Merlin: An AI Tool that Enhances Radiology Workflow for Follow-Up Examinations

Radiology reports often contain follow-up recommendations for further diagnosis and to assist in the evaluation of potentially serious diseases. Over 35% of follow-up imaging recommendations fail to be scheduled.¹  If these recommendations are not tracked in a timely manner, or lost track of entirely, they can damage doctor–patient relationships and result in delayed diagnosis and poor patient outcomes. 

To help minimize these kinds of errors, 1QBit has developed and validated Merlin, a quality-assurance and AI-based tool that uses state-of-the-art natural language understanding (NLU) methods. Merlin was showcased at the 106th Scientific Assembly and Annual Meeting of the Radiological Society of North America (RSNA).

Figure 1: Representation of how Merlin is incorporated into a radiologist’s workflow to automatically detect follow-up recommendations based on what its natural language understanding engine interprets from a medical report.

Merlin extracts data from a radiologist’s report and processes it through an NLU engine to automatically detect follow-up recommendations and generate the details of any recommended follow-up exams. Integrated with the Exa Platform from Konica Minolta, radiology clinics and departments can seamlessly incorporate Merlin into their workflow to automatically pre-order the appropriate follow-up exam for review within a scheduler’s workflow. 

“Many times in clinical practice, we see patients with bad outcomes because of missed follow-ups. Having a robust, intelligent, scalable solution that health providers can use to automate follow-up can ensure that patients do not fall through the cracks. Thanks to Merlin, we can be diligent and precise and keep patients safe.” — Dr. Deepak Kaura, Chief Medical Officer, 1QBit

Reducing errors in the communication and identification of follow-up recommendations can result in better levels of service to referring clinicians. Merlin combines binary classification and entity-extraction models to detect whether a radiologist has recommended a follow-up, and, if so, recognizes the type of follow-up exam, the part of the body concerned, and the recommended date for the exam. Merlin has been validated in its ability to perform this task.

Dr. Deepak Kaura, MD, FRCPC, MBA, is a pediatric interventional radiologist and Chief Medical Officer at 1QBit. As a preeminent figure in the industry, he has led the development and clinical validation of several machine learning products for radiology at 1QBit. Dr. Kaura says about Merlin, “Many times in clinical practice, we see patients with bad outcomes because of missed follow-ups. Having a robust, intelligent, scalable solution that health providers can use to automate follow-up can ensure that patients do not fall through the cracks. Thanks to Merlin, we can be diligent and precise and keep patients safe.”

Deepak Kaura, Chief Medical Officer, 1QBit

While radiologists regularly issue follow-up recommendations, many patients do not return for follow-up, causing them to remain at risk of adverse outcomes due to a delayed diagnosis. Fortunately, Merlin can assist radiologists in tracking follow-ups and helping to improve patient outcomes. It demonstrates that AI has the potential to transform how care is delivered and can support improvements in the radiology workflow and, ultimately, the patient–doctor experience. 

View the Merlin brochure. Visit 1QBit Health Care and follow us on social media.

References

¹ Callen, J. L., Westbrook, J. I., Georgiou, A., and Li, J., “Failure to follow-up test results for ambulatory patients: a systematic review”, Journal of General Internal Medicine, (2011).

Other Blog Articles

Using Natural Language Processing to Enhance Radiology

Natural language processing (NLP) is a family of techniques that often use machine learning to analyze, process, and generate natural language text. It enables computers to process human language and derive meaning from natural language. For example, NLP-driven...

Understanding Winter Trends in Natural Gas Futures

Rising prices? Complacent market? The Market Sentiment Meter can help you look more deeply. Large-scale events tend to affect the futures market. For example, demand for natural gas is affected by the weather. Demand is also affected by the patterns in people’s...

Everything but the Quantum Kitchen Sink

Improving Machine Learning Using Analog Quantum Computing Artificial intelligence is set to become a $126.0B industry by 2025.¹  Machine learning is a major part of artificial intelligence involving algorithms that improve in performance over time given more data....

Andrew Milne, from Physics to Fintech

Andrew Milne immigrated from England to Canada as a child. He grew up partly in Toronto and partly in Vancouver. He began his studies in Computing Science at Simon Fraser University but later transferred to the University of British Columbia (UBC) to obtain his...