Disturbed sex hormone milieu in males and females with major depressive disorder and low-grade inflammation.

Sex hormones have biological effects on inflammation, and these might contribute to the sex-specific features of depression. C-reactive protein (CRP) is the most widely used inflammatory biomarker and consistent evidence shows a significant proportion (20-30 %) of patients with major depressive disorder (MDD) have CRP levels above 3 mg/L, a threshold indicating at least low-grade inflammation. Here, we investigate the interplay between sex hormones and CRP in the cross-sectional, observational Biomarkers in Depression Study. We measured serum high-sensitivity (hs-)CRP, in 64 healthy controls and 178 MDD patients, subdivided into those with hs-CRP below 3 mg/L (low-CRP; 53 males, 72 females) and with hs-CRP above 3 mg/L (high-CRP; 19 males, 34 females). We also measured interleukin-6, testosterone, 17-ß-estradiol (E2), progesterone, sex-hormone binding globulin (SHBG), follicle-stimulating and luteinising hormones, and calculated testosterone-to-E2 ratio (T/E2), free androgen and estradiol indexes (FAI, FEI), and testosterone secretion index. In males, high-CRP patients had lower testosterone than controls (p = 0.001), and lower testosterone (p = 0.013), T/E2 (p < 0.001), and higher FEI (p = 0.015) than low-CRP patients. In females, high-CRP patients showed lower SHGB levels than controls (p = 0.033) and low-CRP patients (p = 0.034). The differences in testosterone, T/E2 ratio, and FEI levels in males survived the Benjamini-Hochberg FDR correction. In linear regression analyses, testosterone (ß = -1.069 p = 0.033) predicted CRP concentrations (R2 = 0.252 p = 0.002) in male patients, and SHBG predicted CRP levels (ß = -0.628 p = 0.009, R2 = 0.172 p = 0.003) in female patients. These findings may guide future research investigating interactions between gonadal and immune systems in depression, and the potential of hormonal therapies in MDD with inflammation.

Higher immune-related gene expression in major depression is independent of CRP levels: results from the BIODEP study.

Compelling evidence demonstrates that some individuals suffering from major depressive disorder (MDD) exhibit increased levels of inflammation. Most studies focus on inflammation-related proteins, such as serum or plasma C-reactive protein (CRP). However, the immune-related modifications associated with MDD may be not entirely captured by CRP alone. Analysing mRNA gene expression levels, we aimed to identify broader molecular immune-related phenotypes of MDD. We examined 168 individuals from the non-interventional, case-control, BIODEP study, 128 with a diagnosis of MDD and 40 healthy controls. Individuals with MDD were further divided according to serum high-sensitivity (hs)CRP levels (n = 59 with CRP <1, n = 33 with CRP 1-3 and n = 36 with CRP >3 mg/L). We isolated RNA from whole blood and performed gene expression analyses using RT-qPCR. We measured the expression of 16 immune-related candidate genes: A2M, AQP4, CCL2, CXCL12, CRP, FKBP5, IL-1-beta, IL-6, ISG15, MIF, GR, P2RX7, SGK1, STAT1, TNF-alpha and USP18. Nine of the 16 candidate genes were differentially expressed in MDD cases vs. controls, with no differences between CRP-based groups. Only CRP mRNA was clearly associated with serum CRP. In contrast, plasma (proteins) IL-6, IL-7, IL-8, IL-10, IL-12/IL-23p40, IL-16, IL-17A, IFN-gamma and TNF-alpha, and neutrophils counts, were all differentially regulated between CRP-based groups (higher in CRP >3 vs. CRP <1 and/or controls), reflecting the gradient of CRP values. Secondary analyses on MDD individuals and controls with CRP values <1 mg/L (usually interpreted as 'no inflammation') confirmed MDD cases still had significantly different mRNA expression of immune-related genes compared with controls. These findings corroborate an immune-related molecular activation in MDD, which appears to be independent of serum CRP levels. Additional biological mechanisms may then be required to translate this mRNA signature into inflammation at protein and cellular levels. Understanding these mechanisms will help to uncover the true immune abnormalities in depression, opening new paths for diagnosis and treatment.

Multivariate profile and acute-phase correlates of cognitive deficits in a COVID-19 hospitalised cohort.

Background: Preliminary evidence has highlighted a possible association between severe COVID-19 and persistent cognitive deficits. Further research is required to confirm this association, determine whether cognitive deficits relate to clinical features from the acute phase or to mental health status at the point of assessment, and quantify rate of recovery. Methods: 46 individuals who received critical care for COVID-19 at Addenbrooke's hospital between 10th March 2020 and 31st July 2020 (16 mechanically ventilated) underwent detailed computerised cognitive assessment alongside scales measuring anxiety, depression and post-traumatic stress disorder under supervised conditions at a mean follow up of 6.0 (± 2.1) months following acute illness. Patient and matched control (N = 460) performances were transformed into standard deviation from expected scores, accounting for age and demographic factors using N = 66,008 normative datasets. Global accuracy and response time composites were calculated (G_SScore & G_RT). Linear modelling predicted composite score deficits from acute severity, mental-health status at assessment, and time from hospital admission. The pattern of deficits across tasks was qualitatively compared with normal age-related decline, and early-stage dementia. Findings: COVID-19 survivors were less accurate (G_SScore=-0.53SDs) and slower (G_RT=+0.89SDs) in their responses than expected compared to their matched controls. Acute illness, but not chronic mental health, significantly predicted cognitive deviation from expected scores (G_SScore (p=​​0.0037) and G_RT (p = 0.0366)). The most prominent task associations with COVID-19 were for higher cognition and processing speed, which was qualitatively distinct from the profiles of normal ageing and dementia and similar in magnitude to the effects of ageing between 50 and 70 years of age. A trend towards reduced deficits with time from illness (r∼=0.15) did not reach statistical significance. Interpretation: Cognitive deficits after severe COVID-19 relate most strongly to acute illness severity, persist long into the chronic phase, and recover slowly if at all, with a characteristic profile highlighting higher cognitive functions and processing speed. Funding: This work was funded by the National Institute for Health Research (NIHR) Cambridge Biomedical Research Centre (BRC), NIHR Cambridge Clinical Research Facility (BRC-1215-20014), the Addenbrooke's Charities Trust and NIHR COVID-19 BioResource RG9402. AH is funded by the UK Dementia Research Institute Care Research and Technology Centre and Imperial College London Biomedical Research Centre. ETB and DKM are supported by NIHR Senior Investigator awards. JBR is supported by the Wellcome Trust (220258) and Medical Research Council (SUAG/051 G101400). VFJN is funded by an Academy of Medical Sciences/ The Health Foundation Clinician Scientist Fellowship. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

The influence of comorbid depression and overweight status on peripheral inflammation and cortisol levels.

BACKGROUND: Depression and overweight are each associated with abnormal immune system activation. We sought to disentangle the extent to which depressive symptoms and overweight status contributed to increased inflammation and abnormal cortisol levels. METHODS: Participants were recruited through the Wellcome Trust NIMA Consortium. The sample of 216 participants consisted of 69 overweight patients with depression; 35 overweight controls; 55 normal-weight patients with depression and 57 normal-weight controls. Peripheral inflammation was measured as high-sensitivity C-Reactive Protein (hsCRP) in serum. Salivary cortisol was collected at multiple points throughout the day to measure cortisol awakening response and diurnal cortisol levels. RESULTS: Overweight patients with depression had significantly higher hsCRP compared with overweight controls (p = 0.042), normal-weight depressed patients (p < 0.001) and normal-weight controls (p < 0.001), after controlling for age and gender. Multivariable logistic regression showed that comorbid depression and overweight significantly increased the risk of clinically elevated hsCRP levels ⩾3 mg/L (OR 2.44, 1.28-3.94). In a separate multivariable logistic regression model, overweight status contributed most to the risk of having hsCRP levels ⩾3 mg/L (OR 1.52, 0.7-2.41), while depression also contributed a significant risk (OR 1.09, 0.27-2). There were no significant differences between groups in cortisol awakening response and diurnal cortisol levels. CONCLUSION: Comorbid depression and overweight status are associated with increased hsCRP, and the coexistence of these conditions amplified the risk of clinically elevated hsCRP levels. Overweight status contributed most to the risk of clinically elevated hsCRP levels, but depression also contributed to a significant risk. We observed no differences in cortisol levels between groups.

Peripheral inflammation is associated with micro-structural and functional connectivity changes in depression-related brain networks.

Inflammation is associated with depressive symptoms and innate immune mechanisms are likely causal in some cases of major depression. Systemic inflammation also perturbs brain function and microstructure, though how these are related remains unclear. We recruited N = 46 healthy controls, and N = 83 depressed cases stratified by CRP (> 3 mg/L: N = 33; < 3 mg/L: N = 50). All completed clinical assessment, venous blood sampling for C-reactive protein (CRP) assay, and brain magnetic resonance imaging (MRI). Micro-structural MRI parameters including proton density (PD), a measure of tissue water content, were measured at 360 cortical and 16 subcortical regions. Resting-state fMRI time series were correlated to estimate functional connectivity between individual regions, as well as the sum of connectivity (weighted degree) of each region. Multiple tests for regional analysis were controlled by the false discovery rate (FDR = 5%). We found that CRP was significantly associated with PD in precuneus, posterior cingulate cortex (pC/pCC) and medial prefrontal cortex (mPFC); and with functional connectivity between pC/pCC, mPFC and hippocampus. Depression was associated with reduced weighted degree of pC/pCC, mPFC, and other nodes of the default mode network (DMN). Thus CRP-related increases in proton density-a plausible marker of extracellular oedema-and changes in functional connectivity were anatomically co-localised with DMN nodes that also demonstrated significantly reduced hubness in depression. We suggest that effects of peripheral inflammation on DMN node micro-structure and connectivity may mediate inflammatory effects on depression.

No evidence for differential gene expression in major depressive disorder PBMCs, but robust evidence of elevated biological ageing.

The increasingly compelling data supporting the involvement of immunobiological mechanisms in Major Depressive Disorder (MDD) might provide some explanation forthe variance in this heterogeneous condition. Peripheral blood measures of cytokines and chemokines constitute the bulk of evidence, with consistent meta-analytic data implicating raised proinflammatory cytokines such as IL6, IL1ß and TNF. Among the potential mechanisms linking immunobiological changes to affective neurobiology is the accelerated biological ageing seen in MDD, particularly via the senescence associated secretory phenotype (SASP). However, the cellular source of immunobiological markers remains unclear. Pre-clinical evidence suggests a role for peripheral blood mononuclear cells (PBMC), thus here we aimed to explore the transcriptomic profile using RNA sequencing in PBMCs in a clinical sample of people with various levels of depression and treatment response comparing it with that in healthy controls (HCs). There were three groups with major depressive disorder (MDD): treatment-resistant (n = 94), treatment-responsive (n = 47) and untreated (n = 46). Healthy controls numbered 44. Using PBMCs gene expression analysis was conducted using RNAseq to a depth of 54.5 million reads. Differential gene expression analysis was performed using DESeq2. The data showed no robust signal differentiating MDD and HCs. There was, however, significant evidence of elevated biological ageing in MDD vs HC. Biological ageing was evident in these data as a transcriptional signature of 888 age-associated genes (adjusted p < 0.05, absolute log2fold > 0.6) that also correlated strongly with chronological age (spearman correlation coefficient of 0.72). Future work should expand clinical sample sizes and reduce clinical heterogeneity. Exploration of RNA-seq signatures in other leukocyte populations and single cell RNA sequencing may help uncover more subtle differences. However, currently the subtlety of any PBMC signature mitigates against its convincing use as a diagnostic or predictive biomarker.

Correction: Whole-blood expression of inflammasome- and glucocorticoid-related mRNAs correctly separates treatment-resistant depressed patients from drug-free and responsive patients in the BIODEP study.

We have corrected this Article post-publication, because Dr. Cattaneo's affiliation details were originally incorrect (she was affiliated with three institutions but is in fact only linked to one: Biological Psychiatry Unit, IRCCS Istituto Centro San Giovanni di Dio Fatebenefratelli, Brescia). These changes reflect in both the PDF and HTML versions of this Article.

Whole-blood expression of inflammasome- and glucocorticoid-related mRNAs correctly separates treatment-resistant depressed patients from drug-free and responsive patients in the BIODEP study.

The mRNA expression signatures associated with the 'pro-inflammatory' phenotype of depression, and the differential signatures associated with depression subtypes and the effects of antidepressants, are still unknown. We examined 130 depressed patients (58 treatment-resistant, 36 antidepressant-responsive and 36 currently untreated) and 40 healthy controls from the BIODEP study, and used whole-blood mRNA qPCR to measure the expression of 16 candidate mRNAs, some never measured before: interleukin (IL)-1-beta, IL-6, TNF-alpha, macrophage inhibiting factor (MIF), glucocorticoid receptor (GR), SGK1, FKBP5, the purinergic receptor P2RX7, CCL2, CXCL12, c-reactive protein (CRP), alpha-2-macroglobulin (A2M), acquaporin-4 (AQP4), ISG15, STAT1 and USP-18. All genes but AQP4, ISG15 and USP-18 were differentially regulated. Treatment-resistant and drug-free depressed patients had both increased inflammasome activation (higher P2RX7 and proinflammatory cytokines/chemokines mRNAs expression) and glucocorticoid resistance (lower GR and higher FKBP5 mRNAs expression), while responsive patients had an intermediate phenotype with, additionally, lower CXCL12. Most interestingly, using binomial logistics models we found that a signature of six mRNAs (P2RX7, IL-1-beta, IL-6, TNF-alpha, CXCL12 and GR) distinguished treatment-resistant from responsive patients, even after adjusting for other variables that were different between groups, such as a trait- and state-anxiety, history of childhood maltreatment and serum CRP. Future studies should replicate these findings in larger, longitudinal cohorts, and test whether this mRNA signature can identify patients that are more likely to respond to adjuvant strategies for treatment-resistant depression, including combinations with anti-inflammatory medications.

Peripheral Blood Cell-Stratified Subgroups of Inflamed Depression.

BACKGROUND: Depression has been associated with increased inflammatory proteins, but changes in circulating immune cells are less well defined. METHODS: We used multiparametric flow cytometry to count 14 subsets of peripheral blood cells in 206 depression cases and 77 age- and sex-matched controls (N = 283). We used univariate and multivariate analyses to investigate the immunophenotypes associated with depression and depression severity. RESULTS: Depression cases, compared with controls, had significantly increased immune cell counts, especially neutrophils, CD4+ T cells, and monocytes, and increased inflammatory proteins (C-reactive protein and interleukin-6). Within-group analysis of cases demonstrated significant associations between the severity of depressive symptoms and increased myeloid and CD4+ T-cell counts. Depression cases were partitioned into 2 subgroups by forced binary clustering of cell counts: the inflamed depression subgroup (n = 81 out of 206; 39%) had increased monocyte, CD4+, and neutrophil counts; increased C-reactive protein and interleukin-6; and more severe depression than the uninflamed majority of cases. Relaxing the presumption of a binary classification, data-driven analysis identified 4 subgroups of depression cases, 2 of which (n = 38 and n = 100; 67% collectively) were associated with increased inflammatory proteins and more severe depression but differed in terms of myeloid and lymphoid cell counts. Results were robust to potentially confounding effects of age, sex, body mass index, recent infection, and tobacco use. CONCLUSIONS: Peripheral immune cell counts were used to distinguish inflamed and uninflamed subgroups of depression and to indicate that there may be mechanistically distinct subgroups of inflamed depression.

Childhood trauma, HPA axis activity and antidepressant response in patients with depression.

Childhood trauma is among the most potent contributing risk factors for depression and is associated with poor treatment response. Hypothalamic-pituitary-adrenal (HPA) axis abnormalities have been linked to both childhood trauma and depression, but the underlying mechanisms are poorly understood. The present study aimed to investigate the link between childhood trauma, HPA axis activity and antidepressant response in patients with depression. As part of the Wellcome Trust NIMA consortium, 163 depressed patients and 55 healthy volunteers were included in this study. Adult patients meeting Structured Clinical Interview for Diagnostic and Statistical Manual Version-5 criteria for major depression were categorised into subgroups of treatment responder (n = 42), treatment non-responder (n = 80) and untreated depressed (n = 41) based on current depressive symptom severity measured by the 17-item Hamilton Rating Scale for Depression and exposure to antidepressant medications established by Antidepressant Treatment Response Questionnaire. Childhood Trauma Questionnaire was obtained. Baseline serum C-reactive protein was measured using turbidimetric detection. Salivary cortisol was analyzed at multiple time points during the day using the ELISA technique. Glucocorticoid resistance was defined as the coexistence of hypercortisolemia and inflammation. Our results show that treatment non-responder patients had higher exposure to childhood trauma than responders. No specific HPA axis abnormalities were found in treatment non-responder depressed patients. Untreated depressed showed increased diurnal cortisol levels compared with patients on antidepressant medication, and higher prevalence of glucocorticoid resistance than medicated patients and controls. The severity of childhood trauma was associated with increased diurnal cortisol levels only in individuals with glucocorticoid resistance. Therefore, our findings suggest that the severity of childhood trauma experience contributes to a lack of response to antidepressant treatment. The effects of childhood trauma on increased cortisol levels are specifically evident in patients with glucocorticoid resistance and suggest glucocorticoid resistance as a target for the development of personalized treatment for a subgroup of depressed patients with a history of childhood trauma rather than for all patients with resistance to antidepressant treatment.

Treatment-resistant depression and peripheral C-reactive protein.

BACKGROUND: C-reactive protein (CRP) is a candidate biomarker for major depressive disorder (MDD), but it is unclear how peripheral CRP levels relate to the heterogeneous clinical phenotypes of the disorder.AimTo explore CRP in MDD and its phenotypic associations. METHOD: We recruited 102 treatment-resistant patients with MDD currently experiencing depression, 48 treatment-responsive patients with MDD not currently experiencing depression, 48 patients with depression who were not receiving medication and 54 healthy volunteers. High-sensitivity CRP in peripheral venous blood, body mass index (BMI) and questionnaire assessments of depression, anxiety and childhood trauma were measured. Group differences in CRP were estimated, and partial least squares (PLS) analysis explored the relationships between CRP and specific clinical phenotypes. RESULTS: Compared with healthy volunteers, BMI-corrected CRP was significantly elevated in the treatment-resistant group (P = 0.007; Cohen's d = 0.47); but not significantly so in the treatment-responsive (d = 0.29) and untreated (d = 0.18) groups. PLS yielded an optimal two-factor solution that accounted for 34.7% of variation in clinical measures and for 36.0% of variation in CRP. Clinical phenotypes most strongly associated with CRP and heavily weighted on the first PLS component were vegetative depressive symptoms, BMI, state anxiety and feeling unloved as a child or wishing for a different childhood. CONCLUSIONS: CRP was elevated in patients with MDD, and more so in treatment-resistant patients. Other phenotypes associated with elevated CRP included childhood adversity and specific depressive and anxious symptoms. We suggest that patients with MDD stratified for proinflammatory biomarkers, like CRP, have a distinctive clinical profile that might be responsive to second-line treatment with anti-inflammatory drugs.Declaration of interestS.R.C. consults for Cambridge Cognition and Shire; and his input in this project was funded by a Wellcome Trust Clinical Fellowship (110049/Z/15/Z). E.T.B. is employed half time by the University of Cambridge and half time by GlaxoSmithKline; he holds stock in GlaxoSmithKline. In the past 3 years, P.J.C. has served on an advisory board for Lundbeck. N.A.H. consults for GlaxoSmithKline. P.d.B., D.N.C.J. and W.C.D. are employees of Janssen Research & Development, LLC., of Johnson & Johnson, and hold stock in Johnson & Johnson. The other authors report no financial disclosures or potential conflicts of interest.

Eligibility for magnetic resonance imaging screening in the United Kingdom: effect of strict selection criteria and anonymous DNA testing on breast cancer incidence in the MARIBS Study.

INTRODUCTION: A UK multicenter study compared the performance of contrast enhanced-magnetic resonance imaging with X-Ray Mammography in women at high-risk of breast cancer commencing in 1997. Selection criteria were used to identify women with at least 0.9% annual risk of breast cancer. METHODS: Women at high breast cancer risk, with a strong family history and/or high probability of a BRCA1/BRCA2/TP53 mutation, were recruited from 22 centers. Those not known as gene carriers were asked to give a blood sample, which was tested anonymously for mutations. Women ages 35 to 49 years were offered annual screening for 2 to 7 years. Study eligibility at entry was assessed retrospectively by detailed examination of pedigrees and overall eligibility accounting for computer risk assessment and mutation results. RESULTS: Seventy-eight of 837 (9%) women entered for screening were ineligible using the strict entry criteria. Thirty-nine cancers were detected in 1,869 women-years in study (incidence 21 per 1,000). Including 3,561 further years follow-up, 28 more breast cancers were identified (12 of 1,000). Incidence rates for 759 eligible women were 22 of 1,000 in study and 13 of 1,000 in total follow-up, compared with 9 of 1,000 and 4 of 1,000, respectively, in 78 ineligible women. Breast cancer rates were higher for BRCA2 than BRCA1 after testing anonymized samples in this selected population at 65 of 1,000 in study and 36 of 1,000 in total follow-up for BRCA2 compared with 44 of 1,000 and 27 of 1,000 for BRCA1. CONCLUSIONS: Strict enforcement of study criteria would have minimally improved the power of the study, whereas testing for BRCA1/2 in advance would have substantially increased the detection rates.

A pilot study of compositional analysis of the breast and estimation of breast mammographic density using three-dimensional T1-weighted magnetic resonance imaging.

PURPOSE: A method and computer tool to estimate percentage magnetic resonance (MR) imaging (MRI) breast density using three-dimensional T(1)-weighted MRI is introduced, and compared with mammographic percentage density [X-ray mammography (XRM)]. MATERIALS AND METHODS: Ethical approval and informed consent were obtained. A method to assess MRI breast density as percentage volume occupied by water-containing tissue on three-dimensional T(1)-weighted MR images is described and applied in a pilot study to 138 subjects who were imaged by both MRI and XRM during the Magnetic Resonance Imaging in Breast Screening study. For comparison, percentage mammographic density was measured from matching XRMs as a ratio of dense to total projection areas scored visually using a 21-point score and measured by applying a two-dimensional interactive program (CUMULUS). The MRI and XRM percent methods were compared, including assessment of left-right and interreader consistency. RESULTS: Percent MRI density correlated strongly (r = 0.78; P < 0.0001) with percent mammographic density estimated using Cumulus. Comparison with visual assessment also showed a strong correlation. The mammographic methods overestimate density compared with MRI volumetric assessment by a factor approaching 2. DISCUSSION: MRI provides direct three-dimensional measurement of the proportion of water-based tissue in the breast. It correlates well with visual and computerized percent mammographic density measurements. This method may have direct application in women having breast cancer screening by breast MRI and may aid in determination of risk.

BRCA1 mutation and young age predict fast breast cancer growth in the Dutch, United Kingdom, and Canadian magnetic resonance imaging screening trials.

PURPOSE: Magnetic resonance imaging (MRI) screening enables early detection of breast cancers in women with an inherited predisposition. Interval cancers occurred in women with a BRCA1 mutation, possibly due to fast tumor growth. We investigated the effect of a BRCA1 or BRCA2 mutation and age on the growth rate of breast cancers, as this may influence the optimal screening frequency. EXPERIMENTAL DESIGN: We reviewed the invasive cancers from the United Kingdom, Dutch, and Canadian MRI screening trials for women at hereditary risk, measuring tumor size at diagnosis and on preceding MRI and/or mammography. We could assess tumor volume doubling time (DT) in 100 cancers. RESULTS: Tumor DT was estimated for 43 women with a BRCA1 mutation, 16 women with a BRCA2 mutation, and 41 women at high risk without an identified mutation. Growth rate slowed continuously with increasing age (P = 0.004). Growth was twice as fast in BRCA1 (P = 0.003) or BRCA2 (P = 0.03) patients as in high-risk patients of the same age. The mean DT for women with BRCA1/2 mutations diagnosed at ages < or =40, 41 to 50, and >50 years was 28, 68, and 81 days, respectively, and 83, 121, and 173 days, respectively, in the high-risk group. Pathologic tumor size decreased with increasing age (P = 0.001). Median size was 15 mm for patients ages < or =40 years compared with 9 mm in older patients (P = 0.003); tumors were largest in young women with BRCA1 mutations. CONCLUSION: Tumors grow quickly in women with BRCA1 mutations and in young women. Age and risk group should be taken into account in screening protocols.

Multiscale analysis of MR-mammography data.

In this work we propose a method for automatically discriminating between different types of tissue in MR mammography datasets. This is accomplished by employing a wavelet-based multiscale analysis. After the data has been wavelet-transformed unsupervised machine learning methods are employed to identify typical patterns in the wavelet domain. To demonstrate the potential of the proposed approach we apply a filtering procedure that extracts the wavelet-based image information related to tumour tissue. In this way we obtain a robust segmentation of suspicious tissue in the MR image.

A test of performance of breast MRI interpretation in a multicentre screening study.

OBJECTIVES: The aim of this study was to assess the consistency and performance of radiologists interpreting breast magnetic resonance imaging (MRI) examinations. MATERIALS AND METHODS: Two test sets of eight cases comprising cancers, benign disease, technical problems and parenchymal enhancement were prepared from two manufacturers' equipment (X and Y) and reported by 15 radiologists using the recording form and scoring system of the UK MRI breast screening study [(MAgnetic Resonance Imaging in Breast Screening (MARIBS)]. Variations in assessments of morphology, kinetic scores and diagnosis were measured by assessing intraobserver and interobserver variability and agreement. The sensitivity and specificity of reporting performances was determined using receiver operating characteristic (ROC) curve analysis. RESULTS: Intraobserver variation was seen in 13 (27.7%) of 47 of the radiologists' conclusions (four technical and seven pathological differences). Substantial interobserver variation was observed in the scores recorded for morphology, pattern of enhancement, quantification of enhancement and washout pattern. The overall sensitivity of breast MRI was high [88.6%, 95% confidence interval (CI) 77.4-94.7%], combined with a specificity of 69.2% (95% CI 60.5-76.7%). The sensitivities were similar for the two test sets (P=.3), but the specificity was significantly higher for the Manufacturer X dataset (P<.001). ROC curve analysis gave an area under the curve of 0.85 (95% CI 0.79-0.92) CONCLUSIONS: Substantial variation in all elements of the scoring system and in the overall diagnostic conclusions was observed between radiologists participating in MARIBS. High overall sensitivity was achieved with moderate specificity. Manufacturer-related differences in specificities possibly occurred because the numerical thresholds set for the scoring system were not optimised for both equipment manufacturers. Scoring systems developed on one equipment software may not be transferable to other manufacturers.

Evaluation of a prospective scoring system designed for a multicenter breast MR imaging screening study.

PURPOSE: To evaluate prospectively the accuracy of a lesion classification system designed for use in a magnetic resonance (MR) imaging high-breast-cancer-risk screening study. MATERIALS AND METHODS: All participating patients provided written informed consent. Ethics committee approval was obtained. The results of 1541 contrast material-enhanced breast MR imaging examinations were analyzed; 1441 screening examinations were performed in 638 women aged 24-51 years at high risk for breast cancer, and 100 examinations were performed in 100 women aged 23-81 years. Lesion analysis was performed in 991 breasts, which were divided into design (491 breasts) and testing (500 breasts) sets. The reference standard was histologic analysis of biopsy samples, fine-needle aspiration cytology, or minimal follow-up of 24 months. The scoring system involved the use of five features: morphology (MOR), pattern of enhancement (POE), percentage of maximal focal enhancement (PMFE), maximal signal intensity-time ratio (MITR), and pattern of contrast material washout (POCW). The system was evaluated by means of (a) assessment of interreader agreement, as expressed in kappa statistics, for 315 breasts in which both readers analyzed the same lesion, (b) assessment of the diagnostic accuracy of the scored components with receiver operating characteristic curve analysis, and (c) logistic regression analysis to determine which components of the scoring system were critical to the final score. A new simplified scoring system developed with the design set was applied to the testing set. RESULTS: There was moderate reader agreement regarding overall lesion outcome (ie, malignant, suspicious, or benign) (kappa=0.58) and less agreement regarding the scored components. The area under the receiver operating characteristic curve (AUC) for the overall lesion score, 0.88, was higher than the AUC for any one component. The components MOR, POE, and POCW yielded the best overall result. PMFE and MITR did not contribute to diagnostic utility. Applying a simplified scoring system to the testing set yielded a nonsignificantly (P=.2) higher AUC than did applying the original scoring system (sensitivity, 84%; specificity, 86.0%). CONCLUSION: Good diagnostic accuracy can be achieved by using simple qualitative descriptors of lesion enhancement, including POCW. In the context of screening, quantitative enhancement parameters appear to be less useful for lesion characterization.

Reading protocol for dynamic contrast-enhanced MR images of the breast: sensitivity and specificity analysis.

PURPOSE: To prospectively determine sensitivity and specificity of breast magnetic resonance (MR) imaging in a screening and symptomatic population by using independent double reading, with histologic or cytologic results or a minimum 18-month follow-up as the standard. MATERIALS AND METHODS: Informed consent and ethical approval were obtained. Reader performance was analyzed in 44 radiologists at 18 centers from 1541 examinations, including 1441 screening examinations in 638 high-risk women aged 24-51 years (mean, 40.5 years) and 100 examinations in symptomatic women aged 23-81 years (mean, 49.2 years). A screening protocol of dynamic T1-weighted three-dimensional imaging and 0.2 mmol/kg gadolinium-based intravenous contrast agent was used. Logistic and Poisson regressions were used to analyze reader performance in relation to experience. Correlation between readers was determined with kappa statistics. Sensitivity and specificity were analyzed according to reader, field strength, machine type, and histologic results. RESULTS: The proportion of studies with lesions analyzed reduced significantly with reader experience (odds ratio, 0.84 per 6 months; P < .001), and number of regions per lesion analyzed also diminished (incidence rate ratio, 0.98 per 6 months; P = .047). The two readers for each study agreed 87% of the time, with a moderately good kappa statistic of 0.52 (95% confidence interval [CI]: 0.45, 0.58). By taking the reading with the highest score (most likely to be malignant) from each double-read study, sensitivity was 91% (95% CI: 83%, 96%) and specificity was 81% (95% CI: 79%, 83%). Single readings had 7% lower sensitivity (95% CI: 4%, 11%) and 7% higher specificity (95% CI: 6%, 7%). Sensitivity did not differ between MR imager manufacturers or between 1.0- and 1.5-T field strength, but there were significant differences in specificity for machine type (P = .001) and for field strength adjusted for manufacturer (P = .001). Specificity, but not sensitivity, was higher in women younger than 50 years (P = .02). CONCLUSION: Independent double reading by 44 radiologists blinded to mammography results showed sensitivity and specificity acceptable for screening; sensitivity was higher when two readings were used, at the cost of specificity. Interreader correlation was moderately good, and evidence of learning was seen. Equipment manufacturer, field strength, and age affected specificity but not sensitivity.

Evaluation of radiological features for breast tumour classification in clinical screening with machine learning methods.

OBJECTIVE: In this work, methods utilizing supervised and unsupervised machine learning are applied to analyze radiologically derived morphological and calculated kinetic tumour features. The features are extracted from dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) time-course data. MATERIAL: The DCE-MRI data of the female breast are obtained within the UK Multicenter Breast Screening Study. The group of patients imaged in this study is selected on the basis of an increased genetic risk for developing breast cancer. METHODS: The k-means clustering and self-organizing maps (SOM) are applied to analyze the signal structure in terms of visualization. We employ k-nearest neighbor classifiers (k-nn), support vector machines (SVM) and decision trees (DT) to classify features using a computer aided diagnosis (CAD) approach. RESULTS: Regarding the unsupervised techniques, clustering according to features indicating benign and malignant characteristics is observed to a limited extend. The supervised approaches classified the data with 74% accuracy (DT) and providing an area under the receiver-operator-characteristics (ROC) curve (AUC) of 0.88 (SVM). CONCLUSION: It was found that contour and wash-out type (WOT) features determined by the radiologists lead to the best SVM classification results. Although a fast signal uptake in early time-point measurements is an important feature for malignant/benign classification of tumours, our results indicate that the wash-out characteristics might be considered as important.

What is the recall rate of breast MRI when used for screening asymptomatic women at high risk?

Breast screening acceptability is dependent on sensitivity and recall rate. We aimed to establish the recall rate for MRI and mammography, separately and together, when screening a cohort of women at high genetic risk. Women aged 35-49 years in the MARIBS study form the cohort. We analysed the recall rate, the number of extra tests and their effectiveness. Wilcoxon Rank test was used to estimate the effect of age and logistic regression with robust variance the effect of mammographic density on recall rates. The first 726 screening studies took place in 415 women. Following 86 of these recall occurred, comprising 140 additional investigations. 28 of the cases were resolved without further MRI, and 18 women had more than 2 additional tests. Neither age nor mammographic density was associated with recall. MRI had a recall of rate of 10.19%, and mammography 4.00%. The two techniques largely recalled different cases and 10 cases only (11.62% of those recalled) were abnormal by both tests. The two together had a recall rate of 11.85%. Recall rates varied widely between centres of the study. Breast MRI in asymptomatic high-risk women age 35-49 years largely recalls different women from mammography. The combined figure of approximately 12% may be acceptable for screening and will be useful for planning similar studies.