Almukhamed Amitov

Students Almukhamed Amitov and Alimzhan Kasabayev claim to have developed an AI system capable of detecting cancer in seven minutes. One of the developers spoke to a NUR.KZ correspondent about the project.

Alimzhan Kasabayev is pursuing a master’s degree at the University of Warwick in the United Kingdom, while Almukhamed Amitov studies at KBTU. The idea for their startup came about by chance. Alimzhan, who studies bioengineering, received a CT scan from his professor along with a simple assignment: “Make something interesting.” A brainstorming session combined with scientific curiosity eventually led the students to develop an AI system for cancer detection.

“We accelerate the CT scanning process from three weeks to seven minutes: a doctor uploads a CT scan to a web platform, and the neural network analyzes it within seven minutes and provides a result. We compared the two approaches: an average oncologist can spend two to three weeks analyzing a CT scan manually, while we reduce the process to seven minutes,” says Almukhamed Amitov.

The developers upload CT scans and “train” the neural network by visually showing it every layer of the scan, while also indicating where the tumor is located and what percentage of the area it covers.

“Initially, when we did not have partnerships with hospitals, we worked with open-source data. But there were only 200 patients, which is far too little for a neural network. Even now, when we have agreements with six hospitals, we still need more data.

We trained the neural network using CT scans from publicly available sources. After we won IT Fest and NURIS, six clinics immediately approached us. We are now training the neural network using local data. These six clinics want to test our development as a pilot project,” Almukhamed Amitov explains.

According to the developers, the neural network can currently detect liver tumors as small as 0.1 mm. By March, the students planned to train it to work with all types of gastrointestinal cancers. This was requested by the same six clinics they are collaborating with.

After that, the students say, the neural network will eventually be able to work with the entire human body.

“There are many legal and technical requirements. We are working on them now: we speak with lawyers every day to resolve these issues. We are also applying for a patent and a license. We were given a list of requirements. It will take approximately one to two years to fully meet all of them,” says Almukhamed.

In addition to Almukhamed and Alimzhan, the team currently includes four developers, all of them Kazakhstani students, as well as three oncologists.

Initially, the students planned to create an application that would allow patients to upload their own CT scans and check them for signs of cancer. However, they have put that idea on hold for now.

“After reviewing the requirements and the legislation, we decided that it would not be entirely appropriate to give people such a medical tool. They could misinterpret the results. This should be handled exclusively by a doctor. However, we are still considering the idea depending on what the legislation allows,” Almukhamed notes.

At present, the neural network works only with CT results. However, the team plans to train it to analyze MRI scans, blood tests, DNA, and chromosome data in the near future.

A doctor would create a patient profile and upload all of this information. According to the developers, the neural network could then potentially identify a person’s predisposition to oncological diseases.

NUR.KZ also asked oncologists about the development during an RSK briefing. Raikhan Bolatbekova, head of the oncogynecology department at the Almaty Oncology Center, said they were aware of the project aimed at detecting cancer at an early stage.

“We wish our IT specialists success and congratulate them on such a development. However, it is still a project. In my opinion, it is not yet fully developed and has not been released for widespread use,” Raikhan Bolatbekova said.

Earlier, during a government meeting, Minister of Science Sayasat Nurbek announced that Kazakhstan was developing an anti-cancer drug that had no analogues in the world. Nurbek also said that the government planned to bring the drug to market within the following year.