The innovation of mediastinal staging in lung cancer with artificial intelligence

Abstract

Lung carcinoma is the leading cause of cancer-related death worldwide. The tumor’s progression, prognosis evaluation, and treatment mainly depend on its staging and histopathological subtype. Available options for mediastinal lymph node staging include imaging studies, with positron emission tomography with computed tomography (PET-CT) being a promising option due to its ability to simultaneously reflect the anatomical structure and functional metabolic information of lesions. At present, the gold standard for mediastinal staging is through histopathology, involving invasive procedures that, besides potential complications, have limitations in diagnosis. Artificial intelligence is a set of advanced computational algorithms that learn patterns from provided data to make predictions about unseen datasets. In radiology, radiomics, which involves the extraction of a wide range of features from medical images, has been implemented. At present, radiomic models have been developed for the diagnosis, staging, and prognosis of lung cancer. The objective of this review is to provide scientific evidence on the use of radiomics in mediastinal staging with PET-CT images in non-small cell lung cancer.

Keywords

• Lung cancer. Staging mediastinal. Radiomics. Artificial intelligence. Positron emission tomography with computed tomography.