How Imidex uses V7 to Detect Pulmonary Nodules, Indicative of Lung Cancer
“We conducted an extensive research of annotation tools and ultimately chose V7 as it fits our needs best due to its customizable workflows, simple setup, consensus stages, and the ability to monitor annotator performance and project progress."
IMIDEX revolutionizes the way medical professionals detect subtle lung nodules in chest x-rays with cutting-edge AI technology. With a vast amount of data and strict medical regulations to navigate, IMIDEX needed a tool that would not only meet FDA and HIPAA standards, but streamline their process—allowing them to build custom data workflow with consensus stages that deliver maximum accuracy, ensuring highest quality ground truth is obtained with unparalleled rigor and efficiency.
IMIDEX uses V7 to annotate chest x-rays and create custom workflows with a consensus stage to train or test the model with the aid of multiple radiologists. With V7's user-friendly interface, advanced data management, and annotation tracking, as well as support from their dedicated customer success team, V7 streamlines the process of preparing training and validation data for their AI solution, making it more efficient and manageable.
With V7 workflows, it’s very easy to see the project's status - I know how far along the labelers are. I know what’s in review and what’s completed. We can see all status changes happening in real-time. That is probably my favorite V7 feature.
About Imidex
IMIDEX is on a mission to save lives lost to lung cancer by leveraging the power of AI. The team of experts in data science, radiology, quality, data curation, and regulatory compliance have created VisiRad XR, the revolutionary software that helps radiologist identify lung nodules in chest x-rays with machine learning accuracy.
Lung nodules, often missed in routing care, are early indicators of lung cancer. VisiRad XR aims to close the gap in diagnosis and treatment, providing hospitals with the tools to catch cancer at an earlier stage when outcomes are much better.
IMIDEX’s Clinical Integrity Team, made up of US board certified radiologists, annotate and label thousands of DICOM images in V7, ensuring the highest possible quality of annotations and representations of the intended-use population. V7’s custom data workflows, including a consensus state with blind reads between multiple annotators, results in improved model accuracy and faster ground truth creation.
Visit www.imidex.com for more information.
Annotating more than 9000 DICOM Images on V7
In order to develop a highly sophisticated computer vision model, IMIDEX must accurately label a vast number of chest x-rays. Their team of in-house radiologists is dedicated to identifying and annotating nodules, utilizing tags for image classification and marking image visibility or additional information associated with specific files.
IMIDEX operates with large batches of data, at times reaching 9,000 files per dataset, which are uploaded and managed on V7. Once all data is labeled and reviewed, it is utilized to train Imidex's model, VisiRad XR.
To ensure the highest possible quality of annotations, IMIDEX utilizes V7 to create custom data workflows tailored to specific projects. These workflows often include a consensus stage, where two radiologists perform a blind read, annotating the same images simultaneously without viewing each other's work. Any discrepancies are then resolved by an independent reviewer.
Easier blind reads and intuitive data workflow setup with V7
The IMIDEX team needed a training data platform that would let them easily set up custom workflows, make pixel-perfect annotations, perform blind reads, handle huge data volumes, and keep track of the annotation progress.
The team chose V7 primarily to improve their model's accuracy by implementing custom workflows and consensus stage blind reads.
V7's workflow system provided a clear, high-level view of the annotation process, aiding in project management. It allowed the team to create ground truth data with more ease. The system's stability and intuitive interface, along with handy keyboard shortcuts for easier navigation also contributed to increased efficiency.
Furthermore, IMIDEX has been really satisfied with V7's customer support, with any emerging issues solved within 24 hours.
V7 is very intuitive and the stability of the platform has been really helpful for our team. It allowed us to perform blind reads, on top of other things, such as setting up different workflows tailored to specific projects, generating reports, and getting full visibility into the project progress at every stage and at any time.
Faster Ground Truth creation and improved model’s accuracy
By using V7 for data labeling and quality assurance, IMIDEX was able to speed up the process of improving the accuracy of their computer vision model - VisiRad XR. This means that the team will be able bring their AI-powered system to the healthcare industry sooner, contributing to better patient outcomes, and ultimately, more lives saved.
Thanks to V7, the IMIDEX team also gained more time that they can allocate to working on other projects. In 2023, they're planning to expand their efforts by acquiring more data and annotating new medical formats, such as CT scans.
Overall, V7 has been a game changer for IMIDEX, streamlining their workflows, increasing efficiency, and laying the foundation for future projects that have the potential to revolutionize the healthcare industry through the use of AI.
Discover how other AI-first companies solved knowledge tasks at scale with V7
