Ross Upton, CEO and Academic co-founder, Ultromics, shares his insight with Hospital Hub…
With recent advancements in artificial intelligence (AI), the MedTech sector is developing new diagnostic techniques and processes. In fact, medical diagnostics is already labelled as one of the most important artificial intelligence (AI) trends of 2019.
Numbers from Accenture indicate that technological innovation in this area could create value of $40bn in robot-assisted surgery, $20bn in virtual nursing solutions, $16bn in reducing dosage errors, and $13bn in terms of identifying prospects for clinical trials. These developments alone would generate potential healthcare value to the tune of $150bn – and other areas are rapidly emerging.
AI has particularly intriguing applications in the field of diagnostics, representing one of the first stages of the care delivery process and perhaps its most critical. A general rule of thumb is with earlier detection comes the potential for easier treatment, as well as an increase in the patient’s prospects of survival and recovery. AI-empowered diagnostics offer the exciting possibility to facilitate more accurate diagnoses and more meaningful, early interventions.
The medical imaging industry stands to benefit hugely from AI. In terms of heart disease, echocardiograms, traditionally read with the naked eye, can now be analysed at a much deeper level. A clinician assisted by AI can diagnose coronary artery disease with much greater accuracy than if they relied on manual techniques alone.
Scanning for disease
The manual analysis of an echocardiogram is subject to human error. A trained clinician will review the scan, look for indicators of heart disease and then make a judgement on whether the patient is at risk of, or suffering from heart disease. The problem is that not all factors that suggest heart disease in an echocardiogram are visible to the naked eye – no matter how well-trained the clinician is.
What this means is that misdiagnosis is a common occurrence. In fact, one in five heart disease patients are incorrectly diagnosed and undergo unnecessary – not to mention costly – treatments. Or worse, there are patients with heart disease who are sent home untreated. Unfortunately, the tragic result of undiscovered illnesses is often that patients suffer at a later stage.
AI-driven diagnostic support tools can help address this problem: technology such as Ultromics’ EchoGo* uses the whole echocardiogram image to build a complex dataset describing the heart. The AI then kicks in, looking for patterns in the data that it knows are associated with heart disease, having seen thousands of previous scans – this means it can process images in a way that goes far beyond the current subjective assessment by a clinician.
Diagnosing the future
A 100 percent diagnostic success rate will likely remain an impossible dream: you can’t always account for anomalies in biology, and heart disease can develop in unexpected ways, so a single image may not capture what happens a few months later. But even if total accuracy remains elusive, there is still the potential for future improvements.
As clinical datasets become larger and more integrated, the possible next step for AI diagnostics will likely be about finding smarter ways to link together holistic patient data. Combining multiple data sources and tracking how the heart changes over time may allow more precise diagnostics and personalised treatment plans, but these approaches are some way off. To deliver this vision, it will be vital to build resources of clinical information to learn from to help boost diagnostic accuracy even further.
Diagnostic success saves lives and money
With all the recent developments in AI, it’s important to remember that a better diagnosis is nothing without clinician input to offer a prognosis and treatment plan. Where the technology works most effectively will be in augmenting and aiding physicians, not eliminating them from the diagnostic pathway.
Nonetheless, common sense and the facts both indicate that improving diagnostic accuracy can save money, time, and most importantly, lives. The more innovation in this area, the better.
*Pending FDA clearance, not available for sale in the United States.
Visit Ultromics to find out more.