Influence of Time of Admission on Endovascular Thrombectomy (EVT): Comparison of Outcomes During Business Hours Versus Off-Business Hours.

Purpose:To investigate the differences in endovascular thrombectomy (EVT) outcomes of patients treated for acute ischaemic stroke (AIS) during business versus off-business hours. Methods: A single-centre retrospective cohort study of patients with AIS treated with EVT from February 1, 2015, to May 31, 2021, was performed at a comprehensive stroke centre (CSC). Patients were divided into business (Monday to Friday, 8 AM-5 PM) versus off-business hours groups. The primary outcome was functional neurological disability, scored using the modified Rankin Scale (mRS) at 90 days. Secondary outcomes included the rate of successful reperfusion and procedural workflow time delays. Differences in proportions were assessed using Fisher's exact and Chi-Square tests as appropriate. For continuous variables, differences in medians between groups were assessed using Mann-Whitney U tests. Results: A total of 676 patients were included, with 399 patients (59%) comprising the off-business-hour group. No significant differences were seen in age, sex, ASPECTS score, or NIHSS at arrival. Off-business hours strokes had a longer delay between CSC arrival to groin puncture (minutes: 81 vs 44, P < .0001) and between imaging to groin puncture (minutes: 67 vs 32, P < .0001) compared to the business hours strokes. There were no differences in the rate of successful reperfusion (mTICI ≥2b) between groups (82% vs 83%, P = .61). At 90 days, 65% of patients in both groups had an mRS ≤2 (P = .91). Conclusion: Despite workflow delays in initiating EVT during off-business hours, there were no differences in the rate of successful reperfusion or functional outcomes.

MRI Markers of Degenerative Disc Disease in Young Patients With Multiple Sclerosis.

Background and Purpose: Evidence has emerged for an association between degenerative disc disease (DDD) and multiple sclerosis (MS). The purpose of the current study is to determine the presence and extent of cervical DDD in young patients (age <35) with MS, an age cohort that is less well studied for these changes. Methods: Retrospective chart review of consecutive patients aged <35 referred from the local MS clinic who were MRI scanned between May 2005 and November 2014. 80 patients (51 female and 29 male) with MS of any type ranging between 16 and 32 years of age (average 26) were included. Images were reviewed by 3 raters and assessed for presence and extent of DDD, as well as cord signal abnormalities. Interrater agreement was assessed using Kendall's W and Fleiss' Kappa statistics. Results: Substantial to very good interrater agreement was observed using our novel DDD grading scale. At least some degree of DDD was found in over 91% of patients. The majority scored mild (grade 1, 30-49%) to moderate (grade 2, 39-51%) degenerative changes. Cord signal abnormality was seen in 56-63%. Cord signal abnormality, when present, occurred exclusively at degenerative disc levels in only 10-15%, significantly lower than other distributions (P < .001 for all pairwise comparisons). Conclusions: MS patients demonstrate unexpected cervical DDD even at a young age. Future study is warranted to investigate the underlying etiology, such as altered biomechanics. Furthermore, cord lesions were found to occur independently of DDD.

MRI texture analysis of acetabular cancellous bone can discriminate between normal, cam positive, and cam-FAI hips.

OBJECTIVES: To compare the MRI texture profile of acetabular subchondral bone in normal, asymptomatic cam positive, and symptomatic cam-FAI hips and determine the accuracy of a machine learning model for discriminating between the three hip classes. METHODS: A case-control, retrospective study was performed including 68 subjects (19 normal, 26 asymptomatic cam, 23 symptomatic cam-FAI). Acetabular subchondral bone of unilateral hip was contoured on 1.5 T MR images. Nine first-order 3D histogram and 16 s-order texture features were evaluated using specialized texture analysis software. Between-group differences were assessed using Kruskal-Wallis and Mann-Whitney U tests, and differences in proportions compared using chi-square and Fisher's exact tests. Gradient-boosted ensemble methods of decision trees were created and trained to discriminate between the three groups of hips, with percent accuracy calculated. RESULTS: Sixty-eight subjects (median age 32 (28-40), 60 male) were evaluated. Significant differences among all three groups were identified with first-order (4 features, all p ≤ 0.002) and second-order (11 features, all p ≤ 0.002) texture analyses. First-order texture analysis could differentiate between control and cam positive hip groups (4 features, all p ≤ 0.002). Second-order texture analysis could additionally differentiate between asymptomatic cam and symptomatic cam-FAI groups (10 features, all p ≤ 0.02). Machine learning models demonstrated high classification accuracy of 79% (SD 16) for discriminating among all three groups. CONCLUSION: Normal, asymptomatic cam positive, and cam-FAI hips can be discriminated based on their MRI texture profile of subchondral bone using descriptive statistics and machine learning algorithms. CLINICAL RELEVANCE STATEMENT: Texture analysis can be performed on routine MR images of the hip and used to identify early changes in bone architecture, differentiating morphologically abnormal from normal hips, prior to onset of symptoms. KEY POINTS: • MRI texture analysis is a technique for extracting quantitative data from routine MRI images. • MRI texture analysis demonstrates that there are different bone profiles between normal hips and those with femoroacetabular impingement. • Machine learning models can be used in conjunction with MRI texture analysis to accurately differentiate between normal hips and those with femoroacetabular impingement.

Clinical Features, Non-Contrast CT Radiomic and Radiological Signs in Models for the Prediction of Hematoma Expansion in Intracerebral Hemorrhage.

PURPOSE: Rapid identification of hematoma expansion (HE) risk at baseline is a priority in intracerebral hemorrhage (ICH) patients and may impact clinical decision making. Predictive scores using clinical features and Non-Contract Computed Tomography (NCCT)-based features exist, however, the extent to which each feature set contributes to identification is limited. This paper aims to investigate the relative value of clinical, radiological, and radiomics features in HE prediction. METHODS: Original data was retrospectively obtained from three major prospective clinical trials ["Spot Sign" Selection of Intracerebral Hemorrhage to Guide Hemostatic Therapy (SPOTLIGHT)NCT01359202; The Spot Sign for Predicting and Treating ICH Growth Study (STOP-IT)NCT00810888] Patients baseline and follow-up scans following ICH were included. Clinical, NCCT radiological, and radiomics features were extracted, and multivariate modeling was conducted on each feature set. RESULTS: 317 patients from 38 sites met inclusion criteria. Warfarin use (p=0.001) and GCS score (p=0.046) were significant clinical predictors of HE. The best performing model for HE prediction included clinical, radiological, and radiomic features with an area under the curve (AUC) of 87.7%. NCCT radiological features improved upon clinical benchmark model AUC by 6.5% and a clinical & radiomic combination model by 6.4%. Addition of radiomics features improved goodness of fit of both clinical (p=0.012) and clinical & NCCT radiological (p=0.007) models, with marginal improvements on AUC. Inclusion of NCCT radiological signs was best for ruling out HE whereas the radiomic features were best for ruling in HE. CONCLUSION: NCCT-based radiological and radiomics features can improve HE prediction when added to clinical features.

Diagnostic Accuracy of Centrally Restricted Diffusion Sign in Cerebral Metastatic Disease: Differentiating Radiation Necrosis from Tumor Recurrence.

Purpose: The centrally restricted diffusion sign of diffusion-weighted imaging (DWI) is associated with radiation necrosis (RN) in treated gliomas. Our goal was to evaluate its diagnostic accuracy to distinguish RN from tumor recurrence (TR) in treated brain metastases. Methods: Retrospective study of consecutive patients with brain metastases who developed a newly centrally necrotic lesion after radiotherapy (RT). One reader placed regions of interest (ROI) in the enhancing solid lesion and the non-enhancing central necrosis on the apparent diffusion coefficient (ADC) map. Two readers qualitatively assessed the presence of the centrally restricted diffusion sign. The final diagnosis was made by histopathology (n = 39) or imaging follow-up (n = 2). Differences between groups were assessed by Fisher's exact or Mann-Whitney U tests. Diagnostic accuracy and inter-reader agreement were evaluated using receiver operating characteristic (ROC) curve analysis and kappa scores. Results: Forty-one lesions (32 predominant RN; 9 predominant TR) were analyzed. An ADC value ≤ 1220 × 10-6 mm2/s (sensitivity 74%, specificity 89%, area under the curve [AUC] .85 [95% confidence interval {CI}, .70-.94] P < .0001) from the necrosis and an ADC necrosis/enhancement ratio ≤1.37 (sensitivity 74%, specificity 89%, AUC .82 [95% CI, .67-.93] P < .0001) provided the highest performance for RN diagnosis. The qualitative centrally restricted diffusion sign had a sensitivity of 69% (95% CI, .50-.83), specificity of 77% (95% CI, .40-.96), and a moderate (k = .49) inter-reader agreement for RN diagnosis. Conclusions: Radiation necrosis is associated with lower ADC values in the central necrosis than TR. A moderate interobserver agreement might limit the qualitative assessment of the centrally restricted diffusion sign.

Aetiology of extracranial carotid free-floating thrombus in a prospective multicentre cohort.

BACKGROUND: Carotid free-floating thrombi (FFT) in patients with acute transient ischaemic attack (TIA)/stroke have a high risk of early recurrent stroke. Management depends on aetiology, which can include local plaque rupture, dissection, coagulopathy, malignancy and cardioembolism. Our objectives were to classify the underlying aetiology of FFT and to estimate the proportion of patients with underlying stenosis requiring revascularisation. METHODS: We prospectively enrolled consecutive patients presenting to three comprehensive stroke centres with acute TIA/stroke and ipsilateral internal carotid artery FFT. The aetiology of FFT was classified as: carotid atherosclerotic disease, carotid dissection, cardioembolism, both carotid atherosclerosis and cardioembolism, or embolic stroke of uncertain source (ESUS). Patients with carotid atherosclerosis were further subclassified as having ≥50% or <50% stenosis. RESULTS: We enrolled 83 patients with confirmed FFT. Aetiological assessments revealed 66/83 (79.5%) had carotid atherosclerotic plaque, 4/83 (4.8%) had a carotid dissection, 10/83 (12%) had both atrial fibrillation and carotid atherosclerotic plaque and 3/83 (3.6%) were classified as ESUS. Of the 76 patients with atherosclerotic plaque (including those with atrial fibrillation), 40 (52.6%) had ≥50% ipsilateral stenosis. CONCLUSIONS: The majority of symptomatic carotid artery FFT are likely caused by local plaque rupture, more than half of which are associated with moderate to severe carotid stenosis requiring revascularisation. However, a significant number of FFTs are caused by non-atherosclerotic mechanisms warranting additional investigations.

Rule-based natural language processing for automation of stroke data extraction: a validation study.

PURPOSE: Data extraction from radiology free-text reports is time consuming when performed manually. Recently, more automated extraction methods using natural language processing (NLP) are proposed. A previously developed rule-based NLP algorithm showed promise in its ability to extract stroke-related data from radiology reports. We aimed to externally validate the accuracy of CHARTextract, a rule-based NLP algorithm, to extract stroke-related data from free-text radiology reports. METHODS: Free-text reports of CT angiography (CTA) and perfusion (CTP) studies of consecutive patients with acute ischemic stroke admitted to a regional stroke center for endovascular thrombectomy were analyzed from January 2015 to 2021. Stroke-related variables were manually extracted as reference standard from clinical reports, including proximal and distal anterior circulation occlusion, posterior circulation occlusion, presence of ischemia or hemorrhage, Alberta stroke program early CT score (ASPECTS), and collateral status. These variables were simultaneously extracted using a rule-based NLP algorithm. The NLP algorithm's accuracy, specificity, sensitivity, positive predictive value (PPV), and negative predictive value (NPV) were assessed. RESULTS: The NLP algorithm's accuracy was > 90% for identifying distal anterior occlusion, posterior circulation occlusion, hemorrhage, and ASPECTS. Accuracy was 85%, 74%, and 79% for proximal anterior circulation occlusion, presence of ischemia, and collateral status respectively. The algorithm confirmed the absence of variables from radiology reports with an 87-100% accuracy. CONCLUSIONS: Rule-based NLP has a moderate to good performance for stroke-related data extraction from free-text imaging reports. The algorithm's accuracy was affected by inconsistent report styles and lexicon among reporting radiologists.

Assessment of left atrial fibrosis progression in canines following rapid ventricular pacing using 3D late gadolinium enhanced CMR images.

BACKGROUND: Atrial fibrillation (AF) is associated with extracellular matrix (ECM) remodelling and often coexists with myocardial fibrosis (MF); however, the causality of these conditions is not well established. OBJECTIVE: We aim to corroborate AF to MF causality by quantifying left atrial (LA) fibrosis in cardiac magnetic resonance (CMR) images after persistent rapid ventricular pacing and subsequent AF using a canine model and histopathological validation. METHODS: Twelve canines (9 experimental, 3 control) underwent baseline 3D LGE-CMR imaging at 3T followed by insertion of a pacing device and 5 weeks of rapid ventricular pacing to induce AF (experimental) or no pacing (control). Following the 5 weeks, pacing devices were removed to permit CMR imaging followed by excision of the hearts and histopathological imaging. LA myocardial segmentation was performed manually at baseline and post-pacing to permit volumetric %MF quantification using the image intensity ratio (IIR) technique, wherein fibrosis was defined as pixels > mean LA myocardium intensity + 2SD. RESULTS: Volumetric %MF increased by an average of 2.11 ± 0.88% post-pacing in 7 of 9 experimental dogs. While there was a significant difference between paired %MF measurements from baseline to post-pacing in experimental dogs (P = 0.019), there was no significant change in control dogs (P = 0.019 and P = 0.5, Wilcoxon signed rank tests). The median %MF for paced animals was significantly greater than that of non-paced dogs at the 5-week post-insertion time point (P = 0.009, Mann Whitney U test). Histopathological imaging yielded an average %MF of 19.42 ± 4.80% (mean ± SD) for paced dogs compared to 1.85% in one control dog. CONCLUSION: Persistent rapid ventricular pacing and subsequent AF leads to an increase in LA fibrosis volumes measured by the IIR technique; however, quantification is limited by inherent image acquisition parameters and observer variability.

Utility of machine learning of apparent diffusion coefficient (ADC) and T2-weighted (T2W) radiomic features in PI-RADS version 2.1 category 3 lesions to predict prostate cancer diagnosis.

PURPOSE: To evaluate if machine learning (ML) of radiomic features extracted from apparent diffusion coefficient (ADC) and T2-weighted (T2W) MRI can predict prostate cancer (PCa) diagnosis in Prostate Imaging-Reporting and Data System (PI-RADS) version 2.1 category 3 lesions. METHODS: This multi-institutional review board-approved retrospective case-control study evaluated 158 men with 160 PI-RADS category 3 lesions (79 peripheral zone, 81 transition zone) diagnosed at 3-Tesla MRI with histopathology diagnosis by MRI-TRUS-guided targeted biopsy. A blinded radiologist confirmed PI-RADS v2.1 score and segmented lesions on axial T2W and ADC images using 3D Slicer, extracting radiomic features with an open-source software (Pyradiomics). Diagnostic accuracy for (1) any PCa and (2) clinically significant (CS; International Society of Urogenital Pathology Grade Group ≥ 2) PCa was assessed using XGBoost with tenfold cross -validation. RESULTS: From 160 PI-RADS 3 lesions, there were 50.0% (80/160) PCa, including 36.3% (29/80) CS-PCa (63.8% [51/80] ISUP 1, 23.8% [19/80] ISUP 2, 8.8% [7/80] ISUP 3, 3.8% [3/80] ISUP 4). The remaining 50.0% (80/160) lesions were benign. ML of all radiomic features from T2W and ADC achieved area under receiver operating characteristic curve (AUC) for diagnosis of (1) CS-PCa 0.547 (95% Confidence Intervals 0.510-0.584) for T2W and 0.684 (CI 0.652-0.715) for ADC and (2) any PCa 0.608 (CI 0.579-0.636) for T2W and 0.642 (CI 0.614-0.0.670) for ADC. CONCLUSION: Our results indicate ML of radiomic features extracted from T2W and ADC achieved at best moderate accuracy for determining which PI-RADS category 3 lesions represent PCa.

Differentiating Carotid Free-Floating Thrombus From Atheromatous Plaque Using Intraluminal Filling Defect Length on CTA: A Validation Study.

OBJECTIVE: To validate a previously proposed filling defect length threshold of >3.8 mm on CT angiography (CTA) to discriminate between free-floating thrombus (FFT) and plaque of atheroma. METHODS: This was a prospective multicenter observational study of 100 participants presenting with TIA/stroke symptoms and a carotid intraluminal filling defect on initial CTA. Follow-up CTA was obtained within 1 week and at weeks 2 and 4 if the intraluminal filling defect was unchanged in length. Resolution or decreased length was diagnostic of FFT, whereas its static appearance after 4 weeks was indicative of plaque. Diagnostic accuracy of FFT length was assessed by receiver operating characteristic analysis. RESULTS: Ninety-five participants (mean [SD] age 68 [13] years, 61 men, 83 participants with FFT, 12 participants with a plaque) were evaluated. The >3.8-mm threshold had a sensitivity of 88% (73 of 83) (95% confidence interval [CI] 78%-94%) and specificity of 83% (10 of 12) (95% CI 51%-97%) (area under the curve 0.91, p < 0.001) for the diagnosis of FFT. The optimal length threshold was >3.64 mm with a sensitivity of 89% (74 of 83) (95% CI 80%-95%) and specificity of 83% (10 of 12) (95% CI 51%-97%). Adjusted logistic regression showed that every 1-mm increase in intraluminal filling defect length is associated with an increase in odds of FFT of 4.6 (95% CI 1.9-11.1, p = 0.01). CONCLUSION: CTA enables accurate differentiation of FFT vs plaque using craniocaudal length thresholds. TRIAL REGISTRATION INFORMATION: ClinicalTrials.gov Identifier: NCT02405845. CLASSIFICATION OF EVIDENCE: This study provides Class I evidence that in patients with TIA/stroke symptoms, the presence of CTA-identified filling defects of lengths >3.8 mm accurately discriminates FFT from atheromatous plaque.

Assessment of Renal Cell Carcinoma by Texture Analysis in Clinical Practice: A Six-Site, Six-Platform Analysis of Reliability.

BACKGROUND. Multiple commercial and open-source software applications are available for texture analysis. Nonstandard techniques can cause undesirable variability that impedes result reproducibility and limits clinical utility. OBJECTIVE. The purpose of this study is to measure agreement of texture metrics extracted by six software packages. METHODS. This retrospective study included 40 renal cell carcinomas with contrast-enhanced CT from The Cancer Genome Atlas and Imaging Archive. Images were analyzed by seven readers at six sites. Each reader used one of six software packages to extract commonly studied texture features. Inter- and intrareader agreement for segmentation was assessed with intraclass correlation coefficients (ICCs). First-order (available in six packages) and second-order (available in three packages) texture features were compared between software pairs using Pearson correlation. RESULTS. Inter- and intrareader agreement was excellent (ICC, 0.93-1). First-order feature correlations were strong (r ≥ 0.8, p < .001) between 75% (21/28) of software pairs for mean intensity and SD, 48% (10/21) for entropy, 29% (8/28) for skewness, and 25% (7/28) for kurtosis. Of 15 second-order features, only cooccurrence matrix correlation, gray-level nonuniformity, and run-length nonuniformity showed strong correlation between software packages (r = 0.90-1, p < .001). CONCLUSION. Variability in first- and second-order texture features was common across software configurations and produced inconsistent results. Standardized algorithms and reporting methods are needed before texture data can be reliably used for clinical applications. CLINICAL IMPACT. It is important to be aware of variability related to texture software processing and configuration when reporting and comparing outputs.

Effect of phase of enhancement on texture analysis in renal masses evaluated with non-contrast-enhanced, corticomedullary, and nephrographic phase-enhanced CT images.

OBJECTIVE: To compare texture analysis (TA) features of solid renal masses on renal protocol (non-contrast enhanced [NECT], corticomedullary [CM], nephrographic [NG]) CT. MATERIALS AND METHODS: A total of 177 consecutive solid renal masses (116 renal cell carcinoma [RCC]; 51 clear cell [cc], 40 papillary, 25 chromophobe, and 61 benign masses; 49 oncocytomas, 12 fat-poor angiomyolipomas) with three-phase CT between 2012 and 2017 were studied. Two blinded radiologists independently assessed tumor heterogeneity (5-point Likert scale) and segmented tumors. TA features (N = 25) were compared between groups and between phases. Accuracy (area under the curve [AUC]) for RCC versus benign and cc-RCC versus other masses was compared. RESULTS: Subjectively, tumor heterogeneity differed between phases (p < 0.01) and between tumors within the same phase (p = 0.03 [NECT] and p < 0.01 [CM, NG]). Inter-observer agreement was moderate to substantial (intraclass correlation coefficient = 0.55-0.73). TA differed in 92.0% (23/25) features between phases (p < 0.05) except for GLNU and f6. More TA features differed significantly on CM (80.0% [20/25]) compared with NG (40.0% [10/25]) and NECT (16.0% [4/25]) (p < 0.01). For RCC versus benign, AUCs of texture features did not differ comparing CM and NG (p > 0.05), but were higher for 20% (5/25) and 28% (7/25) of features comparing CM and NG with NECT (p < 0.05). For cc-RCC versus other, 36% (9/25) and 40% (10/25) features on CM had higher AUCs compared with NECT and NG images (p < 0.05). CONCLUSION: Texture analysis of renal masses differs, when evaluated subjectively and quantitatively, by phase of CT enhancement. The corticomedullary phase had the highest discriminatory value when comparing masses and for differentiating cc-RCC from other masses. KEY POINTS: • Subjectively evaluated renal tumor heterogeneity on CT differs by phase of enhancement. • Quantitative CT texture analysis features in renal tumors differ by phases of enhancement with the corticomedullary phase showing the highest number and most significant differences compared with non-contrast-enhanced and nephrographic phase images. • For diagnosis of clear cell RCC, corticomedullary phase texture analysis features had improved accuracy of classification in approximately 40% of features studied compared with non-contrast-enhanced and nephrographic phase images.

Preoperative Determination of Isocitrate Dehydrogenase Mutation in Gliomas Using Spectral Editing MRS: A Prospective Study.

BACKGROUND: The edited magnetic resonance spectroscopy (MRS) technique has not yet been formally evaluated for the in vivo detection of 2-hydroxyglutarate (2-HG) in patients with gliomas of various grades. PURPOSE: To evaluate the diagnostic accuracy of edited MRS in the preoperative identification of the isocitrate dehydrogenase (IDH) mutation status in patients with gliomas. STUDY TYPE: Prospective. POPULATION: Fifty-eight subjects (31 glioblastomas, 27 grade II and III gliomas). FIELD STRENGTH/SEQUENCE: Mescher-Garwood (MEGA)-PRESS and routine clinical brain tumor MR sequences were used at 3T. ASSESSMENT: Data were analyzed using an advanced method for accurate, robust, and efficient spectral fitting (AMARES) from jMRUI software. The amplitudes of the 2-HG, N-acetyl-aspartate (NAA), choline (Cho), and creatine/phosphocreatine (Cr) resonances were calculated with their associated Cramer-Rao lower bound (CRLB). The IDH1 R132H mutation status was assessed by immunohistochemistry for all patients. Patients with grades II and III gliomas with negative immunohistochemistry underwent DNA sequencing to further interrogate IDH mutation status. STATISTICAL TEST: The differences in 2-HG amplitudes, 2-HG/NAA, 2-HG/Cho, and 2-HG/Cr between IDH-mutant and IDH-wildtype gliomas were assessed using Mann-Whitney U-tests. Receiver operating characteristic curve analysis was performed to evaluate the diagnostic accuracy of each parameter. RESULTS: The 2-HG amplitudes, 2-HG/NAA, and 2-HG/Cho were higher for IDH-mutant gliomas than IDH-wildtype gliomas (P < 0.007). Using a CRLB threshold <30%, a 2-HG cutoff greater than 0 had a sensitivity of 80% (95% confidence interval [CI]: 52-96%) and a specificity of 81% (95% CI: 54-96%) in identifying IDH-mutant gliomas. In the subset of patients with grades II and III gliomas, the sensitivity was 80% (95% CI: 52-96%) and specificity was 100% (95% CI: 40-100%). Among 2-HG ratios, the highest AUC for the identification of IDH mutant status was achieved using the 2-HG/NAA (AUC = 0.8, 95% CI 0.67-.89). DATA CONCLUSION: Preoperative edited MRS appears to be able to help identify IDH-mutant gliomas with high specificity. Level of Evidence 1 Technical Efficacy Stage 2 J. MAGN. RESON. IMAGING 2021;53:416-426.

Importance of phase enhancement for machine learning classification of solid renal masses using texture analysis features at multi-phasic CT.

OBJECTIVE: To compare machine learning (ML) of texture analysis (TA) features for classification of solid renal masses on non-contrast-enhanced CT (NCCT), corticomedullary (CM) and nephrographic (NG) phase contrast-enhanced (CE) CT. MATERIALS AND METHODS: With IRB approval, we retrospectively identified 177 consecutive solid renal masses (116 renal cell carcinoma [RCC]; 51 clear cell [cc], 40 papillary, 25 chromophobe and 61 benign tumors; 49 oncocytomas and 12 fat-poor angiomyolipomas) with renal protocol CT between 2012 and 2017. Tumors were independently segmented by two blinded radiologists. Twenty-five 2-dimensional TA features were extracted from each phase. Diagnostic accuracy for 1) RCC versus benign tumor and 2) cc-RCC versus other tumor was assessed using XGBoost. RESULTS: ML of texture analysis features on different phases achieved mean area under the ROC curve (AUC [SD]), sensitivity/specificity for 1) RCC vs benign = 0.70(0.19), 96%/32% on CM-CECT and 0.71(0.14), 83%/58% on NG-CECT and; 2) cc-RCC vs other = 0.77(0.12), 49%/90% on CM-CECT and 0.71(0.16), 22%/94% on NG-CECT. There was no difference in AUC comparing CECT to NCCT (p = 0.058-0.54) and no improvement when combining data across all three phases compared single-phase assessment (p = 0.39-0.68) for either outcome. AUCs decreased when ML models were trained with one phase and tested on a different phase for both outcomes (RCC;p = 0.045-0.106, cc-RCC; < 0.001). CONCLUSION: Accuracy of machine learning classification of renal masses using texture analysis features did not depend on phase; however, models trained using one phase performed worse when tested on another phase particularly when associating NCCT and CECT. These findings have implications for large registries which use varying CT protocols to study renal masses.

Shape Analysis of Peripheral Zone Observations on Prostate DWI: Correlation to Histopathology Outcomes After Radical Prostatectomy.

OBJECTIVE. The objective of our study was to subjectively and quantitatively assess shape features of peripheral zone (PZ) tumors at DWI compared with pathologic outcomes. MATERIALS AND METHODS. During the study period, 241 consecutive men with PZ dominant prostate tumors underwent 3-T MRI including DWI before undergoing radical prostatectomy. DW images of these patients were retrospectively assessed by two blinded radiologists. The reviewers assigned Prostate Imaging Reporting and Data System (PI-RADS) shape categories (round or oval, crescentic [i.e., conforming to PZ], linear or wedge-shaped) and segmented tumors for quantitative shape analysis. Discrepancies were resolved by consensus. Comparisons were performed with Gleason score (GS) and pathologic stage. RESULTS. Consensus review results were as follows: 63.9% (154/241) of tumors were round or oval; 22.8% (55/241), crescentic; and 13.3% (32/241), linear or wedge-shaped. Agreement for shape assessment was moderate (κ = 0.41). Round or oval tumors were higher grade (GS 6 = 1.3%, GS 7 = 78.0%, GS ≥ 8 = 20.7%) than crescentic tumors (GS 6 = 9.1%, GS 7 = 74.6%, GS ≥ 8 = 16.3%) and linear or wedge-shaped tumors (GS 6 = 6.3%, GS 7 = 78.1%, GS ≥ 8 = 15.6%) (p = 0.011). In addition, round or oval tumors had higher rates of extraprostatic extension (EPE) and seminal vesicle invasion (SVI) (EPE and SVI: 70.1% and 26.0%) than crescentic tumors (67.3% and 9.1%; p = 0.003) and linear or wedge-shaped tumors (40.6% and 9.4%; p = 0.008). Quantitatively, the shape features termed "circularity" and "roundness" were associated with EPE (p < 0.001 and p = 0.003), SVI (p < 0.001 and p = 0.029), and increasing GS (p = 0.009 and p = 0.021), but there was overlap between groups. CONCLUSION. In this study, approximately 10% of resected PZ tumors were linear or wedge-shaped on DWI. PZ tumors that were judged subjectively and evaluated quantitatively to be round or oval were associated with increased prostate cancer aggressiveness.

CT texture analysis of acetabular subchondral bone can discriminate between normal and cam-positive hips.

OBJECTIVES: The purpose of this study was to determine if the CT texture profile of acetabular subchondral bone differs between normal, asymptomatic cam-positive, and symptomatic cam-FAI hips. In addition, the utility of texture analysis to discriminate between the three hip statuses was explored using a machine learning approach. METHODS: IRB-approved, case-control study analyzing CT images in subjects with and without cam morphology from August 2010 to December 2013. Sixty-eight subjects were included: 19 normal controls, 26 asymptomatic cam, and 23 symptomatic cam-FAI. Acetabular subchondral bone was contoured on the sagittal oblique CT images using ImageJ ®. 3D histogram texture features (mean, variance, skewness, kurtosis, and percentiles) were evaluated using MaZda software. Groupwise differences were investigated using Kruskal-Wallis tests and Mann-Whitney U tests. Gradient-boosted decision trees were created and trained to discriminate between control and cam-positive hips. RESULTS: Both asymptomatic and symptomatic cam-FAI hips demonstrated significantly higher values of texture variance (p = 0.0007, p < 0.0001), 90th percentile (p = 0.007, p = 0.006), and 99th percentile (p = 0.009, p = 0.009), but significantly lower values of skewness (p = 0.0001, p = 0.0013) and kurtosis (p = 0.0001, p = 0.0001) compared to normal controls. There were no differences in texture profile between asymptomatic cam and symptomatic cam-FAI hips. Machine learning models demonstrated high classification accuracy for discriminating control hips from asymptomatic cam-positive (82%) and symptomatic cam-FAI (86%) hips. CONCLUSIONS: Texture analysis can discriminate between normal and cam-positive hips using conventional descriptive statistics, regression modeling, and machine learning algorithms. It has the potential to become an important tool in compositional analysis of hip subchondral trabecular bone in the context of FAI, and possibly serve as a biomarker of joint degeneration. KEY POINTS: • The CT texture profile of acetabular subchondral bone is significantly different between normal and cam-positive hips. • Texture analysis can detect changes in subchondral bone in asymptomatic cam-positive hips that are equal to that of symptomatic cam-FAI hips. • Texture analysis has the potential to become an important tool in compositional analysis of hip subchondral bone in the context of FAI and may serve as a biomarker in the study of joint physiology and biomechanics.

Left atrial vortex size and velocity distributions by 4D flow MRI in patients with paroxysmal atrial fibrillation: Associations with age and CHA2 DS2 -VASc risk score.

BACKGROUND: Characterization of left atrial (LA) hemodynamics in paroxysmal atrial fibrillation (PAF) may provide valuable insights for thromboembolic risk. PURPOSE: To evaluate LA vortex formation and velocity distributions by 4D flow MRI and identify associations with age, LA/LV (left ventricle) function, and established risk scores. STUDY TYPE: Prospective clinical. POPULATION: Patients with PAF (n = 45, 46 ± 14 years) and healthy controls (n = 15, 54 ± 9 years) were enrolled. MRI SEQUENCES: 3T standardized cardiac MRI protocol inclusive of 4D flow MRI. ASSESSMENT: Flow analysis planes were prescribed at each pulmonary vein. Velocity distribution analysis and vortex size quantification by the Lambda2 (λ2 ) method were performed in the LA. STATISTICS: Pearson or Spearman's correlation coefficients, r, were calculated to identify relationships between 4D flow-derived LA parameters and age, LA/LV function, and CHA2 DS2 -VASc stroke risk score. Univariate and multivariate determinants of stroke risk were assessed using linear regressions. To compare parameters within multiple groups, one-way analysis of variance or Kruskal-Wallis was used. RESULTS: LA vortice sizes were observed in all subjects using λ2 showing inverse correlations with peak pulmonary vein inflow velocities (P < 0.05), and positive correlations with LA volume (P < 0.05). Vortex size was elevated in PAF at all phases of the cardiac cycle, being most prominent at end early diastole (3.98 ± 1.84 cm3 vs. 6.93 ± 3.11 cm3 , P = 0.001). Velocity distribution analysis showed a greater incidence of flow stasis among patients with PAF (P < 0.05). In univariate regression, vortex size was associated with the CHA2 DS2 -VASc risk score at peak systole (0.457 ± 0.038, P ≤ 0.001). However, in multivariate regression age was the dominant determinant of stroke risk (0.348 ± 0.012, P = 0.006). DATA CONCLUSION: This study demonstrated that LA vortex size is increased among low-risk patients with PAF and is associated with the CHA2 DS2 -VASc risk score. Age remained the dominant determinant of stroke risk. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2020;51:871-884.

Left atrial imaging and registration of fibrosis with conduction voltages using LGE-MRI and electroanatomical mapping.

BACKGROUND AND PURPOSE: Abnormal electrical conduction and excitability associated with fibrosis in the left atrium (LA) may serve as a substrate for atrial fibrillation (AF). Electroanatomical voltage mapping systems (EAMs) have become a dominant facilitator to treat AF with catheter ablation assisted by additional diagnostic imaging modalities. Importantly, AF has been associated with structural changes to the extracellular matrix of the myocardium, including increased collagen deposition-a process known as fibrosis. Late gadolinium enhancement-magnetic resonance imaging (LGE-MRI) may aid in guiding AF cardiac ablation therapy by determination of location of fibrosis in the LA. To locate fibrosis for cardiac ablation, however, accurate registration between EAMs and LGE-MRI data is crucial. The purpose of this work was to develop a method for registering EAMs with late gadolinium enhancement-magnetic resonance (LGE-MR) images of fibrosis. METHODS: Twenty patients with persistent AF, who underwent magnetic resonance imaging scanning and EAMs prior to first-time catheter ablation, participated in the study. In our registration pipeline, LGE-MR images were registered to the left atrial surface on EAMs using manual alignment followed by iterative closest point (ICP), and non-rigid ICP (NICP) algorithm. RESULTS AND CONCLUSIONS: The results demonstrate that NICP provided a substantial reduction in registration error when compared to the use of affine ICP alone. Regions of fibrosis on LGE-MR images identified using the signal threshold to reference mean threshold demonstrated the most regional overlap with low bipolar voltage points on EAMs. Successful co-registration of LGE-MR images to EAMs may assist electro-physiologists in selecting candidate targets for ablation and ultimately, reduce the rate of AF recurrence for patients.

Transition zone prostate cancer: Logistic regression and machine-learning models of quantitative ADC, shape and texture features are highly accurate for diagnosis.

BACKGROUND: The limitation of diagnosis of transition zone (TZ) prostate cancer (PCa) using subjective assessment of multiparametric (mp) MRI with PI-RADS v2 is related to overlapping features between cancers and stromal benign prostatic hyperplasia (BPH) nodules, particularly in small lesions. PURPOSE: To evaluate modeling of quantitative apparent diffusion coefficient (ADC), texture, and shape features using logistic regression (LR) and support vector machine (SVM) models for the diagnosis of transition zone PCa. STUDY TYPE: Retrospective. POPULATION: Ninety patients; 44 consecutive TZ PCa were compared with 61 consecutive BPH nodules (26 glandular/35 stromal). FIELD STRENGTH/SEQUENCE: 3 T/T2 -weighted (T2 W) fast spin-echo, diffusion weighted imaging. ASSESSMENT: A radiologist manually segmented lesions on axial images for quantitative ADC (mean, 10th , 25th -centile-ADC), T2 W-shape (circularity, convexity) and T2 W-texture (kurtosis, skewness, entropy, run-length nonuniformity [RLNU], gray-level nonuniformity [GLNU]) analysis. A second radiologist segmented one-fifth of randomly selected lesions to determine the reproducibility of measurements. The reference standard was histopathology for all lesions. STATISTICAL TESTS: Quantitative features were selected a priori and were compared using univariate and multivariate analysis. LR and SVM models of statistically significant features were constructed and evaluated using receiver operator characteristic (ROC) analysis. Subgroup analysis of TZ PCa vs. only stromal BPH and in lesions measuring <15 mm was performed. Agreement in measurements was assessed using the Dice similarity coefficient (DSC). RESULTS: Mean, 25th and 10th -centile ADC, circularity, and texture (entropy, RLNU, GLNU) features differed between groups (P < 0.0001-0.0058); however, at multivariate analysis only circularity and ADC metrics (P < 0.001) remained significant. LR and SVM models were highly accurate for the diagnosis of TZ PCa (sensitivity/specificity/AUC): 93.2%/98.4%/0.989 and 93.2%/96.7%/0.949, respectively, with no significance difference between the LR and SVM models (P = 0.2271). Reproducibility of segmentation was excellent (DSC 0.84 tumors and 0.87 BPH). Subgroup analyses of TZ PCa vs. stromal BPH (AUC = 0.976) and in <15 mm lesions (AUC = 0.990) remained highly accurate. DATA CONCLUSION: LR and SVM models incorporating previously described quantitative ADC, shape and texture analysis features are highly accurate for the diagnosis of TZ PCa and remained accurate when comparing TZ PCa with stromal BPH and in smaller lesions. LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;50:940-950.

Prospective comparative diagnostic accuracy evaluation of dynamic contrast-enhanced (DCE) vs. dynamic susceptibility contrast (DSC) MR perfusion in differentiating tumor recurrence from radiation necrosis in treated high-grade gliomas.

BACKGROUND: The appearance of a new enhancing lesion after surgery and chemoradiation for high-grade glioma (HGG) presents a common diagnostic dilemma. Histopathological analysis remains the reference standard in this situation. PURPOSE: To prospectively compare the diagnostic accuracy of dynamic contrast-enhanced (DCE) vs. dynamic susceptibility contrast (DSC) in differentiating tumor recurrence (TR) from radiation necrosis (RN). STUDY TYPE: Prospective diagnostic accuracy study. POPULATION: In all, 98 consecutive treated HGG patients with new enhancing lesion. We excluded 32 patients due to inadequate follow-up or technical limitation. FIELD STRENGTH/SEQUENCE: 3 T DCE and DSC MR. ASSESSMENT: Histogram and hot-spot analysis of cerebral blood volume (CBV), corrected CBV, Ktrans , area under the curve (AUC), and plasma volume (Vp). The reference standard of TR and/or RN was determined by histopathology in 43 surgically resected lesions or by clinical/imaging follow-up in the rest. STATISTICAL TESTS: Mann-Whitney U-tests, receiver operating characteristic (ROC) curve, and logistic regression analysis. RESULTS: A total of 68 lesions were included. There were 37 TR, 28 RN, and three lesions with equal proportions of TR and RN. TR had significantly higher CBV, corrected CBV, CBV ratio, corrected CBV ratio, AUC ratio, and Vp ratio (P < 0.05) than RN on hot-spot analysis. CBV had the highest diagnostic accuracy (AUROC 0.71). On histogram analysis, TR had higher CBV and corrected CBV maximal value compared with RN (P = 0.006, AUROC = 0.70). Only CBV on hot-spot analysis remained significant after correction for multiple comparison, with no significant improvement in diagnostic accuracy when using a combination of parameters (AUROC 0.71 vs. 0.76, P = 0.24). DATA CONCLUSION: DSC-derived CBV is the most accurate perfusion parameter in differentiating TR and RN. DSC and DCE-derived parameters reflecting the blood volume in an enhancing lesion are more accurate than the DCE-derived parameter Ktrans . Clinical practice may be best guided by blood volume measurements, rather than permeability assessment for differentiation of TR from RN. LEVEL OF EVIDENCE: 1 Technical Efficacy Stage: 4 J. Magn. Reson. Imaging 2019;50:573-582.