Infervision: Our AI Detects Deadly Coronavirus via CT Lung Screening
The lung cancer detection company has launched a deep learning-based coronavirus detection system in Wuhan.
Infervision (推想科技) described its coronavirus detection model as working through CT lung screening. This means that the new model detects ground-glass opacities (GGO) in the lung that may later be confirmed as one of the complications of the virus.
The firm's model has already been used by Wuhan Tongji Hospital, one of the frontline hospitals that has been fighting with the epidemic in Wuhan. It is set to be deployed in multiple other hospitals in the near future, Infervision said.
“It serves as a surrogate to PCR diagnostic as the lab capacities aren't enough to keep up with the rising number of suspected cases in the afflicted regions," added the company.
Another diagnostics solution support for the rising crisis was launched by Shenzhen-based genomics giant BGI Genomics (300676: SH) on January 28. The company's stocks opened around 10% up as the mainland's stock markets open on January 3, 2020.
“The new coronavirus may cause infections with no symptoms and sicken otherwise healthy people,” postulated the Lancet, showing how hard it is presently for hospitals in Wuhan, which are racing against time to diagnose large numbers of patients.
The new diagnostic solutions possess tremendous importance for the afflicted region, considering the symptoms of the new virus are similar to other fast-spreading diseases like flu, complicating the process and creating excess demand thanks to a large number of concerned patients who need to be assessed. "Symptomatic patients are piling up in the hospital – AI can help triage the patients quickly," said the firm. "Patients have been seen without any noticeable symptoms like fever or cough, but showing large GGOs in the lung,"
The Beijing-based private firm had long been drawing the attention of wealthy venture capitalists, including Qiming Venture Partners and Sequoia China; it has secured around USD 100 million since its inception in January 2016. The team has been adopting deep learning technology broadly in the medical imaging field.
The firm was involved in an early victory in this complex battle: for the first time since the emergence of the new virus in late 2019, the number of new suspected coronavirus cases in China started to drop, over two consecutive days, on December 31 and January 1, 2020 - mostly driven by faster diagnostic solutions. PCR diagnostics had shown that it could play a crucial role.