Infervision is a high-tech company specialized in medical AI and committed to applying deep learning technology to assist medical image diagnosis. Infervision products are designed to increase diagnosis efficiency and accuracy. They perfectly integrate into the clinical workflow, and comprehensively cover the imaging diagnosis for several body parts (lung, heart, brain, skeleton) to provide an ultimate medical imaging AI solution.
Infervision has several consolidated products in the market and new product introductions:
1. InferRead CT Lung is a quick and accurate AI solution for lung cancer screening and management, which is fully integrated into clinical routines of +280 hospitals worldwide (assisting in +26,000 lung cancer reports every day). InferRead CT Lung automatically identifies different types of lung nodules (e.g. solid, calcified, partial calcified or ground glass). This solution provides comprehensive and accurate information on nodules, including position, size, density and malign rate. Co-registration and pre-processing of actual and older images for volume doubling time calculation is also incorporated.
2. InferRead DR Chest detects +20 chest lesions from conventional x-ray scans (including lung nodules, fractures, pneumothorax, pleural effusion, pneumonia, emphysema, among others). Clinical evidences in hospitals shows that InferRead DR Chest reduces missed diagnosis rate and increases detection rates for lung cancer screening and other chest pathologies, thereby redefining the value of conventional chest x-ray.
3. InferRead CT Bone detects bone fractures from chest CT scans. It is capable of diagnosing bone lesions in sternum, rib cage and shoulder blade. InferRead CT Bone is designed for various cases (e.g. polytrauma patients) and it is highly sensitive for diagnosing tiny lesions. Metastasis and bone destruction are also detected.
4. InferRead CT Stroke was developed to provide an accurate assessment of haemorrhagic stroke. It locates the bleeding area and accurately measures bleeding volume to determine cerebral palsy.
5. InferRead CT Chest is a new product concept that detects 4 different conditions from a single chest CT scan. The new InferRead CT Chest allows radiologists to perform multiple disease screenings simultaneously: lung nodules, chest fractures, bone metastases, bone tumour, chronic lung diseases (e.g. emphysema) and cardiac calcification.
It is worth noting that big hospitals have started to adopt AI as their standard practice because they can increase accuracy of diagnosis and improve efficiency. At current stage, Infervision solutions help in the detection, quantification and characterisation of specific lesions. The aim is to relieve radiologists form laborious and repetitive work and enable them to focus on more meaningful clinical cases. Infervision solutions also work as triage systems, by prioritising urgent cases in the radiology worklist. Infervision products automatically generate structured reports and contain several features to improve follow-up practices.
Besides, Lung Cancer Screening Programs are not yet enforced in Europe due to the various reasons. The high costs involved is one of the main hurdles that prevent the implementation of such potentially beneficial programs. Infervision’s CT Lung has become the trend of future and it is able to help doctors detecting early stage lung cancer with a much higher productivity, thereby reducing the cost involved in screenings.
AI solutions are mainly focused on image interpretation tasks and therefore, they do not help in many other tasks performed by radiologists (e.g. correlate findings with medical records or test, advising other doctors on diagnoses and treatments, interventional radiology, communication with patients, etc.). Nevertheless, AI solutions will evolve to provide more capabilities. For instance, AI could be useful during image acquisition, by alerting whether the image has been correctly taken. Another industry challenge is that some diseases cannot be detected with AI if there is little availability of high-quality images with confirmed diagnosis (e.g. tuberculosis). Providing universal access to radiology services through AI is another very complex task. In this regard, Infervision is working on closing the gap between cities and rural areas across China.
A challenge faced by medical AI companies lies in the fact that rapid product innovation and iteration require to break through the existing boundaries within organizational forms and to strengthen teamwork. Medical AI products need constant improvements within the product life cycle, so all departments need to carry out product iteration together. Such cyclical improvement is not automatic, but relies on high-intensity cross-team coordination. In such cycle, whoever iterates faster and has the strongest ability to continually correct themselves, will have more chances to succeed in the market. At Infervision, the “Marketing – Operations – Data – Software Development – Engineering – Marketing” teams closed-loop is crucial to integrate the doctor’s feedback into new product versions. Therefore, the core competitiveness does not lie in any single team, but rather in the performance of this closed loop.
Infervision was founded in 2015 by Kuan Chen and Shaokang Wang in Beijing (China). The founding team spent one year working at Sichuan Provincial People’s Hospital, learning how the AI tools they were developing could accurately identify signs of lung cancer growth. After this first successful pilot, Infervision secured investment and expanded to cover several other Chinese hospitals. Today, Infervision’s products have been deployed in more than 300 hospitals worldwide, including China, the United States, Germany, Japan, Spain, Switzerland and Austria. In China, Infervision is the company with the largest market share: 70% of top-tier hospitals use the InferRead CT Lung solution for lung cancer screening. Infervision currently employs around 320 people.
Infervision will be part of the ECR 2019 Artificial Intelligence Exhibition (AIX) in expo X1, booth AI-10.
Website: www.infervision.com/en
