Radiomics enables high-throughput extraction of information from images, including features not easily visible or quantifiable to the clinician, and analyzes them to yield visual quantitative parameters. Immunotherapy, an emerging treatment for lung cancer, has significantly altered the traditional treatment paradigm due to its exceptional therapeutic effects. However, the absence of objective and precise tools for evaluating efficacy and adverse events hampers the comprehensive scientific assessment of immunotherapy’s benefits and risks. Despite being in its nascent stages, numerous studies have demonstrated the utility of radiomics in early diagnosis, prognosis prediction, and guidance for personalized lung cancer treatment. This review summarizes the progress of radiomics in evaluating the efficacy and adverse effects of immunotherapy for non-small cell lung cancer, aiming to maximize its efficacy and minimize its risks, and discusses its clinical significance, future goals, and challenges.