Digital technology and AI: allies of 21st-century anatomical pathology
The role of anatomical pathology is constantly evolving and gaining importance. To support this change, digitalization has become essential. And logically, so has Artificial Intelligence (AI). The Tribun Health team discusses this necessity with us.
Anatomical Pathology: An Increasingly Central Role
“Precision medicine begins with the accuracy of diagnosis.” Based on this observation, Jean-François Pomerol, CEO of Tribun Health, argues for the need to effectively equip anatomical pathology laboratories. This specialty is playing an increasingly central role in the patient’s healthcare journey. With the increasing number of patients, the decreasing number of pathologists, and the discovery of new types of cancer, the need for rapid, accurate, and reliable diagnoses is growing. Furthermore, the development of holistic medicine, which integrates the patient’s clinical, radiological, pathological, and genomic data, leads professionals to combine a wealth of information. “Precision medicine therefore requires digitization,” and AI complements the support provided by medical services because, as Jean-François Pomerol points out, “the pathologist’s value isn’t in counting. Algorithms free them from all the time-consuming and quantitative tasks, allowing them to gain a comprehensive understanding of the various elements presented to them.” For this reason, the company develops algorithms that support professionals at every stage of their work: workflow management, diagnosis, and prediction.
A broad scope for AI
To achieve this, Tribun Health has a team of 10 data scientists specializing in computer vision, and therefore image-oriented. At the workflow level, the goal is “to ensure that the examination arrives at the physician’s workstation ready for analysis.” To this end, the AI has already analyzed the type of sample, detected the target organ, and performed quality control checks on the slide preparation, digitization, and labeling. It can then direct the different cases to the appropriate specialist. Next, AI comes into play to refine the diagnosis. In this area, the algorithms have “no limits”: cell segmentation, counting, measurements, registration, etc. “We thus considerably reduce the workload of doctors on less complex tasks, allowing them to go further.” This approach also has numerous applications in the field of tele-expertise, which has become widespread since the COVID crisis in particular and allows for obtaining an expert medical opinion remotely. Finally, prediction and prognosis are also among the tools offered. “These tools allow us to correlate the patient’s history with clinical results,” explains Saima Ben Hadj, Director of AI and Computer Vision. “Our algorithms, trained on cohorts of thousands of patients for each question addressed, link images, genomic data, and clinical results, and allow us to predict the progression of the disease or its response to targeted therapies.”
A promising market
The company expects new opportunities. Currently, around twenty centers in France have made the leap to digitization. Once this step is completed, AI plays a crucial role very quickly. “The era of digitization pioneers is behind us,” observes Jean-François Pomerol. “Adoption is going to accelerate. Purchases are now driven by IT departments that want to implement a comprehensive data management system within their institutions. We are ultimately witnessing the same boom as in radiology 30 years ago, but in a different ecosystem, with the cloud, high-speed internet, and AI!” These prospects allow the software publisher to establish an ambitious roadmap, aiming to cover 90% of a laboratory’s activities with its artificial intelligence solutions within three years, thanks in part to the PortrAIt consortium project funded by BPI France. In collaboration with Owkin, Gustave Roussy, the Léon Bérard Cancer Center, Unicancer, and Cypath, the PortrAIt project aims to create around fifteen AI algorithms within five years to improve cancer diagnosis, discover new treatment biomarkers, and predict outcomes for patients in hospitals across France.
Marion Bois
[pdf-embedder url=“https://www.sih-solutions.fr/wp-content/uploads/2023/07/013SIHMAG_27.pdf” title=“SIH Solutions Magazine”]