THE HEALTH CARE TEAM
Providing high-quality care with limited resources is one of the most important challenges a clinician faces. As a global leader in quantum software and AI, 1QBit is uniquely positioned to provide advanced solutions and superior results. Our seasoned group of health care experts, machine learning scientists, and software developers are passionate about creating tools that empower frontline care workers with scalable intelligence. By integrating leading-edge technology and rigorous science in health care, we are determined to establish a new standard on how diagnostics and treatment are performed worldwide, and bring high-quality care to everyone, everywhere.
1QBit is working with partners from around the world to deploy XrAI and support them in their fight against COVID-19 and other lung diseases. XrAI is the first product in 1QBit’s suite of machine learning health care solutions. Our long-term aim is to empower physicians in all imaging and lab diagnostics with intuitively designed and technically advanced tools that enhance their uniquely human ability for critical thinking and optimize their use of time and resources.
WHAT IS XrAI?
XrAI is a quality-control and clinical-decision support tool that uses deep learning to expedite the way frontline care workers identify lung-related abnormalities by highlighting regions of abnormality in chest X-ray images. This tool was developed as a “co-pilot” for clinicians to improve the accuracy and timeliness of diagnosing lung and pleural abnormalities on chest radiographs.
The solution empowers clinicians by providing immediate analysis of a chest radiograph, enhancing their ability to form an accurate diagnosis at the time treatment is prescribed.
XrAI has been validated in clinical trials in Canada and is the first AI tool for radiology to be approved by Health Canada as a Class III Certified Medical Device. The landmark clinical trial was dramatically accelerated and prioritized by the health care community in response to the COVID-19 pandemic. After seeing the impact this tool has for optimizing patient care, health care organizations across Canada and internationally mobilized to get XrAI into the hands of clinicians.
HOW XrAI WORKS
Rather than employing a traditional black box system, XrAI provides physicians with transparency into the conclusions made by the underlying algorithm and an intuitive interface to view accurate results and inform their final diagnoses.
The confidence dial, which allows clinicians to increase or decrease the confidence level in the identified potential abnormalities, is unique to XrAI and has been widely praised by users as a critical feature that is intuitive to the way radiologists share their opinions with other medical professionals.
XrAI integrates into existing clinical information systems with no workflow disruption and little to no training required. The results of XrAI’s analysis are made available within a clinician’s standard X-ray viewer, displaying the machine learning algorithm’s findings along with its level of confidence in the provided results.
XrAI was built using diverse data with extensibility and applicability across Canada and around the world. The algorithm was trained on 250,000 radiology cases taken from more than 500,000 anonymized radiograph images from Canadian health care organizations and open subscription–based datasets, covering different medical conditions across diverse populations.
The technology was tested and validated in an independent, single-blind, randomized control clinical trial. The study results confirm that XrAI improves the quality of the clinical interpretation of chest X-rays by up to 20% for the participating physician groups: radiologists, emergency doctors, family doctors, pulmonologists, and radiology residents.
WHAT IS MERLIN?
Follow-up management can be overwhelming. Losing track of follow-up recommendations is not only dangerous for patient outcomes—it can damage the important relationship between a health care provider and a patient. 1QBit has developed and validated an easy, intuitive way to minimize these kinds of errors using state-of-the-art natural language understanding (NLU) methods. Merlin is a quality-assurance and workflow enhancement tool that automatically detects recommendations for follow-ups in radiologists’ reports.
Combining binary classification and entity-extraction models for follow-up identification, Merlin detects whether a radiologist 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 interpret proposed follow-up durations and detect and understand recommended exam types. Merlin is now available for purchase and deployment, ready to improve clinical workflow.
Merlin has capabilities to directly integrate into Radiology Information Systems (RIS). This has been shown with Konica Minolta’s EXA RIS-PACS integration, resulting in dramatic improvements in workflow. Radiology reports, once generated by a radiologist, are instantaneously routed to Merlin, where an analysis is performed in search of follow-up recommendations. If such a recommendation is detected, a scheduling queue is automatically populated with the follow-up details, including the modality, body type, and time frame. This is an important step toward enhancing the quality of care that radiology practices can provide to their patients and referring clinicians.
NLU IN HEALTH CARE IMPROVES OUTCOMES
Natural language understanding uses neural net–based models for training and understanding the context of the text being processed. Natural language understanding models are able to analyze and understand an entire radiology report and produce an accurate result that will help clinicians provide world-class care and improve patient outcomes. Negations, uncertainty, and irrelevant summarization are not a problem for Merlin, which identifies recommendations in the report and presents findings using standardized labels. Merlin, a trusted partner in radiology, ensures that patients receive the follow-up care they need by identifying whether a follow-up is recommended and extracting the relevant information.
Have Any Questions?
For more information on implementing XrAI into your workflow or about our other health efforts, please contact us.