The National Cancer Center Hospital is developing medical AI.
To develop high-quality medical AI, it is important to create high-quality training data. However, the work to create high-quality training data is time-consuming.

This article introduces the reasons why Dr. Mototaka Miyake (MD), the assistant chief of Department of Diagnostic Radiology, introduced the i3 series "CL-S1200" and the effectiveness of introducing the CL-S1200.

• Your specialty is diagnosis of gastrointestinal tract and bone and soft tissue tumor, and you are working on the latest diagnostic imaging research.

• Yes.
  I am working on the development of AI for them.
  The CL-S1200 has been very useful in the process of AI development.

• Please tell us what you are developing.

• We are developing medical image AI using machine learning techniques including deep learning. AI development to detect the region of cancer requires a large dataset of the region of interest (ROI). By using a medical-grade diagnostic monitor, we can develop AI in the same environment as in our daily work.
  The quality of AI is highly dependent on the quality of the dataset. In order to prepare a good quality dataset of ROI, we need an environment that allows us to concentrate and work efficiently for long periods of time with little fatigue.

• What kind of monitor do you need for such an environment?

• A monitor with good gamma correlation is required.
  When creating ROIs, multiple images are displayed on the same monitor.
  CL-S1200 is a large and a high-resolution monitor, and the CL-S1200's dynamic gamma function can display each image with the appropriate gamma.

• Is it clearly different from the monitor you used before?

• Yes.
  The reason I decided to introduce the CL-S1200 was that I noticed a difference between ROIs created using a general monitor and those created using the CL-S1200.
  In addition, this model can display appropriate images not only for CT and MRI images, but also for mammograms and radiographs. This model can also display not only medical images but also various development tools at the same time without causing any problems.

• What do you think the relationship between physicians and AI will be in the future?

• Several AI-assisted radiological diagnostic imaging software products have already been approved. We, diagnostic radiologists, are now in the era of managing AI as part of our work. We need to recognize that medical-grade diagnostic monitors are also important devices for AI management.

• AI will "grow" as it is used in clinical practice, won't it?

• Ah... I should not say "growing" but "adapting to changes in the external environment". AI does not necessarily show the same performance against new diagnostic equipment as it did in the past. For AI to maintain its performance, it must continue to learn new data. In addition, long-term verification is necessary to examine whether the output of AI was correct or not. We need to continue to evaluate and maintain the performance of AI in parallel with our daily clinical work.
  In the future, I expect that medical-grade diagnostic monitors will occupy an important position not only as an appropriate display of AI detection results, but also as an important management tool in the maintenance of dataset.