Radiomics in Oncology: A Practical Guide

Shur, Joshua; Doran, Simon J.; Kumar, Santosh; Dafydd, Derfel Ap; Downey, Kate; O’Connor, James P.B.; Papanikolaou, Nikolaos; Messiou, Christina; Koh, Dow‐Mu; Orton, Matthew

doi:10.1148/rg.2021210037

Cited by 214 publications

(202 citation statements)

References 54 publications

Supporting

Mentioning

199

Contrasting

Unclassified

Order By: Relevance

“…It promises a non-invasive, personalized medicine and is applied primarily in an oncological context for diagnosis, survival prediction, and other purposes [ 3 ]. Radiomics is often performed using a well-established machine learning pipeline; generic features from the images are first extracted before feature selection methods and classifiers for modeling are employed [ 4 , 5 ]. One of the main concerns related to radiomics is whether the extracted features have biological meaning [ 6 ].…”

Section: Introductionmentioning

confidence: 99%

Evaluation of the dependence of radiomic features on the machine learning model

Demircioğlu

2022

Insights Imaging

View full text Add to dashboard Cite

Background In radiomic studies, several models are often trained with different combinations of feature selection methods and classifiers. The features of the best model are usually considered relevant to the problem, and they represent potential biomarkers. Features selected from statistically similarly performing models are generally not studied. To understand the degree to which the selected features of these statistically similar models differ, 14 publicly available datasets, 8 feature selection methods, and 8 classifiers were used in this retrospective study. For each combination of feature selection and classifier, a model was trained, and its performance was measured with AUC-ROC. The best-performing model was compared to other models using a DeLong test. Models that were statistically similar were compared in terms of their selected features. Results Approximately 57% of all models analyzed were statistically similar to the best-performing model. Feature selection methods were, in general, relatively unstable (0.58; range 0.35–0.84). The features selected by different models varied largely (0.19; range 0.02–0.42), although the selected features themselves were highly correlated (0.71; range 0.4–0.92). Conclusions Feature relevance in radiomics strongly depends on the model used, and statistically similar models will generally identify different features as relevant. Considering features selected by a single model is misleading, and it is often not possible to directly determine whether such features are candidate biomarkers.

show abstract

Section: Introductionmentioning

confidence: 99%

Evaluation of the dependence of radiomic features on the machine learning model

Demircioğlu

2022

Insights Imaging

View full text Add to dashboard Cite

show abstract

“…In general, radiomics is the study of data from imaging processes that can provide appropriate and efficient prognostic and diagnostic information about cancer. Therefore, the information of this branch of omics science can pave the way for accurate identification of the type of cancer and subsequently, the use of appropriate treatment for different cancers (Shur et al, 2021) which can be extracted via different techniques such as magnetic resonance imaging (MRI), computed tomography (CT), and positron-emission-tomography (PET) (van Timmeren et al, 2020). However, the branches of omics studied in the field of cancer are not limited to the mentioned cases and can also cover wide areas of the realm of omics, such as lipidomics (Yan et al, 2018), glycomics (Drake, 2015), pathomics , phosphoproteomics (Harsha and Pandey, 2010), immunomics (Basharat et al, 2018), interactomics (Vallet et al, 2021) etc.…”

Section: The Integration Of Omics Science With Cancer Researchmentioning

confidence: 99%

Machine Learning: A New Prospect in Multi-Omics Data Analysis of Cancer

et al. 2022

View full text Add to dashboard Cite

Cancer is defined as a large group of diseases that is associated with abnormal cell growth, uncontrollable cell division, and may tend to impinge on other tissues of the body by different mechanisms through metastasis. What makes cancer so important is that the cancer incidence rate is growing worldwide which can have major health, economic, and even social impacts on both patients and the governments. Thereby, the early cancer prognosis, diagnosis, and treatment can play a crucial role at the front line of combating cancer. The onset and progression of cancer can occur under the influence of complicated mechanisms and some alterations in the level of genome, proteome, transcriptome, metabolome etc. Consequently, the advent of omics science and its broad research branches (such as genomics, proteomics, transcriptomics, metabolomics, and so forth) as revolutionary biological approaches have opened new doors to the comprehensive perception of the cancer landscape. Due to the complexities of the formation and development of cancer, the study of mechanisms underlying cancer has gone beyond just one field of the omics arena. Therefore, making a connection between the resultant data from different branches of omics science and examining them in a multi-omics field can pave the way for facilitating the discovery of novel prognostic, diagnostic, and therapeutic approaches. As the volume and complexity of data from the omics studies in cancer are increasing dramatically, the use of leading-edge technologies such as machine learning can have a promising role in the assessments of cancer research resultant data. Machine learning is categorized as a subset of artificial intelligence which aims to data parsing, classification, and data pattern identification by applying statistical methods and algorithms. This acquired knowledge subsequently allows computers to learn and improve accurate predictions through experiences from data processing. In this context, the application of machine learning, as a novel computational technology offers new opportunities for achieving in-depth knowledge of cancer by analysis of resultant data from multi-omics studies. Therefore, it can be concluded that the use of artificial intelligence technologies such as machine learning can have revolutionary roles in the fight against cancer.

show abstract

“…In particular, it could be employed to provide a signature (that is, a combination of features) that is not necessarily visible to the naked eye—even an expertly trained one. Although radiomics has shown promising preliminary results in identifying tumor subtypes and aggressiveness, as well as in predicting responses to therapy and the outcomes for patients with various cancers, most of these results have been obtained in small, retrospective, monocentric cohorts, often employing different methods for lesion segmentation and feature extraction [ 80 ]. It is currently difficult to generalize the results obtained in published papers due to excessive data heterogeneity; indeed, the application of AI in clinical practice requires absolute methodological harmonization [ 81 ].…”

Section: Imaging Of Nens With Radiolabeled Somatostatin Analoguesmentioning

confidence: 99%

Radiolabeled Somatostatin Analogues for Diagnosis and Treatment of Neuroendocrine Tumors

Ambrosini

Zanoni

Filice

et al. 2022

Cancers

View full text Add to dashboard Cite

Neuroendocrine neoplasms (NENs) are rare and heterogeneous tumors that require multidisciplinary discussion for optimal care. The theranostic approach (DOTA peptides labelled with [68Ga] for diagnosis and with 90Y or 177Lu for therapy) plays a crucial role in the management of NENs to assess disease extension and as a criteria for peptide receptor radionuclide therapy (PRRT) eligibility based on somatostatin receptor (SSTR) expression. On the diagnostic side, [68Ga]Ga-DOTA peptides PET/CT (SSTR PET/CT) is the gold standard for imaging well-differentiated SSTR-expressing neuroendocrine tumors (NETs). [18F]FDG PET/CT is useful in higher grade NENs (NET G2 with Ki-67 > 10% and NET G3; NEC) for more accurate disease characterization and prognostication. Promising emerging radiopharmaceuticals include somatostatin analogues labelled with 18F (to overcome the limits imposed by 68Ga), and SSTR antagonists (for both diagnosis and therapy). On the therapeutic side, the evidence gathered over the past two decades indicates that PRRT is to be considered as an effective and safe treatment option for SSTR-expressing NETs, and is currently included in the therapeutic algorithms of the main scientific societies. The positioning of PRRT in the treatment sequence, as well as treatment personalization (e.g., tailored dosimetry, re-treatment, selection criteria, and combination with other alternative treatment options), is warranted in order to improve its efficacy while reducing toxicity. Although very preliminary (being mostly hampered by lack of methodological standardization, especially regarding feature selection/extraction) and often including small patient cohorts, radiomic studies in NETs are also presented. To date, the implementation of radiomics in clinical practice is still unclear. The purpose of this review is to offer an overview of radiolabeled SSTR analogues for theranostic use in NENs.

show abstract

Radiomics in Oncology: A Practical Guide

Cited by 214 publications

References 54 publications

Evaluation of the dependence of radiomic features on the machine learning model

Evaluation of the dependence of radiomic features on the machine learning model

Machine Learning: A New Prospect in Multi-Omics Data Analysis of Cancer

Radiolabeled Somatostatin Analogues for Diagnosis and Treatment of Neuroendocrine Tumors

Contact Info

Product

Resources

About