• Ashrika Gaikwad

Teleradiology, Enabled by AI

Teleradiology is one of the most successful components of telemedicine. It uses various technologies such as Digital Imaging and Communications in Medicine (DICOM), Picture Archiving and Communication System (PACS), and Radiology Information System (RIS) to transfer clinical data. With access to the internet and recent advancements in cloud computing and data mining technologies, imaging experts can provide diagnostic image interpretation services remotely.



Why Teleradiology


Access to quality and timely diagnosis and treatment is limited not just in rural areas, but also in urban cities. For example, in the UK, a radiologist takes 21 days on average to submit an MRI report.

India has only one radiologist for every 100,000 people. This shortage of professionals has created a market for teleradiology supported by artificial intelligence, which can solve the non-availability issue, making consultation through experts available even in remote areas. Teleradiology is poised to see explosive growth in the next few years, now that the COVID-19 pandemic and advancements in technologies like AI are encouraging the digitization of telemedicine services. Teleradiology can ensure 24x7 coverage and pave the way for timely diagnosis and triage.



Value-Add of DeepTek's Teleradiology Services


Even if teleradiology ensures access to quality radiology services to all sections of society, the problem of shortage of experts remains. The demand for radiologists has grown in the last few years, but the availability of expertise has not kept up. AI adds value to teleradiology in precisely this area. It assists radiologists by minimizing human errors in diagnosis caused by burnouts and maintains workflow efficiency by automating several chunks of the clinical workflow. In most cases, AI halves the reporting time. AI-based teleradiology can ensure all sections of society have access to quality radiology services.



Modalities Covered: X-ray, CT, and MRI

DeepTek’s Teleradiology Services supports X-ray, CT, and MRI modalities. The platform supports scans in the widely used DICOM format.



Human Experts-in-the-Loop - the Best Experts and a Team of Quality Assurance Radiologists

Medical imaging data from different locations have different distributions, and it is impractical to train and retrain a machine learning model on every new dataset it encounters to ensure it performs well across the board. Thus, AI models in deployment often face setbacks in maintaining performance and fail to generalize. We side-track this problem by keeping an expert-in-the-loop. These experts review the AI outputs, and the AI learns from this feedback in a continuous fine-tuning loop on all new datasets it encounters. The feedback loop helps make the model more robust and adaptable. Our teleradiology service also supports quality assurance checks of reports by radiologists.


Instant and Detailed Structured Reports, Smart Notifications, and Dashboard

The AI-powered workflow and decision support platform automatically triages and prioritizes studies. It presents the high-priority studies requiring immediate attention from the radiologist first on a clean dashboard. Other features include instant detailed and structured reports compliant with international standards, smart notifications to follow up with patients, and an analytics dashboard to visualize workload. Radiologists can quickly review and approve the AI-generated report.

The platform connects to PACS (picture archiving and communication system) and RIS (radiology information system), which are integrated into the teleradiology system, to ensure standardized communication of findings to stakeholders.


Reports generated using the smart and structured radiology reporting tool


The Advantage to Hospital and CIO - Cost-Efficacy and No Overheads/Upfront Costs

No upfront costs and overheads are required. Hospitals can outsource scans from overworked radiologists to save the cost of hiring additional experts, and they can increase capacity without investing in expensive technologies. Hospitals can easily consult subspecialists with expertise in specific modalities and sub-fields for higher-quality second opinions. This way, radiology services remain available 24x7, throughout the year.

The Advantage to Radiologists - Better Remuneration, Minimal Efforts, Fewer Burnouts

Trained radiologists report dozens of scans daily which can be exhausting. Using AI solutions integrated into teleradiology can allow radiologists to report more scans and get better remuneration, with minimal efforts and errors. This can potentially reduce burnouts.

The Advantage to Small Clinics - Instant Access to a Digital Setup and AI with No Extra Costs

IT requirements to set up teleradiology are now less complex because of the long strides made in cloud-based technology. The scalable teleradiology platform can be set up at small clinics to expand their capacity without burdening them with expensive equipment. Small hospitals or clinics need not worry about acquiring a server or an IT team to support this activity; it is easy to scale directly to the cloud. The clinics get access to state-of-the-art smart structured reporting tools and the entire bouquet of services enabled by full-scale digitization without much effort.

Availability of Subspecialty-Trained Experts Subspecialty trained experts for neuroimaging, body imaging, women’s imaging need not wait to visit the imaging clinic or for the imaging clinic to share the images physically; this is now possible on a simple-to-use platform that makes reporting completely virtual and seamless and gives access to cutting edge tools at a click of a button.

Data Transfer and Storage Security

Data privacy, security, and confidentiality concerns often accompany PACS/RIS teleradiology practices. We make use of a machine learning technique called federated learning (a distributed learning method) that preserves the privacy of raw medical data. Raw data never leaves the imaging center; instead, the machine learning model communicates for training to the center and back via secure servers and private cloud.

Onboarding Clinics, Hospital or Imaging Center for Teleradiology

The onboarding process is quick and takes around 20 minutes. It starts with the implementation team connecting with the imaging unit and accessing the local area network using a team viewer and installing a gateway app. Following this, the center can share images for teleradiology. A small video presentation educates the team about the methodology to upload history and clinical documentation correctly. We also share other documentation needed to formalize the MoU and begin the process of reporting.



The complete teleradiology solution automates several processes to help save radiologists’ time and effort and make radiology services more accessible. Ultimately, the digitization and automation telemedicine offer will bring in transparency, trust, accessibility, accountability, and democratization.
10 views

DeepTek’s mission is to make quality radiology services more affordable and accessible by leveraging the power of AI.

It is amongst very few Radiology AI companies which have successfully adopted its technology in a commercial mode – creating clear and quantifiable value for patients, hospitals, and radiologists. Currently, it is servicing over 50 hospitals and imaging centers and helping governments (India and APAC region) in their Tuberculosis, COVID, and other public health screening programs.

© 2017-2020 DeepTek.ai, Inc