Vuno has been attracting attention by providing artificial intelligence-based CT solutions to diagnose complications and illnesses in the chest and lung areas. Now the company is expanding its technology to pathology.

Already, Vuno has made some advancement in the field by developing VUNO Med-PathLab, its AI-using pathology research platform.

The company conducted various researches using the device. It presented one of such studies analyzing the spatial distribution of the tumor and matrix in the tumor microenvironment in the cancer tissue slide in colorectal cancer tissue slides at the American Association for Cancer Research (AACR) conference last year.

This year, the company expanded the research into liver cancer to confirm a biomarker highly associated with the survival rate of liver cancer patients and presented the results at the 2021 American Society of Clinical Oncology (ASCO) conference.

Vuno's research team analyzed the tissue slides of 351 liver cancer patients using VUNO Med-PathLab. The platform divides the tissues into malignant cells, lymphocytes, mucus, and normal tissues. It analyzes pathological images by detecting and classifying cells.

The results showed that cell density per lymphocyte segmented area (CDpLA) was the primary variable predicting the survival rate of liver cancer patients. In addition, the study found that the higher the CDpLA, the higher the median overall survival of liver cancer patients.

Korea Biomedical Review met with two Vuno executives – Chief Technology Officer Jung Kyu-hwan and Senior Researcher Kim Kyung-doc -- to discuss the company's presentations at ASCO in detail and its plans and goals.

KBR: Could you explain VUNO Med-PathLab?

Jung: VUNO Med-PathLab is a platform for quantifying digital pathology slides and extracting biomarkers.

A whole H&E slide is a very high-resolution image with a wide surface. Therefore, it is nearly impossible for a person to separate each cell type, draw a tissue area, and quantify how many and where a specific cell is in the entire cell area.

After teaching the AI to do those jobs, however, it can provide researchers with such information.

This is important as such cell and tissue distribution is important when making a diagnosis. For example, if patients have many immune cells, such as lymphocytes in the tumor, they are regarded as having a good prognosis.

If we can use VUNO Med-PathLab to find new biomarkers, physicians will use the data to make a diagnosis, prognosis, and prediction.

KBR: Does VUNO Med-PathLab drastically shorten the analysis time compared to humans?

Jung: Definitely. There is a limit to what humans can do as giant analytical hardware, but AI uses better hardware, and paralleled uses are also possible.

VUNO Med-PathLab needs one hour if we run it completely naively and 20 minutes when set a characteristic. Humans could need several or tens of hours to produce the same results.

KBR: Did the company face any problems securing statistical significance on the platform with a data set of 351 patient samples used in this liver cancer study?

Kim: We created and analyzed the recent study model using a public data set, and we could obtain clear results without any particular statistical deficiency.

Of course, the company believes that it will be better to validate the data using the data from actual hospitals. Therefore, we are planning additional researches in the area.

Jung: Digital pathology slides are more difficult to obtain than X-rays or CT screening images.

Even if we collect all the genomic data and pathology slides with the patient's consent, the patient pool usually does not surpass hundreds in such studies.

In the future, when we are trying to validate our results on actual patients, we will also try to predict the response of immunotherapy.

However, as these drugs are pretty new to the medical field, it will be difficult to obtain patient data and even more so for patients who saw an efficacy from receiving the treatment.

Therefore, the size of this data is not small compared to other similar studies.

KBR: You said you could develop a precision medical solution by predicting each patient's prognosis and treatment response. Does this mean the VUNO Med-PathLab can provide information on customized treatment for each patient?

Jung: Yes. As of now, the device can classify patients in detail to how they will react to anticancer treatment.

However, rather than simply classifying patients based on pathology slides, we believe that the platform will tell whether or not this person will react to a specific anticancer drug in the future.

Currently, when we divide patients based on existing immune checkpoint inhibitor or biomarkers, the results are not completely accurate as patients who expected good response to a treatment sees bad results, while patients believed to have bad results showed good response after treatment.

This is because even if the patients have identical genetic markers, the expression of such markers may differ by the patient.

Therefore, if we can well quantify how well the immune cells of the morphological tumor are expressed in each patient, we will judge what the best customized treatment is for each patient.

Our ultimate goal is to combine the existing diagnostic biomarkers and the image biomarkers extracted from VUNO Med-PathLab and allow physicians to choose which treatment will be best for the patient accurately.

Besides, we plan to expand the use of this device to help pharmaceutical companies recruit certain patients to conduct better clinical trials.

KBR: Do you think VUNO Med-PathLab will change the paradigm for cancer treatment?

Jung: Of course. Cancer treatment is moving toward personalized medicine, and I believe that our AI, including VUNO Med-PathLab, can be used to analyze all possible data of a patient and provide more precise and personalized treatment.

KBR: Do you plan to expand VUNO Med-PathLab to other cancers as well?

Kim: Yes. We developed VUNO Med-PathLab to help achieve personalized treatment and not to focus on specific cancers. In the future, we plan to use I plan to conduct researches for VUNO Med-PathLab in other major cancers, including lung cancer.

KBR: Are there any other platforms under development in the field of pathology?

Kim: We are developing a variety of pathology-related software.

VUNO Med-PathQuant is software that detects a specific number of cells at the beginning of a diagnosis and can reduce the burden of doctors.

The second is diagnostic assistance software, which assists physicians in diagnosing cancer through biopsy. The company is developing two software -- VUNO Med-PATH GC AI and VUNO Med-Path Pros AI.

PATH GC AI is software that helps pathologists find cancer, adenoma, and other inflammations from a biopsy sample when performing gastroscopy, and Path Pros AI is a product that assists in diagnosing prostate cancer.