products and then used them to successfully complete their projects.
This goes back to the point I made earlier about momentum, because a company starts to establish a track record of delivering products that work and of having good customer success stories. That always helps make the subsequent projects a little bit easier to pursue.
JT – So, where does Indica go from here? You stated earlier that successful companies continually remodel themselves and find new strategic avenues. How are you planning your next phase of growth for the company? What do you think are the new opportunities in the market?
SH – Well, obviously I don’t want to give away all of our strategic plans, but I can say that we see ourselves staying ahead of the market a little, but not too far ahead, because the digital pathology space is a very conservative one and there is something of a graveyard of companies out there that have tried to move too quickly by offering products that the market wasn’t ready for. We need to meet the customer where they are now, build products that they need today and help them take manageable steps towards the leading edge and the future state.
JT – Well, the pathology lab is going through a period of dramatic change at the moment, so the ‘leading edge’ that you describe, must also be moving forward quickly. I think we both understand that we are very much in a second phase of the pathology revolution now. The first part over a decade ago, was centred on the introduction of digital slide scanning and whole slide imaging, but we are now witnessing a whole new phase built around a plethora of image analysis and AI applications. Indica Labs is very much at the center of this new revolution, so let me ask you how you feel all this is ultimately going change the nature of the pathology lab and the role of the pathologist?
SH – Anatomic pathology has become far more complex than it used to be with all sorts of new technologies, new tests, new companion diagnostics etc. Pathologists are being asked to do a lot more and to be conversant with a load more techniques than they were 20 or 30 years ago. I think this leaves the pathologist really with no choice but to rely on some of the computational tools that are available today and the new tools that will be developed in the future. Otherwise, they just won’t be able to answer the questions presented to them or be able to keep up with the workload that’s being asked of them.
This sort of thing has already happened in the research space where we’ve seen an evolution in the types of research questions that are being investigated and the questions that pathologists are expected to answer. There are also some research techniques that can’t be done with the unaided human eye. Multiplexing or assessing the spatial relationships between cell populations, for example. These tasks are just too challenging and too tedious to do manually.
The same will be true in the diagnostic space, and this is already happening with some applications. I think it’s only a matter of time before most clinicians will have to use digital slides in conjunction with AI and image analysis to meet the demands of the job.
JT – Well, we could also postulate the existence of AI-driven biomarkers, or at least tissue patterns indicating disease which are undetectable to the human eye and can only be assessed by an algorithm. If AI biomarkers at some point became a component of the regulatory environment that would pretty much force the adoption of digital imaging and simultaneously cause a heavy reliance on the use of the AI.
SH – Yes, for diagnostics and maybe also for quality control. There are all sorts of areas where AI can improve patient outcomes, laboratory efficiency, quality of results etc.
JT – Do you think there is the potential here for the pathologist to be pushed to the side lines and find himself or herself almost working for the algorithms?
SH – No, I think it ultimately becomes a synergy between human and algorithm.
The algorithms are tools which aid efficiency and accuracy. They can help pathologists to focus on the right cases, to look at the right slides in the case, or the right regions within the slide. AI may even help them sleep better at night if they’re worried about having missed something. Patients too may feel better knowing that they had a virtual second opinion QC their diagnosis. Who knows, someday it may even be mandatory.
JT – Well, that leads me to a final question Steven, because the pathologist’s brain is doing much more than analysing and diagnosing. It is also identifying the tissue type and then using a great deal of previous analytical experience together with an acute understanding of both disease progression and likely outcomes to reach a judgement. At the moment, we see the development of supervised AI, a series of point solutions designed for specific diseases, and I wonder how you view the notion of unsupervised AI. Do we reach a point in the future where the algorithm also determines the tissue type?
SH – Yeah, we’re exploring these types of applications already in the field of toxicological pathology, where the goal is to design models that can detect problems in certain organ types that haven’t necessarily been defined in a training model. Ideally the model would be able to identify ‘abnormal’ specimens even though it’s never been trained on that specific pathological representation.
Technically, it’s a challenging problem, but it would be hugely applicable, not just in toxicological pathology but also in clinical diagnostics.
JT – Well thank you Steven and thank you for your time today. Do you have any further comments you’d like to make?
SH – Well, I’d just like to say thank you and Pathology News for the service that you provide. There’s a lot of information out there and to be able to sum it up and boil it down as well as you guys have is really, really beneficial, not just to businesses, but to employees and to end users and also for investors who are looking at this space. So, thank you and keep up the great work!
JT – Thank you, Steven, and good luck to Indica Labs!