These two master’s students used quantum computing and AI to produce deeper, more detailed images to study.

Krishna Kalavala and Tejaswini Thirumalaraju weren’t even thinking about breast cancer research that day a year or so ago, but sitting in a quantum computing class, their scientific curiosity was challenged.

They were pursuing master’s degrees at Texas Tech University when Lin Zhou, vice president for information technology and professor of computer science, opened windows of scholarly possibility for his students.

“Initially, I didn’t have any idea about what quantum computing is or how it works in the real world,” said Kalavala, who is now a staff network support technician on campus. “We decided to work with Dr. Zhou, and he gave us the opportunity to work on any project.”

Zhou said he wants students solving meaningful problems through a multidisciplinary approach, and he offers them the freedom to select their own research topic.

“This was meaningful research,” he said. “Breast cancer is one of the most prevalent diseases for women, and this research could potentially represent the next frontier for improving patient care. This was also an opportunity to really leverage what quantum computing and what it can provide.”

Kalavala and his research partner, Thirumalaraju, broadly began to look at healthcare and the Quantum Cure initiative, which emphasizes quantum computing and artificial intelligence (AI) to accelerate medical discoveries. In turn, that led them to breast cancer and ultrasound imaging. 

“We were learning something completely new,” Kalavala said. “We wanted to see if we could use it in a meaningful problem-solving way. The curiosity was our motivation.”

Both have family or friends who have been affected by breast cancer, so they saw the human factor as well as the sobering and compelling statistics. Approximately 1 in 8 women (13%) in the U.S. will develop invasive breast cancer at some point in their life.

According to the World Health Organization, some 2.3 million cases of breast cancer are diagnosed worldwide with almost 300,000 new cases in the U.S. and approximately 19,000 in Texas.

The global economic cost, or value of lost well-being (VLW), due to breast cancer reached approximately $2.54 trillion, accounting for 1.65% of the global GDP. Within the United States alone, the National Cancer Institute (NCI) and CDC record direct breast cancer medical spending at $29.8 billion.

“We realized how significant the problem is globally,” said Thirumalaraju. “Breast cancer is one of the major health challenges worldwide. That’s why we were drawn to breast cancer.”

The two decided to try the Quantum Cure approach to see if they could develop a screening process that might be more accessible, more affordable and more accurate than existing technologies. 

For example, traditional mammography detects early tumors, but there are possibilities for radiation exposure and false positives. Likewise, current ultrasound imaging is radiation-free and relatively low-cost, but it requires expert interpretation and can be subject to errors with accuracy at just more than 80%.

“There is already a lot of excellent research using classical AI and deep learning for breast cancer image analysis,” Thirumalaraju said. “We did not want to ignore that work, so we started from classical approaches because we wanted to understand what they could already accomplish.”

In their research, they added a quantum component to the process, using a convolutional neural network that employs deep learning to analyze images and extract spatial features.

“A major transition is coming – AI on classic computing to AI on quantum computing to take advantage of the unprecedented computational power,” Zhou said “We must prepare our student for that now. The combination of quantum and AI together is our brand. It’s not just quantum, and it’s not just AI. It is both of them together. We even spell them as one word. We are in a place technology-wise where there is even more enablement to work on these types of sophisticated problems and continue to improve upon what was done before.”

Using the quantum component, they were able to encode ultrasound images in a new way that could then be processed further for additional data. It was a hybrid approach, using the strengths of classical deep learning to leverage visual data and then discovering what a quantum model could contribute to the process.

The researchers took all the breast ultrasound images and fed the data into a computer, which would learn the various images and classify them as to what type of breast cancer the image contained. Eventually, the idea is to develop a tool providing additional information to support clinicians making decisions.

Thirumalaraju stressed this research is in the earliest stage, although the results were encouraging. They consider the findings to be proof of concept and worthy of deeper analysis with broader data sets. For their project they used more than 1,500 publicly available images across three breast cancer categories (benign, malignant, normal).

“Texas Tech has been extremely important for this project,” Kalavala said. “Dr. Lin played a huge role because he was the person who introduced us to quantum computing. What started as a class project became much bigger because after we presented our work, Dr. Zhou saw the potential and encouraged us to continue developing it.”

Zhou said his students are challenged over the 15-week term to use different disciplines in helping them attack their material.

“You start with physics,” he said. “Then you talk about the device and how they make quantum computers. That’s engineering. They have to learn the quantum information and how to process the data. That’s linear algebra and mathematics. Then you get into programming, and that’s computer science. It’s not just one discipline making it work. They cover a broad spectrum in my class over 15 weeks.”

The students and Lin hope to see the project continue, looking for a possible collaboration with the Texas Tech University Health Sciences Center. He said one of the next steps would be to potentially look at real patient data. They also have sent a funding proposal to Microsoft for a $200,000 grant.

“We have started thinking beyond building a model and are looking at how this could eventually become clinically meaningful,” Kalavala said.

Other future possibilities include establishing a physical quantum computer at Texas Tech to support real-time quantum experimentation, reducing waiting times on cloud-based platforms and extending the current framework to include additional diseases using other imaging techniques or exploring commercialization opportunities.

“Seeing something that started as a course project grow into a large research effort has been really rewarding,” Thirumalaraju said. “I have also learned that research is very collaborative. We have the computer science and AI side, but then we needed help from healthcare to tell us if what we’re doing actually makes sense clinically.”

Kalavala said it has been an interesting journey from classroom theory to research reality, which has consistently stoked that scientific curiosity that helped get this project going in the first place.

“We don’t see this as just work or something like that,” he said. “We take it personally because it’s been a long way. This project went from zero to all the way up, and that keeps motivating us to work.