Today in New Delhi, India
Oct 14, 2018-Sunday
-°C
New Delhi
  • Humidity
    -
  • Wind
    -

#startupsaturday: Pune’s DeepTek has AI on the job in medical rooms across the country

Apart from prioritising scans in a diagnostic centre, DeepTek algorithm also helps search the database for similar infections, so if a patient is affected with lung infection it will search the entire database for similar cases in the past. It also acts like a prediction tool for the radiologist. It searches for relevant cases and compares it to the report in hand predicting whether this particular case has a 50% or 80% chance of being a lung infection. This improves the efficiency of the doctor by about 30%.

pune Updated: Oct 14, 2018 14:03 IST
Namita Shibad
Namita Shibad
Hindustan Times, Pune
Pune,Artificial intelligence,Money
(Left to Right) Dr Amit Kharat, Aniruddh Pant and Ajit Patil, co-founders of DeepTek.(HT FILE PHOTO)

Given that there is a huge proliferation of machines that read your body parts and an equally big shortage of people mandated to read these reports, DeepTek decided to do something about it. “We thought why not use artificial intelligence to help the radiologist? Typically when a patient goes to a diagnostic centre or a hospital, there are a lot of scanned reports waiting to be read. Say a patient has pneumonia; his X-ray may be placed 64th in the scheme of things, but he needs urgent attention. What our algorithm does is it takes all filed scans are prioritises those that require urgent attention.” says Dr Amit Kharat one of the founders.

Dr Amit Kharat , radiologist said, “the use of MRIs and CT scans has dramatically improved, worldwide.While it helps doctors get better and more accurate diagnosis, we do not have enough radiologists to help decipher the scripts. The current workload has profoundly increased for radiologists”.

SMARTER, FASTER, BETTER
With the use of algorithms to quicken the process, are we looking at theonset of the medical revolution?
CONCEPT
A software designed to read X-rays, MRI reports and bring urgent cases to the doctor’s attention hence leading to quicker diagnosis.
FUNCTION
1. Decipher medical transcripts
2. Prioritise cases by indentifying those that require immediate attention.
3. Comparing similar current to cases to past ones draw up diagnosis and treatment.
4. Reduce workload on radiologist.
FUNDING
1. Capital invested by co-founders
2. Received funding from NTT Data software company
Artificial intelligence, also known as AI, is a software application that runs automated tasks (scripts). It typically, performs tasks that are both simple and structurally repetitive, at a much higher rate than would be possible for a human alone.

This problem later gave birth to DeepTek.ai Pvt Ltd , the brain child of three friends - Dr Amit Kharat, Ajit Patil and Aniruddh Pant. Says Dr Kharat, “Radiology reports should not be sent to consultants as there is legal binding requiring it to be read by a radiologist, however, during time of emergencies this is a general practise.”

Despite legal requirements Dr Kharat says that there are thousands of reports that go ‘unreported’, that is not signed off by a radiologist.

What does DeepTek do?

Apart from prioritising scans in a diagnostic centre, DeepTek algorithm also helps search the database for similar infections, so if a patient is affected with lung infection it will search the entire database for similar cases in the past. It also acts like a prediction tool for the radiologist. It searches for relevant casesand compares it to the report in hand predicting whether this particular case has a 50% or 80% chance of being a lung infection. This improves the efficiency of the doctor by about 30%.”

What role can DeepTek play in emergencies?

Says Dr Kharat, “First of all it releases the doctor from routine tasks. Say a brain scan shows a tumour. DeepTek will not just be able to prioritise the scan, but also give measurements and density of the image. In the case of a haemorrhage,not only will it give information as to which part of the brain has been affected, but also the size so the radiologist has to only ratify it. In case there is no radiologist available our tool can help the doctor basic information.”

So will it replace the radiologist?

Says Dr Kharat, “No. It cannot replace the radiologist. AI technologies have a long way to go. It needs much more historical and diverse data to get refined. Our tool is a second eye for the radiologist. It aims to help an already overburdened doctor. Moreover it informs the hospital about a particular case.”

Building the AI tool

When the trio got together to build this tool they had their jobs clearly marked out. Says Dr. Kharat, “Since I am a radiologist, I provided the domain expertise. Anirudh has a data science background and has worked on algorithms for medical science in the past, so he gave the tech backbone to our project. Ajit manages the operations and human resources. We have 10 people on board right now. Since we cannot afford to pay them what MNCs do, we offer them ESOPs (employee stock ownership plans). Also working on a patented technology can do wonders to your CV. So far we have a good team and have managed to achieve quite a lot in the past year.”

Funding

At the start DeepTek began with founder’s money. Dr Kharat refuses to disclose the details. They are currently filing a patent which will be clearer by the end of this month. A month ago they realized they needed more funds to get to the next level. This is when they received money from NTT Data, a software company. Says Dr Kharat,“In Japan, the government is pushing for ai in different sectors. Moreover their usage of imaging technologies is far greater than the entire world put together. Their shortfall of radiologists is acute and our tool is a solution.” Dr Kharat would not disclose how much money they received.”

Going to market

Perhaps the biggest challenge to a startup is selling their wares. DeepTek has recognized the shortfall of radiologists in the medical field, but how do you sell this to them? Dr Kharat, “Our plan is to approach hospitals and diagnostic centres across the world. We are working with our funding partners on approaching several medical centres across the globe. The fact that the radiology sector has always been very proactive about new technology and radiology machines across the world speak the same language is an added bonus for us.

They are looking to get proof of concept deals with different centres across the world. Again it’s a secret. “We will tell you once we have signed up for PoCs.” What about giants like Siemens, Philips and GE, leaders in imaging technologies and surely they would have worked on AI tools?

Says Dr Kharat, “Yes of course. Also these MNCs have the wherewithal and power that a startup like ours does not. They can hire the best people in the world but I think we score in agility. We can quickly move in a new area, say spines, unlike a giant like GE. Our size and agility is our strength.”

First Published: Oct 13, 2018 14:57 IST