Clinical MRIs Research Articles

Toward More Accessible Fully Automated 3D Volumetric MRI Decision Trees for the Differential Diagnosis of Multiple System Atrophy, Related Disorders, and Age-Matched Healthy Subjects.

Differentiating multiple system atrophy (MSA) from related neurodegenerative movement disorders (NMD) is challenging. MRI is widely available and automated decision-tree analysis is simple, transparent, and resistant to overfitting. Using a retrospective cohort of heterogeneous clinical MRIs broadly sourced from a tertiary hospital system, we aimed to develop readily translatable and fully automated volumetric diagnostic decision-trees to facilitate early and accurate differential diagnosis of NMDs. 3DT1 MRI from 171 NMD patients (72 MSA, 49 PSP, 50 PD) and 171 matched healthy subjects were automatically segmented using Freesurfer6.0 with brainstem module. Decision trees employing substructure volumes and a novel volumetric pons-to-midbrain ratio (3D-PMR) were produced and tenfold cross-validation performed. The optimal tree separating NMD from healthy subjects selected cerebellar white matter, thalamus, putamen, striatum, and midbrain volumes as nodes. Its sensitivity was 84%, specificity 94%, accuracy 84%, and kappa 0.69 in cross-validation. The optimal tree restricted to NMD patients selected 3D-PMR, thalamus, superior cerebellar peduncle (SCP), midbrain, pons, and putamen as nodes. It yielded sensitivities/specificities of 94/84% for MSA, 72/96% for PSP, and 73/92% PD, with 79% accuracy and 0.62 kappa. There was correct classification of 16/17 MSA, 5/8 PSP, 6/8 PD autopsy-confirmed patients, and 6/8 MRIs that preceded motor symptom onset. Fully automated decision trees utilizing volumetric MRI data distinguished NMD patients from healthy subjects and MSA from other NMDs with promising accuracy, including autopsy-confirmed and pre-symptomatic subsets. Our open-source methodology is well-suited for widespread clinical translation. Assessment in even more heterogeneous retrospective and prospective cohorts is indicated.

The case for simul-omics: Identifying IO biomarkers of ganitumab-metformin pathological complete response (pCR).

e12551 Background: During the I-SPY2 clinical trial, ganitumab (G) immunotherapy (IO) was co-administered with metformin (GM) to counteract G-induced hyperglycemia. The GM + standard of care (SOC) combination showed promising results but did not meet the statistical threshold for phase III. Predicting individual patient response to IO is a key limitation for clinical use and a reason why IO phase II to III oncology trial transitions fail. Thus, we used SimBioSys TumorScope (TS), a 4D spatiotemporal multiscale biophysical model, to identify predictive IO biomarkers of pCR in response to GM+SOC. Methods: Patients that received GM+SOC therapy (n=41) and SOC (n=41) matched controls from the I-SPY2 trial were included in analyses. Using SOC clinical data and DCE MRIs, we generated biophysical simulations of each individual patient’s response to SOC therapy. Derived from the TS simulation object of each patient, our “simul-omics” data represent tumor morphology, breast tissue proportions, drug response and delivery, and microvasculature-related feature sets. We also had access to the pre-treatment tumor biopsy-derived transcriptome microarray data. Predictive modeling was performed in a modular 3-step fashion: 1. clinical-only data to predict pCR, 2. clinical data plus candidate simul-omics data to predict pCR, and 3. available feature space for features that interact with treatment to provide the strongest predictive model of pCR. Results: The SOC data model (HR status, grade, treatment, age, race) predicted pCR with an accuracy (acc) of 0.67, sensitivity (sens)=0.71, specificity (spec)=0.66, and Cohen’s kappa=0.27. Simul-omic features median tumoral Kt at sim week 1 (acc=0.73), and tumoral pre-contrast small area emphasis (acc=0.77) improved model performance over SOC. When included in the same model, they increased predictive acc of the model to 0.79 (sens=0.64, spec=0.82, kappa=0.41). Simulation-derived features improved predictive power beyond SOC data. Exploratory search of the feature-space identified simul-omic features that would generate the most predictive models with a significant treatment interaction effect. Two simul-omics features added predictive strength to the model, with near-nominal interaction effects: phi_quant_90_regimen_wk22 (late timepoint mv density) (acc=0.84, kappa=0.58), and slopemap_tumor_glcm_clustershade (acc=0.86, kappa=0.60). We tested transcript expression as interaction terms in the model. The best performing model with an interaction effect was SHISA4 (acc=0.88, sens=0.86, spec=0.88, kappa=0.67). Conclusions: Our results show value in: 1. predicting pCR, the additive value of TS simulation data over just clinical/SOC data, 2. identification of potentially mechanistic drug targets facilitating GM response, and 3. potential to stratify patient populations into responder vs. non-responder categories prior to treatment.

Dietary oily fish intake and progression of diffuse subcortical damage of vascular origin: A longitudinal prospective study in community-dwelling older adults.

Oily fish intake may reduce the progression of white matter hyperintensities (WMH) of presumed vascular origin due to their high content of omega-3 polyunsaturated fatty acids and other nutrients. However, information on this relationship is limited. We aimed to assess the association between oily fish intake and WMH progression in older adults living in rural coastal Ecuador. Participants of the Atahualpa Project Cohort received baseline clinical interviews and brain MRIs. Oily fish intake was calculated at every annual door-to-door survey from enrollment to the end of the study. Individuals who also received a follow-up brain MRI were included. Poisson regression models were fitted to assess the incidence rate ratio (IRR) of WMH progression according to the amount of oily fish intake, after adjusting for demographics, level of education and traditional vascular risk factors. The study included 263 individuals of Amerindian ancestry aged ⩾60 years (mean age: 65.7 ± 6.2 years; 57% women). The mean oily fish intake was 8.3 ± 4 servings per week. Follow-up MRIs demonstrated WMH progression in 103 (39%) individuals after a median follow-up of 6.5 years. A multivariate Poisson regression model showed an inverse relationship between oily fish intake and WMH progression (IRR: 0.89; 95% CI: 0.84-0.95; p < 0.001). A similar model also revealed an inverse relationship between tertiles of oily fish intake and probabilities of WMH progression, which became significant when individuals allocated to the third tertile were compared to those in the first and second tertiles. Study results show an inverse relationship between the amount of oily fish intake and WMH progression in frequent fish consumers of Amerindian ancestry.

Decreased Nighttime Heart Rate Variability and Progression of White Matter Hyperintensities of Presumed Vascular Origin. A Prospective Study in Community-Dwelling Older Adults

ObjectivesEvidence on the role of autonomic dysfunction on white matter hyperintensities (WMH) progression is limited. This study aims to assess the impact of a low nighttime heart rate variability (HRV) on WMH progression in community-dwelling older adults. Materials and methodsFollowing a prospective longitudinal study design, all individuals aged ≥60 years enrolled in the Atahualpa Project Cohort from 2012 to 2019 were invited to receive baseline HRV determinations through 24-h Holter monitoring, together with clinical interviews and brain MRIs. These individuals were periodically followed by means of annual door-to-door surveys, and those who also received brain MRIs at the end of the study (May 2021) were included in the analysis. Poisson regression models, adjusted for relevant confounders, were fitted to assess the incidence rate ratio (IRR) of WMH progression according to nighttime standard deviation of normal-to-normal R-R intervals (SDNN). ResultsThis study included 254 individuals aged ≥60 years (mean age: 65.4 ± 5.9 years; 55% women). The mean nighttime SDNN was 116.8 ± 36.3 ms. Follow-up MRIs showed WMH progression in 103 (41%) individuals after a median follow-up of 6.5 years. In unadjusted analyses, nighttime SDNN was lower among participants who developed WMH progression than in those who did not (p < 0.001). A Poisson regression model, adjusted for relevant covariates, disclosed a significantly inverse association between nighttime SDNN and WMH progression (IRR: 0.99; 95% C.I.: 0.98–0.99; p = 0.014). ConclusionsStudy results show an inverse association between nighttime SDNN and WMH progression, and provide support for the role of sympathetic overactivity in this relationship.

T2 FLAIR Hyperintensity Volume Is Associated With Cognitive Function and Quality of Life in Clinically Stable Patients With Lower Grade Gliomas.

Survival outcomes for patients with lower grade gliomas (LrGG) continue to improve. However, damage caused both by tumor growth and by the consequences of treatment often leads to significantly impaired cognitive function and quality of life (QoL). While neuropsychological testing is not routine, serial clinical MRIs are standard of care for patients with LrGG. Thus, having a greater understanding of MRI indicators of cognitive and QoL impairment risk could be beneficial to patients and clinicians. In this work we sought to test the hypothesis that in clinically stable LrGG patients, T2 FLAIR hyperintensity volumes at the time of cognitive assessment are associated with impairments of cognitive function and QoL and could be used to help identify patients for cognitive and QoL assessments and interventions. We performed anatomical MR imaging, cognitive testing and QoL assessments cross-sectionally in 30 clinically stable grade 2 and 3 glioma patients with subjective cognitive concerns who were 6 or more months post-treatment. Larger post-surgical T2 FLAIR volume at testing was significantly associated with lower cognitive performance, while pre-surgical tumor volume was not. Older patients had lower cognitive performance than younger patients, even after accounting for normal age-related declines in performance. Patients with Astrocytoma, IDH mutant LrGGs were more likely to show lower cognitive performance than patients with Oligodendroglioma, IDH mutant 1p19q co-deleted LrGGs. Previous treatment with combined radiation and chemotherapy was associated with poorer self-reported QoL, including self-reported cognitive function. This study demonstrates the importance of appreciating that LrGG patients may experience impairments in cognitive function and QoL over their disease course, including during periods of otherwise sustained clinical stability. Imaging factors can be helpful in identifying vulnerable patients who would benefit from cognitive assessment and rehabilitation.

Hyperintense vessels on imaging account for neurological function independent of lesion volume in acute ischemic stroke

In acute ischemic stroke, reported relationships between lesion metrics and behavior have largely focused on lesion volume and location. However, hypoperfusion has been shown to correlate with deficits in the acute stage. Hypoperfusion is typically identified using perfusion imaging in clinical settings, which requires contrast. Unfortunately, contrast is contraindicated for some individuals. An alternative method has been proposed to identify hypoperfusion using hyperintense vessels on fluid-attenuated inversion recovery (FLAIR) imaging. This study aimed to validate the clinical importance of considering hypoperfusion when accounting for behavior in acute stroke and demonstrate the clinical utility of scoring the presence of hyperintense vessels to quantify it.One hundred and fifty-three participants with acute ischemic stroke completed a battery of commonly-used neurological and behavioral measures. Clinical MRIs were used to determine lesion volume and to score the presence of hyperintense vessels seen on FLAIR images to estimate severity of hypoperfusion in six different vascular regions. National Institutes of Health Stroke Scale (NIHSS) scores, naming accuracy (left hemisphere strokes), and language content produced during picture description were examined in relation to lesion volume, hypoperfusion, and demographic variables using correlational analyses and multivariable linear regression.Results showed that lesion volume and hypoperfusion, in addition to demographic variables, were independently associated with performance on NIHSS, naming, and content production. Specifically, hypoperfusion in the frontal lobe independently correlated with NIHSS scores, while hypoperfusion in parietal areas independently correlated with naming accuracy and content production.These results correspond to previous reports associating hypoperfusion with function, confirming that hypoperfusion is an important consideration—beyond lesion volume—when accounting for behavior in acute ischemic stroke. Quantifying hypoperfusion using FLAIR hyperintense vessels can be an essential clinical tool when other methods of identifying hypoperfusion are unavailable or time prohibitive.

Automatic Comprehensive Radiological Reports for Clinical Acute Stroke MRIs

Automatic Comprehensive Radiological Reports for Clinical Acute Stroke MRIs

Subjective loss of clinical response to TNFi in axSpA relates to recurrence of MRI bone marrow oedema particularly with long-acting agents.

Subjective loss of response immediately prior to routine TNFi therapy can occur in axial spondyloarthritis (axSpA). We investigated clinical outcomes in patients taking the first three licenced TNFis and correlated this with recurrence of MRI bone marrow oedema (MRI-BMO). Proof-of-concept study including axSpA patients established on etanercept (ETA), adalimumab (ADA) or infliximab (IFX) reporting symptom deterioration prior to next dose. MRI/clinical data were collected prior to scheduled dose (v1), 4 days post-dose (v2) and at the time of patient-reported symptom return (v3). MRI spine/sacroiliac joints utilizing 3 T were scored using the semi-quantitative Leeds MRI scoring system. A total of 113 clinical assessments and MRIs were performed in 38 participants (ADA = 16, ETA = 12, IFX = 10), mean age 42.1 years ± 24.4(2SD, n = 38), 71.1% male (n = 27/38), 69.7% HLA-B27 positive (n = 23/33). At v1, all patients had high disease activity [ASDAS-CRP = 3 (2.7-3.7)] and 57.9% had MRI-BMO (number of MRI-BMO: ETA = 26, ADA = 59, IFX = 28). Improved clinical responses were seen at v2 [ASDAS-CRP -0.41(-0.81 - 0.30), P =0.018; BASDAI -0.58(-2.2 - 0.52), P =0.024]. Despite just a 4-day interval between v1 and v2, a numerical reduction in MRI-BMO lesions between v1/v2 was observed (ETA = -6, ADA = -10, IFX = -3). By v3, comparatively fewer new BMO lesions were detected in the ETA and ADA groups compared with IFX (ETA = -1, ADA = +3, IFX = +8), although the numbers were too small to enable testing for statistical significance. Short-lived fluctuations in MRI-BMO were commoner with longer-acting agents and corresponded with subjective loss of clinical response before next scheduled TNFi dose. Larger studies are needed to confirm the possible pathogenic implications of this phenomenon.

The triangular fibrocartilage complex on high-resolution 3\xa0T MRI in healthy adolescents: the thin line between asymptomatic findings and pathology

ObjectiveThe objective of the study is to provide a reference for morphology, homogeneity, and signal intensity of triangular fibrocartilage complex (TFCC) and TFCC-related MRI features in adolescents.Materials and methodsProspectively collected data on asymptomatic participants aged 12–18 years, between June 2015 and November 2017, were retrospectively analyzed. A radiograph was performed in all participants to determine skeletal age and ulnar variance. A 3-T MRI followed to assess TFCC components and TFCC-related features. A standardized scoring form, based on MRI definitions used in literature on adults, was used for individual assessment of all participants by four observers. Results per item were expressed as frequencies (percentages) of observations by all observers for all participants combined (n = 92). Inter-observer agreement was determined by the unweighted Fleiss’ kappa with 95% confidence intervals (95% CI).ResultsThe cohort consisted of 23 asymptomatic adolescents (12 girls and 11 boys). Median age was 13.5 years (range 12.0–17.0). Median ulnar variance was −0.7 mm (range − 2.7–1.4). Median triangular fibrocartilage (TFC) thickness was 1.4 mm (range 0.1–2.9). Diffuse increased TFC signal intensity not reaching the articular surface was observed in 30 (33%) observations and a vertical linear increased signal intensity with TFC discontinuation in 19 (20%) observations. Discontinuation between the volar radioulnar ligament and the TFC in the sagittal plane was seen in 23 (25%) observations. The extensor carpi ulnaris was completely dislocated in 10 (11%) observations, more frequent in supinated wrists (p = 0.031). Inter-observer agreement ranged from poor to fair for scoring items on the individual TFCC components.ConclusionMRI findings, whether normal variation or asymptomatic abnormality, can be observed in TFCC and TFCC-related features of asymptomatic adolescents. The rather low inter-observer agreement underscores the challenges in interpreting these small structures on MRI. This should be taken into consideration when interpreting clinical MRIs and deciding upon arthroscopy.

INTEGRATIVE RADIOMICS MODELS TO PREDICT BIOPSY RESULTS FOR NEGATIVE PROSTATE MRI.

Multi-parametric MRI (mpMRI) is a powerful non-invasive tool for diagnosing prostate cancer (PCa) and is widely recommended to be performed before prostate biopsies. Prostate Imaging Reporting and Data System version (PI-RADS) is used to interpret mpMRI. However, when the pre-biopsy mpMRI is negative, PI-RADS 1 or 2, there exists no consensus on which patients should undergo prostate biopsies. Recently, radiomics has shown great abilities in quantitative imaging analysis with outstanding performance on computer-aid diagnosis tasks. We proposed an integrative radiomics-based approach to predict the prostate biopsy results when pre-biopsy mpMRI is negative. Specifically, the proposed approach combined radiomics features and clinical features with machine learning to stratify positive and negative biopsy groups among negative mpMRI patients. We retrospectively reviewed all clinical prostate MRIs and identified 330 negative mpMRI scans, followed by biopsy results. Our proposed model was trained and validated with 10-fold cross-validation and reached the negative predicted value (NPV) of 0.99, the sensitivity of 0.88, and the specificity of 0.63 in receiver operating characteristic (ROC) analysis. Compared with results from existing methods, ours achieved 11.2% higher NPV and 87.2% higher sensitivity with a cost of 23.2% less specificity.

Prevalence of Imaging Targets in Patients With Minor Stroke Selected for IV tPA Treatment Using MRI: The Treatment of Minor Stroke With MRI Evaluation Study (TIMES).

To determine the IV tissue plasminogen activator (tPA) treatment rate of patients with minor acute ischemic stroke (mAIS) at our centers and compare the frequency of MRI targets by treatment stratification and clinical severity, we evaluated clinical characteristics and baseline MRIs for tPA-treated and untreated patients. Patients with ischemic stroke from 2015 to 2017 with admit NIH Stroke Scale (NIHSS) <6 were considered. The treated cohort received standard IV tPA and was screened with baseline MRI. The untreated cohort received no acute intervention and baseline MRI was <4 hours from onset. Patients were stratified into "clearly" and "not clearly" disabling deficits by NIHSS elements. Baseline MRI was evaluated by independent raters for AIS targets, with frequencies compared between groups. Of 255 patients with mAIS ≤4.5 hours from onset, 140 (55%) received IV tPA, accounting for 46% of all IV tPA patients (n = 305). Eighty-five percent (n = 119) were screened with baseline MRI and had significantly more frequent imaging targets compared to those untreated (n = 90). Of this treated cohort, 75% (n = 89) were not clearly disabling. Except for perfusion-diffusion mismatch (81% clearly disabling vs 56% not clearly disabling [p = 0.036]), there were no significant differences in the frequency of imaging targets across the treated cohort stratified by clinical severity. In MRI-screened mAIS, imaging targets were more frequently seen in patients treated with IV tPA, with similar frequencies even in those without clearly disabling deficits. MRI targets could be used to guide thrombolytic therapy in patients with mAIS; however, a randomized trial is needed to demonstrate efficacy.

Automated multiclass tissue segmentation of clinical brain MRIs with lesions

Delineation and quantification of normal and abnormal brain tissues on Magnetic Resonance Images is fundamental to the diagnosis and longitudinal assessment of neurological diseases. Here we sought to develop a convolutional neural network for automated multiclass tissue segmentation of brain MRIs that was robust at typical clinical resolutions and in the presence of a variety of lesions. We trained a 3D U-Net for full brain multiclass tissue segmentation from a prior atlas-based segmentation method on an internal dataset that consisted of 558 clinical T1-weighted brain MRIs (453/52/53; training/validation/test) of patients with one of 50 different diagnostic entities (n = 362) or with a normal brain MRI (n = 196). We then used transfer learning to refine our model on an external dataset that consisted of 7 patients with hand-labeled tissue types. We evaluated the tissue-wise and intra-lesion performance with different loss functions and spatial prior information in the validation set and applied the best performing model to the internal and external test sets. The network achieved an average overall Dice score of 0.87 and volume similarity of 0.97 in the internal test set. Further, the network achieved a median intra-lesion tissue segmentation accuracy of 0.85 inside lesions within white matter and 0.61 inside lesions within gray matter. After transfer learning, the network achieved an average overall Dice score of 0.77 and volume similarity of 0.96 in the external dataset compared to human raters. The network had equivalent or better performance than the original atlas-based method on which it was trained across all metrics and produced segmentations in a hundredth of the time. We anticipate that this pipeline will be a useful tool for clinical decision support and quantitative analysis of clinical brain MRIs in the presence of lesions.

Icobrain ms 5.1: Combining unsupervised and supervised approaches for improving the detection of multiple sclerosis lesions

Multiple sclerosis (MS) is a chronic autoimmune, inflammatory neurological disease of the central nervous system. Its diagnosis nowadays commonly includes performing an MRI scan, as it is the most sensitive imaging test for MS. MS plaques are commonly identified from fluid-attenuated inversion recovery (FLAIR) images as hyperintense regions that are highly varying in terms of their shapes, sizes and locations, and are routinely classified in accordance to the McDonald criteria. Recent years have seen an increase in works that aimed at development of various semi-automatic and automatic methods for detection, segmentation and classification of MS plaques. In this paper, we present an automatic combined method, based on two pipelines: a traditional unsupervised machine learning technique and a deep-learning attention-gate 3D U-net network. The deep-learning network is specifically trained to address the weaker points of the traditional approach, namely difficulties in segmenting infratentorial and juxtacortical plaques in real-world clinical MRIs. It was trained and validated on a multi-center multi-scanner dataset that contains 159 cases, each with T1 weighted (T1w) and FLAIR images, as well as manual delineations of the MS plaques, segmented and validated by a panel of raters. The detection rate was quantified using lesion-wise Dice score. A simple label fusion is implemented to combine the output segmentations of the two pipelines. This combined method improves the detection of infratentorial and juxtacortical lesions by 14% and 31% respectively, in comparison to the unsupervised machine learning pipeline that was used as a performance assessment baseline.

Intrinsic connectomes underlying response to trauma-focused psychotherapy in post-traumatic stress disorder

Although trauma-focused cognitive behavior therapy (TF-CBT) is the frontline treatment for post-traumatic stress disorder (PTSD), up to one-half of patients are treatment nonresponders. To understand treatment nonresponse, it is important to understand the neural mechanisms of TF-CBT. Here, we used whole-brain intrinsic functional connectivity analysis to identify neural connectomic signatures of treatment outcome. In total, 36 PTSD patients and 36 healthy individuals underwent functional MRI at pre-treatment baseline. Patients then underwent nine sessions of TF-CBT and completed clinical and follow-up MRIs. We used an established large-scale brain network atlas to parcellate the brain into 343 brain regions. Pairwise intrinsic task-free functional connectivity was calculated and used to identify pre-treatment connectomic features that were correlated with reduction of PTSD severity from pretreatment to post treatment. We formed a composite metric of intrinsic connections associated with therapeutic outcome, and then interrogated this composite metric to determine if it distinguished PTSD treatment responders and nonresponders from healthy control status and changed post treatment. Lower pre-treatment connectivity for the cingulo-opercular, salience, default mode, dorsal attention, and frontoparietal executive control brain networks was associated with treatment improvement. Treatment responders had lower while nonresponders had significantly greater connectivity than controls at pretreatment. With therapy, connectivity significantly increased for responders and decreased for nonresponders, while controls remain unchanged over this time period. We provide evidence that the intrinsic functional architecture of the brain, specifically connectivity within and between brain networks associated with external vigilance, self-awareness, and cognitive control, may characterize a positive response to TF-CBT for PTSD.

Effects of White Matter Hyperintensities on 90-Day Functional Outcome after Large Vessel and Non-Large Vessel Stroke

Introduction: White matter hyperintensity (WMH) burden is a critically important cerebrovascular phenotype related to the diagnosis and prognosis of acute ischemic stroke. The effect of WMH burden on functional outcome in large vessel occlusion (LVO) stroke has only been sparsely assessed, and direct LVO and non-LVO comparisons are currently lacking. Material and Methods: We reviewed acute ischemic stroke patients admitted between 2009 and 2017 at a large healthcare system in the USA. Patients with LVO were identified and clinical characteristics, including 90-day functional outcomes, were assessed. Clinical brain MRIs obtained at the time of the stroke underwent quantification of WMH using a fully automated algorithm. The pipeline incorporated automated brain extraction, intensity normalization, and WMH segmentation. Results: A total of 1,601 acute ischemic strokes with documented 90-day mRS were identified, including 353 (22%) with LVO. Among those strokes, WMH volume was available in 1,285 (80.3%) who had a brain MRI suitable for WMH quantification. Increasing WMH volume from 0 to 4 mL, age, female gender, a number of stroke risk factors, presence of LVO, and higher NIHSS at presentation all decreased the odds for a favorable outcome. Increasing WMH above 4 mL, however, was not associated with decreasing odds of favorable outcome. While WMH volume was associated with functional outcome in non-LVO stroke (p = 0.0009), this association between WMH and functional status was not statistically significant in the complete case multivariable model of LVO stroke (p = 0.0637). Conclusion: The burden of WMH has effects on 90-day functional outcome after LVO and non-LVO strokes. Particularly, increases from no measurable WMH to 4 mL of WMH correlate strongly with the outcome. Whether this relationship of increasing WMH to worse outcome is more pronounced in non-LVO than LVO strokes deserves additional investigation.

Using a simplified version of a common surgical grading scale for acetabular labral tears improves the utility of preoperative hip MRI for femoroacetabular impingement.

To evaluate whether a commonly used surgical grading scale, when applied to acetabular labral findings on MRI, could improve preoperative planning and counseling for patients undergoing hip arthroscopy. We evaluated 76 clinical MRIs performed on patients with femoroacetabular impingement. Three musculoskeletal radiologists and one musculoskeletal fellow reviewed each scan in a blinded fashion, classifying the acetabular labrum from 12:00 to 4:00 using the Beck scale, a common surgical grading scale. Clinical correlation was provided via surgical examination and classification. Reliability was determined between readers and between reader and surgical data using Cohen's kappa and Krippendorff's alpha at each clock position and for the worst grading for each scan. In addition, a simplified version of the scale comprised of only two grades, potentially reparable and not potentially reparable, was evaluated. When the scale was simplified into categories of potentially reparable and not potentially reparable, the sensitivity was excellent, ranging from 85.5 to 96%. Observer agreement when using individual Beck grades was found to range from poor to fair; Kappa ranged from 0.03 to 0.19, and Alpha ranged from - 0.27 to 0.22. The simplified version of the Beck labral scale when applied to MRI is a highly sensitive predictor of potentially reparable labral pathology while excluding normal and grossly degenerative tissue. Use of this scale provides clinically relevant information that can drive preoperative planning and improve patient counseling. It does so in a standardized fashion that can be applied across practice sites and without additional cost.

Perfusion kinetics from clinical DCE mris increase the accuracy of predictions of tumor response to chemotherapy.

e12651 Background: Nutrient and drug penetration into any solid tumor are critical determinants of the tumor's response to treatment. They depend on both the density of microvasculature within the tumor microenvironment, as well as the exchange rates of nutrients between the microvasculature and the extracellular space. But these parameters are heterogenous, varying considerably from location to location within the tumor and surrounding tissues. The Toft's model and its analogues date back to the early 1990s, and have been used to estimate vascular density, exchange rates, and extracellular-extravascular volume in a spatially-resolved manner using dynamic contrast enhaced (DCE) MRI's. Unfortunately, accurately extracting kinetic parameters from a DCE time-series requires the images to have a time-resolution of just a few seconds, which is rarely done in clinical practice. Methods: We employ a custom designed parallel algorithm to fit DCE MRI data to an exactly-solved ODE model of tissue perfusion kinetics. Results: Here we describe a simplified model of tissue perfusion that can be fit to DCE time traces with temporal resolutions of 90 seconds or more. We show that for many breast tumors, the vascular density and tissue-vascular exchange rate are such that they give rise to a halo of fast-perfusing tissue on the tumor periphery, and slower-perfusing tissue inside. We then use this model as part of a more comprehensive tumor simulation methodology to predict how different patients will respond to neoadjuvant chemotherapy (NACT). We find that the incorporation of our microvascular model gives rise to significantly more accurate predictions of post-treatment tumor volume. Conclusions: Performing perfusion kinetics analyses on clinical MRIs is both challenging, but critical for accurately predicting how a patient will respond to treatment. Our model, which relaxes the requirement for fine DCE temporal resolution, allows for these analyses to be performed on a larger swath of patients without the need for small volumes of interest, or ultra-fast MRI techniques. Moreover, when used within a broader tumor-modeling framework, our model increases the accuracy of predictions of tumor response to NACT.

Improvement in MR quality control workflow and outcomes with a web-based database.

PurposeTo describe a custom‐built, web‐based MR Quality Control (QC) database, and to assess its impact on the QC workflow and outcomes in a large U.S. academic medical center.MethodsThe MR QC database was built with Microsoft Access 2010 and published on a Microsoft Sharepoint website owned and maintained by the authors' institution. Authorized users can access the database remotely with mainstream web browsers on any institutional computers. QC technologists were granted access to add, review, and print daily and weekly QC records. Qualified medical physicists (QMPs) were granted additional access to edit, review, and approve existing QC records and to change tolerance limits. A macro was utilized to conduct an automatic weekly review of QC status and to email the results to a QMP. This web‐based QC database was implemented on 17 clinical MRIs at the authors' institution. Weekly ACR QC findings within one year before and after implementation were compared.ResultsWe analyzed 158 QC issues detected by the web‐based database and 127 QC issues identified in conventional paper records before we implemented the database. The web‐based database significantly reduced the number of QC issues due to technologist error (before/after: 59/24 cases, P < 0.0001) but did not affect the number of QC issues related to scanner performance (before/after: 49/46 cases, P = 1). Further analysis revealed that the web‐based database significantly reduced the average time for the QMPs to identify a QC issue (before/after: 177 ± 110/2 ± 2 days, P < 0.0001) and time to correction (before/after: 81 ± 102/7 ± 8 days, P < 0.0001). The correction rate also significantly increased (before/after: 22%/99%, P < 0.0001).ConclusionThe web‐based QC database provides a positive impact on our MR QC workflow and outcomes. It simplifies QC workflow, enables early detection of quality issues, and facilitates quick resolution of problems that may affect the quality of clinical MRI studies.

Effect of radiation therapy on cerebral cortical thickness in glioma patients: Treatment-induced thinning of the healthy cortex.

BackgroundWith overall survival of brain tumors improving, radiation induced brain injury is becoming an increasing issue. One of the effects of radiation therapy (RT) is thinning of the cerebral cortex, which could be one of the factors contributing to cognitive impairments after treatment. In healthy brain, cortex thickness varies between 1 and 4.5 mm. In this study, we assess the effect of RT on the thickness of the cerebral cortex and relate the changes to the local dose.MethodsWe identified 28 glioma patients with optimal scan quality. Clinical CTs and MRIs at baseline and 1 year post-RT were collected and coregistered. The scans were processed via an automated image processing pipeline, which enabled measuring changes of the cortical thickness, which were related to local dose.ResultsThree areas were identified where significant dose-dependent thinning occurred, with thinning rates of 5, 6, and 26 μm/Gy after 1 year, which corresponds to losses of 5.4%, 7.2%, and 21.6% per 30 Gy per year. The first area was largely located in the right inferior parietal, supramarginal, and superior parietal regions, the second in the right posterior cingulate and paracentral regions, and the third almost completely in the right lateral orbital frontal region.ConclusionsWe have identified three areas susceptible to dose-dependent cortical thinning after radiation therapy. Should future prospective studies conclude that irradiation of these areas lead to cognitive decline, they need to be spared in order to prevent this debilitating consequence of treatment.

Convolutional Neural Network for Automated FLAIR Lesion Segmentation on Clinical Brain MR Imaging.

Most brain lesions are characterized by hyperintense signal on FLAIR. We sought to develop an automated deep learning-based method for segmentation of abnormalities on FLAIR and volumetric quantification on clinical brain MRIs across many pathologic entities and scanning parameters. We evaluated the performance of the algorithm compared with manual segmentation and existing automated methods. We adapted a U-Net convolutional neural network architecture for brain MRIs using 3D volumes. This network was retrospectively trained on 295 brain MRIs to perform automated FLAIR lesion segmentation. Performance was evaluated on 92 validation cases using Dice scores and voxelwise sensitivity and specificity, compared with radiologists' manual segmentations. The algorithm was also evaluated on measuring total lesion volume. Our model demonstrated accurate FLAIR lesion segmentation performance (median Dice score, 0.79) on the validation dataset across a large range of lesion characteristics. Across 19 neurologic diseases, performance was significantly higher than existing methods (Dice, 0.56 and 0.41) and approached human performance (Dice, 0.81). There was a strong correlation between the predictions of lesion volume of the algorithm compared with true lesion volume (ρ = 0.99). Lesion segmentations were accurate across a large range of image-acquisition parameters on >30 different MR imaging scanners. A 3D convolutional neural network adapted from a U-Net architecture can achieve high automated FLAIR segmentation performance on clinical brain MR imaging across a variety of underlying pathologies and image acquisition parameters. The method provides accurate volumetric lesion data that can be incorporated into assessments of disease burden or into radiologic reports.