Overcoming scar inspires student to develop tumour-detecting AI

A fourth year University of Namibia (Unam) student says the scar from his brain tumour surgery made him feel inferior to his classmates, inspiring him to develop an artificial intelligence (AI) program that detects brain tumours.

“I didn’t take school seriously until I got to Grade 8. People thought I was stupid because of my scar and I believed it,” he says.

Douglas Mashonganyika (23) was diagnosed with a brain tumour at the age of two after developing balance problems.

Tumours are cancerous or non-cancerous abnormal growths of tissue in the body.

Mashonganyika says the medical algorithm (MedAlgo) AI model was inspired by his mother seeking a second opinion, despite his first doctor blaming the balance issues on an infection.

MENTORS
Nalina Suresh, senior computing lecturer at Unam, says the model has the potential to revolutionise the healthcare sector by integrating technology and intelligence to accurately detect and classify brain tumours early.

Suresh, who was Mashonganyika’s supervisor and suggested the project, says the current magnetic resonance imaging (MRI) and computed tomography (CT) scan machines use radio waves to produce detailed images used to detect tumours inside the body.

These machines are operated manually and often result in delays and errors that are detrimental to a patient’s success rate.

“I firmly believe that I made the right decision in proposing this research direction for [Mashonganyika], as his dedication, creativity, critical thinking and his effort in developing this artefact have been outstanding,” she says.

He says although he does not remember the operation itself, the weight of his scar made him feel alienated from his classmates. He credits his motivation to take school seriously and develop an interest in subjects like mathematics to his parents’ reassurance that his scar was the mark of a survivor.

“When I started taking school seriously I realised I love problem-solving subjects like mathematics, and started teaching myself coding programs,” he says.

CHALLENGES

The computer science student says there is a lack of information on the over 120 types of brain tumours for the model.

He adds that he faced challenges with the server to host his model. Servers are powerful machine, that store information which can be accessed by other computers.

“The stronger the servers, the stronger the AI. The servers needed to host the MedAlgo AI model are expensive,” he says.

THE FUTURE

Mashonganyika says he wants the AI model to be accessible to all medical professionals to improve the detection of various types of cancers.

“I believe the model can help speed up the diagnosis of tumours, breast cancers and other types of cancers,” he says.

Suresh adds that a neurosurgeon from Lady Pohamba Private Hospital has invited Mashonganyika to present the pilot testing of his MedAlgo AI model.

“This is a significant milestone showcasing the growing interest and potential impact of his AI-driven solution in real-world clinical settings,” she says.