Augmenting inter-rater concordance of radiologic extranodal extension in HPV-positive oropharyngeal carcinoma: A multicenter study.

OBJECTIVES: To assess intra- and inter-institutional concordance and identify methods to increase precision in radiologic extranodal extension (rENE) ascertainment in HPV+ oropharyngeal carcinoma. METHODS: Six radiologists, blinded to clinical outcomes, from three centers assessed rENE in two phases: Phase-I (20 cases) utilized each individual's a priori appreciation of the literature. Phase-II (30 additional cases) was performed after deliberating experience and consolidating operating definitions. Intra- and inter-institutional Kappa were calculated at >50% and >75% certainty levels, respectively. RESULTS: The Phase-I intra-institutional kappa was 0.76, 0.32, and 0.44 at >50% certainty and improved to 0.89, 0.61, and 0.66 at >75% certainty. Inter-institutional Fleiss' kappa also improved with higher certainty (from 0.40 to 0.57, p = 0.039).  The Phase-II inter-rater kappa was significantly higher than Phase-I at the same certainty level (both p < 0.001). CONCLUSION: A learning curve exists for rENE assessment. Strategies to augment reliability include high certainty for declaration, consolidated operating definitions, and sharing experience among radiologists.

Increase in blood-brain barrier leakage in healthy, older adults.

Blood-brain barrier (BBB) breakdown can disrupt nutrient supply and waste removal, which affects neuronal functioning. Currently, dynamic contrast-enhanced (DCE) MRI is the preferred in-vivo method to quantify BBB leakage. Dedicated DCE MRI studies in normal aging individuals are lacking, which could hamper value estimation and interpretation of leakage rate in pathological conditions. Therefore, we applied DCE MRI to investigate the association between BBB disruption and age in a healthy sample. Fifty-seven cognitively and neurologically healthy, middle-aged to older participants (mean age: 66 years, range: 47-91 years) underwent MRI, including DCE MRI with intravenous injection of a gadolinium-based contrast agent. Pharmacokinetic modeling was applied to contrast concentration time-curves to estimate BBB leakage rate in each voxel. Subsequently, leakage rate was calculated in the white and gray matter, and primary (basic sensory and motor functions), secondary (association areas), and tertiary (higher-order cognition) brain regions. A difference in vulnerability to deterioration was expected between these regions, with especially tertiary regions being affected by age. Higher BBB leakage rate was significantly associated with older age in the white and gray matter, and also in tertiary, but not in primary or secondary brain regions. Even in healthy individuals, BBB disruption was stronger in older persons, which suggests BBB disruption is a normal physiologically aging phenomenon. Age-related increase in BBB disruption occurred especially in brain regions most vulnerable to age-related deterioration, which may indicate that BBB disruption is an underlying mechanism of normal age-related decline.Netherlands Trial Register number: NL6358, date of registration: 2017-03-24.

White matter hyperintensities mediate the association between blood-brain barrier leakage and information processing speed.

Blood-brain barrier (BBB) leakage is considered an important underlying process in both cerebral small vessel disease (cSVD) and Alzheimer's disease (AD). The objective of this study was to examine associations between BBB leakage, cSVD, neurodegeneration, and cognitive performance across the spectrum from normal cognition to dementia. Leakage was measured with dynamic contrast-enhanced magnetic resonance imaging in 80 older participants (normal cognition, n = 32; mild cognitive impairment, n = 34; clinical AD-type dementia, n = 14). Associations between leakage and white matter hyperintensity (WMH) volume, hippocampal volume, and cognition (information processing speed and memory performance) were examined with multivariable linear regression and mediation analyses. Leakage within the gray and white matter was positively associated with WMH volume (gray matter, p = 0.03; white matter, p = 0.01). A negative association was found between white matter BBB leakage and information processing speed performance, which was mediated by WMH volume. Leakage was not associated with hippocampal volume. WMH pathology is suggested to form a link between leakage and decline of information processing speed in older individuals with and without cognitive impairment.

The interobserver agreement in the detection of recurrent HNSCC using MRI including diffusion weighted MRI.

INTRODUCTION: For the detection of local recurrences of head and neck squamous cell carcinomas (HNSCC) after (chemo)radiation, diagnostic imaging is generally performed. Diffusion weighted magnetic resonance imaging (DW-MRI) has been proven to be able to adequately diagnose the presence of cancer. However evaluation of DW-MR images for recurrences is difficult and could be subject to individual interpretation. AIM: To determine the interobserver agreement, intraobserver agreement and influence of experience of radiologists in the assessment of DW-MRI in patients clinically suspected of local recurrent HNSCC after (chemo)radiation. METHODS: Ten experienced head and neck radiologists assessed follow-up MRI including DW-MRI series of 10 patients for the existence of local recurrence on a two point decision scale (local recurrence or local control). Patients were clinically suspected for a recurrence of laryngeal (n = 3), hypopharyngeal (n = 3) or oropharyngeal (n = 4) cancer after (chemo)radiation with curative intent. Fleiss' and Cohen's Kappa were used to determine interobserver agreement and intraobserver agreement, respectively. RESULTS: Interobserver agreement was κ = 0.55. Intraobserver agreement was κ = 0.80. Prior experience within the field of radiology and with DW-MRI had no significant influence on the scoring. CONCLUSION: For the assessment of HNSCC recurrence after (chemo)radiation by DW-MRI, moderate interobserver agreement and substantial intraobserver agreement was found.

DWI for Assessment of Rectal Cancer Nodes After Chemoradiotherapy: Is the Absence of Nodes at DWI Proof of a Negative Nodal Status?

OBJECTIVE: When considering organ preservation in patients with rectal cancer with good tumor response, assessment of a node-negative status after chemoradiation therapy (CRT) is important. DWI is a very sensitive technique to detect nodes. The study aim was to test the hypothesis that the absence of nodes at DWI after CRT is concordant with a ypN0 status. MATERIALS AND METHODS: A retrospective study was performed of 90 patients with rectal cancer treated with CRT followed by restaging MRI at 1.5 T, including DWI (highest b value, 1000 s/mm2). Two independent readers counted the number of nodes visible in the mesorectal compartment on DW images obtained after CRT. The number of nodes on DWI (0 vs ≥ 1) was compared with the number of metastatic nodes at histopathology or long-term clinical follow-up (yN0 vs yN-positive status). RESULTS: Seventy-one patients had a yN0 status, and 19 had a yN-positive status. For 10 patients, no nodes were observed at DWI, which was concordant with a yN0 status in 100% of cases. In the other 61 patients with a yN0 status, the median number of nodes detected at DWI was three (range, 1-17 nodes). To differentiate between yN0 and yN-positive status, sensitivity was 100%, specificity was 14%, the positive predictive value was 24%, and the negative predictive value was 100%. CONCLUSION: Although the absence of nodes at DWI is not a frequent finding, it appears to be a reliable predictor of yN0 status after CRT in patients with rectal cancer. DWI may thus be a helpful adjunct in assessing response after CRT and may help select patients for organ-saving treatment.

Cerebral atrophy in elderly with subjective memory complaints.

PURPOSE: To evaluate ventricular shape differences along the complete surface of the lateral and third ventricles of persons with subjective memory complaints (MC). MATERIALS AND METHODS: We included 28 controls and 21 persons with MC. FLAIR, T2, and PD-weighted brain MRI scans were acquired at 1.5 Tesla, followed by semi-automated segmentation of the lateral and third ventricles, and local shape difference analysis based on growing and adaptive meshes. Ventricular meshes were used to highlight local areas with significant differences between controls and persons with MC, determined by permutation tests with a predefined threshold (P = 0.01). RESULTS: Compared with control subjects, relevant differences were found in the shape of the ventricular surface adjacent to the thalamus and corona radiata in persons with MC. Before correction for multiple comparisons, relevant differences were also found in the shape of the ventricular surface adjacent to the corpus callosum, hippocampus, and amydala. CONCLUSION: Our findings suggest the presence of localized structural brain differences in patients with subjective memory complaints in the thalamus and the corona radiata.

Ventricular dilation: association with gait and cognition.

OBJECTIVE: Normal pressure hydrocephalus is characterized by gait impairment, cognitive impairment, and urinary incontinence, and is associated with disproportionate ventricular dilation. Here we report the distribution of ventricular volume relative to sulcal cerebrospinal fluid (CSF) volume, and the association of increasing ventricular volume relative to sulcal CSF volume with a cluster of gait impairment, cognitive impairment, and urinary incontinence in a stroke-free cohort of elderly persons from the general population. METHODS: Data are based on 858 persons (35.4% men; age range, 66-92 years) who participated in the Age, Gene/Environment Susceptibility-Reykjavik Study. Gait was evaluated with an assessment of gait speed. Composite scores representing speed of processing, memory, and executive function were constructed from a neuropsychological battery. Bladder function was assessed with a questionnaire. Magnetic resonance brain imaging was followed by semiautomated segmentation of intracranial CSF volume. White matter hyperintensity (WMH) volume was assessed with a semiquantitative scale. For the analysis of ventricular dilation relative to the sulcal spaces, ventricular volume was divided by sulcal CSF volume (VV/SV). RESULTS: Disproportion between ventricular and sulcal CSF volume, defined as the highest quartile of the VV/SV z score, was associated with gait impairment (odds ratio [OR], 1.9; 95% confidence interval [CI], 1.1-3.3) and cognitive impairment (OR, 1.8; 95% CI, 1.1-3.0). We did not find an association between the VV/SV z score and bladder dysfunction. INTERPRETATION: The prevalence and severity of gait impairment and cognitive impairment increases with ventricular dilation in persons without stroke from the general population, independent of WMH volume.

Ventricular shape biomarkers for Alzheimer's disease in clinical MR images.

The aim of this work was to identify ventricular shape-based biomarkers in MR images to discriminate between patients with Alzheimer's disease (AD) and healthy elderly. Clinical MR images were collected for 58 patients and 28 age-matched healthy controls. After normalizing all the images the ventricular cerebrospinal fluid was semiautomatically extracted for each subject and an innovative technique for fully automatic shape modeling was applied to generate comparable meshes of all ventricles. The search for potential biomarkers was carried out with repeated permutation tests: results highlighted well-defined areas of the ventricular surface being discriminating features for AD: the left inferior medial temporal horn, the right medial temporal horn (superior and inferior), and the areas close to the left anterior part of the corpus callosum and the head of the right caudate nucleus. The biomarkers were then used as features to build an intelligent machine for AD detection: a Support Vector Machine was trained on AD and healthy subjects and subsequently tested with leave-1-out experiments and validation tests on previously unseen cases. The results showed a sensitivity of 76% for AD, with an overall accuracy of 84%, proving that suitable biomarkers for AD can be detected in clinical MR images.

MMSE scores correlate with local ventricular enlargement in the spectrum from cognitively normal to Alzheimer disease.

In this work, we aimed at correlating focal atrophy in periventricular structures with cognitive function, in the spectrum from healthy subjects to severe Alzheimer disease: 28 subjects with normal cognition and 84 patients presenting various degrees of cognitive impairment were included in the study. The cognitive level of each subject was assessed with the Mini-Mental State Examination (MMSE). Atrophy in periventricular structures was inferred by modeling and analyzing local shape variations of brain ventricles: for a given subject, we distinguished between the severity of atrophy, estimated as local enlargement (in mm) of the ventricular surface relative to an average normal subject, and the extent of atrophy, defined as the percentage of the ventricular surface (global or per anatomical region) significantly different from an average control. Linear regression across subjects was performed to evaluate the correlation between atrophy and MMSE score. The severity of atrophy showed good correlation with MMSE score in the left thalamus, the left temporal horn, the left corona radiata, and the right caudate nuclei. The extent of atrophy showed no significant correlations. In conclusion, the MMSE scores correlate with localized depth of atrophy in well-defined periventricular structures.

GAMEs: growing and adaptive meshes for fully automatic shape modeling and analysis.

This paper presents a new framework for shape modeling and analysis, rooted in the pattern recognition theory and based on artificial neural networks. Growing and adaptive meshes (GAMEs) are introduced: GAMEs combine the self-organizing networks which grow when require (SONGWR) algorithm and the Kohonen's self-organizing maps (SOMs) in order to build a mesh representation of a given shape and adapt it to instances of similar shapes. The modeling of a surface is seen as an unsupervised clustering problem, and tackled by using SONGWR (topology-learning phase). The point correspondence between point distribution models is granted by adapting the original model to other instances: the adaptation is seen as a classification task and performed accordingly to SOMs (topology-preserving phase). We thoroughly evaluated our method on challenging synthetic datasets, with different levels of noise and shape variations. Finally, we describe its application to the analysis of a challenging medical dataset. Our method proved to be reproducible, robust to noise, and capable of capturing real variations within and between groups of shapes.

Shape differences of the brain ventricles in Alzheimer's disease.

The brain ventricles are surrounded by gray and white matter structures that are often affected in dementia in general and Alzheimer's disease (AD) in particular. Any change of volume or shape occurring in these structures must affect the volume and shape of the ventricles. It is well known that ventricular volume is significantly higher in AD patients compared to age-matched healthy subjects. However, the large overlap between the two volume distributions makes the measurement unsuitable as a biomarker of the disease. The purpose of this work was to assess whether local shape differences of the ventricles can be detected when comparing AD patients and controls. In this work, we captured the ventricle's shape and shape variations of 29 AD subjects and 25 age-matched controls, using a fully automatic shape modeling technique. By applying permutation tests on every single node of a mesh representation of the shapes, we identified local areas with significant differences. About 22% of an average surface of the ventricles presented significant difference (P < 0.05) ( approximately 14% of the left against approximately 7% of the right side). We found out that in patients with Alzheimer disease, not only the lateral horns were significantly affected, but also the areas adjacent to the anterior corpus callosum, the splenium of the corpus callosum, the amygdala, the thalamus, the tale of the caudate nuclei (especially the left one), and the head of the left caudate nucleus.