Early detection not only saves lives, but it also improves the length and quality of life that people can benefit from. In the UK, this amount of quality-adjusted life year (QALY) is calculated by estimating the years of life remaining for a patient following a particular treatment or intervention and weighted with the ability to carry out the activities of daily life free from pain and constraint.&nbsp;&nbsp;&nbsp;

Artificial Intelligence (AI) has the power to detect anatomical anomalies fast, supporting clinicians with streamlining and redesigning diagnostic pathways, to assure patients the highest standard of care and better life outcomes.&nbsp;

AI to optimise lung cancer pathways&nbsp;

A recent roundtable discussion, chaired by Professor Sir Mike Richards, Chair of the UK National Screening Committee brought together experts to discuss the <a href="https://newmarket-strategy.com/wp-content/uploads/2023/12/Transformation_of_the_lung_cancer_pathway.pdf" target="_blank">transformation of the lung cancer pathway in the NHS</a>, highlighting the crucial role of AI and other digital health innovations to redesign healthcare. It suggested that new technologies should be seen as an opportunity to streamline existing processes and championed solutions that detect lung nodules from chest X-rays or determine malignancy risk or monitor growth via CT, particularly when used in partnership with Cancer Alliances and NHS Trusts. This approach improves early detection rates and streamlines the diagnostic process, reducing the burden on healthcare professionals.&nbsp;

Many high-profile UK studies are now looking into improving the diagnostic pathway for lung cancer:&nbsp;

Furthermore, a <a href="https://www.qure.ai/news_press_coverages/major-milestone-met-in-ai-powered-fight-against-lung-cancer" target="_blank">pioneering partnership between Qure and AstraZeneca</a> has made considerable progress in the fight against lung cancer on a global scale. This collaboration is testament to the transformative potential of AI, innovating beyond traditional lung cancer screening profiles and improving patient outcomes on a large scale. It's a glimpse into the future of healthcare.&nbsp;&nbsp;

The latest figures suggest 2 million people in 28 countries have been screened using Qure's AI-powered chest X-ray analysis for the incidental detection of high-risk lung nodules. This proactive lung health AI commitment casts the net wider, beyond traditional lung cancer screening based on age or smoking history profiles. The initiative has found over 220,000 lung nodules, identifying close to 3,000 potential early lung cancer cases.&nbsp;&nbsp;

AI to create triage efficiencies for neurocritical CT&nbsp;

AI can also help healthcare professionals make timely decisions and ensure that patients with Traumatic Head Injuries (TBI) or suspected Strokes receive prompt attention. This is particularly important in the context of triage in acute or frontline emergency care where high patient numbers and resource shortages can impact timely patient care.&nbsp;&nbsp;

The ACcEPT (Assess the Clinical Effectiveness in Prioritising CT Heads) study commenced during winter 2023/24 at NHS Greater Glasgow and Clyde, the first of four planned NHS sites across the UK, utilising <a href="https://www.qure.ai/product/qer" target="_blank">qER, Qure’s head CT AI</a>. In the first initial weeks, qER analysed 651 non-contrast head CTs, detecting 128 head injuries including cranial fractures, intracranial haemorrhage (ICH), mass effect and mid-line shift within the brain. The intracranial pathologies were identified and prioritised by the qER AI solution as critical findings for radiologist review, confirmation, treatment planning and / or discharge out of the Emergency Department.&nbsp;

Tackling NHS acute imaging workloads through teleradiology has also been powered by AI to redesign processes and flow. Medica, a leading UK teleradiology provider, uses qER as its <a href="https://www.qure.ai/impact_stories/medica-qure-partnership" target="_blank">24/7 emergency reporting diagnostic tool to support the delivery of acute care</a> in many NHS Trusts.&nbsp;&nbsp;

Redesigning healthcare for better patient outcomes&nbsp;

Optimising lung cancer pathways and creating triage efficiencies for neurocritical CT with the power of diagnostic AI will help redesign care and treatment pathways. The key to AI activation and deployment will be the cross collaboration of stakeholders to share the challenges, champion the successes and steer the NHS ecosystems safely and securely into the future of improved patient care.&nbsp;&nbsp;

Meet Qure.ai this summer to hear more about how AI can redesign healthcare. We’ll be at <a href="https://www.ukio.org.uk/" target="_blank">UKIO</a> in Liverpool, 10-12 June 2024, stand B8/9 and at the <a href="https://www.nhsconfedexpo.org/" target="_blank">NHS ConfedExpo</a> in Manchester, 11-12 June 2024 stand D5a.&nbsp;

large_Optimising LC and Neuro.webp

medium_Optimising LC and Neuro.webp

small_Optimising LC and Neuro.webp

thumbnail_Optimising LC and Neuro.webp

Optimising LC and Neuro.webp

Optimising Lung Cancer Pathways and Neurocritical CT Triage with AI  

The advent of Large Language Models (LLMs) in healthcare represents a paradigm shift in medical informatics, diagnosis, and patient care. These models, trained on extensive datasets, offer unprecedented capabilities in understanding and processing complex medical information.&nbsp;&nbsp;

The fine-tuning of such models for specific diseases like lung cancer exemplifies a bespoke approach to healthcare analytics, ensuring high accuracy and relevancy in their outputs. LLMs hold great potential for enhancing the occupational state of healthcare workers and by extension, the quality of care that patients receive.

Foundational Models and Fine-Tuning in Healthcare&nbsp;

Foundational models serve as the bedrock for specialized LLMs in healthcare. These models are trained on a vast corpus of data, encompassing a broad spectrum of knowledge. This comprehensive training, as suggested by <a href="https://arxiv.org/abs/2302.14691" target="_blank">research from Cornell University</a>, equips them with a general understanding of language and concepts, making them versatile in various applications. However, the versatility of foundational models is a double-edged sword. While they possess extensive knowledge, the breadth of their training can sometimes lead to diluted expertise in specific domains, such as medical specializations. This is where the concept of fine-tuning becomes paramount.&nbsp;

Fine-tuning involves the adaptation of a foundational LLM to a specific domain or task. In healthcare, this means training the model on datasets highly relevant to medical conditions or areas of practice. For instance, a model fine-tuned for lung cancer would focus on data related to lung cancer symptoms, treatments, epidemiology, and related research.&nbsp;&nbsp;

By narrowing the focus, these specialized models become more adept at handling queries and providing information related to their expertise. The benefit is twofold: first, the model retains more relevant knowledge without being distracted by unrelated data; second, it reduces the likelihood of generating false or irrelevant information, a critical factor in healthcare applications.&nbsp;

Expanding Roles of LLMs in Medical Fields&nbsp;

The role of LLMs in healthcare is multifaceted and continually evolving.  LLMs can assist healthcare professionals in diagnosis by providing rapid, data-driven insights. They can analyze patient data, compare it with vast medical records, and suggest possible diagnoses or flag potential issues requiring further investigation. This capability is precious in areas with limited access to specialist medical knowledge or where healthcare systems are overwhelmed.&nbsp;

These models can recommend tailored treatment plans by analyzing a patient's medical history, genetic information, and current health status. This approach could lead to more effective treatments with fewer side effects, as the recommendations are based on a comprehensive analysis of relevant data. LLMs also hold significant promise in medical research, particularly in understanding complex diseases and developing new treatments. By processing and analyzing vast amounts of research data, these models could identify patterns, propose hypotheses, and could even in the future predict the efficacy of new drugs. This could accelerate the pace of medical research, potentially leading to quicker discoveries and innovations in healthcare.&nbsp;

Another critical area where LLMs can contribute is in public health. Models trained on epidemiological data could aid in predicting disease outbreaks, understanding transmission patterns, and formulating public health strategies. This is particularly relevant during global health crises, such as the COVID-19 pandemic, where timely and data-driven responses are crucial.&nbsp;

LLMs: Bridging Gaps and Setting New Standards in Healthcare&nbsp;

According to the journal <a href="https://www.nature.com/articles/s43856-023-00370-1" target="_blank">Nature</a>, LLMs have demonstrated considerable potential in enhancing patient care through their ability to augment core medical competencies. Notably, these models possess substantial semantic medical knowledge, enabling them to engage in medical reasoning. This capability is exemplified by their performance in medical licensing exams, indicating a proficiency that can be further enhanced through fine-tuning with specific medical data.&nbsp;&nbsp;

Furthermore, one of the transformative applications of LLMs in healthcare is improving communication between healthcare staff and patients. By simplifying complex medical jargon and facilitating translations, LLMs can bridge language barriers and make medical knowledge more accessible. This is particularly significant for conditions associated with social stigma, where digital tools like chatbots can offer discreet and informative assistance. The ongoing advancement of LLMs promises to overcome current technical constraints and limited acceptance by healthcare practitioners.&nbsp;

Another critical role of LLMs is in easing the documentation burden on clinicians. By converting unstructured notes into structured formats, LLMs can streamline administrative tasks, allowing healthcare providers to devote more time to patient care. This functionality can extend to clinical trials, where accurate and efficient documentation is crucial. In medical research, LLMs can democratize access to scientific evidence while presenting risks of misinformation and facilitating scientific misconduct. Their ability to summarize and critically appraise up-to-date, high-quality, peer-reviewed evidence can significantly aid researchers. However, the quality of these summaries depends on the training data.&nbsp;&nbsp;

The integration of LLMs in medical education heralds a shift in teaching methodologies. While they can serve as valuable tools in providing customized educational content and facilitating interactive learning experiences, there is a risk of undermining critical thinking skills.&nbsp; Integrating LLMs into healthcare is an exciting development, heralding a new era of innovation and efficiency in medical practice. These models offer remarkable potential to enhance medical knowledge, facilitate communication, and optimize administrative processes. Their application spans various aspects of healthcare, promising significant improvements in patient care, research, and education.&nbsp;

This technological advancement represents a substantial leap forward in making healthcare more accessible and practical. With their sophisticated data processing and analysis capabilities, LLMs are poised to play a crucial role in the future of medicine. As we continue to explore and harness the capabilities of LLMs, the prospects for their contribution to healthcare are boundless. This innovative time holds great promise for improving the quality of patient care and advancing medical knowledge, setting a new standard for healthcare in the digital age.&nbsp;

large_image (15).png

medium_image (15).png

small_image (15).png

thumbnail_image (15).png

image (15).png

Role of Large Language Models (LLMs) in Healthcare

<a href="https://www.auntminnie.com/imaging-informatics/artificial-intelligence/article/15623418/6-ways-ai-will-enhance-musculoskeletal-imaging" target="_blank">Musculoskeletal (MSK) radiology</a> is rapidly evolving, with artificial intelligence (AI) playing an increasingly important role. AI-powered tools can <a href="https://jmai.amegroups.org/article/view/5742/html#:~:text=AI%20and%20ML%20advances%20for,by%20potentially%20improving%20patient%20outcomes." target="_blank">improve the efficiency, accuracy, and accessibility of MSK imaging</a> while providing new insights into diagnosing and treating MSK conditions.&nbsp;

<a href="https://mss-ijmsr.com/artificial-intelligence-in-musculoskeletal-radiology-past-present-and-future/#:~:text=Based%20on%20the%20studies%20discussed,this%20field%20%5BTable%201%5D." target="_blank">One of the most essential benefits of AI in MSK radiology</a> is its ability to improve diagnostic accuracy.&nbsp;&nbsp;

AI is also making MSK imaging more accessible to patients. Radiologists can use AI to develop new imaging protocols that reduce scan time and radiation exposure. This is particularly important for patients who require frequent imaging, such as those with chronic MSK conditions.&nbsp;

The relationship between AI and MSK radiology is anchored in <a href="https://pubmed.ncbi.nlm.nih.gov/33983500/#:~:text=Artificial%20intelligence%20and%20deep%20learning,Keywords%3A" target="_blank">myriad applications that span</a> image reconstruction, tissue segmentation, workflow augmentation, and disease detection​.&nbsp;

The following outlines the significant avenues through which AI is sculpting the future of MSK radiology.&nbsp;

AI, especially deep learning, has catalyzed innovation in osteoarthritis applications. It has propelled advancements in image classification, lesion detection, cartilage segmentation, and more.

One of the <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7416509/#:~:text=Magnetic%20resonance%20imaging%20,for%20accelerated%20MRI" target="_blank">significant drawbacks</a> of MRI as a diagnostic tool has been the lengthy data acquisition time. However, AI has addressed this challenge head-on by accelerating MRI acquisition and reconstruction, thus reducing the long acquisition times that were previously a bottleneck.&nbsp;&nbsp;

This acceleration is crucial for large-scale longitudinal studies and emergency diagnostics where time is of the essence​​.&nbsp;

The 3D nature and soft-tissue contrast of MRI, which are indispensable for osteoarthritis research, are being <a href="https://www.nature.com/articles/s41584-021-00719-7#:~:text=The%203D%20nature%20and%20soft,The%20recent" target="_blank">significantly enhanced</a> with the application of AI. AI facilitates the elucidation of disease pathogenesis and progression, making MRI an invaluable tool in osteoarthritis research. This enhancement in imaging quality provides a clearer view of the anatomical structures and the pathological changes occurring in osteoarthritis​.&nbsp;

AI enables the automated detection of anatomically relevant health markers at the hip, knee, and ankle. This facilitates radiological monitoring of the progression of <a href="https://link.springer.com/referenceworkentry/10.1007/978-3-030-64573-1_292#:~:text=AI%20relies%20on%20deep%20learning,as%20osteoarthritis%20of%20the%20knee" target="_blank">various bone diseases</a>, such as knee osteoarthritis, thus providing a longitudinal view​.&nbsp;

The application of AI, particularly machine learning and deep learning, is known to improve multiple stages of MRI, including acquisition, processing, and post-processing steps. This, in turn, minimizes the chances of diagnostic errors and enhances the overall efficacy and efficiency of osteoarthritis research​​.&nbsp;

Enriching Disease Understanding and Diagnosis&nbsp;

The <a href="https://pubs.rsna.org/doi/full/10.1148/ryai.2019180057" target="_blank">automated models</a> developed through AI have shown promise in staging knee osteoarthritis severity from radiographs, where the performance of these models has been comparable to that of musculoskeletal radiologists.&nbsp;

This automation ensures high diagnostic accuracy, critical for determining the appropriate treatment interventions.&nbsp;

Using AI and machine learning algorithms, <a href="https://www.nih.gov/news-events/nih-research-matters/machine-learning-detects-early-signs-osteoarthritis" target="_blank">researchers</a> have identified early signs of osteoarthritis three years before diagnosis through MRI scans. This capability is essential for understanding the disease's developmental trajectory and identifying individuals who would likely benefit from early interventions.&nbsp;

AI improves the overall quality and efficiency of imaging, which is crucial for a more accurate diagnosis. It aids in assessing the appropriateness of imaging orders, thereby improving image quality, patient centricity, imaging efficiency, and, eventually, diagnostic accuracy.&nbsp;&nbsp;

Particularly in cases of advanced knee osteoarthritis requiring surgical intervention like Total Knee Arthroplasty (TKA), <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7296289/" target="_blank">AI and machine learning modeling</a> serve as indispensable tools in pre-surgical planning, patient selection, prediction of disease progression, and estimation of treatment outcomes. This is critical for tailoring the surgical approach to the individual needs and conditions of the patient, ensuring optimal post-surgical results.&nbsp;

<a href="https://www.semanticscholar.org/paper/Studying-osteoarthritis-with-artificial-to-applied-Kovacs/2a6c6e9b3a1e8e7eae930f0f432c0e8d5a14d097" target="_blank">Deep learning methods</a> have been employed to improve the understanding of osteoarthritis.&nbsp;&nbsp;

AI transcends conventional diagnostic paradigms by facilitating a more comprehensive understanding of osteoarthritis by discovering novel image features indicative of short-term and long-term progression. This multi-faceted insight is crucial for a more patient-specific prognostic approach, ensuring the diagnostic deliberations are nuanced and rooted in a deeper understanding of the disease.&nbsp;

Enhancing Workflow and Clinical Operations&nbsp;

AI is identified as a pivotal tool that could <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7881801/" target="_blank">significantly affect every step</a> in the imaging value chain. The early integration of AI into radiology has already demonstrated its potential value, particularly in MSK imaging. The journey begins with placing the imaging requisition, traversing the entire imaging cycle, and ensuring well-oiled operational machinery.&nbsp;

The benefits of AI extend to workflow support, which is paramount for optimizing the efficiency and productivity of radiological practices. AI and deep learning offer multiple avenues of workflow support, including image reconstruction and transformation, tissue segmentation, and disease detection, which are intrinsic to <a href="https://www.semanticscholar.org/paper/Deep-learning-and-its-applications-to-machine-Pelt-Goh/eaed97bd8e46646dcbf35d32a98a4009c9a1d5e1" target="_blank">MSK radiology</a>.&nbsp;

Specifically, in spine imaging—a crucial part of MSK radiology—the incorporation of AI has led to significant advancements. Innovations rooted in AI and big data have considerably impacted musculoskeletal radiology, hinting at a future where a holistic approach could redefine every element of the <a href="https://pubs.rsna.org/doi/10.1148/ryai.2019180057" target="_blank">radiology service value chain, from request to report</a>.&nbsp;

From the broader radiology perspective, AI applications are being developed to span all steps of the imaging chain, from test ordering to report communication.&nbsp;

AI applications' traction across the radiology space underscores a promising horizon where workflow and delivery are redefined, thereby catalyzing a more efficient and streamlined operational framework in <a href="https://pubs.rsna.org/doi/10.1148/ryai.2019180057" target="_blank">MSK radiology</a>.&nbsp;

Amidst the evolving landscape of MSK radiology, the emergence of specialized AI tools drives nuanced advancements in diagnostic accuracy and operational efficiency.&nbsp;&nbsp;

A notable example technologically in this trend is Qure.ai's <a href="https://www.qure.ai/product/qmsk" target="_blank">qMSK software</a>, engineered to augment the detection and localization of bone fractures and joint dislocations across multiple body regions on MSK X-rays, delivering insights in under a minute​.&nbsp;

The comprehensive nature of qMSK extends to accommodating 15 anatomical views, thereby fostering a holistic understanding of potential musculoskeletal anomalies. This breadth of analysis is executed with a processing time of less than 20 seconds, testifying to the software's computational alacrity. Having been trained on over one million scans, qMSK exudes a sensitivity exceeding 0.9 in detecting signs of fractures, underscoring its diagnostic precision​​.&nbsp;

Moreover, qMSK is engineered to transcend conventional diagnostic thresholds by mitigating the 'satisfaction of search' bias - a phenomenon where the identification of one anomaly may overshadow or inhibit the detection of others.&nbsp;&nbsp;

By facilitating the recognition of multiple fractures, qMSK amplifies the diagnostic gaze, fostering a more comprehensive understanding of the patient's condition.&nbsp;

This capability is further enhanced by qMSK's proficiency in working across multiple views, thereby reducing diagnostic errors associated with overlooking subtle fractures​​.&nbsp;

The clinical journey of a patient in MSK radiology often necessitates a longitudinal examination of their condition, particularly in the context of fracture healing.&nbsp;

qMSK is tailored to track the progress in fracture healing, thus providing a dynamic diagnostic narrative pivotal for personalized patient care. Furthermore, the software significantly curtails the processing time to under 20 seconds per scan, translating to substantial time savings and augmented workflow efficiency​when compounded over numerous cases.&nbsp;

In a domain where every second can be critical, qMSK's deployment flexibility resonates profoundly.&nbsp;

The software can be deployed on the cloud or on-premises, offering a versatile solution that aligns with the varied infrastructural and operational frameworks across different healthcare settings​​.&nbsp;

The articulation of qMSK's capabilities reflects a broader narrative of how AI bridges the technical and operational difficulties in MSK radiology by enhancing diagnostic acuity, streamlining the workflow, and catalyzing a more patient-centric approach.&nbsp;

Regulatory Recognition and Trajectory Forward&nbsp;

The institutional recognition of AI's efficacy is highlighted by the increasing number of AI products being cleared by the Food and Drug Administration (FDA) for utilization in clinical radiology, encompassing tasks such as quantification, identification, and even diagnosis.&nbsp;

This legitimizes AI's role in clinical radiology and hints at a trajectory of deeper integration shortly, with a <a href="https://pubmed.ncbi.nlm.nih.gov/37087138/#:~:text=There%20is%20a%20rapidly%20increasing,in%20neuroimaging%20and%20musculoskeletal%20radiology" target="_blank">broader spectrum of AI applications awaiting regulatory approval for clinical deployment</a>​​.&nbsp;

Integrating AI in MSK radiology represents an enhancement and a fundamental transformation in diagnosing and treating musculoskeletal disorders.&nbsp;&nbsp;

This transformative alliance between AI and MSK radiology sets the stage for marked improvements in diagnostic precision, operational fluidity, and patient-centric care quality.&nbsp;

large_MSK radiology.webp

medium_MSK radiology.webp

small_MSK radiology.webp

thumbnail_MSK radiology.webp

MSK radiology.webp

The Importance of AI in MSK Radiology

Partnerships stand out as a driving force in the activation and deployment of Artificial Intelligence (AI) for next-generation radiology. The essence of forging close working relationships across industry, academia and hospitals lies in the shared vision of achieving more together, to deliver a truly integrated and efficient future. 

Leveraging tech partnerships for advanced health AI integration&nbsp;&nbsp;

The most critical level of partnership to deploy AI is technology collaborations. These alliances are instrumental in leveraging diverse expertise, fostering innovation, and ensuring that AI technologies are accessible, scalable, and tailored to meet the nuanced demands of radiology workforces.&nbsp;

At the very first level, seamless integration with modality equipment such as CT or X-ray and associated applications from Original Equipment Manufacturers (OEMs) is a must. This marriage between image capturing hardware and innovative AI software is the cornerstone of propelling diagnostic detection.&nbsp;

Then comes integration with other systems in the medical imaging ecosystem. This includes Picture Archiving and Communication Systems (PACS) to create workflow and elevate patient care across clinical specialties and wider Hospital Information Systems (HIS), Electronic Medical Records and Radiology Information Systems (RIS).&nbsp;&nbsp;

Furthermore, flexible on-premises and in-cloud deployment options of AI deployment tailor to the diverse infrastructure requirements at hospitals. This ensures enhanced accessibility and scalability for AI activation planning.&nbsp;&nbsp;

Digital agility in radiology is a collaborative process&nbsp;

As a vendor-agnostic AI solution with over 30 global partnerships with platform, OEM and wider third-party tech infrastructure providers, Qure's AI will dovetail neatly into other systems. This is also tested vigorously at integration stages to enhance, not impact, other IT infrastructure.&nbsp;

Recent integration in action examples include <a href="https://www.qure.ai/news_press_coverages/chest-x-ray-ai-live-at-university-hospitals-of-leicester-nhs-trust" target="_blank">University Hospitals of Leicester NHS Trust</a>. This UK project is using Qure.ai’s flagship <a href="https://www.qure.ai/product/qxr" target="_blank">qXR AI</a> integrating into the Trust's Radiology Information System (RIS). Close partnership working with the RIS provider has created a seamless AI install for Leicester’s radiologists, powering productivity and enabling smooth switching between Chest X-ray and CT reporting workflows.&nbsp;

To support AI deployment momentum across the UK and Europe, a host of other technical collaborations are growing in number. This includes inclusion in innovative <a href="https://www.qure.ai/news_press_coverages/qure-to-boost-global-healthcare-with-advanced-ai-integration" target="_blank">AI orchestrator platforms</a> and <a href="https://www.qure.ai/news_press_coverages/empowering-care-teams-deepc-teams-up-with-qure-ai-for-ai-enhanced-patient-care-workflows" target="_blank">radiology operating systems</a> to streamline diagnostic processes; collaborations with <a href="https://www.qure.ai/news_press_coverages/qure-and-xwave-technologies-announce-strategic-collaboration" target="_blank">clinical decision support</a> players to optimise and co-develop solutions that guide clinical referrers to order the most appropriate diagnostic test; distributors of <a href="https://www.qure.ai/news_press_coverages/qure-ai-and-therapixel-announce-partnership-to-improve-access-to-breast-cancer-detection-through-ai" target="_blank">complementary AI solutions</a>; and so on.&nbsp;

Automating healthcare processes to re-humanise patient care&nbsp;

As AI continues its coming of age in radiology around the world, the emphasis on strategic partnerships will drive activation. By creating an army of AI advocates and activators, technology, expertise, and vision will converge. This unleashes the full power of radiology AI into the clinical environment and onwards, delivering better, quicker patient care.&nbsp;&nbsp;

large_Power of partnerships blog - web creative.webp

medium_Power of partnerships blog - web creative.webp

small_Power of partnerships blog - web creative.webp

thumbnail_Power of partnerships blog - web creative.webp

Power of partnerships blog - web creative.webp

The Power of AI Partnerships  

According to the <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1071277/" target="_blank">Western Journal of Medicine</a>, snowboarding injury rates have significantly increased, ranging from 3.5 to 40 per 1,000 snowboarding days. &nbsp;

Pneumothorax, or collapsed lung, occurs when air collects in the pleural space, causing the lung to collapse. According to the <a href="https://jtd.amegroups.org/article/view/27617/21158" target="_blank">Journal of Thoracic Disease</a>, Primary Spontaneous Pneumothorax is common among healthy individuals aged 10-30, with an annual incidence of 7.4–18 (age-adjusted incidence) and 1.2–6 cases per 100,000 population, among men and women respectively.&nbsp;

This condition is particularly prevalent in high-impact sports like snowboarding –so much so that researchers from the Western Journal of Medicine say that “pneumothorax should be suspected in any snowboarder with blunt chest trauma, no matter how severe, who complains of pleuritic chest pain and dyspnea.”&nbsp;

Advancing trauma care in winter sports requires innovations designed to meet the exigencies of emergencies. AI emerges as a paramount tool, uniquely equipped to address these critical needs with precision and efficiency.&nbsp;

The Growing Challenge of Snowboarding Injuries

The Western Journal of Medicine emphasizes the need for rapid diagnosis in cases of blunt chest trauma, common in snowboarding accidents. The nature of these injuries, which often involve high speeds and hard impacts, makes timely and accurate detection of pneumothorax crucial for effective treatment.&nbsp;

The advent of Qure.ai's qXR --especially in detecting pneumothorax-- offers hope in this scenario.&nbsp;

Recently cleared by the FDA for clinical use, qXR is a testament to the advancement in medical imaging and AI. <a href="https://appliedradiology.com/Articles/qure-ai-nets-fda-clearance-for-ai-enabled-pneumothorax-and-pleural-effusion-triage-solution" target="_blank">Applied Radiology</a> reports that Qure's qXR can triage pneumothorax and pleural effusion, which pose severe challenges in emergency rooms and intensive care units.&nbsp;

The relevance of qXR becomes even more pronounced considering the nature of snowboarding injuries. According to the Western Journal of Medicine, the sport is associated with various injuries, including sprains, fractures, concussions, and crucially, pulmonary air leaks resulting from blunt chest trauma.&nbsp;&nbsp;

These injuries can often lead to pneumothorax, where air collects in the pleural space, causing the lung to collapse. The journal emphasizes the need for suspicion of pneumothorax in any snowboarder with blunt chest trauma, highlighting the criticality of rapid and accurate diagnosis.&nbsp;

The time needed to transfer patients to trauma centers may lead to dangerous delays. Implementing AI solutions like qXR in rural settings –as <a href="https://www.itnonline.com/content/minxray-sets-base-camp-mount-everest" target="_blank">Qure already does at Everest Base Camp in Nepal</a>-- could significantly improve outcomes for athletes injured in remote areas.&nbsp;

Looking ahead, continued research into the biomechanical forces involved in snowboarding crashes could allow innovations in protective gear to prevent chest trauma before it occurs.&nbsp;&nbsp;

For example, advanced airbag technologies are being explored to absorb impact when riders collide with the ground or terrain park fixtures. However, AI-assisted diagnostics remain essential for identifying trauma once it happens on the slopes.&nbsp;

qXR: An Advanced AI Device for Pneumothorax Detection&nbsp;

qXR demonstrates exemplary clinical accuracy in triaging pneumothorax, offering rapid passive notifications to healthcare professionals. qXR takes an average of just 10 seconds to alert medical staff. This feature is indispensable in critical settings like ICUs and emergency rooms.&nbsp;

qXR, using a globally sourced training dataset, integrates seamlessly into the existing standard of care workflow. It serves as a passive notification system for worklist prioritization, enhancing the efficiency of medical professionals.&nbsp;

Dr. Subba R. Digumarthy, senior thoracic radiologist at Massachusetts General Hospital, <a href="https://appliedradiology.com/Articles/qure-ai-nets-fda-clearance-for-ai-enabled-pneumothorax-and-pleural-effusion-triage-solution" target="_blank">reports up</a> to 96% sensitivity and 100% specificity for the qXR algorithm in a multicenter publication on missed and mislabeled chest radiography findings.&nbsp;

Despite these advancements, room for improvement stays. A study in the <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7028462/" target="_blank">Annals of Emergency Medicine</a> finds that the conventional use chest X-rays –that is, unassisted by artificial intelligence-- have limited sensitivity in diagnosing pneumothorax, missing over half of cases.&nbsp;&nbsp;

This highlights the need for continued innovation in radiographic technology to improve diagnostic accuracy further, making advancements like qXR even more salient for winter sports injuries.&nbsp;

Furthermore, Neil Roy, MD, MBA, FACEP, CPE, Chief Medical Officer at Shady Grove Medical Center, underscores the <a href="https://appliedradiology.com/Articles/qure-ai-nets-fda-clearance-for-ai-enabled-pneumothorax-and-pleural-effusion-triage-solution" target="_blank">transformative impact of qXR</a> in emergency medicine. The AI's ability to rapidly triage pneumothorax improves the speed and efficiency of diagnosis. It adds a layer of safety by quickly naming one of the most time-sensitive radiographic findings.&nbsp;

The FDA clearance of the qXR algorithm is a significant milestone in Qure's commitment to perfecting healthcare delivery in time-sensitive settings like the ICU and ER. This technology is designed to integrate seamlessly into current healthcare pathways, improving ICU &amp; ER workflows and hastening patient care.&nbsp;

Qure's qXR stands for a significant leap forward in managing chest trauma, particularly in winter sports like snowboarding. Its ability to detect pneumothorax streamlines the diagnostic process accurately and rapidly and enhances the overall efficacy of trauma care.&nbsp;

Moreover, spreading awareness of pneumothorax risks and symptoms among snowboarding communities can encourage prompt medical care after crashes. Often, athletes ignore warning signs of lung collapse, losing crucial time before seeking treatment. Educational initiatives targeting winter sports participants may promote proactive monitoring following high-velocity collisions on the slopes.&nbsp;

large_FEB24 - Advancing Trauma Care in Winter Sports.webp

medium_FEB24 - Advancing Trauma Care in Winter Sports.webp

small_FEB24 - Advancing Trauma Care in Winter Sports.webp

thumbnail_FEB24 - Advancing Trauma Care in Winter Sports.webp

FEB24 - Advancing Trauma Care in Winter Sports.webp

Advancing Trauma Care in Winter Sports

Accelerating the deployment of medical imaging Artificial Intelligence (AI) to diagnose patients quickly and support radiology workforces continues at pace.&nbsp;&nbsp;

Many NHS hospitals are considering radiology AI and are analysing individual challenges, building business cases and exploring AI solution options.&nbsp;&nbsp;&nbsp;

Advice &amp; guidance across the AI activation continuum&nbsp;

To simplify the AI journey, Qure’s cross-functional and UK-based team has produced a range of AI activation tools.&nbsp;&nbsp;

These supportive resources cover the entire radiology AI activation continuum, from guidance on initial planning and decision-making, to business case development, and support on physical deployment within the healthcare facility.&nbsp;&nbsp;

Latest additions to the roster of NHS planning assets include an AI Value Calculator to scenario plan the financial value deployments can deliver to departments, and an AI Deployment Guide answering key questions facing clinicians and wider deployment stakeholders.&nbsp;&nbsp;

large_Support Tools social.webp

medium_Support Tools social.webp

small_Support Tools social.webp

thumbnail_Support Tools social.webp

Support Tools social.webp

Support tools to simplify healthcare AI decision-making & activation in the UK

Insights from the Westchester Medical Center Case:

In the wake of a <a href="https://radiologybusiness.com/topics/healthcare-management/healthcare-policy/hospital-hook-120m-jury-verdict-after-residents-miss-signs-stroke-ct" target="_blank">$120 million malpractice verdict</a> against a hospital north of New York City, the healthcare sector stands at a critical juncture.&nbsp;

This case, resulting from a misinterpretation of stroke signs in CT images by resident physicians at Westchester Medical Center, underscores the critical need for more reliable diagnostic tools in medical imaging.&nbsp;

Such incidents highlight the challenges faced by medical professionals, especially in high-pressure environments where the likelihood of diagnostic errors increases.&nbsp;&nbsp;

This context is not an indictment of physicians' competence but rather a reflection of the strenuous circumstances under which they work and the limitations of current methodologies.&nbsp;

Radiological Diagnostic Errors: A Persistent Challenge in Need of Solutions&nbsp;

AI provides a reliable and accurate diagnostic tool, adeptly compensating for human limitations.&nbsp;

Research indicates that AI algorithms can reduce error rates in image interpretation. For instance, a study published in <a href="https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0000404" target="_blank">PLOS Digital Health</a> found that AI could identify and localize fractures on X-rays with an accuracy comparable to senior orthopedic surgeons.&nbsp;

While AI's role in enhancing diagnostic accuracy is clear, its integration into medical practice must be subtle yet impactful. Technology should not overshadow human expertise but rather complement it.&nbsp;&nbsp;

This symbiotic relationship between AI and human expertise can significantly reduce the time to diagnosis, which is crucial in stroke management, where every minute counts.&nbsp;

According to research from <a href="https://radiologykey.com/error-in-radiology/" target="_blank">Radiology Key</a>, error rates in radiologists' interpretations range from 3.5% to 4%, potentially leading to three or four errors per day per radiologist​​.&nbsp;

Moreover, research from the journal <a href="https://insightsimaging.springeropen.com/articles/10.1186/s13244-021-00986-8" target="_blank">Insights into Imaging</a> highlights various error types, including history-related errors, location-related errors, and satisfaction of search, where radiologists may overlook additional findings after identifying a primary issue​​.&nbsp;

AI offers a genuine solution in response to these challenges. Qure.ai, a front-runner in AI for medical imaging, has considerably enhanced diagnostic accuracy.&nbsp;

A <a href="https://www.qure.ai/evidence/performance-of-a-chest-radiography-ai-algorithm-for-detection-of-missed-or-mislabeled-findings-a-multicenter-study" target="_blank">comprehensive multi-center study</a> involving Qure's qXR algorithm revealed impressive results: the AI demonstrated high sensitivity (96%), specificity (100%), and accuracy (96%) in detecting overlooked or incorrectly labeled findings in chest X-rays (CXRs).&nbsp;&nbsp;

This study, which analyzed an extensive database of 13 million reports spanning 22 years from esteemed institutions like Massachusetts General Hospital and Brigham Women's Hospital, significantly highlights AI's capacity to reduce errors in radiographic findings.&nbsp;&nbsp;

Qure's study aimed to determine if its qXR algorithm could effectively identify these missed or mislabeled CXR findings. The outcome was promising: the AI successfully identified about 90% of the critical abnormalities in missed or mislabeled CXRs, recording zero false positives. T&nbsp;

These results demonstrate the high accuracy of Qure's qXR and underscore the profound potential of AI in revolutionizing radiology diagnostics, paving the way for more reliable and precise medical imaging practices in a landscape where radiologists need all the help they can get.&nbsp;

Case Studies: Demonstrating AI's Impact in Radiology&nbsp;

Qure's contributions extend beyond CXRs to other radiological domains.&nbsp;&nbsp;

In addressing the critical need for stroke diagnosis improvements, Qure's technology has shown efficacy in systematically measuring brain hemorrhages accurately and rapidly, setting the stage for AI to become a standard in stroke and trauma patient care​​​​.&nbsp;&nbsp;

These case studies demonstrate AI's ability to enhance diagnostic processes, offering an antidote to issues like those encountered in the Westchester Medical Center case.&nbsp;

Qure's impact in transforming healthcare through AI is exemplified in a case study involving their qXR system in the <a href="https://www.qure.ai/blog/scaling-up-tb-screening-with-ai-deploying-automated-x-ray-screening-in-remote-regions" target="_blank">Philippines' tuberculosis (TB) screening program</a>. The ACCESS TB project, utilizing mobile vans equipped with X-ray machines, aimed to screen high-risk individuals for TB across various cities.&nbsp;&nbsp;

Post qXR implementation, significant improvements were observed. The qXR system, a deep learning solution on the cloud, was integrated into the existing workflow to facilitate the rapid screening and diagnosis of TB.&nbsp;&nbsp;

The implementation of qXR notably enhanced the efficiency of the TB screening program. The technology significantly reduces the time and costs incurred in the diagnostic process by providing rapid, accurate TB screening results.&nbsp;&nbsp;

It reduced the reliance on conventional, time-consuming tests like sputum culture testing, thereby decreasing the risk of dropouts in report collection and treatment enrolment.&nbsp;&nbsp;

This case study demonstrates the transformative potential of AI in healthcare, particularly in areas with limited resources. It emphasizes Qure's commitment to making healthcare more accessible and efficient through innovative AI solutions​​.&nbsp;

The Westchester Medical Center incident is an important lesson regarding the limitations and risks inherent in conventional radiological practices. As exemplified by Qure's innovations and their practical use cases, AI presents a viable solution to these challenges, offering enhanced accuracy, reduced error rates, and faster diagnosis.&nbsp;&nbsp;

Integrating AI into radiological practices is pivotal in transforming patient care and setting new standards in medical diagnostics. The future of radiology lies in a synergistic relationship between AI and human expertise, delivering the highest level of patient care and safety.&nbsp;

large_Qure - Blog Post - Feb 5.webp

medium_Qure - Blog Post - Feb 5.webp

small_Qure - Blog Post - Feb 5.webp

thumbnail_Qure - Blog Post - Feb 5.webp

Qure - Blog Post - Feb 5.webp

The Imperative of AI for Improving Radiological Accuracy

Every day, countless people rise out of chairs or beds and wince at the pain in their lower backs. Lower back pain is considered the leading cause of disability in 160 countries. It is also an indicator of potential musculoskeletal (MSK) conditions.&nbsp;

In 2022, the World Health Organisation estimated that 21% of the world's population suffered from musculoskeletal ailments. MSK-related issues limit movement and flexibility and are a key factor in disability, early retirement, poorer well-being, and a diminished ability to contribute to or participate in societal activities.&nbsp;&nbsp;

Musculoskeletal fractures are a common occurrence in clinical practice, with over 2 million fractures occurring annually in the United States alone. Meanwhile, the UK, France, Germany, Italy, Spain, and Sweden (EU6 nations) see 2.7 million fractures taking place yearly. The currently associated healthcare expenditure is €37 billion; it is projected to increase to €47 billion (an increase of 23%) by 2030 and a staggering €77 billion by 2050.&nbsp;&nbsp;

Detecting musculoskeletal fractures is essential for providing appropriate treatment, as untreated fractures can lead to long-term disabilities. Diagnosing fractures typically involves X-rays, allowing doctors to visualize the bone and determine the extent of the injury. Emergency physicians must usually review MSK radiographs to determine a treatment approach before radiologists can review the images and provide a formal report.&nbsp;&nbsp;

Medical imaging has revolutionised healthcare by providing non-invasive diagnostic tools. But the shortage of radiologists worldwide, along with the growing number of studies and images per study, have proportionately increased the workloads of radiologists as well. It has resulted in the possibility of lower efficiency and a higher risk of errors that accompany burnout. The most common diagnostic mistakes made by clinicians interpreting musculoskeletal radiographs involve missed fractures. These errors cause treatment delays, avoidable medical costs, legal issues, and morbidity that could have been prevented.&nbsp;

This is where Artificial Intelligence has a role to play. AI algorithms can be trained to recognize patterns in X-ray images that may indicate the presence of a fracture. These algorithms can also be used to classify the type of fracture and predict the likelihood of it healing successfully. AI tools like qMSK by Qure.ai have the potential to improve the accuracy and efficiency of musculoskeletal imaging diagnosis, deliver imaging results to clinicians and provide emergency physicians with more accurate imaging information at the time of patient contact, improving patient care and triage.&nbsp;&nbsp;

It should be emphasized that the quality and quantity of the training data dramatically influence the accuracy of AI algorithms for fracture detection. Therefore, continuous efforts to collect and annotate large radiograph datasets are imperative to enhance the precision of these algorithms. In addition, radiologists possess the expertise to identify and incorporate suitable AI software to assist non-radiologists in various situations. Moreover, they are equipped to provide necessary oversight and training to other medical professionals regarding using AI tools for fracture detection.&nbsp;

The potential impact of an AI tool augmenting the efforts of healthcare stakeholders to identify MSK fractures cannot be emphasised enough. In particular, emergency physicians could benefit significantly from AI-powered imaging technology, which can act as a secondary pair of eyes or reader, providing clinicians with more accurate imaging information at the time of patient contact. The AI tool can also bolster their original diagnosis or even urge them to revisit their interpretation of the image before treating patients.&nbsp;&nbsp;

In essence, the AI's input, combined with the clinical examination and the physician's own judgment, can help render a more correct diagnosis and treatment approach the first time. It can empower physicians to treat patients based on more accurate imaging information, help triage patients effectively and render a more correct diagnosis and treatment approach the first time, saving valuable time in the treatment pathway.&nbsp;&nbsp;

While research into using AI tools for musculoskeletal fracture detection is still nascent, it does have tremendous potential to impact stakeholders across the healthcare value chain. Therefore, it is imperative that the use of AI and deep learning solutions for musculoskeletal issues be pursued with due perseverance and rigor.&nbsp;&nbsp;

<a href="https://www.who.int/news-room/fact-sheets/detail/musculoskeletal-conditions" target="_blank">https://www.who.int/news-room/fact-sheets/detail/musculoskeletal-conditions</a>&nbsp;&nbsp;&nbsp;

<a href="https://www.osteoporosis.foundation/sites/iofbonehealth/files/2019-06/1.%202018_EU6_Report_BrokenBonesBrokenLives_English.pdf" target="_blank">https://www.osteoporosis.foundation/sites/iofbonehealth/files/2019-06/1.%202018_EU6_Report_BrokenBonesBrokenLives_English.pdf</a>&nbsp;&nbsp;

<a href="https://www.osteo-pharma.com/bone-fracture-facts/" target="_blank">https://www.osteo-pharma.com/bone-fracture-facts/</a>&nbsp;&nbsp;

<a href="https://journals.lww.com/clinorthop/Fulltext/2023/03000/Deep_Learning_Assistance_Closes_the_Accuracy_Gap.25.aspx" target="_blank">https://journals.lww.com/clinorthop/Fulltext/2023/03000/Deep_Learning_Assistance_Closes_the_Accuracy_Gap.25.aspx</a>&nbsp;&nbsp;&nbsp;&nbsp;

<a href="https://pubs.rsna.org/doi/10.1148/rg.220037" target="_blank">https://pubs.rsna.org/doi/10.1148/rg.220037</a>&nbsp;&nbsp;&nbsp;

<a href="https://www.rsna.org/news/2022/may/Global-Radiologist-Shortage" target="_blank">https://www.rsna.org/news/2022/may/Global-Radiologist-Shortage</a>&nbsp;&nbsp;&nbsp;

<a href="https://www.linkedin.com/pulse/how-ai-can-help-address-global-shortage-radiologists-colangelo/" target="_blank">https://www.linkedin.com/pulse/how-ai-can-help-address-global-shortage-radiologists-colangelo/</a>&nbsp;&nbsp;&nbsp;

<a href="https://www.rsna.org/news/2022/march/AI-Helps-Detect-Bone-Fractures" target="_blank">https://www.rsna.org/news/2022/march/AI-Helps-Detect-Bone-Fractures</a>&nbsp;&nbsp;

Insights onArtificial Intelligence &Healthcare Technology

Insights on
Artificial Intelligence &
Healthcare Technology