Simon Mugisha is developing a software that is making early detection for breast cancer more accessible and affordable for all. PHOTO/DEOGRATIUS WAMALA
In a world where many young minds are content with following the status quo, Mr Simon Mugisha stands out as a disruptor.
He is not just dreaming about change; he is putting his energy into creating it.
With an eye for innovation and a passion for problem-solving, Mr Mugisha is taking on a challenge that most would shy away from—revolutionising early detection for breast cancer.
Breast cancer in Uganda is a silent and deadly threat. Data from Uganda’s Health ministry reveals that up to 89 percent of Ugandan women are diagnosed at stage III or IV, where treatment is harder and survival rates are low.
The country’s age-standardised incidence and mortality rates sit at 21.3 and 10.3 per 100,000 people, respectively.
Behind these numbers are not just statistics but real lives and families facing unimaginable challenges.
What makes breast cancer so insidious is that early detection can significantly improve outcomes, yet there are formidable barriers.
A dysfunctional referral system, limited knowledge among both patients and service providers, and diagnostic delays compound the problem.
Even mammography, an effective tool for early detection, is rare in Uganda, with only two units available, priced out of reach for many.
Mr Mugisha, a fresh graduate from Ernest Cook University in Australia, is developing a software that is making early detection for breast cancer more accessible and affordable for all.
By leveraging his background in biomechanics, he is exploring creative, practical solutions to this. It’s ambitious, yes.
But with a world that needs change, why settle for anything less?
Using Artificial Intelligence (AI), Mr Mugisha is developing a model that analyses patient data and alerts doctors to early warning signs of breast cancer.
This isn’t a high-tech machine you’d find in every clinic; it’s a software solution powered by datasets.
“People fear going to hospitals for cancer tests or worry about contracting it there,” says Mr Mugisha. “I’ve been to the [Uganda] Cancer Institute for meetings or to pick up items and just looking at the conditions there makes me scared.”
Quicker turnaround
His realisation that Uganda hasn’t fully leveraged AI to help these cancer patients fuelled his mission.
“We don’t have the standard technology, and the machines we do have are often outdated.
Even programming languages like Python, which we’re just starting to learn, may not produce exactly what we want but can still help build a functional model,” he says.
The software he is building works by taking input from doctors when a patient reports symptoms like swollen nipples or uneven breast size or any other.
The doctor enters this data into the model, which compares it against its vast dataset to calculate a probability.
For instance, it might assess that a patient has an 85 percent chance of having cancer or only 20 percent, based on patterns it has learned. And that’s where the medication starts.
This is a game changer from the current testing methods.
When you use mammography, for example, finding a small breast cancer with a screening mammogram can take one to three weeks.
This includes scheduling the mammogram, having it performed, and getting the results.
When you use biopsy, a breast biopsy sample is sent to a lab for analysis, and the results are usually available within one to 10 days.
However, it can take longer depending on the complexity of the case and the tissue sample.
And after a cancer diagnosis, additional tests may be needed to determine the stage of the cancer.
These tests can include a CT scan, MRI scan, PET scan, or bone scan.
I ask Mr Mugisha why he picked breast cancer over other diseases, he chuckles and says: “Adding more diseases to an AI dataset isn’t as simple as throwing in extra toppings on a pizza.”
He adds: “If I were to expand the model to include diseases like malaria, I’d need to code an entirely new section, which is complicated. Even so, sourcing reliable medical data is difficult. I’ve been asked questions like, ‘Are you an expert in data collection?’”
Despite the challenges, he remains undeterred. “It’s tough, but I won't stop,” he says.
The data problem
Currently, he’s training his model using data he collects online, but even that has its barriers.
Many datasets are behind paywalls or restricted by copyright, and there are few research papers on breast cancer-specific to Uganda.
He’s turned to non-profit organisations for data, but even their resources are now becoming harder to access.
“Collecting datasets is costly, akin to extracting 'new oil' in today’s data-driven world. Go to hospitals or the Uganda Cancer Institute, and you’ll find some doctors won’t even let you touch patient information. But we need that data, accurate data,” he says.
He is now figuring out how to collect good data that can be fed into his AI model because models like that are as good as the data they’re trained on.
“And as time passes, it needs fresh data to update itself. It learns, improves, gets better—that’s how it works,” he adds.
It turns out Mr Mugisha isn’t the only one wrestling with this data dilemma.
Nesta Paul Katende of Ortic Foundation and Ernest Mwebaze, the chief executive officer and founder of Sunbird AI, shared similar stories.
In October, he told me thus: “Building the infrastructure we need for local AI systems requires good policies, funding, skilled personnel, and facilities that Uganda just doesn’t have at the moment.”
He added, apocalyptically: “If we don’t have local data centres, we risk creating products that don’t meet Ugandans’ needs.”
No National Data Centre
Currently, Uganda lacks a national data centre, a gap the government is attempting to fill with a ¥1b (Shs500.3b) deal with the Export-Import Bank of China.
Until then, innovators like Mr Mugisha are left scrambling for data housed in far-off places like California, US or foot their bill to collect data home.
Mr Mwebaze has been in the game for five years, long before ChatGPT turned AI into everyone’s favourite dinner party topic.
Sunbird AI aims to leverage data technology for social good.
Its projects include collecting noise level data in Kampala for city planning and creating Uganda’s only free text-to-speech model for Luganda, thanks to crowd-sourced data from Common Voice.
But there’s a hitch, Mr Mwebaze says: “Without a clear AI policy, there’s uncertainty for entrepreneurs, and innovation can stagnate.”
It’s also challenging when you think about it.
“There’s always this trade-off between heavy regulations and letting innovation thrive,” he explained at Uganda’s Deep Tech Summit in early October.
“We need to figure out where the boundaries are, especially when it comes to data privacy and the ethics of AI,” he says.
Mr Mugisha couldn’t agree more, saying: “Once policymakers get this sorted, the sky’s the limit for Uganda’s innovators.”
He added: “We need more innovators in fields like [AI]. It shouldn’t be this hard. We’re just trying to change the world, but there are no designated places to support us. Securing government support can be difficult without influential connections.”
“But think about it,” Mugisha continues, leaning forward with conviction, “If we didn’t have innovators, would we have these cars we see everywhere? No. It’s because someone sat down and thought of something that didn’t exist. Now we’re getting into electric cars, robotics... all that jazz. And how do you do that? You need deep technology.”
Building capacity
I ask him what it would take to have more people like him leading this tech revolution.
Education. “Schools in Uganda? We’re light-years away. There’s a massive gap. What we really need is serious investment in education—labs full of robots, coding classes, decent Internet access. Not just dusty benches for mixing chemicals that won’t save the world. At the university level, we’re still copying code off the Internet for our Arduino projects. By the time you’re balancing bills, coding, and editing, it’s too late to cram everything you need,” he says.
The key, he says, is collaboration—universities, the government, and the private sector joining forces. But how do we make it happen?
“Look at Silicon Valley,” he says. “Great minds, great exposure, constant idea swapping. We need more partnerships here, more ways to cross-pollinate knowledge.”
I chuckle. “Two heads are better than one, right?”
“Yes,” he nods, smiling.
“But in Uganda, innovation hubs and accelerators are like borrowing someone else's brain. We get the foreign tech, but we don’t make it our own. Instead of just selling us a finished product, why not show us how it’s done? Let us learn, adapt, and then innovate.”
