News | Press Releases

July 15, 2020

Start of Clinical Evaluation of New Technology to Improve the Diagnosis of Novel Coronavirus Pneumonia

Canon Medical Systems Corporation
Fujita Health University

 Canon Medical Systems Corporation (Headquarters: Otawara, Tochigi Prefecture, Japan; President and CEO: Toshio Takiguchi) and Fujita Health University (Toyoake, Aichi Prefecture, Japan; President: Eiichi Saitoh) are starting clinical evaluation of the usefulness of new technology to improve the diagnosis of novel coronavirus pneumonia based on CT image analysis focusing on diffuse pulmonary diseases. This technology has been jointly developed through academic-industrial cooperation in which Professor Yoshiharu Ono of the Department of Radiology at Fujita Health University worked in partnership with Canon Medical Systems Corporation.

It has been reported that some patients with novel coronavirus pneumonia exhibit CT findings which are characteristic of interstitial pneumonia. In particular, the detection of high-density shadows such as ground-glass opacities and infiltration and the identification of specific findings related to the shape and location of such abnormal areas are useful for the diagnosis of novel coronavirus-specific pneumonia. The ability to quantitatively assess the characteristics of pneumonia associated with novel coronavirus infection should be extremely helpful in improving diagnosis, prognosis, and treatment outcomes. This technology is also expected to provide new information that should be of great value in the diagnosis of other pulmonary diseases.

By undertaking clinical evaluation of this new technology in partnership with Fujita Health University (which has extensive clinical experience in the treatment of COVID-19 patients, including those who became infected on the cruise ship Diamond Princess), we hope to develop effective solutions at the front lines of medical care.

The CT texture analysis technology to be clinically evaluated focuses on diffuse pulmonary diseases. This new technology has been developed through collaborative research conducted by Professor Yoshiharu Ono of Fujita Health University working with Canon Medical Systems Corporation. The analysis technology, which employs advanced machine learning techniques, supports the quantitative analysis of seven types of morphological characteristics based on the shadows observed and also permits detailed evaluation of the morphological changes caused by various pulmonary diseases. The characteristics which are analyzed include normal lung, ground-glass opacities, infiltration, honeycomb lung, granular shadows, and emphysematous changes. The results of this basic research have already been presented at conferences hosted by various academic societies, demonstrating strong correlations between the degree of severity of interstitial pneumonia and the results of quantitative analysis1,2 as well as good agreement between the findings reported by experienced radiologists and those identified by the analysis technology.3,4 At the end of July 2020, we will begin to examine patients with suspected novel coronavirus pneumonia using this technology as a part of its initial clinical evaluation.

The new analysis technology will be incorporated into X-ray CT systems manufactured by Canon Medical Systems Corporation, which feature outstanding image quality at low exposure doses. It is therefore expected that reliable analysis results will be obtained while minimizing the burden on patients and ensuring efficient workflow throughout the entire process from CT scanning to image analysis. As well as undertaking clinical evaluation in partnership with Fujita Health University, Canon Medical Systems Corporation is also conducting clinical studies together with other medical institutions in Japan and overseas in order to achieve quicker and more accurate diagnosis.

Professor Yoshiharu Ono of Fujita Health University has summarized his thoughts concerning this clinical evaluation as follows.
It has been observed that there is some variation between physicians in the interpretation of CT images in cases of pulmonary diseases, including novel coronavirus pneumonia. This variation may lead to difficulties in ensuring a consistent standard of medical care, including patient management. For this reason, medical professionals in all parts of the world have high expectations for the development and clinical introduction of diagnostic systems based on AI and other advanced technologies.

The application of CT texture analysis technology for the characterization of various pulmonary diseases is expected to provide the following benefits:

  1. To permit the identification of individuals with COVID-19 who are asymptomatic or PCR false-negative when they visit the hospital.
  2. To improve diagnostic accuracy and help ensure appropriate care for patients in whom the differential diagnosis of COVID-19 pneumonia and other types of infection is difficult.
  3. To allow estimation of the time from onset, evaluation of the degree of severity, and selection of appropriate treatment based on the quantitative evaluation of CT findings.
  4. To assist in patient management by permitting treatment effects to be quantitatively assessed based on CT findings.
By applying this analysis technology in the clinical setting, this study is expected to promote safe and effective medical care to not only for patients with novel coronavirus pneumonia, but also for patients with many other diseases who visit the hospital. At the same time, it is my hope that the findings of this study will also help to protect the safety of medical professionals, allowing them to implement effective measures for hospital infection control as they provide essential healthcare services.

Diagnostic CT images and analysis results in a case of novel coronavirus pneumonia
(upper row: original images; lower row: analysis results).

[Conference Presentations]
1) 3D Computer-Aided Diagnosis System for Thin-Section CT: Utility for Pulmonary Functional Loss and Treatment Response Assessments in Connective Tissue Disease Patients. 2016, Radiological Society of North America (RSNA).
2) Utility of 3D Computer-Aided Diagnosis System for Pulmonary Functional Loss and Treatment Response Assessments in Connective Tissue Disease Patients. 2017, European Congress of Radiology (ECR).
3) 3D Computer-Aided Diagnosis System on Thin-Section CT: Quantitative Assessment of Disease Severity and Therapeutic Response in Patients with Polymyositis/Dermatomyositis. 2018, Japan Radiological Society (JRC).
4) 3D Computer-Aided CT Texture Analysis with Machine Learning: Capability to Play as Second Reader for Radiological Finding Assessment in Patients with Interstitial Lung Disease. 2019, European Congress of Radiology (ECR).

Contact Us