Mesothelioma is the cancer of the thin tissue layer covering the lungs and the thoracic cavity walls. Most cases of mesothelioma are caused by exposure to asbestos.1 This article explains how Prof. Papotti and Prof. Duregon used Aiforia®️ Create to build an AI model for mesothelioma subtype recognition with reticulin stain detection.
The laboratory where Prof. Papotti and Prof. Duregon work has a historical link to the investigation of mesothelioma. As Prof. Papotti explains: “In our region, mesothelioma is very common because our region hosted, for over 70 years, the largest asbestos mine in Western Europe and also a large fibre cement plant, which was active until 1986, but still is causing a lot of problems in terms of mesothelioma development.”
According to Prof. Papotti, the prognosis of mesothelioma is heavily impacted by the histopathological subtype, which is recognized at the microscope by the pathologist. “Usually, the cytology or biopsy materials that we receive in our daily activity are relatively limited, so the final diagnosis is sometimes challenging,” he explains.
“The tumor structure is regulated, among others, by a connective tissue scaffold made of collagen and reticuline fibers, which arrange into different patterns depending on the various mesothelioma histological subtypes. As recent reports show that the reticulin stain helps to distinguish these different reticulin patterns, this technique may support a further stratification of mesothelioma and possibly the prediction of different therapeutic options.”
With this in mind, the research team selected a large series of aggressive mesotheliomas of the biphasic and sarcomatoid subtypes and attempted a revised classification of these subtypes using the reticulin stain to highlight the tumor's structure in combination with other markers.
At this point of their investigation, the team saw the opportunity to enrich the microscopic examination with modern tools and started to look for alternatives.
The academic institution where Prof. Papotti and Prof. Duregon work started implementing digital pathology about a year ago. “We were looking for a solution with an AI- and cloud-based platform that could support our research and, at the same time, offer diagnostic algorithms for our clinical practice. Last but not least, we wanted to be able to use the same platform during our lessons with the students,” Prof. Duregon explains.
The research team hypothesized that AI could help improve objectivity in their research by reducing inter-observer variability. “After adequate training and careful validation, AI systems or platforms can support an automatic and objective recognition of different parameters, which would otherwise be heavily influenced by the single pathologist’s attitude or experience,” Prof. Papotti describes.
After comparing different possible diagnostic solutions, the team decided on Aiforia®️ Create, the most versatile tool for developing deep learning AI models for image analysis. “The cooperation with Aiforia has been really positive and awesome. I would say it was open, very knowledgeable, on demand, and without time restrictions. The team who supported us has been truly helpful in supporting us and turning our research idea into a reality.” Prof. Duregon comments on the collaboration.
“Aiforia Create is a great solution for pathologists with no data science or specific coding expertise willing to start working hands-on with AI in their research. The cloud-based platform is very user-friendly, and the open and on-demand communication with their knowledgeable team of scientists is a perk.” – Eleonora Duregon, Assistant Professor of Pathology at the Department of Oncology of the University of Turin
Dr. Duregon and Dr. Papotti are eager to see where AI-assisted research takes them. “I envision that the AI-driven detection and quantification of mesothelioma will first allow us to better understand mesothelioma disease itself. And as a consequence, as pathologists, we will be able to provide our clinician colleagues and the patient with the diagnostic reports, which have more impactful and robust prognostic and predictive information,” Dr. Duregon summarizes.
Moving forward, the research team has a strong interest in endocrine pathology, especially with the adrenal gland and neuroendocrine tumors. “We would like to develop multiple parallel projects supported by the Aiforia Platform and work by matching the special morphological information we are getting from the whole slide image with the genetic molecular and clinical information in order to be able to stratify and identify better the more biologically aggressive and clinically aggressive cases.”
Listen to Prof. Papotti and Prof. Duregon’s interview in the video below:
References:
U.S. Centers for Disease and Control Prevention (CDC). (2024, August 16). Mesothelioma. https://www.cdc.gov/mesothelioma/about/index.html