Radiomics: Data Are Also Images

Abstract

The aim of this review is to provide readers with an update on the state of the art, pitfalls, solutions for those pitfalls, future perspectives, and challenges in the quickly evolving field of radiomics in nuclear medicine imaging and associated oncology applications. The main pitfalls were identified in study design, data acquisition, segmentation, feature calculation, and modeling; however, in most cases, potential solutions are available and existing recommendations should be followed to improve the overall quality and reproducibility of published radiomics studies. The techniques from the field of deep learning have some potential to provide solutions, especially in terms of automation. Some important challenges remain to be addressed but, overall, striking advances have been made in the field in the last 5 y.

Voxel‐based dosimetry is superior to mean absorbed dose approach for establishing dose‐effect relationship in targeted radionuclide therapy

“…With a "reduction to absurdity", neglecting the reliability of single voxel counts would invalidate both concepts of SUV max and radiomics. This conceptual link could also indicate that, in order to obtain reliable voxel dosimetry results, careful methodological check and standardization could be necessary as in radiomics …”

Combining positron emission tomography/computed tomography, radiomics, and sentinel lymph node mapping for nodal staging of endometrial cancer patients

“…It has been shown for various neoplasms that tumor characteristics on images correlate with tumor molecular phenotype, genotype, and prognosis. The extraction of a large number of features from medical images and the successive mining of these features by means of machine-learning approaches is called radiomics 9. Radiomics analysis of the uterine primary tumor on pre-operative 18F-FDG PET images may help predict the presence of metastatic lymph nodes, even when they are too small to be detected by PET, thus reducing false-negative results and increasing the sensitivity of the technique.…”

Early Prediction of Tumor Response to Neoadjuvant Chemotherapy and Clinical Outcome in Breast Cancer Using a Novel FDG-PET Parameter for Cancer Stem Cell Metabolism

“…Advanced tumors are thus comprised of cancer cells with varying levels of differentiation, consequently leading to cancer cell subsets exerting differential biological traits [5]. Recently, radiomics has been used to measure the intratumoral heterogeneity using texture analysis [6,7], but it has a major drawback. The results are not consistent across the studies [6,8].…”

“…Recently, radiomics has been used to measure the intratumoral heterogeneity using texture analysis [6,7], but it has a major drawback. The results are not consistent across the studies [6,8]. The lack of reproducibility in radiomics modeling is primarily attributable to the radiomic features that cannot fully represent the true tumor characteristics.…”