[Edaily Reporter KIM SAE-MI ] “I’ve been a pathologist for over 30 years, but I was never able to predict a patient’s future based solely on a slide. Even though I could tell how severe the disease was, I didn’t know what would happen to that patient a few years down the road. But I thought artificial intelligence (AI) would be able to do that.”

Cho Nam-hoon, Chief Medical Officer (CMO; hereinafter referred to as “President”) of MTS Company, made these remarks in an interview with Edaily on the 20th at the “International Hospital & Health Tech Expo (KHF 2026)” held at COEX in Seoul. In the view of President Cho, a former president of the Korean Society of Pathology, the role of AI goes beyond simply replacing the eyes and hands of a pathologist. He believes that the ultimate goal of digital pathology AI is to identify features in pathological images that are difficult for the human eye to detect and use them to predict patient prognosis and treatment response.

A Pathologist Who

Chose MTS Over a Hospital… Why?

President Cho is a former professor at Yonsei University College of Medicine who led a government-funded digital pathology AI project for five years. After leaving his professorship, he had planned to return to a hospital. He even went so far as to say, “I had absolutely no intention of joining a company.” What steered President Cho toward MTS Company was the company’s technological capabilities—which he had witnessed firsthand while collaborating on the government-funded project—and his desire to see the transition to digital pathology through to completion.

A decisive factor was the timing: his retirement from his professorship coincided with the pathological AI—which had been developed and refined over five years with secured research funding—entering the commercialization phase. President Cho explained, “I realized it would be difficult to properly refine the product by providing feedback from outside the company,” adding, “I decided I needed to be inside the company to hear how the product was evaluated by hospitals after it received regulatory approval and entered the market, and to make improvements based on that feedback.” He went on to emphasize, “I joined this company to see this through to the end.”

The “end” President Cho refers to goes beyond the commercialization of the pathology AI currently under development. His ultimate goal is to complete the “Physical Transformation (PX),” in which robots perform the physical tasks of a pathology lab—going beyond the “Digital Transformation (DX),” which converts microscopes into digital images, and the “AI Transformation (AX),” where AI assists with diagnosis and prognosis prediction. CEO Cho estimates this journey will take about 10 years.

His relationship with MTS Company did not begin with a firm belief in the company’s technical capabilities from the start. President Cho first encountered the company while collaborating on a data project related to cervical cancer, and what caught his attention above all else was the work ethic of its employees.

“Even when doing the same tasks, some people stand out as particularly reliable, and I got the sense that the employees at MTS Company were truly dedicated to their work,” he recalled. “I didn’t have high expectations at first, but after reviewing the interim results from the third year of the government-funded project, my perspective changed completely. From that point on, I viewed MTS Company as a core development partner, began participating directly in weekly meetings, and started collaborating in earnest.”

Pathology AI Headed Toward Prediction… 60,000 Data Points as the Foundation The area

President Cho emphasizes most is

“Predictive

AI.” While much of today’s medical AI focuses on auxiliary tools that reduce repetitive tasks for medical staff—such as image interpretation or counting—he believes we must take it a step further.

“When doctors and AI companies meet, the first request is usually, ‘This task is too difficult—please do it for us,’” he noted. “While AI that reduces repetitive tasks and workload is necessary, not all AI needs to be developed solely for that purpose,” he added.

To this end, from the very beginning of the government-funded project, he built a dataset that tracked not only pathology slides but also patients’ age, gender, disease stage, family history, treatment history, and survival status. It sometimes took two full days just to track medical records to verify the treatment progress of a single patient. The data collected in this way totals approximately 60,000 cases, of which about 20,000 are breast cancer cases.

President Cho explained, “Even when comparing slides from living patients to those from deceased patients, I couldn’t see any differences with my own eyes,” adding, “However, when we trained the AI, probabilistic differences emerged from combinations of features invisible to the human eye.” He continued, “I’ve been working in pathology for over 30 years, but I couldn’t determine a patient’s prognosis just by looking at a slide.” He added, “I thought AI could do that, and that was my vision from the very beginning when designing the dataset.”

Following the approval of its AI for assisting in the diagnosis of pathologies such as breast cancer, MTS Company is expanding its scope to develop AI that predicts prognosis and treatment response. The company is developing models that predict homologous recombination deficiency (HRD) test results for ovarian cancer and the presence of lymph node metastasis for breast cancer based solely on pathological images.

An HRD test checks for abnormalities in the ability of cancer cells to repair damaged DNA. While it is used to determine treatment strategies—such as PARP inhibitors—for ovarian cancer patients, its high cost and the fact that it takes several months to receive results are cited as limitations. CEO Cho aims to replace this with AI that predicts HRD status using only pathological images.

CEO Cho stated, “Existing HRD testing is expensive and takes a long time to produce results,” adding, “After training the AI on pathology images and HRD test results, it achieved an accuracy rate of over 85% from the outset.” He continued, “I want to bring hope to patients suffering from ovarian cancer by commercializing this product as soon as possible.”

In the case of breast cancer, the company is refining AI that predicts lymph node metastasis using only pathology images. The current accuracy stands at approximately 78–79%, and work is underway to raise it to over 80%.

“Color Pathology AI Is MTS’s Strength… We Will Complete the DX→AX→PX Journey”

President Cho identifies MTS Company’s core competitive advantage as its ability to handle complex color pathology images. While radiological images such as CT scans are primarily composed of grayscale, pathology images involve the overlay of RGB and various staining combinations, making the information far more complex.

“Analyzing pathology images is similar to solving high-dimensional equations,” he explained. “Since MTS Company has a relatively small workforce, a single person has been responsible for developing the technology while understanding a significant portion of the process from start to finish—and this has actually become a strength in the field of pathology.”

He continued, “AI that handles color is a very specialized field,” and expressed confidence, saying, “We must not let this technology slip away but continue to develop it to become the undisputed leader. MTS Company possesses truly excellent technology.”

President Cho’s vision for the next step is “Physical AI,” which integrates robotics. In pathology laboratories, a significant amount of manpower is still required for the process of washing, cutting, and transferring surgical specimens to create slides. The plan is to introduce Physical AI to reduce specimen mix-ups and errors that can occur during manual handling and to standardize the process.

“The next step is, without a doubt, Physical AI,” he emphasized, adding, “If we’ve moved from Digital Transformation (DX) to AI Transformation (AX), the next step is Physical Transformation (PX). I want to complete the transition from DX to AX and then on to PX.”

However, he stressed that the relevant market will not open up through technology development alone. President Cho identified regulations and reimbursement rates as the biggest barriers. This is because, even if approval is granted, it is difficult to achieve business viability without receiving appropriate reimbursement rates.

“If reimbursement rates are based on working hours, it’s difficult for AI to receive appropriate compensation because it completes an interpretation in just one second,” President Cho pointed out. “We must first establish new standards for how to reflect the cost and value of AI technology in reimbursement rates.”

In the future envisioned by President Cho, AI is not a technology that will eliminate the jobs of pathologists. It is a mutually beneficial structure where AI handles repetitive tasks, allowing medical staff to focus on more sophisticated judgment.

He predicted, “Once predictions become possible, the pathology department will no longer be limited to providing diagnoses but will become a department that supports both diagnosis and treatment,” adding, “If pathology can even provide solutions for drug selection, the department’s standing will actually grow even stronger.”