Radiomics and deep learning in lung cancer
- PDF / 521,854 Bytes
- 9 Pages / 612.419 x 808.052 pts Page_size
- 26 Downloads / 193 Views
REVIEW ARTICLE
Radiomics and deep learning in lung cancer Michele Avanzo1 · Joseph Stancanello2 · Giovanni Pirrone1 · Giovanna Sartor1 Received: 31 March 2020 / Accepted: 15 April 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Lung malignancies have been extensively characterized through radiomics and deep learning. By providing a three-dimensional characterization of the lesion, models based on radiomic features from computed tomography (CT) and positron-emission tomography (PET) have been developed to detect nodules, distinguish malignant from benign lesions, characterize their histology, stage, and genotype. Deep learning models have been applied to automatically segment organs at risk in lung cancer radiotherapy, stratify patients according to the risk for local and distant recurrence, and identify patients candidate for molecular targeted therapy and immunotherapy. Moreover, radiomics has also been applied successfully to predict side effects such as radiation- and immunotherapy-induced pneumonitis and differentiate lung injury from recurrence. Radiomics could also untap the potential for further use of the cone beam CT acquired for treatment image guidance, four-dimensional CT, and dose-volume data from radiotherapy treatment plans. Radiomics is expected to increasingly affect the clinical practice of treatment of lung tumors, optimizing the end-to-end diagnosis–treatment–follow-up chain. The main goal of this article is to provide an update on the current status of lung cancer radiomics.
Keywords Artificial Intelligence · Image biomarkers · Quantitative Imaging · Machine learning · PET · CT
Introduction Because lung cancer is the leading cause of cancer-related mortality worldwide [1], improvements in diagnosis at an early, potentially curable stage would have a major impact on human health [2]. Over time, radiologists have identified a relatively small number of qualitative visual physical characteristics to differentiate benign and malignant lesions, for example, the historical use of speculation, lesion size, attenuation, and perilesional cystic airspaces. Radiomic features are image-based descriptors able to quantitatively capture shape, size or volume, and texture of tumor or normal tissue regions; they can be combined with artificial intelligence applications into predictive and prognostic models [3, 4]. The current challenge for radiomic applications to the thorax is twofold: First, systems must be developed to accurately extract phenotypic characteristics from images, and second, systems must determine which features, among Michele Avanzo
[email protected] 1
Department of Medical Physics, Centro di Riferimento Oncologico di Aviano (CRO) IRCCS, Via F. Gallini 2, 33081 Aviano, PN, Italy
2
Guerbet SA, Villepinte, France
thousands of phenotypic characteristics, correlate with the underlying genotype and disease behavior, thus aiding in the prognosis and clinical management of disease [5]. Perhaps one of the earliest works showing correlation of imaging biomarkers, or “
Data Loading...
