V. Hingot, Ph.D., CTO of Resolve Stroke.
The digital transformation has reshaped numerous industries, making software ubiquitous. Traditionally, software and hardware operated separately, with software development appealing to a wider range of talents and resources due to faster cycles, quicker deployments, and lower infrastructure investments. This division has caused hardware-centered fields, such as manufacturing and healthcare, to lag in innovation and agility.
Complementarity of Hardware and Software
In Software-Defined Everything, Thomas Renaudin highlights a crucial limitation of software-only technologies: they require a physical interface for user interaction and data collection. The success of major software companies, including Google and OpenAI, relies on the widespread availability of hardware like smartphones and computers. Without these hardware mediums, their reach and impact would be significantly diminished.
This synergy between hardware and software depends on two key factors. First, a mature hardware ecosystem is essential to support integration and deployment through accessible APIs, enabling seamless interaction between the two. Second, the software must deliver enough end-user value to justify this investment, encouraging deeper integration with the hardware.
A Virtuous Synergy
A hardware object serves not only as a medium for user interaction and data presentation but also as the entry point for acquiring new data. In today’s data-driven landscape, hardware providers play a crucial role in acquiring and consolidating databases. Software technologies rely on these datasets to generate value for end users by creating new endpoints, automating measurements, and streamlining workflows. Ultimately, the synergy between hardware and software can yield results that far exceed their individual contributions.
This establishes a virtuous cycle: as more data is collected, the overall value increases, leading to greater adoption, which, in turn, drives further data collection. This “data river”, described by Alan Cohen and Zachary Bogue in The TechMed movement drives a constant stream of updates with each new input improving the system’s performance. This creates a win-win scenario where hardware providers are incentivized to grant access to their APIs, as it can boost their revenue.
Integration: Challenges and Opportunities
In healthcare, integration remains a significant challenge, as many manufacturers are hesitant to grant broader API access due to concerns over intellectual property and cultural skepticism regarding the potential value of software. However, healthcare data holds vast potential for generating new use cases, especially when accessible to third-party software for analysis.
Recently, software driven healthcare companies, such as Gleamer, Viz.AI, or Rapid.AI, have bypassed manufacturers and acquired databases directly from hospitals. This approach has allowed for the development of over $1 billion market segment specifically for medical image analysis, demonstrating the medical and economic value that software technologies can bring.
Software-Defined Ultrasound
Unlike other medical imaging technologies, ultrasound is a versatile modality that accommodates a wide range of sensors, personalized settings, and post-processing, with a field of view often limited to a single 2D frame. This limitation has historically made examinations highly dependent on user skill and has limited the creation of large and homogeneous databases, slowing the advancement of image analysis technologies.
For ultrasound, unlocking access to image analysis technologies and AI requires two key aspects. First, reducing user dependence by shifting to volume imaging, which captures entire organs rather than just selected planes. Second, working with raw data instead of processed images, as processing strips away valuable information and adds a layer of proprietary modifications from manufacturers.
The good news is that hardware has finally caught up. Today’s 3D probes and systems can capture and process hundreds of gigabytes of raw data in real-time at very high volume rates, all on compact and affordable scanners. Furthermore, most manufacturers are transitioning to software-based architectures, equipping their systems with increased data throughput and storage capabilities while also embedding GPUs for advanced data processing.
The Potential to Create New Use Cases
Ultrasound is inherently safe and accessible compared to many other medical imaging technologies. It does not require special shielding from radiation or magnetic fields, nor does it demand dedicated infrastructure, allowing it to be performed by individuals with minimal training. Ultrasound is widely used across nearly all areas of modern medicine and is valued for its simplicity and ease of use.
Despite its widespread availability, many unmet medical needs persist, often due to complex workflows, a lack of dedicated technology, or a combination of both. These challenges can be addressed with Software-Defined Ultrasound.
A recent example of this is the democratization of echocardiography. Traditionally, obtaining valuable insights required extensive training and high-end systems. However, AI-based solutions like Caption Health and DIA Imaging have transformed the workflow by guiding users in positioning, automating measurements, and creating a new paradigm for using echocardiography at the point of care.
So What’s Next?
The rise of programmable research platforms, along with a strong OEM ecosystem and newcomers like Butterfly and Clarius providing access to their APIs, makes it easier to develop new software technologies and explore innovative applications. This creates opportunities in areas where clinical use cases exist but where current technology is lacking. For example, in neurocritical care, regular checks of cerebral blood flow are important but can be complex, as they rely on indirect methods that often require CT scan confirmation.
As technology becomes more accessible, new applications are bound to emerge. Key focus areas include neurotechnologies, which are thriving due to an aging population and limited options for frequent check-ups. Other important areas are those where traditional imaging is difficult, such as in perioperative and postoperative situations or pediatrics. Additionally, functional studies for metabolic diseases and cancers present exciting opportunities for growth and progress in the field.
An Exciting Time to Work in Ultrasound!
—
Our Vision at Resolve Stroke
At Resolve Stroke, we aim to lead the Software-Defined Ultrasound movement by developing software engines with high medical and economic value. Our goal is to identify significant unmet medical needs that ultrasound imaging could address but currently cannot due to existing technological and workflow limitations.
Our focus is on building a diverse portfolio of applications, particularly within the promising neurotech sector. We are inherently open to collaborating with hardware manufacturers on a case-by-case basis, aiming to rapidly provide access to our technologies to a wide audience.