There’s no doubt that cancer screening, early detection, and swift intervention are invaluable for improving patients’ health outcomes. In this process, once a clinician has determined a need for a tissue biopsy after detecting an abnormality, diagnostic accuracy and efficiency become crucial factors in the treatment journey.
We’ve seen great advancements when it comes to screening and targeted therapies. However, these treatments cannot occur without an accurate and timely diagnosis. Once a screening tool, such as a mammogram, detects an abnormality, tissue biopsies are taken for analysis by a pathologist. In this critical process, patients and providers alike expect to receive accurate results in the shortest possible amount of time. However, there is an ongoing global pathologist shortage, and this, coupled with increased caseloads as the volume of cancer diagnoses ticks upwards, leads to overburdened pathologists with case backlogs and patients without timely, high-quality diagnostic results. Implementing AI solutions to support pathologists offers promising opportunities to enhance both diagnostic accuracy and efficiency, reducing the incidence of misdiagnosis and increasing pathologists’ confidence in decision-making.
The value of AI-powered cancer diagnostics for patients and pathologists
For all patients, especially those awaiting potentially life-altering diagnoses, prioritizing diagnostic accuracy and efficiency is essential. AI-powered diagnostic tools can play a crucial role by providing support in this complex process. Pathologists review tissue biopsies to distinguish between various types of cancer and other relevant clinical features. To be truly effective, AI tools must be trained on extensive datasets from diverse sources. When enriched with insights from clinical experts, these tools ensure high accuracy, even for rare or challenging cases.
It’s important to note that the effectiveness of AI is closely linked to the quality of the data used for its training. There are concerns about a potential future where excessive reliance on AI could lead to a decline in pathologists’ skills, thereby compromising diagnostic quality. It’s essential to maintain high standards in pathologist training to prevent a situation where the interpretation by AI becomes the de facto gold standard and ensure that AI complements rather than replaces the expertise of pathologists.
AI tools can act as digital assistants, flagging discrepancies between AI findings and pathologist diagnoses, which helps detect potential errors. This setup reduces error rates and provides pathologists with additional confidence. Furthermore, AI enhances efficiency by assisting with tasks like distinguishing between benign or cancerous slides, quantifying cells, as well as providing tumor grading and identifying cancer subtypes. These tools help manage workloads and reduce diagnostic delays, which is crucial given the global pathologist shortage and rising cancer cases.
The work of a pathologist is inherently complex, and AI-powered diagnostic tools must be trained responsibly to deliver value. AI that is trained in extensive, diverse datasets and enriched with expert knowledge can accurately mimic the pathologist’s role, serving both as a digital assistant and a quality control tool. By identifying discrepancies and supporting decision-making, these tools help to reduce error rates to nearly zero while instilling confidence among pathologists.
Additionally, AI enhances efficiency by helping pathologists anticipate and order additional ancillary tests, thereby reducing the time patients must wait for their diagnosis. According to the American Cancer Society, while some routine biopsy and cytology results might be ready as soon as a day or two after the sample reaches the lab, it often takes longer for a patient to receive their results. Patients naturally feel concerned, stressed and/or anxious when they’re awaiting a possible cancer diagnosis. By helping to triage cases, urgent cases are then prioritized, and diagnoses are streamlined.
Barriers to adoption
AI in pathology is not the future – it’s today. However, while the benefits of AI-powered diagnostic solutions are clear, according to KLAS, fewer than ten percent of U.S. organizations have adopted digital pathology for clinical use and less than five percent of cases are signed out digitally. Although adopting digital pathology can provide many benefits to patients and providers alike, there are two key barriers to adoption to consider.
The first barrier to adoption is twofold: the upfront costs of adoption and a lack of reimbursement from payers for digitization. In the U.S., there has been a general lack of awareness on the payer side of the improvements and quality of care that can be achieved through digitization in the pathology lab. Without reimbursement in place from payer organizations, this requires upfront investments in technology from pathology labs that wish to adopt the technology.
Another barrier to adoption is the integration of technology in pathology labs. Digital transformation in labs requires changes to IT, clinical workflows, and the medical practices themselves, which can be a daunting leap for pathologists, who might feel as though they do not have the time or resources in place to support such a transformation. However, with the many clinical and operational benefits that AI provides, it’s clear that pathologists should embrace an AI-first mindset to overcome existing challenges.
Although these barriers exist, several key forces are accelerating adoption, including collaboration between industry and healthcare professionals, work driven by individual organizations (including the Digital Pathology Association and the College of American Pathologists), and successful ongoing advocacy for adding new CPT codes.
Championing the art and science of pathology
The future is now with digital pathology. Despite existing barriers, the integration of AI-powered diagnostic tools into pathology labs presents strong opportunities for early adopters. The technology offers a competitive advantage through increased diagnostic accuracy and efficiency, leading to enhanced patient care and tangible ROI for labs.
The key is to champion both the art and science of pathology while effectively leveraging AI. Furthermore, AI should enhance, not undermine, the expertise of pathologists. To avoid potential pitfalls, we must ensure that AI is trained on high-quality data and used responsibly.
By embracing AI thoughtfully, we achieve better patient outcomes and strengthen the practice of pathology. Beyond the business case, the industry must come together to transform pathology. Every patient deserves an accurate, timely, and personalized cancer diagnosis. AI in pathology is not a distant vision but a current reality, already making a substantial impact for both patients and pathologists alike.
About Joseph Mossel
Joseph Mossel is the CEO of Ibex Medical Analytics. His career in the tech industry spans more than 20 years, starting with software development and product management and ending with leadership positions in startups, large multinational corporations, and non-profits. Joseph has led products from inception to maturity as multi-million dollar businesses.