HOPPR demonstrates an AI-powered multimodal basic model for medical imaging
HOPPR is a technology company developing a multimodal basic model for medical imaging. The company is backed by Health2047, the Silicon Valley venture studio powered by the American Medical Association.
Today at HIMSS24, HOPPR will demonstrate this artificial intelligence-powered model – which can provide diagnostic, clinical and operational insights from medical imaging data – in AWS’s booth 1561 in the South Hall.
We interviewed HOPPR CEO Dr. Khan Siddiqui to gain a better understanding of this basic model and what it could mean for healthcare.
Q. You discuss your development of a basic multimodal model for medical imaging at HIMSS24. Please tell conference participants what that is.
A. HOPPR develops a comprehensive generative model for all medical imaging modalities. Let’s see what that means.
The HOPPR base model is available through an API service and can be used by developers, radiological image archiving and communication systems (PACS) vendors and AI companies to develop their own applications.
Rather than spending a year or more developing AI algorithms from scratch, application developers can fine-tune HOPPR’s pre-trained Large Vision model to meet their exact specifications and compress the development process into about a month. In this way, the HOPPR model significantly shortens the time to market for much-needed AI-powered medical imaging applications.
The HOPPR base model is multimodal and built on large amounts of high-quality medical data from various sources, including images from computed tomography and magnetic resonance imaging scans, X-rays and other medical imaging scans. The model can be used to create products that dramatically improve the way radiologists, technicians and support staff interact with medical images.
Applications built on the HOPPR core model promise to improve the experience of both patients and physicians. For example, imagine you are a neurosurgeon planning an aneurysm coil procedure. Treatment planning for these procedures is highly dependent on medical imaging.
Applications powered by HOPPR’s API service can help fully characterize the aneurysm and predict which vascular approach, catheter type, and coil choice will be most conducive to success. Statistics, previous procedures with similar anatomy, and published research can be called upon as necessary to further support decision making.
But the benefits don’t end with treatment planning. Once a plan is finalized, scheduling and allocating sufficient OR time, staffing, and materials can be automated to help maximize results and cost-efficiency.
AI holds tremendous promise to advance medical imaging and improve patient outcomes. HOPPR wants to unlock that potential by enabling developers to bring AI applications to market faster.
Q. You are supported by Health2047 and UHG. Please tell us about these organizations and why their support is important to this development.
A. In November 2023, HOPPR announced a landmark investment from Health2047, the Silicon Valley venture studio powered by the American Medical Association and founded to overcome systemic dysfunction in US healthcare. Their goal is to have a meaningful and measurable impact on healthcare before the American Medical Association’s 200th anniversary in 2047.
Health2047 is transforming healthcare at a system level and seeks powerful ideas, industry partners and entrepreneurs to address systemic transformation in data, chronic disease and productivity.
Health2047’s deep relationships with both the AMA and its network of strategic partners create a unique force multiplier that helps drive informed, large-scale change in healthcare.
Q. You’ll talk to attendees about how you say HOPPR enables end users to unlock diagnostic, clinical and operational value from medical imaging data. Explain how this is done and what the expected results are.
A. The model can provide diagnostic, clinical and operational insights from medical imaging data.
Diagnostic value: At scale, HOPPR’s base model will be trained on more than a petabyte of medical imaging research data, allowing it to distinguish nuances and detect abnormalities beyond the capabilities of the human eye. While many current AI tools are developed by downsampling the grayscale to 256 shades, HOPPR sees 65,000 shades of gray. HOPPR has developed proprietary vision transformers to train its model on full resolution, full bit depth images, ensuring no loss of useful data in the images.
Clinical value: The model can be refined for use in applications that allow users to interact with medical imaging studies about findings, alternative imaging views, proposed surgical procedures, and treatment protocols. With mammography, a patient can be informed at the point of care – before leaving the imaging center – whether additional diagnostic imaging is needed, reducing patient stress, streamlining care delivery and improving clinical outcomes.
Operational value: From an operational perspective, a possible application of the technology is to automatically complete radiology reports. Much of what we dictate in radiology is repetitive, and research has shown that doing manual work in addition to analyzing images causes cognitive dissonance. Using AI to automate some elements of the work – for example pre-filling a preliminary report that the doctor can complete after review – would create more time for image review and improve the doctor’s experience.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.