Radiomics Repeatability Pitfalls in a Scan-Rescan MRI Study of                    Glioblastoma

Hoebel, Katharina; Patel, Jay; Beers, Andrew; Chang, Ken; Singh, Praveer; Brown, James M.; Pinho, Marco C.; Batchelor, Tracy T.; Gerstner, Elizabeth R.; Rosen, Bruce R.; Kalpathy–Cramer, Jayashree

doi:10.1148/ryai.2020190199

Cited by 41 publications

(42 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Image pre-processing of the two-minute postcontrast-T1W images was performed after tumor segmentation using an in-house developed pipeline and using a widely used proposed pre-processing method by Pyradiomics [ 35 , 36 ]. The in-house developed pipeline started first by applying bias field correction to every image using MIM software (version 6.9.4, Cleveland, OH, USA) to correct for nonuniform grayscale intensities in the MRI caused by field inhomogeneities.…”

Section: Methodsmentioning

confidence: 99%

MRI-Based Radiomics Analysis for the Pretreatment Prediction of Pathologic Complete Tumor Response to Neoadjuvant Systemic Therapy in Breast Cancer Patients: A Multicenter Study

Granzier

Ibrahim

Primakov

et al. 2021

Cancers

View full text Add to dashboard Cite

This retrospective study investigated the value of pretreatment contrast-enhanced Magnetic Resonance Imaging (MRI)-based radiomics for the prediction of pathologic complete tumor response to neoadjuvant systemic therapy in breast cancer patients. A total of 292 breast cancer patients, with 320 tumors, who were treated with neo-adjuvant systemic therapy and underwent a pretreatment MRI exam were enrolled. As the data were collected in two different hospitals with five different MRI scanners and varying acquisition protocols, three different strategies to split training and validation datasets were used. Radiomics, clinical, and combined models were developed using random forest classifiers in each strategy. The analysis of radiomics features had no added value in predicting pathologic complete tumor response to neoadjuvant systemic therapy in breast cancer patients compared with the clinical models, nor did the combined models perform significantly better than the clinical models. Further, the radiomics features selected for the models and their performance differed with and within the different strategies. Due to previous and current work, we tentatively attribute the lack of improvement in clinical models following the addition of radiomics to the effects of variations in acquisition and reconstruction parameters. The lack of reproducibility data (i.e., test-retest or similar) meant that this effect could not be analyzed. These results indicate the need for reproducibility studies to preselect reproducible features in order to properly assess the potential of radiomics.

show abstract

Section: Methodsmentioning

confidence: 99%

MRI-Based Radiomics Analysis for the Pretreatment Prediction of Pathologic Complete Tumor Response to Neoadjuvant Systemic Therapy in Breast Cancer Patients: A Multicenter Study

Granzier

Ibrahim

Primakov

et al. 2021

Cancers

View full text Add to dashboard Cite

show abstract

“…Besides, the influence of some brain MRI preprocessing methods have also been considered. Previous papers [38][39][40][41][42][43] have investigated the impact of various image preprocessing methods on brain MRI radiomic feature reproducibility. In our study, we only consider the impact of N4 bias field correction and image resampling (resampling image pixels to 1 mm × 1 mm × 1 mm) which were used as standardization in paper [19].…”

Section: Introductionmentioning

confidence: 99%

Impact of Preprocessing and Harmonization Methods on the Removal of Scanner Effects in Brain MRI Radiomic Features

Ammari

Balleyguier

et al. 2021

Cancers

View full text Add to dashboard Cite

In brain MRI radiomics studies, the non-biological variations introduced by different image acquisition settings, namely scanner effects, affect the reliability and reproducibility of the radiomics results. This paper assesses how the preprocessing methods (including N4 bias field correction and image resampling) and the harmonization methods (either the six intensity normalization methods working on brain MRI images or the ComBat method working on radiomic features) help to remove the scanner effects and improve the radiomic feature reproducibility in brain MRI radiomics. The analyses were based on in vitro datasets (homogeneous and heterogeneous phantom data) and in vivo datasets (brain MRI images collected from healthy volunteers and clinical patients with brain tumors). The results show that the ComBat method is essential and vital to remove scanner effects in brain MRI radiomic studies. Moreover, the intensity normalization methods, while not able to remove scanner effects at the radiomic feature level, still yield more comparable MRI images and improve the robustness of the harmonized features to the choice among ComBat implementations.

show abstract

“…We further improved this approach by using the iterative projection under the guidance of tract orientation 30 . We also applied a histogram matching normalization method to reduce the variations across the subjects and imaging protocols 31, 32 . The skeletonized FA maps of patients and controls ( Fig.…”

Section: Resultsmentioning

confidence: 99%

Structural connectome quantifies tumor invasion and predicts survival in glioblastoma patients

Wei

Cui

et al. 2021

Preprint

View full text Add to dashboard Cite

Background: Glioblastoma is characterized by extensive invasion into brain parenchymal tissue through white matter tracts. Systematically quantifying invasion, however, is limited by the conventional imaging tools, and could potentially be achieved by a structural connectome approach. Methods: Two prospective patient cohorts of newly diagnosed glioblastoma were included for network construction. A fiber template was firstly derived by employing probabilistic tractography on healthy subjects. Through performing tract-based spatial statistics in patients and age-matched controls, the connectivity strength of each fiber was estimated in patients for network construction. Contrast-enhancing and non-enhancing tumors were segmented and overlaid to the network to identify connectome disruption in lesion and distant areas. The connectome disruption probabilities were calculated across all patients. Disruption indices and network topological features were examined using survival models. Results: Compared to healthy controls, patient networks demonstrated lower clustering coefficient, but higher characteristic path length (each P < 0.001). Higher distant area disruption (HR = 1.43, P = 0.027) and characteristic path length (HR = 1.59, P = 0.031) were associated with worse survival, while higher clustering coefficient (HR = 0.59, P = 0.016) was associated with prolonged survival. Conclusion: The occult invasion in glioblastoma could be identified and quantified using structural connectome, which may confer benefits to precise patient management.

show abstract

Radiomics Repeatability Pitfalls in a Scan-Rescan MRI Study of Glioblastoma

Cited by 41 publications

References 27 publications

MRI-Based Radiomics Analysis for the Pretreatment Prediction of Pathologic Complete Tumor Response to Neoadjuvant Systemic Therapy in Breast Cancer Patients: A Multicenter Study

MRI-Based Radiomics Analysis for the Pretreatment Prediction of Pathologic Complete Tumor Response to Neoadjuvant Systemic Therapy in Breast Cancer Patients: A Multicenter Study

Impact of Preprocessing and Harmonization Methods on the Removal of Scanner Effects in Brain MRI Radiomic Features

Structural connectome quantifies tumor invasion and predicts survival in glioblastoma patients

Contact Info

Product

Resources

About