News

Toshiba Digital Solutions and Chiba University’s Center for Frontier Medical Engineering Begin Joint Research on Lymph Node Metastasis Detection for Stomach Cancer Using AI

January 31, 2018
Toshiba Digital Solutions Corporation

KAWASAKI, Japan―Toshiba Digital Solutions Corporation (TDSL) today announced the commencement of joint research on an AI*1-based pathological diagnostic system for lymph node metastasis of stomach cancer with the following faculty members of Chiba University: Professor Hideki Hayashi (Center for Frontier Medical Engineering), Professor Hisahiro Matsubara (Department of Frontier Surgery) and Assistant Professor Jun Matsushima (Medical Diagnostic Pathology) and their Research Groups.

Cancer is one of the major causes of death for both men and women in Japan, claiming the lives of 28.5% of those people. Stomach cancer is the second leading cause of death for men, and the fourth for women*2, as well as the most prevalent malignant disease among men and third among women. Meanwhile, the surgical approach has been shown to be the most effective for the treatment of stomach cancer, and the five-year relative survival rate*3 of the disease has reached 65.3% for men and 63.0% for women*4 in Japan. Accurate and fast-track diagnosis for lymph node metastasis is essential to provide minimal invasiveness without lowering the curability of treatments.

This joint research includes AI analyses for microscopic images of a lymph node section stained with H-E *5, and a plan to verify the feasibilities of AI-assisted diagnoses for the metastasis with accuracies similar to those of expert pathologists. The final goal of this project is to the reduce physical and psychological burden of pathologists engaged in surgical oncology, and to provide higher chances of minimally-invasive and function-preserving surgical treatments for the people with this disease expecting much improved post-operative Quality of Life (QOL).

TDSL aims to digitize the diverse diagnostic insights of expert pathologists with this joint research, and contribute to total cancer-treatment solution with the use of Toshiba analytics AI “SATLYS™.”

Hideki Hayashi, Professor at Chiba University, Center for Frontier Medical Engineering , commented:

“AI analysis of images could contribute to a broad range of medical fields because it has the capability to process enormous amount of information within a very short period of time. Recently, pathologists have been frequently requested to conduct ‘rapid diagnoses’ to reveal the presence of metastasis in surgical samples during surgery. Even a micrometastasis cannot be overlooked within such a short period of time since it has been shown to have biological activities. If AI could assist in checking the total H-E stained sections of surgical samples for metastasis with similar accuracies as expert pathologists, it would be an ideal procedure since the final pathological diagnosis is made based on such H-E stained sections. Furthermore, it would reduce the workload of pathologists and increase the number of chances to provide minimally-invasive and function-preserving surgeries for patients. It would contribute to establishing not ‘disease-oriented’ but ‘patient-oriented’ treatments, as well.”

Fig. 1 Image

Figure 1: Lymph node metastasis of stomach cancer (virtual slide scanner image of a HE-stained lymph node section: closed solid lines indicate metastatic regions)

Notes:

  • *1 AI: Artificial Intelligence.
  • *2 Japan Ministry of Health, Labor and Welfare’s Outline of Vital Statistics (final data) for 2016.
  • *3   Five-year period relative survival rate: a ratio of the survival rates of patients following 5 years from diagnosis of a disease to those of general population with the same sex and age distributions. One hundred percent of relative survival indicates that the disease is completely curable, and 0% implies that no treatments are effective in curing the disease.
  • *4 Cancer Statistics published by the National Cancer Center Japan’s information service.
  • *5 HE staining: Staining methods using hematoxylin and eosin for cell nuclei and cell cytoplasm visualization, respectively.

About Toshiba “SATLYS™” Analytics AI

Toshiba “SATLYS™” analytics AI is a service that applies Toshiba’s 140 years of insights and achievements in manufacturing to AI design to realize high-precision recognition, forecasting, cause prediction, anomaly detection, malfunction prediction and behavior prediction. By offering solutions that use SATLYS to analyze inspection data, sensor data, work data and behavioral data, Toshiba is accelerating the digital transformation of its customers. For more information, please see the following website (in Japanese):
https://www.toshiba-sol.co.jp/pro/satlys/

  • * SATLYS is a registered trademark or trademark of Toshiba Digital Solutions Corporation in Japan and other countries.
  • * Other company names and product names in this document may be trademarks or registered trademarks of the respective companies.