Interview with Gabriel Falcão

Colorectal cancer (CRC) accounts for approximately two million new cases diagnosed annually, according to the World Health Organization. His research proposes the use of quantum computing to enhance artificial intelligence models for the diagnosis and prevention of gastrointestinal cancer. 

By developing new quantum neural networks, quantum support vector machines, and quantum reservoir computing models, the project aims to accelerate image analysis and classification processes while improving diagnostic accuracy.

“We want colorectal cancer diagnosis to be faster, more accessible, and more sustainable. Quantum computing enables us to achieve that goal with less energy and greater intelligence,” explains Gabriel Falcão.

To learn more about this technology, we caught up with Gabriel Falcão. 

1 - Could you explain how quantum computing is being used to detect gastrointestinal cancer?

Not yet. What we anticipate is that it will be feasible within a few years. My research aims to develop new AI algorithms capable of extracting more features from images, and, ideally, from other types of sensors expected to become available soon, such as quantum sensors. By leveraging multidimensional data, we expect to achieve better pattern separation and classification, ultimately leading to improved algorithms for the prevention of certain pathologies.

2 - What differentiates your model from conventional image-based diagnostic systems?

Current technologies rely entirely on classical computers for processing 2D or 3D data. At this stage, we can extract information from data in two or three dimensions, but not much beyond that. Our models, however, will be able to access higher-dimensional data spaces.

It is also important to note that classical machines will eventually be unable to keep pace with the processing speed of quantum computers once these devices reach a significant number of qubits (whereas classical computing processes bits, quantum computing processes qubits, allowing for several orders of magnitude more computational power). The technology we are developing should therefore be capable of processing vastly larger amounts of data, supporting medical teams in diagnosing many more patients.

3 - How have you managed to combine higher diagnostic accuracy with lower energy demands?

At the current number of qubits available, quantum computers are not yet energy-efficient enough to be competitive. However, as hardware advances, energy efficiency is expected to improve alongside computational power.

4 - What potential impact could this have on healthcare systems, particularly in colorectal cancer screening?

The goals behind this project are extraordinarily ambitious, and I would highlight two in particular. First, suppose we can demonstrate that technology is scalable, which will require quantum computers supporting hundreds of thousands to millions of qubits. In that case, it will become possible to process an amount of medical imaging data that is intractable for classical computation. Such capability could enable massive-scale health monitoring for billions of people.

Second, quantum AI algorithms could extract richer and more diverse features from training data, leading to higher classification rates, potentially even for medical conditions not yet identified. This latter perspective is what I find most exciting about the project.

5 - Do you believe quantum computing will become a key technology in medicine in the coming years?

I strongly believe so, both in terms of computational applications and the adoption of quantum sensors in medicine. Interestingly, the same principle that complicates the development of quantum computers (quantum noise) is extremely useful for creating highly sensitive sensors, since these technologies operate within the domain of quantum mechanics. Breakthroughs are expected in both areas.

6 - Are you exploring applications of your solution to other types of cancer or diseases?

Yes, it is in our roadmap to extend this work to other types of gastrointestinal cancers, as an initial step.