A follow-up study on the novel use of contrast-enhanced susceptibility-weighted imaging for extremity desmoid fibromatosis response assessment

Undifferentiated pleomorphic sarcomas (UPSs) demonstrate therapy-induced hemosiderin deposition, granulation tissue formation, fibrosis, and calcification. We aimed to determine the treatment-assessment value of morphologic tumoral hemorrhage patterns and first- and high-order radiomic features extracted from contrast-enhanced susceptibility-weighted imaging (CE-SWI). This retrospective institutional review board-authorized study included 33 patients with extremity UPS with magnetic resonance imaging and resection performed from February 2021 to May 2023. Volumetric tumor segmentation was obtained at baseline, postsystemic chemotherapy (PC), and postradiation therapy (PRT). The pathology-assessed treatment effect (PATE) in surgical specimens separated patients into responders (R; ≥90%, n = 16), partial responders (PR; 89%-31%, n = 10), and nonresponders (NR; ≤30%, n = 7). RECIST, WHO, and volume were assessed for all time points. CE-SWI T2* morphologic patterns and 107 radiomic features were analyzed. A Complete-Ring (CR) pattern was observed in PRT in 71.4% of R (P = 7.71 × 10-6), an Incomplete-Ring pattern in 33.3% of PR (P = .2751), and a Globular pattern in 50% of NR (P = .1562). The first-order radiomic analysis from the CE-SWI intensity histogram outlined the values of the 10th and 90th percentiles and their skewness. R showed a 280% increase in 10th percentile voxels (P = .061) and a 241% increase in skewness (P = .0449) at PC. PR/NR showed a 690% increase in the 90th percentile voxels (P = .03) at PC. Multiple high-order radiomic texture features observed at PRT discriminated better R versus PR/NR than the first-order features. CE-SWI morphologic patterns strongly correlate with PATE. The CR morphology pattern was the most frequent in R and had the highest statistical association predicting response at PRT, easily recognized by a radiologist not requiring postprocessing software. It can potentially outperform size-based metrics, such as RECIST. The first- and high-order radiomic analysis found several features separating R versus PR/NR.

Routine radiologic reporting (RRR) often considers progressive desmoid tumors to have a higher proportion of T2-hyperintense and T1-shortened-enhancing components, while responsive or mature collagenized tumors demonstrate a higher proportion of T2-hypointense-non-enhancing components. We aim to determine the utility of the novel use of contrast-enhanced susceptibility-weighted imaging (CE-SWI) in Desmoid-Tumor treatment response assessment, distinguishing between the T1-shortening-enhancing/T2-hyperintense immature components from the T2-hypointense mature collagenized components. This pilot study included 10 single-lesion extremity desmoid fibromatosis patients undergoing standard-of-care magnetic resonance imaging, including CE-SWI. Three-dimensional (3D) tumor segmentation was performed using MIM software in 48 volumes of interest. Maximum diameter, volume, and modified Choi (mChoi) measurements were computed from CE-SWI and T2-weighted image (T2-WI). Five first-order radiomic features, including mean, skewness, kurtosis, and 10th and 90th percentiles, were calculated using in-house developed software (CARPI-AF). (i) RECIST Progression: We observed two cases of progression according to the T2-WI-based Response Evaluation Criteria in Solid Tumors standard (RECIST). Interestingly, CE-SWI-based-volume and CE-SWI-based-mChoi predicted the same assessment 4.5 months earlier than T2-WI-based-RECIST. RRR assessed both cases as progression; (ii) RECIST Stability: Out of the eight patients classified as having stable disease by T2-WI-based-RECIST, four discrepant progressions were determined: three patients showed an increase greater than 25% of T2-WI-based-volume, and two patients showed an increase greater than 25% of CE-SWI-based-volume. Moreover, from the RECIST stable group, four discrepant-positive responses were predicted by CE-SWI-based-mChoi (three patients) and T2-WI-based-mChoi (four patients). RRR only assessed one patient as having progressive disease; (iii) First-Order Radiomics: CE-SWI detected 23% more 90th-percentile voxels than T2-WI, while T2-WI demonstrated 8.5% more 10th-percentile voxels than CE-SWI. Notably, expected first-order response/progression-related changes in 10th-percentile, 90th-percentile, mean, and skewness were present in 90% of cases. In conclusion, CE-SWI-based-volume and CE-SWI-based-mChoi measurements could improve the prediction of response/progression in desmoid tumors, enhancing the ability in discriminating between T2*- hypointense-collagenized-mature and T1-shortened-enhancing immature components, respectively, in predominant mature responsive and immature progressive tumors, respectively. RRR is relatively insensitive to volumetric tumor changes before RECIST progression and tends to be better tuned with T2* signal and enhancement changes.

This study explores whether textural features from initial non-contrast CT scans of infarcted brain tissue are linked to hemorrhagic transformation susceptibility. Stroke patients undergoing thrombolysis or thrombectomy from Jan 2012 to Jan 2022 were analyzed retrospectively. Hemorrhagic transformation was defined using follow-up magnetic resonance imaging. A total of 94 radiomic features were extracted from the infarcted tissue on initial NCCT scans. Patients were divided into training and test sets (7:3 ratio). Two models were developed with fivefold cross-validation: one incorporating first-order and textural radiomic features, and another using only textural radiomic features. A clinical model was also constructed using logistic regression with clinical variables, and test set validation was performed. Among 362 patients, 218 had hemorrhagic transformations. The LightGBM model with all radiomics features had the best performance, with an area under the receiver operating characteristic curve (AUROC) of 0.986 (95% confidence interval [CI], 0.971-1.000) on the test dataset. The ExtraTrees model performed best when textural features were employed, with an AUROC of 0.845 (95% CI, 0.774-0.916). Minimum, maximum, and ten percentile values were significant predictors of hemorrhagic transformation. The clinical model showed an AUROC of 0.544 (95% CI, 0.431-0.658). The performance of the radiomics models was significantly better than that of the clinical model on the test dataset (p < 0.001). The radiomics model can predict hemorrhagic transformation using NCCT in stroke patients. Low Hounsfield unit was a strong predictor of hemorrhagic transformation, while textural features alone can predict hemorrhagic transformation. Using radiomic features extracted from initial non-contrast computed tomography, early prediction of hemorrhagic transformation has the potential to improve patient care and outcomes by aiding in personalized treatment decision-making and early identification of at-risk patients. • Predicting hemorrhagic transformation following thrombolysis in stroke is challenging since multiple factors are associated. • Radiomics features of infarcted tissue on initial non-contrast CT are associated with hemorrhagic transformation. • Textural features on non-contrast CT are associated with the frailty of the infarcted tissue.

Susceptibility-weighted imaging (SWI) is based on a 3D high-spatial-resolution, velocity-corrected gradient-echo MRI sequence that uses magnitude and filtered-phase information to create images. It SWI uses tissue magnetic susceptibility differences to generate signal contrast that may arise from paramagnetic (hemosiderin), diamagnetic (minerals and calcifications) and ferromagnetic (metal) molecules. Distinguishing between calcification and blood products is possible through the filtered phase images, helping to visualize osteoblastic and osteolytic bone metastases or demonstrating calcifications and osteoid production in liposarcoma and osteosarcoma. When acquired in combination with the injection of an exogenous contrast agent, contrast-enhanced SWI (CE-SWI) can simultaneously detect the T2* susceptibility effect, T2 signal difference, contrast-induced T1 shortening, and out-of-phase fat and water chemical shift effect. Bone and soft tissue lesion SWI features have been described, including giant cell tumors in bone and synovial sarcomas in soft tissues. We expand on the appearance of benign soft-tissue lesions such as hemangioma, neurofibroma, pigmented villonodular synovitis, abscess, and hematoma. Most myxoid sarcomas demonstrate absent or just low-grade intra-tumoral hemorrhage at the baseline. CE-SWI shows superior differentiation between mature fibrotic T2* dark components and active enhancing T1 shortening components in desmoid fibromatosis. SWI has gained popularity in oncologic MSK imaging because of its sensitivity for displaying hemorrhage in soft tissue lesions, thereby helping to differentiate benign versus malignant soft tissue tumors. The ability to show the viable, enhancing portions of a soft tissue sarcoma separately from hemorrhagic/necrotic components also suggests its utility as a biomarker of tumor treatment response. It is essential to understand and appreciate the differences between spontaneous hemorrhage patterns in high-grade sarcomas and those occurring in the therapy-induced necrosis process in responding tumors. Ring-like hemosiderin SWI pattern is observed in successfully treated sarcomas. CE-SWI also demonstrates early promising results in separating the T2* blooming of healthy iron-loaded bone marrow from the T1-shortened enhancement in bone marrow that is displaced by the tumor. SWI and CE-SWI in MSK oncology learning objectives: SWI and CE-SWI can be used to identify calcifications on MRI. Certain SWI and CE-SWI patterns can correlate with tumor histologic type. CE-SWI can discriminate mature from immature components of desmoid tumors. CE-SWI patterns can help to assess treatment response in soft tissue sarcomas. Understanding CE-SWI patterns in post-surgical changes can also be useful in discriminating between residual and recurrent tumors with overlapping imaging features.

BackgroundThe pathological sub-classification of lung cancer is crucial in diagnosis, treatment and prognosis for patients. Quick and timely identification of pathological subtypes from imaging examinations rather than histological tests could help guiding therapeutic strategies. The aim of the study is to construct a non-invasive radiomics-based model for predicting the subtypes of lung cancer on brain metastases (BMs) from multiple magnetic resonance imaging (MRI) sequences.MethodsOne hundred and sixty-one patients of primary lung cancer with synchronous BMs [121 with adenocarcinoma (AD); 40 with small cell lung carcinoma (SCLC)] were enrolled in the study (129 and 32 in the training set and validation set). A total of 960 radiomics features were extracted from multiple MRI sequences [fluid attenuated inversion recovery (FLAIR), diffusion weighted imaging (DWI), contrast-enhanced T1 weighted imaging (CE-T1WI) and contrast-enhanced susceptibility weighted imaging (CE-SWI)] and four clinical features were recorded. Forty-one key features were selected by the least absolute shrinkage selection operator (LASSO). The machine learning (ML) models for predicting AD and SCLC with radiomics features alone and with radiomics features plus clinical features were constructed using classifiers of logistic regression (LR), random forest (RF), support vector machine (SVM) and extreme gradient boosting (XGBoost). The prediction performance of models was evaluated by accuracy (ACC), sensitivity (SEN), specificity (SPE), F1 score and area under the curves (AUC).ResultsThe AUCs of LR, RF, SVM and XGBoost models were 0.8177 vs. 0.7604, 0.8177 vs. 0.7839, 0.4792 vs. 0.8594 and 0.9062 vs. 0.8750, respectively, when using radiomics features alone and radiomics features plus clinical features. In the best-performing model using XGBoost, combination of conventional MRI sequences and CE-SWI had better sub-classification performance than conventional MRI sequences.ConclusionsRadiomics of BMs from multiple MRI sequences provides high discriminatory performance in predicting AD and SCLC using classifiers of LR, RF and XGBoost, and can serve as a potential useful tool to non-invasively distinguish pathological subtypes of lung cancer.

Intracranial lesions exhibit clear contrast enhancement in T1-weighted imaging, but the mechanism whereby contrast-enhanced susceptibility-weighted imaging (CE-SWI) generates signals remains unclear. Contrast enhancement patterns cannot be reliably predicted. To explore the mechanism of CE-SWI contrast enhancement. Fifty-five patients were retrospectively enrolled. All of the imaging employed a clinical 3T magnetic resonance imaging (MRI) system fitted with a 32-channel head coil. Minimum-intensity projection reformatted images were evaluated. Intracranial lesions and brain parenchymal intensities were explored using SWI and CE-SWI. signal intensity rates were calculated by dividing the lesional intensity by the white matter intensity, after which the SWI and CE-SWI signal intensity rate were compared. Two observers independently performed intralesional susceptibility signal analysis. After contrast medium administration, malignant and extra-axial tumors exhibited obvious contrast enhancement on CE-SWI (P < 0.001 and P = 0.013, respectively). The signal intensity of white matter was significantly reduced. The signal intensity rates rose significantly in the benign, malignant, and extra-axial groups (P < 0.001). Between-radiologist agreement in terms of intralesional susceptibility signal assessment was strong (kappa = 0.8, P < 0.001). Contrast media can either reduce or increase SWI signal intensities. The dual contrast feature of CE-SWI can be useful when exploring intracranial disorders.

BackgroundUndifferentiated pleomorphic sarcoma (UPS) is the largest subgroup of soft-tissue sarcomas. It demonstrates post-therapeutic hemosiderin deposition, granulation tissue formation, fibrosis, and calcification. Our research aims to establish the multiparametric MRI (mp-MRI) value for predicting UPS treatment response.MethodsAn IRB-approved retrospective study included 33 extremity UPS patients with pre-operative mp-MRI, including diffusion-weighted imaging (DWI), contrast-enhanced susceptibility-weighted imaging (CE-SWI), and perfusion-weighted imaging with dynamic contrast-enhancement (PWI/DCE), and surgical resection between February 2021 and May 2023. Lesions were visually classified on CE-SWI into one of 6 morphology patterns. On PWI/DCE, lesions were classified into one of 6 patterns, and time-intensity curves (TICs) were classified as types I-V. Patients were categorized into three groups based on the percentage of pathology-assessed treatment effect (PATE) in the surgical specimen: Responders (> = 90% PATE, n = 16), partial-responders (31–89% PATE, n = 10), and non-responders (< = 30% PATE, n = 7).ResultsAt post-radiation therapy (PRT), a CE-SWI Complete-Ring pattern was observed in 71% of responders (p = 7.71 × 10–6). On PWI/DCE images, 79% of responders displayed a Capsular pattern (p = 1.49 × 10–7), and 100% demonstrated a TIC-type II (p = 8.32 × 10–7). ROC analysis comparing responders (n = 14) vs. partial/non-responders (n = 16) at PRT showed that the model combining PWI/DCE TIC-type II, PWI/DCE Capsular pattern, and CE-SWI Complete-Ring pattern yielded the highest classification performance (AUC = 0.99), outperforming PWI/DCE Capsular + TIC-type II (AUC = 0.97), PWI/DCE Capsular (AUC = 0.89), PWI/DCE TIC-type II (AUC = 0.88), and CE-SWI Complete Ring (AUC = 0.79). Contrary to prior reports, DWI/ADC played a secondary role in predicting response: ADC mean & skewness (AUC = 0.63). RECIST demonstrated 100% stability at PRT and 100% pseudo-progression at PC in responders and partial/non-responders (AUC = 0.47).ConclusionMp-MRI-derived features are valuable in assessing UPS treatment response. A pre-operative model that combines PWI/DCE TIC-type II, PWI/DCE Capsular pattern, and CE-SWI Complete Ring pattern can reliably predict successfully treated UPS with > = 90% PATE, outperforming RECIST, which was proven unreliable in separating responders from partial/non-responders. Institutions that have not yet implemented CE-SWI can rely on a single-sequence approach based on PWI/DCE, combining the presence of TIC II and Capsular enhancement as criteria for response prediction.

Given the limited repeatability and reproducibility of radiomic features derived from weighted magnetic resonance imaging (MRI), there may be significant advantages to using radiomics in conjunction with quantitative MRI. This study introduces a novel physics-informed discretization (PID) method for reproducible radiomic feature extraction and evaluates its performance using quantitative MRI sequences including magnetic resonance fingerprinting (MRF) and apparent diffusion coefficient (ADC) mapping. A multiscanner, scan-rescan dataset comprising whole-brain 3D quantitative (MRF T1, MRF T2, and ADC) and weighted MRI (T1w MPRAGE, T2w SPACE, and T2w FLAIR) from 5 healthy subjects was prospectively acquired. Subjects underwent 2 repeated acquisitions on 3 distinct 3 T scanners each, for a total of 6 scans per subject (30 total scans). First-order statistical (n = 23) and second-order texture (n = 74) radiomic features were extracted from 56 brain tissue regions of interest using the proposed PID method (for quantitative MRI) and conventional fixed bin number (FBN) discretization (for quantitative MRI and weighted MRI). Interscanner radiomic feature reproducibility was measured using the intraclass correlation coefficient (ICC), and the effect of image sequence (eg, MRF T1 vs T1w MPRAGE), as well as image discretization method (ie, PID vs FBN), on radiomic feature reproducibility was assessed using repeated measures analysis of variance. The robustness of PID and FBN discretization to segmentation error was evaluated by simulating segmentation differences in brainstem regions of interest. Radiomic features with ICCs greater than 0.75 following simulated segmentation were determined to be robust to segmentation. First-order features demonstrated higher reproducibility in quantitative MRI than weighted MRI sequences, with 30% (n = 7/23) features being more reproducible in MRF T1 and MRF T2 than weighted MRI. Gray level co-occurrence matrix (GLCM) texture features extracted from MRF T1 and MRF T2 were significantly more reproducible using PID compared with FBN discretization; for all quantitative MRI sequences, PID yielded the highest number of texture features with excellent reproducibility (ICC > 0.9). Comparing texture reproducibility of quantitative and weighted MRI, a greater proportion of MRF T1 (n = 225/370, 61%) and MRF T2 (n = 150/370, 41%) texture features had excellent reproducibility (ICC > 0.9) compared with T1w MPRAGE (n = 148/370, 40%), ADC (n = 115/370, 32%), T2w SPACE (n = 98/370, 27%), and FLAIR (n = 102/370, 28%). Physics-informed discretization was also more robust than FBN discretization to segmentation error, as 46% (n = 103/222, 46%) of texture features extracted from quantitative MRI using PID were robust to simulated 6 mm segmentation shift compared with 19% (n = 42/222, 19%) of weighted MRI texture features extracted using FBN discretization. The proposed PID method yields radiomic features extracted from quantitative MRI sequences that are more reproducible and robust than radiomic features extracted from weighted MRI using conventional (FBN) discretization approaches. Quantitative MRI sequences also demonstrated greater scan-rescan robustness and first-order feature reproducibility than weighted MRI.

e12612 Background: Computationally derived quantitative imaging (radiomic) features that describe tumor phenotypes at the pixel level have demonstrated associations with clinical characteristics in early investigations of other cancers. This implies that molecular differences among tumors may be reflected in their structure on the scales probed by 3D magnetic resonance imaging (MRI). We investigated whether radiomic features computed over tumor volumes from pre-treatment breast MRI could predict risk factors in triple-negative breast cancer (TNBC). Methods: We evaluated breast tumors on pre-treatment, post-contrast T1-weighted MRI from 156 patients with non-metastatic TNBC who underwent neoadjuvant chemotherapy. Tumor regions of interest were segmented by a convolutional neural network algorithm, with validation by breast radiologists. Features quantifying tumor shape and texture were extracted for the largest tumor present in each patient. We identified 23 principal components (PCs) describing these data within the original 364-dimensional feature space for further analysis. Tumor volume was also extracted for comparison with the shape and texture PCs, clinical variables and outcomes, but was kept separate from other radiomic features, since it directly correlates with clinical stage. We compiled for the cohort clinical variables including demographics, stage, grade, and, where available, absolute lymphocyte count (ALC) and Ki-67, a cellular proliferation index routinely used in clinical practice. We then performed a series of univariate and multivariate regression analyses to identify radiomic PCs and clinical variables that significantly predict patient outcomes, and radiomic PCs that predict established risk factors. Our multivariate analyses utilized 5-fold cross-validation and Monte-Carlo determination of p-values (based on 3000 random samplings from the null hypothesis), to ensure statistical rigor in identifying predictive relationships while correcting for multiple hypothesis testing. Results: Our univariate analyses confirmed expected correlations between: overall survival and pre-treatment tumor volume (p = 0.010); survival and ALC (p = 0.002); and clinical stage and tumor volume (p = 1.2⨉10-7). From our multivariate analysis, shape and texture radiomic features were predictive of: tumor volume (p &lt; 0.001); clinical stage (p &lt; 0.001); and Ki-67 (p = 0.02). We confirmed that Ki-67 was predictive of post-treatment residual cancer (p = 0.014), as has been previously reported. Conclusions: Radiomic features predict breast cancer risk factors that are significant for determining outcomes for TNBC patients. Combinations of radiomic shape and texture features track closely with tumor volumes, stage, and proliferative activity, potentially reflecting underlying molecular evolution.

Reproducibility assessment of radiomics features in various ultrasound scan settings and different scanner vendors

This study investigates the feasibility of using texture radiomics features extracted from mammography images to distinguish between benign and malignant breast lesions and to classify benign lesions into different categories and determine the best machine learning (ML) model to perform the tasks. Six hundred and twenty-two breast lesions from 200 retrospective patient data were segmented and analysed. Three hundred fifty radiomics features were extracted using the Standardized Environment for Radiomics Analysis (SERA) library, one of the radiomics implementations endorsed by the Image Biomarker Standardisation Initiative (IBSI). The radiomics features and selected patient characteristics were used to train selected machine learning models to classify the breast lesions. A fivefold cross-validation was used to evaluate the performance of the ML models and the top 10 most important features were identified. The random forest (RF) ensemble gave the highest accuracy (89.3%) and positive predictive value (66%) and likelihood ratio of 13.5 in categorising benign and malignant lesions. For the classification of benign lesions, the RF model again gave the highest likelihood ratio of 3.4 compared to the other models. Morphological and textural radiomics features were identified as the top 10 most important features from the random forest models. Patient age was also identified as one of the significant features in the RF model. We concluded that machine learning models trained against texture-based radiomics features and patient features give reasonable performance in differentiating benign versus malignant breast lesions. Our study also demonstrated that the radiomics-based machine learning models were able to emulate the visual assessment of mammography lesions, typically used by radiologists, leading to a better understanding of how the machine learning model arrive at their decision.

MRI pulse sequences and imaging parameters substantially influence the variation of MRI radiomics features, thus impose a critical challenge on MRI radiomics reproducibility and reliability. This study aims to prospectively investigate the impact of various imaging parameters on MRI radiomics features in a 3D T2-weighted (T2W) turbo-spin-echo (TSE) pulse sequence for MR-guided-radiotherapy (MRgRT). An anthropomorphic phantom was scanned using a 3D-T2W-TSE MRgRT sequence at 1.5T under a variety of acquisition imaging parameter changes. T1 and T2 relaxation times of the phantom were also measured. 93 first-order and texture radiomics features in the original and 14 transformed images, yielding 1,395 features in total, were extracted from 10 volumes-of-interest (VOIs). The percentage deviation (d%) of radiomics feature values from the baseline values and intra-class correlation coefficient (ICC) with the baseline were calculated. Robust radiomics features were identified based on the excellent agreement of radiomics feature values with the baseline, i.e., the averaged d% <5% and ICC >0.90 in all VOIs for all imaging parameter variations. The radiomics feature values changed considerably but to different degrees with different imaging parameter adjustments, in the ten VOIs. The deviation d% ranged from 0.02% to 321.3%, with a mean of 12.5% averaged for all original features in all ten VOIs. First-order and GLCM features were generally more robust to imaging parameters than other features in the original images. There were also significantly different radiomics feature values (ANOVA, P<0.001) between the original and the transformed images, exhibiting quite different robustness to imaging parameters. 330 out of 1395 features (23.7%) robust to imaging parameters were identified. GLCM and GLSZM features had the most (42.5%, 153/360) and least (3.8%, 9/240) robust features in the original and transformed images, respectively. This study helps better understand the quantitative dependence of radiomics feature values on imaging parameters in a 3D-T2W-TSE sequence for MRgRT. Imaging parameter heterogeneity should be considered as a significant source of radiomics variability and uncertainty, which must be well harmonized for reliable clinical use. The identified robust features to imaging parameters are helpful for the pre-selection of radiomics features for reliable radiomics modeling.

To prospectively investigate the impact of image reconstruction on MRI radiomics features. An anthropomorphic phantom was scanned at 1.5 T using a standardized sequence for MR-guided radiotherapy under SENSE and compressed-SENSE reconstruction settings. A total of 93 first-order and texture radiomics features in 10 volumes of interest were assessed based on (1) accuracy measured by the percentage deviation from the reference, (2) robustness on reconstruction in all volumes of interest measured by the intraclass correlation coefficient, and (3) repeatability measured by the coefficient of variance over the repetitive acquisitions. Finally, reliable and unreliable radiomics features were comprehensively determined based on their accuracy, robustness, and repeatability. Better accuracy and robustness of the radiomics features were achieved under SENSE than compressed-SENSE reconstruction. The feature accuracy under SENSE reconstruction was more affected by acceleration factor than direction, whereas under compressed-SENSE reconstruction, accuracy was substantially impacted by the increasing denoising levels. Feature repeatability was dependent more on feature types than on reconstruction. A total of 45 reliable features and 13 unreliable features were finally determined for SENSE, compared with 22 reliable and 26 unreliable features for compressed SENSE. First-order and gray-level co-occurrence matrix features were generally more reliable than other features. Radiomics features could be substantially affected by MRI reconstruction, so precautions need to be taken regarding their reliability for clinical use. This study helps the guidance of the preselection of reliable radiomics features and the preclusion of unreliable features in MR-guided radiotherapy.

The purpose of this study was to quantitatively assess the longitudinal acquisition repeatability of MRI radiomics features in a three-dimensional (3D) T1-weighted (T1W) TSE sequence via a well-controlled prospective phantom study. Thirty consecutive daily datasets of an ACR-MRI phantom were acquired on two 1.5T MRI simulators using a 3D T1W TSE sequence. Images were blindly segmented by two observers. Post-acquisition processing was minimized but an intensity discretization (fixed bin size of 25). One hundred and one radiomics features (shape n=12; first order n=16; texture n=73) were extracted. Longitudinal repeatability of each feature was evaluated by Pearson correlation and coefficient of variance (CV68% ). Interobserver feature value agreement was also quantified using intraclass correlation coefficient (ICC) and Bland-Altman analysis. A most repeatable radiomics feature set on both scanners was determined by feature coefficient of variance (CV68% <5%), ICC (>0.75), and the ratio of the interobserver difference to the interobserver mean δ<5%. No trend of radiomics feature value changed with time. Longitudinal feature repeatability CV68% ranged 0.01-38.60% (mean/median: 12.5%/9.9%), and 0.01-40.47%, (8.49%/7.34%) on the scanners A and B. Shape features exhibited significantly better repeatability than first-order and texture features (all P<0.01). Significant longitudinal repeatability difference was observed in texture features (P<0.001) between the two scanners, but not in shape and first-order features (P>0.30). First-order and texture features had smaller interobserver-dependent variation than acquisition-dependent variation. They also showed good interobserver agreement on both scanners (A:ICC=0.80±0.23; B:ICC=0.80±0.22), independent of acquisition repeatability. The repeatable radiomics features in common on both scanners, including 12 shape features, 0 first-order features, and 3 texture features, were determined as the most repeatable MRI radiomics feature set. Radiomics features exhibited heterogeneous longitudinal repeatability, while the shape features were the most repeatable, in this phantom study with a 3D T1W TSE acquisition. The most repeatable radiomics feature set derived in this study should be helpful for the selection of reliable radiomics features in the future clinical use.

Prediction of intracranial aneurysm rupture is important in the management of unruptured aneurysms. The application of radiomics in predicting aneurysm rupture remained largely unexplored. This study aims to evaluate the radiomics differences between ruptured and unruptured aneurysms and explore its potential use in predicting aneurysm rupture. One hundred twenty-two aneurysms were included in the study (93 unruptured). Morphological and radiomics features were extracted for each case. Statistical analysis was performed to identify significant features which were incorporated into prediction models constructed with a machine learning algorithm. To investigate the usefulness of radiomics features, three models were constructed and compared. The baseline model A was constructed with morphological features, while model B was constructed with addition of radiomics shape features and model C with more radiomics features. Multivariate analysis was performed for the ten most important variables in model C to identify independent risk factors. A simplified model based on independent risk factors was constructed for clinical use. Five morphological features and 89 radiomics features were significantly associated with rupture. Model A, model B, and model C achieved the area under the receiver operating characteristic curve of 0.767, 0.807, and 0.879, respectively. Model C was significantly better than model A and model B (p < 0.001). Multivariate analysis identified two radiomics features which were used to construct the simplified model showing an AUROC of 0.876. Radiomics signatures were different between ruptured and unruptured aneurysms. The use of radiomics features, especially texture features, may significantly improve rupture prediction performance. • Significant radiomics differences exist between ruptured and unruptured intracranial aneurysms. • Radiomics shape features can significantly improve rupture prediction performance over conventional morphology-based prediction model. The inclusion of histogram and texture radiomics features can further improve the performance. • A simplified model with two variables achieved a similar level of performance as the more complex ones. Our prediction model can serve as a promising tool for the risk management of intracranial aneurysms.