Predictive Analysis – The Next Evolution for AI in Medical Imaging

Published 23/05/2019

Cranfield, UK, 23rd May 2019 – Written by Dr. Sanjay M. Parekh – As mentioned in our last insight (Unlocking the Value of AI in Medical Imaging) one of the key challenges for companies developing AI solutions for medical imaging is demonstrating their clinical value, and the return on investment for healthcare providers.

Detect-Only

Initially, the approach of many companies in this field was to automate detection of anomalies for the attention of the radiologist. This has improved the radiology workflow, improving productivity and reducing the number of miss and missed diagnoses.

Although detect-only solutions are certainly beneficial, it is becoming increasingly clear that AI can offer much greater benefits, both to the radiology department and the wider healthcare institution.

Detect + Quantify

More recently, there was been an increase in the number of AI solutions providing quantification in addition to detection. These solutions offer automated measurements of lesions, including size, volume, and texture, reducing the time consuming task of obtaining these measurements manually. This further enhances radiologists’ productivity and provides more abundant and accurate data for the diagnostic decision-making process.

Automated quantification is also an enabler of precision medicine, which uses patient data and imaging biomarkers to personalize treatment plans and improve patient outcomes. This data-driven approach enables more efficient and cost-effective diagnostic and treatment pathways, potentially reducing costs for the healthcare provider.

These newer AI solutions which detect and quantify have a stronger value proposition and offer a potentially greater return on investment for healthcare providers, extending beyond the radiology department, further enhancing the value of AI in medical imaging.

Detect + Quantify + Predict

The next stage in the evolution of AI solutions for medical imaging is expected to be the addition of predictive analysis, in addition to detect and/or quantify. These solutions offer real-time diagnostic decision support, such as likelihood of malignancy (following lesion/nodule detection and quantification), or a differential diagnosis (based on a confidence score), leading to earlier detection and diagnosis.

Such solutions will enable radiologists to prioritize their focus on cases of greater clinical need and to determine with greater confidence whether follow-up invasive procedures (for example, biopsy) are required. Although the final diagnosis will require the radiologist’s own judgement, predictive analysis will greatly enhance the value of AI solutions to the radiologist.

A broader, and potentially greater, benefit of predictive analysis is risk stratification. This is a population-health management approach, which groups patients according to their risk of disease. At-risk patients are identified earlier and then placed on the appropriate clinical pathway, which improves their clinical outcome, and improves cost-savings for the healthcare provider.

Summary

With over 150 companies developing AI solutions for medical imaging, across all modalities and applications (organs or body regions), there is no shortage of options for radiologists to choose from. However, this nascent market is very crowded, and companies need to differentiate their software solutions to ensure they offer ‘something more’ for the radiologist. At the same time, healthcare providers need to justify their investment in medical imaging AI.

While there is certainly a market opportunity for solutions that detect anomalies, and arguably a bigger opportunity for solutions that both detect and quantify, looking forward we expect that the predictive capabilities of AI will unlock its full potential in medical imaging. Solutions that provide value beyond the radiology department will make the greatest impact on both the healthcare provider’s and software developer’s bottom lines.

About Signify Research

Signify Research is an independent supplier of market intelligence and consultancy to the global healthcare technology industry. Our major coverage areas are Healthcare IT, Medical Imaging and Digital Health. Our clients include technology vendors, healthcare providers and payers, management consultants and investors. Signify Research is headquartered in Cranfield, UK.

For further information, please contact:

Sanjay.Parekh@signifyresearch.net

+44 1234 436150