Konica Minolta Jointly Develops AI for Interpreting Chest Radiographs to Detect Abnormalities
Accelerating Development of AI for Primary Care in Cooperation with Enlitic, Inc. and Marubeni Corporation
May 9, 2019
Konica Minolta, Inc. (Konica Minolta) has entered into an agreement on the joint development of artificial intelligence (AI) with Enlitic, Inc. (Enlitic), a pioneering startup company which specializes in developing AI for medical image analysis, and Marubeni Corporation (Marubeni). As the first step under this agreement, work will start on developing AI for interpreting chest radiographs.
Through this joint development project, Konica Minolta hopes to accelerate the development of AI to support primary care and cancer screening services by combining its product development capabilities with Enlitic’s technical prowess.
Enlitic, which was named one of the 50 Smartest Companies by MIT Technology Review magazine in 2016, enjoys a leading position in the field of medical image analysis technology based on deep learning. Furthermore, the company has already collected a large amount of data for around one million cases, and has applied AI to learn these cases. Marubeni is a shareholder of Enlitic, and is expected to leverage its business bases around the world to rapidly expand sales channels. Meanwhile, Konica Minolta possesses know-how in the development and evaluation of technologies that serve clinical needs, including the development of bone suppression processing*1 and temporal subtraction processing*2 techniques for chest radiographs. The partnership among these three companies will accelerate the development of valuable AI solutions.
Konica Minolta has been developing the healthcare business centering on medical imaging information services designed for hospitals and clinics, focusing on the three core areas of X-ray diagnostic imaging systems, diagnostic ultrasound systems and medical IT. Recently, recognizing the importance of primary care, the company has been turning its attention to providing related products and services.
Primary care requires the ability to deal with various diseases treated by different clinical departments. Konica Minolta considers that by applying AI that has learned the skills of medical specialists to primary care, the burden on patients, doctors and technological radiologists will be eased, and that its AI-based product will serve as a gatekeeper to primary care.
Konica Minolta also believes that its AI-based product will reduce the burden on doctors during cancer screening programs, as they have to see many patients in a limited time.
Development of AI for Interpreting Chest X-ray Radiographs
Chest radiographs, which are used widely and taken in large numbers, are easy to take and can help detect various lesions at a time. However, these radiographs capture multiple tissues overlapping one another, making it difficult to identify lesions.
As the first stage of the joint development project, therefore, Konica Minolta will develop AI for interpreting chest radiographs to support primary care and cancer screening intended to detect lesions at an early stage.
Under the brand proposition “Giving Shape to Ideas,” Konica Minolta will accelerate efforts to offer solutions that meet various needs of medical personnel globally and enhance the value of medical services.
*1:Bone suppression processing is an image processing technique to suppress the signals of the anterior and posterior ribs and clavicles based on a proprietary database of chest radiographs using an advanced algorithm. This technology enables visualization of the rib- and clavicle-suppression image in the mind of a doctor and improves the visibility of lesions behind bones in the lungs. Thus it could greatly enhance the accuracy of interpreting chest radiographs.
*2:Temporal subtraction processing is an image processing technique by which a previous chest radiograph is subtracted from a current radiograph using a specific algorithm to correct any differences in position. The temporal subtraction image thus obtained makes chronological changes visible and enables doctors to see how lesions have progressed, have been cured, or have otherwise changed. This technique could improve the accuracy of diagnosis and reduce the time required for diagnosis, helping doctors interpret chest radiographs.