Artificial intelligence improves early lung cancer detection

A team from the Faculty of Medicine and the University Hospital of Cologne, led by Privatdozent Dr. Yuri Tolkach and Professor Dr. Reinhard Büttner, has developed a platform that opens up new diagnostic possibilities. Using artificial intelligence (AI) methods, this platform enables fully automated and precise analysis of tissue sections from patients with lung cancer. The platform is based on the now common digitization of existing histological data.

Software replaces microscopes in pathology

In recent years, pathology has undergone a digital revolution that has made the use of microscopes superfluous. Slides and microscopes have been replaced by the digitization of tissue sections, which can be analyzed directly on the computer. This digitization is now being further developed. Based on a method known as a multi-class tissue segmentation algorithm, the AI platform enables significantly faster analysis and evaluation of the available data.

Where digital images were previously sorted and evaluated manually, they are now selected and pre-evaluated by the AI. This new validation process has proven to be much more effective and accurate than conventional methods. The optical evaluation of digital images obviously misses aspects that only become apparent in the comparative analysis by the AI. By using the new platform, pathologists gain additional information about the cancer that would have remained hidden without this evaluation technique.

Privatdozent Dr. Yuri Tolkach, senior physician and head of the study at the Institute of General Pathology and Pathological Anatomy at Cologne University Hospital, emphasized to Informationsdienst Wissenschaft (IDW) that the platform has the potential to develop completely new clinical tools. "The new tools can not only improve the quality of diagnosis, but also provide new types of information about the disease, such as how the patient responds to therapy," explains Tolkach.

The research team plans to prove the broad applicability of the platform in a validation study. The researchers are working closely with other pathology institutes in Germany, Austria and Japan.

Lung cancer is one of the most common forms of cancer and has a very high mortality rate. The pathological examination plays a decisive role in the selection of the right therapy. The research results to date have been published in the journal Cell Reports Medicine under the title "Next generation lung cancer pathology: development and validation of diagnostic and prognostic algorithms".

(nie)