Artificial Intelligence to predict the best treatment in patients with lung cancer

Small cell lung cancer represents 15% of cancer cases in this organ and has one of the worst prognoses, with an average five-year survival rate of less than 10%, and only 20% for cases diagnosed at an early stage. One of the key factors that determines the chances of survival is the treatment that the affected person will receive.

Currently, immunotherapy is being incorporated to varying degrees as part of treatment. But chemotherapy remains a fundamental component. The main problem is that there is no formula to ensure the suitability of a given combination of therapies.

Increasing survival rates

Hospital del Mar is working on the development of a tool based on artificial intelligence (AI) to provide a solution to this problem. It is hoped that an algorithm will be able to determine the prognosis and treatment of a patient from the analysis of digitised images of biopsies – extraction of cells from a tissue sample to detect the presence of cancer cells – and survival data.

(Photo: The research team is part of IMIM)

Prognosis and treatment prediction

The team is made up of Edurne Arrional – section head of the Medical Oncology services and researcher at the Instituto Hospital del Mar de Investigaciones Médicas (IMIM-Hospital del Mar) – and Joan Gibert, bioinformatician at the Anatomical Pathology service and researcher at the IMIM Molecular Cancer Therapy Research Group. Currently, they have data from 350 patients, collected over the last 10 years, to start developing the functionalities of the algorithm, such as the identification of patterns in anatomical pathology images that are impossible to detect with the naked eye.

The tool can also determine the technical capacity of a liquid biopsy to deliver a prognosis and the prediction of a treatment based on the specific mutations of each tumor. The first results are expected in 2023.