Fujifilm’s New AI-based^*1 Technology for Lung Nodule Detection Now Approved for Use in Japan

TOKYO, May 19, 2020 —FUJIFILM Corporation (President: Kenji Sukeno) is pleased to announce the company’s Artificial Intelligence (AI)-based lung nodule detection technology— an algorithm developed to assist diagnostic imaging by automatically detecting lung nodules in chest computed tomography (CT) scans— has been cleared by the Pharmaceutical and Medical Device Act^*3.  The algorithm will be released in Japan on June 1, 2020 through FUJIFILM Medical Co., Ltd. (President: Akio Niinobe). 

The algorithm detects lung nodules, which appear as white shadows in the lung on an X-ray or CT scan, and identifies them for physicians to further evaluate as they may indicate lung cancer or other diseases. It was developed to support the criticality of early detection and treatment for lung-related diseases, to provide physicians with the diagnostic confidence that comes with AI-support, and to alleviate some of the manual effort required of physicians reviewing CT scans.

The technology uses Fujifilm’s 3D analysis technology to deliver a more detailed view, allowing physicians to better understand the nodule’s location in areas that would be difficult to identify with cross-sectional CT images alone. It features “organ recognition technology” which is used to identify the pulmonary region before searching for nodule candidates, enabling the detection of nodules adjacent to the chest wall. Physicians can also leverage a lesion tracking function to efficiently monitor the progression of lesions and nodules. 

Fujifilm’s lung nodule detection algorithm operates within the company’s Synapse SAI Viewer, an image interpretation viewer that utilizes AI to support diagnostic workflow on its Synapse 5 PACS system in Japan. It is part of a growing number of AI-enabled technologies designed to assist in diagnostic imaging and facilitate efficient workflow under REiLI— the company’s global Medical Imaging and Informatics AI technology initiative. 

The company plans to introduce the technology in the United States, Europe and other markets once required regulatory clearances are obtained.

• *1 The function was designed with Deep Learning, which is a form of AI technology, and does not automatically change system performance or accuracy once installed.
• *2 Pulmonary nodules appear as white shadows in the lungs in X-ray and CT scan images, and may be an indication of lung cancer or other diseases. Pulmonary nodules are up to 3cm in size and could be as small as 5mm or below.
• *3 The Pharmaceutical and Medical Device Act is a law concerning the assurance of quality, efficacy and safety of pharmaceuticals and medical devices.

June 1, 2020

Detecting lung nodules
The lung nodule detection is a function that assists the identification of lung nodules. A doctor who has examined a scanned image can re-check nodule candidates detected by the function to mitigate oversight. It can be combined with the lesion tracking function, which is used to monitor progression of lesions and nodules, thereby assisting comparison and reducing doctors’ workload.