Loading...

How 5C Network is using AI to solve India's radiology accessibility woes

How 5C Network is using AI to solve India's radiology accessibility woes
Kalyan Sivasailam, co-founder, 5C Network  |  Photo Credit: 5C Network
2 Jun, 2021

In April, when India was clocking over three lakh Covid-19 cases a day, Bengaluru based startup 5C Network launched a platform called ask5c.com. The platform would allow anyone with mild symptoms to upload a link to their CT scan report and get a second opinion from a network of radiologists in a few hours.  

Chest CT (computed tomography) scans have emerged as an important tool in the diagnosis and management of Covid-19 pneumonia. However, doctors have cautioned people against rushing for CT scans as it exposes them to unnecessary radiation.  

For 5C, which helps hospitals and diagnostic centers get immediate access to radiologists, the pandemic opened up new channels for growth. The company says it has seen a 3.5X spike in the number of scans per month since March 2020.  

“At its peak, we received in excess of 5,500 Covid-19 scans on a daily basis, while ensuring turnaround times (TAT) as low as 24 minutes. This is in comparison to the average TAT for a CT Thorax scan (used to diagnose Covid-19) is around 3-4 hours at a hospital/diagnostic center due to the overwhelmed healthcare system,” 5C Network founder Kalyan Sivasailam told TechCircle. 

Since 2016, 5C has been building a horizontal diagnostics layer to make diagnostics a utility for hospitals and diagnostic centres. Founded by Sivasailam and Syed Ahmed, it is a digital platform that enables the storage, sharing and interpretation of radiology images. It facilitates AI- assisted diagnosis of X-rays, CTs and MRIs. 

Last month, the company developed ATMAN AI, an artificial intelligence (AI) algorithm that can detect the presence of COVID-19 disease in chest X-rays. The AI product is used for chest X-ray screening as a triaging tool in Covid-19 diagnosis, a method for rapid identification and assessment of lung involvement. The algorithm was developed in collaboration with the Centre for Artificial Intelligence and Robotics (CAIR), part of Defence Research Development Organisation. 

A new lens on radiology 

Even without a raging pandemic, radiologists are hard come by in India. According to a 2014 report by National Center for Biotechnology Information, US National Library of Medicine, India has approximately one radiologist for every 100,000 people compared to the US where the corresponding ratio is 1:10000. 

What’s even more alarming is that the country has only one interventional radiology (IR) expert per 2.18 lakh population, according to a Hindustan Times report from 2018 that cited professor Shiv Kumar of the department of radio-diagnosis, Sanjay Gandhi Post Graduate Institute of Medical Sciences. 

As opposed to radiologists who diagnose and treat a medical condition using radiologic imaging, IR experts are specialists who use minimally invasive image-guided procedures.   

Radiology is a vast field and radiologists are increasingly shying away from specialising in modalities such as CT or magnetic resonance (MR), choosing organ system specialisation instead. With fewer generalists in the field and increased workloads, access to quality diagnostic imaging interpretation services has deteriorated. This is where 5C’s solutions come in.

The company primarily targets hospital and diagnostic centres in tier II and III cities that do not have ready access to general radiologists, let alone specialists in the field. For patients in remote areas, the platform from 5C works like an on-demand radiology department.

However, 5C’s solution isn’t limited to smaller cities and towns. Today, the company works with some of the largest hospital chains in India including Apollo and ASTER as well as standalone X-ray centres. The platform currently works with close to 1,000 hospitals and diagnostic centres across 27 states in India, has a network of 400 daily active radiologists who process about 7,500 cases every day.

5C also helps hospitals cut down on expenses. Offered on a pay-per-case basis, the service helps hospitals and diagnostic centres save anywhere between 23-40% on costs incurred when compared to working with in-house radiologists, Sivasailam claimed.

AI powered diagnostics 

AI capabilities are applied throughout the platform, from a triaging tool that helps detect and flag scans for specific pathologies to reporting verification mechanism that runs the completed reports against an AI models to verify absence of any additional pathology. It also offers a workflow optimizer that detects and flags bleeds or clinically relevant artifacts to report clinical cases in a timely manner, and a quality assurance tool where AI can help digital health providers flag and cross check reports with high probability of error. 

The process begins with technicians uploading patient information on a system called Scribe. Sivasailam said the interface helps technicians reduce upload time by 70%. 

“Every case is then automatically routed to the optimum radiologist through a system called Maxflow which considered several factors including TAT, radiologist’s expertise and more. A tool named Sigma flags cases with high error probability to the QA team and SID generates completed report based on inputs from AI,” he explained. 

It has deployed algorithms for each specialty. ‘Chester’ checks for abnormalities in chest x-rays, ‘Cerebro’ checks for brain scan, ‘CAL’ for fractures in hand, and ‘Bacon’ tracks and identifies disease incidences across geographies. 

Sivasailam said the company is growing at “over significant double-digit month-on-month”, on-boarding close to 70-80 clients every month, and reading over five cases a minute. He claimed the platform has directly led to over 80,000 fast interventions and did a little over 2 million reports over the years. 

5C has so far raised a total of $1.9 million from a slew of investors such as Unitus Ventures, Axilor Ventures, and the Centre for Innovation Incubation and Entrepreneurship (CIIE), which comes under the Indian Institute of Management.

“If we are able to build a horizontal diagnostics layer in India, we can potentially be the holder of diagnostics data for over 100 million Indians which can help us build a valuable digital asset,” Sivasailam said. 

Sivasailam has previously co-founded Healthizen, a grievance redressal app of the Health Department, Government of Karnataka. The platform focused on preventive healthcare through public amenities.