Tuberculosis (TB) is the second deadliest infectious killer, after COVID-19, which claimed 1.5 million lives in 2020 but is now largely under control. Meanwhile, multi-drug-resistant TB remains a public health crisis and a health security threat. The World Health Organization confirms that the COVID-19 pandemic could start to unravel years of progress in the fight against tuberculosis. This is largely a result of disruption to access to TB services and a drop in resources, which has led to a fall in the detection of new cases. Due to restricted access to diagnostics and lockdowns imposed to contain the COVID-19 pandemic, 4.1 million cases went undiagnosed. India was the worst (41%) with Indonesia (14%) and the Philippines (12%) following next.
Viewed against the milestone of a 35% reduction in TB deaths by 2020, detailed in The End TB Strategy the global reduction in the corresponding time period has only been 9.2%.
Prevention and early diagnosis of tuberculosis are key to its treatment
To achieve the targets set forth in The End TB Strategy, patients must be put at the heart of service delivery, and early diagnosis and prevention is the first step. A robust infrastructure for testing and an adequate and trained workforce are the essential tenets needed to achieve the same. The 2021 Global TB report, however, finds that spending on TB diagnostic, treatment and prevention services fell from $5.8 billion to $5.3 billion, which is less than half of the global target for fully funding the tuberculosis response of $13 billion annually by 2022.
The current situation, along with the importance of data, has triggered a rising awareness and acceptance of the need for evolution in our approach to healthcare workflows. This acknowledgement has been made easier by the rapid strides taken by machine learning and artificial intelligence (AI) driven solutions specifically designed to address medical needs.
AI technology can help detect tuberculosis
AI’s role in diagnostics is growing rapidly. The broad areas in which it can assist hospitals and clinicians include efficient and accurate clinical decision-making, medical image recognition, workflow streamlining via the automation of repetitive tasks, relieving administrative burdens and treatment management. In particular, the field of radiology has been swift to embrace the use of AI solutions. This is because the field is data-driven and diagnosis depends on visual confirmation and interpretation of chest X-rays by trained radiologists. This is where a significant challenge lies.
The global shortage of radiologists is one of healthcare’s unspoken predicaments. More than two-thirds (5.2 billion) of the 7.9 billion people on earth do not have access to one. The shortage of this skillset is a key factor behind the exacerbating issues in lung healthcare and it is an area that AI solutions can impact by reducing the pressure on time and resource-strapped medical imaging professionals, assisting them to process considerable volumes of imaging data, triage critical cases and create reports.
There are various organizations developing AI solutions for medical imaging. One of them is Qure.ai, which has obtained FDA/CE clearances to highlight and prioritise abnormalities in chest X-rays. Let’s look at an example of how Qure.ai’s solution was deployed and contributed to alleviating issues and enhancing existing tuberculosis systems.
Streamlining TB diagnostics in Rajasthan, India
According to the WHO, India faces one of the highest TB burdens in the world. In such a densely populated nation, even hospitals in city centres struggle to manage the diagnosis-to-treatment cycle of this highly infectious disease. One of the major concerns for clinicians in urban facilities is TB triage, as resource constraints often lead to missing potentially infected patients before a proper diagnosis is rendered.
In the north-western state of Rajasthan, the Baran District Hospital caters to a region of 1.2 million residents. It is a tertiary care facility with a dedicated tuberculosis centre and a series of radiology services and capabilities. It receives patient references from the local population and migrants from neighbouring states. In 2019, the percentage of newly identified TB cases had crossed 80% of the total notification, an increase from previous years. As a result, chest physicians at the hospital struggled with the large TB patient base and were in dire need of assistance.
Qure.ai teamed with the hospital to begin comprehensive deployment and real-time testing of its AI-powered chest X-ray solution. Its integration into the diagnostic workflow positively impacted clinical efficiencies in several key areas. There was a 33% increase in the notification rate and the number of drop-outs of presumptive cases reduced from 72% to 53%.
From possible to probable: AI beyond TB
The use of AI tools for TB screening is a watershed moment. The Global TB report recommends increasing investments in tuberculosis research to drive technological breakthroughs and the rapid uptake of innovation. AI-based interventions are one lever to achieving this.
At the same time, AI has a critical part to play in the diagnosis and treatment of the deadliest cancer in the world – lung cancer. Approximately 75% of patients die within five years of diagnosis because symptoms are detected in the disease’s later stages when it is harder to treat. About 35% of lung nodules are missed at the initial screening and initial symptoms tend to be innocuous and often dismissed.
When physiological indicators of lung cancer are identified earlier, the outcomes for patients improve dramatically. A radiologist’s key task in the screening workflow is to search for pulmonary nodules and assess their malignancy risk based on size, shape, structure, type, location and growth. These can also be evaluated by AI solutions, which can scan CTs and detect lung nodules that may not be visible to the naked eye as well.
Thus, AI can play a vital role as a parallel diagnostic tool, automating select repetitive processes, augmenting the efforts of physicians and operating as a second pair of eyes to ensure no treatment delays. Eventually, AI’s potential to impact healthcare and benefit stakeholders is endless, limited only by our imagination.
Source: weforum.org
