While breast cancer treatment can be highly effective, women across the globe face dramatically different outcomes depending on where they live.
According to research compiled by World Health OrganizationSurvival rates for at least five years after diagnosis range from more than 90% in high-income countries to just 66% in India and 40% in South Africa.
Geetha Manjunath, founder and CEO of Bengaluru, India-based Niramai Health Analytix, has set out to improve access to screening when a close family member dies of cancer breast in their 40s not long after receiving the diagnosis. The company recently participated in Impact M2D2 accelerators at the University of Massachusetts Lowell and received FDA 510(k) certification this early year.
Manjunath sat down with MobiHealthNews discusses how Niramai’s AI-powered screening system works, the importance of explainability when using AI in healthcare, and what’s next for the company.
MobiHealthNews: Can you tell me a little bit about how the Thermalytix system works in breast cancer screening?
Geetha Manjunath: I’ll set up a bit of context. If you look at the mortality rates in different countries, there is a big variation on breast cancer survivors. To prevent these deaths, we need regular screening, but that’s not feasible today. One, because of economic constraints. Such a large initiative is often restricted to women about 45 years old and older, because there is a relationship with age. In addition, mammography, which is the standard for detecting breast cancer, does not work well for young women under 45 years of age, because what they have called dense breasts. Actually, in almost 50% of women over 40 years old There’s the problem of density.
In countries like India, China, Philippines, affordability of the machine itself is a big problem for the government as well as small diagnostic centers or private hospitals. So with all this in mind, what Niramai has developed is an affordable, accessible method of detecting breast cancer in women of all ages and of all breast densities. In addition, the machine is really comfortable in the hand. You can get the test done in the hospital. You can also take it out to take the test in remote areas, rural villages as well as corporate offices. We also offer home breast cancer screening.
The woman entered a small room, like a small pavilion. She goes in, she closes the door and then she takes off her clothes in front of this device. With no one inside, it was like a dressing room. For example, no one saw her or touched her during the check-up, which is not the same as the mammogram experience.
It uses an imaging technique known as thermal imaging, can be controversial. Traditionally, thermal imaging was used for anomaly detection. However, it has never been accurate enough to be used or recommended in a hospital, because we are measuring, say, 400,000 temperature points per person. It is very difficult for the human eye to distinguish between different shades of yellow, different shades of orange, etc.
We’ve developed our AI-powered smart software that analyzes this temperature distribution over the chest area and translates it into a cancer report. That’s completely automated with grading indicating the level of anomalies. That is our main value proposition, AI algorithms to convert temperature distribution into a cancer report.
MHN: So the cancer report doesn’t say, you 100% have breast cancer. The idea that it highlights potential concerns and you get extra tests?
Manjunath: Sure. It’s a screening test, which means that out of every 100 women screened, we identify 9 or 10 women who need a follow-up diagnostic test – possibly another mammogram, or 3D mammogram, or more complex breast MRI, or ultrasound breast.
MHN: AI is becoming a lot more popular in healthcare, especially for Picture. How do you balance worries about introducing bias or not understanding how the AI is making its recommendations?
Manjunath: AI is a machine, and a machine works the way you train it. So the training phase is very, very important. What kind of pattern do you use for training, make sure that the training team is dealing with many anomalies. For example, in breast cancer, we look at pregnant women, we look at people who are menstruating, we look at people with fibroids. All different categories and subcategories of potential abnormalities must be included. You definitely need to work with a medical professional to really make sure that your training is unbiased. It’s truly multidisciplinary, because domain experts and tech professionals have to come together.
And the part that can be explained is also extremely important. So, for example, in the beginning, we just said it would look at a patient and say, cancerous or non-cancerous. But the doctor said, “What do I do about this? I can’t take any action on this. You just said cancer, but which breast and what happened?” So now we have an auto-generated three-page PDF report, giving scores for left and right breasts. We automatically mark on the breast, saying this is where you want to check again.
MHN: You recently received an FDA 510(k) license in the US What are the next steps for the company?
Manjunath: We recently received the US FDA license, we have only just completed the registration of the device, even though we launched in beta mode last month. So I was looking for a partner. To start, we will work with temperature analysts who already use thermal imaging. Our current license from the FDA is to use this method as an adjunct to mammograms, so we would love to work with these imaging centers to provide this facility.
At the same time, we are working on our next device, which is a bit more complex than our current device, to get FDA approval. We needed a multisite clinical study in the US, so we identified hospitals in New Jersey and Arizona, and possibly Florida.
Meanwhile, we’ve received a huge response from low and middle income countries because of its affordability and accessibility. So in countries like Philippines, UAE, India, Indonesia, we are working with distributors in the local domestic market to come up with solutions for the developing world. And we are also allowed to use in Europe.
So I’m very pleased. I was trying to solve a very local problem of trying to get Indian women diagnosed with cancer. We have now screened 60,000 women in India alone, which is a significant number, given this is a new medical device. We launched in Kenya. So I’m thrilled to have the opportunity to make a difference in the lives of women around the world.
