Anthony Horsman

Observer variability in the interpretation of contrast enhanced MRI of the breast

The purpose of this study was to determine observer variability in the interpretation of contrast enhanced breast MRI and to evaluate its effect on the detection and differentiation of breast cancer. 57 women underwent breast MRI using spin echo and dynamic spoiled gradient-recalled sequences. Images were independently reviewed by three radiologists, two experienced and one newly trained in breast MRI interpretation. One of the experienced readers reviewed all examinations twice. Interpretation was based on lesion conspicuity, signal intensity, contour and enhancement pattern. Contrast uptake was assessed using region of interest (ROI) analysis of the dynamic images and calculation of a maximum enhancement index. Sensitivity and specificity in the diagnosis of malignancy irrespective of disease extent, and in the diagnosis of multifocal malignancy were estimated. 113 lesions were reported. Kappa coefficient estimations showed only a moderate agreement between the two experienced readers in rating morphological characteristics; the agreement between the newly trained reader and the experienced readers was even worse. Moreover, there was significant interobserver and intraobserver variation in the enhancement index measurements. Weighted kappa values indicated good agreement between the experienced readers in lesion and overall interpretation, excellent intraobserver agreement, but substantial disagreement between the newly trained reader and both experienced readers. All readers showed good sensitivity in cancer detection, but specificity was substantially lower. There is significant observer variability and a substantial learning curve in the interpretation of breast MRI, and variability in the ROI analysis of dynamic data. Further efforts to improve the reliability of ROI analysis and image interpretation are needed to help MRI realise its full potential in the clinical management of breast cancer.

Dynamic MRI of invasive breast cancer: assessment of three region-of-interest analysis methods

PURPOSE In this study, three region-of-interest (ROI) analysis methods based on operator-defined and semiautomated sampling of pharmacokinetic breast maps of contrast uptake are described. The observer variability and impact of the methods on the estimated enhancement characteristics of invasive cancer are also presented. METHOD Fifty-four women with invasive breast cancer underwent dynamic Gd-DTPA-enhanced MRI. ROIs were drawn by two observers on parametric images obtained from compartmental modeling of the dynamic data. Three methods were used: (a) An irregular ROI was drawn to include as much of the enhancing part of the tumor as possible (large ROI); (b) a 12 pixel circular ROI was placed at the most rapidly enhancing part of the large region (small ROI); and (c) a computer algorithm interrogated the large region pixel by pixel using a 9 pixel square mask and selected the region with the highest mean parameter value (semiautomated ROI). RESULTS Significant observer variability and bias were found in the enhancement measurements using the large ROI method. There was no observer bias associated with the other methods, but the variability of the small ROI method was substantial. An almost perfect observer agreement was achieved using the semiautomated method. The small and semiautomated ROI methods produced significantly higher enhancement ratios than the large ROI method, especially in grade III carcinomas. CONCLUSION Variability is inherent in subjective ROI analysis, but the semiautomated method of ROI selection and sampling of parameter images of the breast is an efficient and reliable alternative that may allow better standardization of the MR technique.

Dynamic MR imaging of the breast combined with analysis of contrast agent kinetics in the differentiation of primary breast tumours

OBJECTIVE To assess dynamic Gd-DTPA-enhanced magnetic resonance (MR) imaging in the diagnosis of primary breast pathology, and to test the hypothesis that analysis of contrast agent kinetics increases specificity. METHODS Forty-seven women underwent breast MR imaging using three-dimensional and dynamic spoiled gradient-recalled sequences. Image interpretation was based on the evaluation of lesion conspicuity, signal intensity, contour and enhancement pattern from the static acquisitions. Assessment of contrast kinetics was based on pixel-by-pixel analysis of the dynamic data. A two-compartment model described by three parameters (amplitude of uptake, exchange rate and washout rate), and a three-compartment model described by two parameters (permeability and exchange rate) were used. Regions of interest were drawn for all lesions found in the dynamic sections. Mean regional pixel values were calculated for each parameter and tested for diagnostic efficacy. RESULTS Twenty-two malignant and 36 benign lesions were examined. Fibroadenomas accounted for 86% of the benign tumours. Image interpretation had a sensitivity of 0.95 and specificity of 0.86. The fat-suppressed post-contrast images permitted good visualization of the contour and matrix characteristics of fibroadenomas, but all non-fibroadenomatous benign lesions were classified as indeterminate or suspicious. Significant differences were found between benign and malignant lesions in the amplitude of uptake (P = 0.0008) and exchange rate (P < 0.00005) of the two-compartment model, and permeability (P=0.0001) and exchange rate (P < 0.00005) of the three-compartment model. However, image interpretation was superior to the isolated use of quantitative indices (P=0.02). The most discriminating parameters were the exchange rates of both models, with no significant difference between them. CONCLUSION Assessment of lesion morphology is essential and probably sufficient for the differentiation of fibroadenomas from malignant tumours. However, specificity of conventional MR imaging may be much lower for other types of primary benign breast pathology. Analysis of Gd-DTPA kinetics improves the specificity obtained using simple enhancement measurements and can be used to produce parametric images that provide information about lesion heterogeneity, permeability and vascularity.

Dynamic gradient-echo and fat-suppressed spin-echo contrast-enhanced MRI of the breast.

PURPOSE To evaluate heavily T2-weighted, dynamic contrast-enhanced and fat-suppressed magnetic resonance imaging (MRI) of the breast in comparison with conventional imaging and fine needle aspiration cytology (FNAC). PATIENTS AND METHODS Fifty patients with surgically/pathologically proven breast disease were examined pre-operatively by MRI. The majority, 45 patients, had invasive carcinoma. T1-weighted spin-echo, T2-weighted fast spin-echo (with chemical-shift-selective fat-suppression in 20 cases), rapid dynamic contrast-enhanced gradient-echo and post-contrast fat-suppressed T1-weighted images were obtained. Signal intensity changes during dynamic scanning were assessed qualitatively and quantitatively. Comparison was made with the results of X-ray mammography, ultrasound and fine needle aspiration cytology. RESULTS Unenhanced MRI was inadequate for determining the location, extent or nature of most lesions even when fat-suppressed T2-weighted images were obtained. Following contrast injection, there was significantly greater enhancement of invasive carcinomas than normal parenchyma. Invasive carcinomas of ductal and lobular subtypes did not differ significantly in their enhancement profiles. Prominent enhancement of the lesion periphery, which was a feature in 33 out of 50 cases (the majority of which were invasive carcinomas) was not due to central tumour necrosis. In four cases, invasive carcinomas which were clearly visible on early dynamic scans could not be identified on post-contrast fat-suppressed images. Lesions that were more numerous or extensive than had been recognised clinically or mammographically were revealed by MRI in 14 patients, though MRI was no more specific than conventional assessment. Invasion of the chest wall was accurately predicted by MRI in three cases. There was excellent correlation between tumour size shown by MRI and histopathology. FNAC was accurate in predicting the final histological diagnosis except in those cases where samples were unsatisfactory. CONCLUSIONS Contrast-enhanced MRI appears useful in the assessment of suspected malignant breast masses, especially in younger women with predominantly glandular breast tissue. Specificity is no better than FNAC but tumour extent and multifocality are more accurately disclosed than by conventional imaging techniques. Contrast-enhanced chemical-shift-selective fat-suppressed images are sub-optimal in a small proportion of cases (particularly where lesions are located posteriorly) and some benign breast disease may appear misleadingly conspicuous on such images. Morphological features and quantification of lesion enhancement during dynamic scanning presently offer only limited prospects for discrimination between various pathologies. Heavily T2-weighted sequences appear of marginal value.

Peripheral Enhancement and Spatial Contrast Uptake Heterogeneity of Primary Breast Tumours: Quantitative Assessment with Dynamic Mri

PURPOSE The purpose of our study was to determine if quantitative analysis of peripheral enhancement and spatial contrast uptake heterogeneity is useful in the characterisation of abnormalities seen at breast MRI. METHOD Ninety-one women underwent dynamic gadolinium-enhanced breast MRI. Regions of interest were processed by a parallel thinning algorithm to define central and peripheral subregions of lesions. Peripheral enhancement was quantified at every time point of the dynamic sequence as a signal difference-to-noise ratio. Moreover, a radiologist assessed the images for the presence of peripheral enhancement and classified the regional and subregional time-intensity profiles of each lesion. RESULTS Sixty-four invasive carcinomas and 30 benign lesions were analysed. Significant differences were found between benign and malignant lesions in peripheral enhancement as determined from the dynamic images (p = 0.0002; sensitivity, 0.34; specificity, 1.00) in time-intensity profiles (p < 0.000005; sensitivity, 0.67; specificity, 0.93) and in peripheral percentage signal changes at 1 min postcontrast (p = 0.001). There was a much higher relative signal increase centrally than peripherally (p < 0.0005), but peripheral signal changes had greater diagnostic value than central ones (Az = 0.72 vs. 0.63; p = 0.02). Carcinomas showed higher peripheral enhancement than benign lesions (p = 0.001). Peripheral enhancement reached maximum diagnostic value at 4 min postcontrast (Az = 0.80) and performed best as a highly sensitive but moderately specific diagnostic index. CONCLUSION Quantification of peripheral enhancement is diagnostically useful and offers insight into the enhancement mechanisms encountered in breast MRI. Primary breast tumours show substantial spatial contrast uptake heterogeneity. Lesion differentiation based on percentage signal changes is improved by restricting sampling to the periphery of tumours.

Dynamic contrast-enhanced and fat suppressed magnetic resonance imaging in suspected recurrent carcinoma of the breast: preliminary experience

20 women with suspected recurrent breast cancer who had undergone previous breast-conserving operations were investigated using dynamic contrast-enhanced gradient echo (GRE) and fat suppressed spin echo (SE) magnetic resonance (MR) imaging. Histologically confirmed recurrent tumour was readily recognized on dynamic GRE scans by virtue of rapid, early and avid enhancement. Benign scars enhanced more slowly and reached lower magnitudes of enhancement. Fat suppressed SE images, which were typically acquired 10 min after contrast administration, were sensitive for the detection of tumour recurrence but lacked specificity. Early scanning after contrast administration offers the best prospects for distinguishing tumour recurrence from benign scarring. The criteria used to distinguish these two entities are highly dependent on the scan technique and the time at which images are obtained post-contrast.

Effect of temporal resolution on the diagnostic efficacy of contrast-enhanced MRI in the conservatively treated breast

PURPOSE Our goal was to assess the effect of image acquisition rate on the diagnostic efficacy of contrast-enhanced MRI in the conservatively treated breast. METHOD Sixty-seven women, treated with breast-conserving surgery and radiotherapy for early-stage breast cancer, were imaged at 1.5 T using a dynamic contrast-enhanced sequence with a temporal resolution of 12 s. Enhancement was recorded over time for the dominant lesion in each patient. Data were subsequently removed to simulate various acquisition rates and the enhancement indexes of benign and malignant lesions were compared. RESULTS Seventeen patients had confirmed local recurrence and 50 remained disease-free. There were significant differences in the enhancement indexes of benign and malignant lesions 24-264 s after contrast agent administration. Acquisition rate had a negligible effect upon diagnostic efficacy. CONCLUSION Two image data sets collected before contrast agent administration and between 1.5 and 3.5 min afterward may be sufficient to differentiate recurrent and benign disease.

Localization of MR-detected breast cancer using a prototype stereotactic guidance MR system

Abstract We report the case of an invasive tubular breast carcinoma detected solely by MR imaging, and subsequently localized for excision biopsy with an MR-compatible, spring-hook wire positioning marker using a prototype MR-guided stereotactic localization system. This report emphasises the potential of interventional MR imaging in the management of breast cancer.

3D MRI and angiography of human extremities using a local gradient coil

The limits of the spatial resolution achievable on a standard GE Signa 1.5-T imaging machine have been explored by interfacing a 12-cm-diameter local gradient set to such a system. Using a range of purpose-built radio-frequency (RF) coils, 3D gradient echo images with a spatial resolution of 100×100×500 µ have been routinely obtained on the finger. Also produced were 100×200×500 µ, resolution angiograms of the finger detailing submillimeter vessels, with a velocity encoding of 6 cm s−1. High-resolution fat/water suppressed images and isotropic data sets with a resolution of 170 µ are also demonstrated. With regard to the investigation of the mechanics of the human finger, isotropic data sets are desirable if accurate segmentation is to be implemented. A preliminary study illustrating the potential for producing a magnetic-resonance-based computer simulation of the mechanics of the human finger shows the distal interphalangeal joint extended and then in flexion.

Quantitative analysis of GD-DTPA enhanced dynamic MR images by simplex minimization

The introduction of fast clinical imaging techniques enables the dynamic sampling of contrast agent kinetics. The quantity of data generated by such techniques demands the application of automated analysis tools to provide objective assessments of contrast washin/washout characteristics across an entire image. Using a simple compartmental model, the washin/washout characteristics may be defined in terms of a biexponential function. Curve fitting to this nonlinear function may be approached using a nonlinear least squares technique. We demonstrate an alternative technique, simplex minimization, to be more robust in a comparative study using Monte Carlo simulation. The application of the technique is illustrated by pixel-by-pixel analysis of dynamic data obtained from a patient with primary carcinoma of the breast.