Augmento ROP is a smart cloud-based platform for radiologists to evaluate and report X-ray, CT, and MRI scans. By intelligently optimizing radiology workflows, Augmento reduces radiologist effort, improves productivity, and reduces reporting times. This ultimately leads to higher-quality reports and lower costs.
Pathologies that Augmento's AI can detect from radiology images.
Augmento is a comprehensive solution. Its AI models can accurately diagnose over 20 different pathologies. These models pre-screen all studies and present their results to the radiologist who can approve or modify them at the click of a button. A standardized high-quality report is instantly generated and made available to the radiologist for a final sign-off. The AI organizes all cases to be reported as a smart worklist that presents critical cases first. This ensures prompt service to those who need attention the most. The analytics dashboard provides a visual overview of cases reported, patient demographics, imaging modalities, etc. The inbuilt notification system supports sending push notifications to physicians, administrators, and patients.
AI-enabled radiology solutions require large volumes of diverse and well-annotated medical imaging data. While hospitals often hold retrospective radiology data, its annotation usually requires substantial manual intervention by radiologists. Augmento not only supports daily radiology workflow but also simultaneously and automatically creates annotated data in the background.
In our experiments, a group of radiologists assisted by Augmento demonstrated an 86% improvement in productivity and a 46% reduction in reporting times over the control group not using Augmento.
The DeepTek Solution Architecture
Augmento comes with two deployment options that make up its on-ground component.
Cloud: The platform is rendered through a secure cloud service such that it can receive scans from different geographical locations. Radiologists and administrators in different locations can log into the platform to access its services. Augmento is currently in use in more than sixty hospitals and medical imaging centres. A gateway software ensures that images are shared with the platform. It can queue, push, or pull after discussion with the stakeholders. All the analysis and pre-screening happens on the cloud, and the results are sent to various stakeholders. The smart notification system updates and alerts healthcare professionals as well as patients. Being a cloud platform, it's scalable at a moment’s notice and highly elastic with negligible downtime. It gives way to a truly virtualized, AI-infused radiology department.
On-Premise: A few large hospitals and imaging clinics typically opt for the on-premise version. It has the same features as the cloud version but acts within a fixed perimeter of a given local area network with the server hosted locally.
Augmento’s second component is the AI housed in the cloud. Two models work behind the scenes. The horizontal taskbar at the bottom of the page houses the results from the first one, a classification model. This model gives classification labels according to what the AI suspects. The second is a localization model, which is responsible for box-like localisations on each image.
The platform is versatile and caters to various users such as radiologist, technologist, front end receptionist, physician, floor managers, administrators, and nodal officers. Users get controlled administrative privileges to monitor and use specific areas of the platform. These can be customized to adhere to hospital security requirements and rules set by regulatory authorities.
Performance Audit and Tuning by a Unique Expert-in-the-Loop
In our workflow, expert radiologists audit the predictions generated by our AI models. Incorrect predictions are collected and feedback to the algorithms is ingrained in the system through a feedback loop. This ensures that the models not only keep improving but also evolve and adapt to changes in the input data. The Augmento platform gets better and better with use and gets customized to the users based on their preferences.
Report Critical Studies Instantly
Radiologists can access critical studies first, as Augmento prioritizes studies almost magically using AI. Radiologists can sort and search for studies by pathologies predicted by AI and criticality.
Once the radiologist agrees with a finding predicted by the AI, the platform generates a complete radiology report in compliance with RSNA/ACR guidelines. The radiologist can modify and regenerate an incomplete report using the tools inbuilt in the platform as a microservice. It is as if the platform understands beforehand what the imaging expert is going to pen down in the report. A report edited and approved by a radiologist is sent directly to the hospital.
Quick compare allows comparison of two or more studies of the same patient. This comparison, which typically takes 3 to 5 minutes, can be condensed into 30 seconds to 1 minute by Augmento.
Augmento automatically visualizes data to give insights on patient demographics and location, image modalities, AI predictions, and radiologist workload.
The geolocation map below visualizes areas with patients diagnosed with community-acquired diseases like COVID-19 and tuberculosis. This map acts as ground truth to confirm the disease burden in a particular region and can help make decisions on containment zones.
The Report tab has a green check mark for a reported case and a red cross for an unreported case. It also displays four other icons. Users can preview, view, download the report, and share it with a clinic or referring physician using the other buttons.
The yellow triangles indicate the triage status of each study. The radiologist has the option to raise alerts, which manifest as blinking yellow triangles. Treating physicians, floor administrators, and even patients are notified automatically of these alerts via secure chat/email or SMS.
Gender Conflict Alert
The gender alert feature helps avoid gender mismatch in delivered reports. During review and reporting, if a radiologist accidentally uses a male template in a female patient’s report, a pop-up alerts the radiologist to go back and correct the issue before submitting the report.