Identification of cerebral ischemic lesions in rat using Eigenimage filtered magnetic resonance imaging

Integrated radiomic framework for breast cancer and tumor biology using advanced machine learning and multiparametric MRI

Radiomics deals with the high throughput extraction of quantitative textural information from radiological images that not visually perceivable by radiologists. However, the biological correlation between radiomic features and different tissues of interest has not been established. To that end, we present the radiomic feature mapping framework to generate radiomic MRI texture image representations called the radiomic feature maps (RFM) and correlate the RFMs with quantitative texture values, breast tissue biology using quantitative MRI and classify benign from malignant tumors. We tested our radiomic feature mapping framework on a retrospective cohort of 124 patients (26 benign and 98 malignant) who underwent multiparametric breast MR imaging at 3 T. The MRI parameters used were T1-weighted imaging, T2-weighted imaging, dynamic contrast enhanced MRI (DCE-MRI) and diffusion weighted imaging (DWI). The RFMs were computed by convolving MRI images with statistical filters based on first order statistics and gray level co-occurrence matrix features. Malignant lesions demonstrated significantly higher entropy on both post contrast DCE-MRI (Benign-DCE entropy: 5.72 ± 0.12, Malignant-DCE entropy: 6.29 ± 0.06, p = 0.0002) and apparent diffusion coefficient (ADC) maps as compared to benign lesions (Benign-ADC entropy: 5.65 ± 0.15, Malignant ADC entropy: 6.20 ± 0.07, p = 0.002). There was no significant difference between glandular tissue entropy values in the two groups. Furthermore, the RFMs from DCE-MRI and DWI demonstrated significantly different RFM curves for benign and malignant lesions indicating their correlation to tumor vascular and cellular heterogeneity respectively. There were significant differences in the quantitative MRI metrics of ADC and perfusion. The multiview IsoSVM model classified benign and malignant breast tumors with sensitivity and specificity of 93 and 85%, respectively, with an AUC of 0.91.

An automated system for analyzing magnetic resonance imaging (MRI) can differentiate benign from malignant breast tumors with high accuracy. Vishwa S. Parekh and Michael A. Jacobs 
from Johns Hopkins University School of Medicine in Baltimore, Maryland, USA, developed an algorithm for extracting textural information from MRI scans that are not visually perceivable to radiologists using machine learning and Radiomic features. Their model combines different MRI parameters to produce so-called radiomic feature maps. The researchers tested their mapping framework on a retrospective cohort of 124 patients, 26 of whom had benign breast lesions and 98 had malignant tumors. They found statistical differences in certain MRI and radiomic metrics. Moreover, they demonstrated quantitative ADC map values and Dynamic contrast pharmacokinetic modeling to characterize the radiomic features. Overall, the method identified a breast lesion as benign or malignant with 93% sensitivity and 85% specificity, suggesting that radiomic feature mapping could aid in diagnosing and characterizing the disease correctly and tailoring therapy accordingly.

Automated Ischemic Lesion Segmentation in MRI Mouse Brain Data after Transient Middle Cerebral Artery Occlusion

Magnetic resonance imaging (MRI) has become increasingly important in ischemic stroke experiments in mice, especially because it enables longitudinal studies. Still, quantitative analysis of MRI data remains challenging mainly because segmentation of mouse brain lesions in MRI data heavily relies on time-consuming manual tracing and thresholding techniques. Therefore, in the present study, a fully automated approach was developed to analyze longitudinal MRI data for quantification of ischemic lesion volume progression in the mouse brain. We present a level-set-based lesion segmentation algorithm that is built using a minimal set of assumptions and requires only one MRI sequence (T2) as input. To validate our algorithm we used a heterogeneous data set consisting of 121 mouse brain scans of various age groups and time points after infarct induction and obtained using different MRI hardware and acquisition parameters. We evaluated the volumetric accuracy and regional overlap of ischemic lesions segmented by our automated method against the ground truth obtained in a semi-automated fashion that includes a highly time-consuming manual correction step. Our method shows good agreement with human observations and is accurate on heterogeneous data, whilst requiring much shorter average execution time. The algorithm developed here was compiled into a toolbox and made publically available, as well as all the data sets.

3D reconstruction of 2D fluorescence histology images and registration with in vivo MR images: application in a rodent stroke model

To validate and add value to non-invasive imaging techniques, the corresponding histology is required to establish biological correlates. We present an efficient, semi-automated image-processing pipeline that uses immunohistochemically stained sections to reconstruct a 3D brain volume from 2D histological images before registering these with the corresponding 3D in vivo magnetic resonance images (MRI). A multistep registration procedure that first aligns the "global" volume by using the centre of mass and then applies a rigid and affine alignment based on signal intensities is described. This technique was applied to a training set of three rat brain volumes before being validated on three normal brains. Application of the approach to register "abnormal" images from a rat model of stroke allowed the neurobiological correlates of the variations in the hyper-intense MRI signal intensity caused by infarction to be investigated. For evaluation, the corresponding anatomical landmarks in MR and histology were defined to measure the registration accuracy. A registration error of 0.249 mm (approximately one in-plane voxel dimension) was evident in healthy rat brains and of 0.323 mm in a rodent model of stroke. The proposed reconstruction and registration pipeline allowed for the precise analysis of non-invasive MRI and corresponding microstructural histological features in 3D. We were thus able to interrogate histology to deduce the cause of MRI signal variations in the lesion cavity and the peri-infarct area.

Magnetic resonance imaging of ischemia viability thresholds and the neurovascular unit

Neuroimaging has improved our understanding of the evolution of stroke at discreet time points helping to identify irreversibly damaged and potentially reversible ischemic brain. Neuroimaging has also contributed considerably to the basic premise of acute stroke therapy which is to salvage some portion of the ischemic region from evolving into infarction, and by doing so, maintaining brain function and improving outcome. The term neurovascular unit (NVU) broadens the concept of the ischemic penumbra by linking the microcirculation with neuronal-glial interactions during ischemia reperfusion. Strategies that attempt to preserve the individual components (endothelium, glia and neurons) of the NVU are unlikely to be helpful if blood flow is not fully restored to the microcirculation. Magnetic resonance imaging (MRI) is the foremost imaging technology able to bridge both basic science and the clinic via non-invasive real time high-resolution anatomical delineation of disease manifestations at the molecular and ionic level. Current MRI based technologies have focused on the mismatch between perfusion-weighted imaging (PWI) and diffusion weighted imaging (DWI) signals to estimate the tissue that could be saved if reperfusion was achieved. Future directions of MRI may focus on the discordance of recanalization and reperfusion, providing complimentary pathophysiological information to current compartmental paradigms of infarct core (DWI) and penumbra (PWI) with imaging information related to cerebral blood flow, BBB permeability, inflammation, and oedema formation in the early acute phase. In this review we outline advances in our understanding of stroke pathophysiology with imaging, transcending animal stroke models to human stroke, and describing the potential translation of MRI to image important interactions relevant to acute stroke at the interface of the neurovascular unit.

Monitoring of neoadjuvant chemotherapy using multiparametric, ²³Na sodium MR, and multimodality (PET/CT/MRI) imaging in locally advanced breast cancer

We prospectively investigated using advanced magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT) to identify radiological biomarkers for treatment response in patients receiving preoperative systemic therapy (PST) for locally advanced breast cancer. Patients with a stage II or III breast cancer receiving PST were selected and underwent positron emission tomography (PET), magnetic resonance imaging (MRI), and breast biopsies at baseline and after the first cycle of PST (days 7-8) during the full course of treatment. PET/CT was acquired after injection of 2-deoxy-2-[18F]-fluoro-D-glucose (¹⁸FDG, 0.22 mCi/kg) and quantified with standardized uptake value assessment (SUV). Diagnostic breast MRI and sodium (²³Na) was acquired at 1.5 T. Total tissue sodium concentration (TSC), response criteria in solid tumors (RECIST), and volumes were quantified. Treatment response was determined by pathological assessment at surgery. Immunohistochemistry values of the proliferative index (Ki-67) were performed on biopsy specimens. Six of nineteen eligible women (43 ± 11 years) who received PST underwent radiological imaging of ¹⁸FDG-PET/CT and MRI for at least two cycles of treatment. Five patients had a pathological partial response (pPR) and one had pathological non-response (pNR). TSC decreased 21% in responders with increases in the non-responder (P = 0.03). Greater reduction in SUV was observed in responders (38%) compared to the non-responder (22%; P = 0.03). MRI volumes decreased after cycle 1 by 42% (responders) and 35% (non-responder; P = 0.11). Proliferation index Ki-67 declined in responders in the first cycle (median = 47%, range = 29-20%), but increased (4%) in the non-responder. Significant decreases in TSC, SUV, and Ki-67 were observed in responders with increases in TSC and Ki-67 in non-responders. Our results demonstrate the feasibility of using multi-modality proton, ²³Na MRI, and PET/CT metrics as radiological biomarkers for monitoring response to PST in patients with operable breast cancer.

Multiparametric magnetic resonance imaging, spectroscopy and multinuclear (²³Na) imaging monitoring of preoperative chemotherapy for locally advanced breast cancer

RATIONALE AND OBJECTIVES

The aim of this prospective study was to investigate using multiparametric and multinuclear magnetic resonance imaging during preoperative systemic therapy for locally advanced breast cancer.

MATERIALS AND METHODS

Women with operable stage 2 or 3 breast cancer who received preoperative systemic therapy were studied using dynamic contrast-enhanced magnetic resonance imaging, magnetic resonance spectroscopy, and ²³Na magnetic resonance. Quantitative metrics of choline peak signal-to-noise ratio, total tissue sodium concentration, tumor volumes, and Response Evaluation Criteria in Solid Tumors were determined and compared to final pathologic results using receiver-operating characteristic analysis. Hormonal markers were investigated. Statistical significance was set at P < .05.

RESULTS

Eighteen eligible women were studied. Fifteen responded to therapy, four (22%) with pathologic complete response and 11 (61%) with pathologic partial response. Three patients (17%) had no response. Among estrogen receptor-positive, HER2-positive, and triple-negative phenotypes, observed frequencies of pathologic complete response, pathologic partial response, and no response were 2, 5, and 0; 1, 4, and 0; and 1, 1, and 3, respectively. Responders (pathologic complete response and pathologic partial response) had the largest reductions in choline signal-to-noise ratio (35%, from 7.2 ± 2.3 to 4.6 ± 2; P < .01) compared to nonresponders (11%, from 8.4 ± 2.7 to 7.5 ± 3.6; P = .13) after the first cycle. Total tissue sodium concentration significantly decreased in responders (27%, from 66 ± 18 to 48.4 ± 8 mmol/L; P = .01), while there was little change in nonresponders (51.7 ± 7.6 to 56.5 ± 1.6 mmol/L; P = .50). Lesion volume decreased in responders (40%, from 78 ± 78 to 46 ± 51 mm³; P = .01) and nonresponders (21%, from 100 ± 104 to 79.2 ± 87 mm³; P = .23) after the first cycle. The largest reduction in Response Evaluation Criteria in Solid Tumors occurred after the first treatment in responders (18%, from 24.5 ± 20 to 20.2 ± 18 mm; P = .01), with a slight decrease in tumor diameter noted in nonresponders (17%, from 23 ± 19 to 19.2 ± 19.1 mm; P = .80).

CONCLUSIONS

Multiparametric and multinuclear imaging parameters were significantly reduced after the first cycle of preoperative systemic therapy in responders, specifically, choline signal-to-noise ratio and sodium. These new surrogate radiologic biomarkers maybe able to predict and provide a platform for potential adaptive therapy in patients.

Comparison between diffusion-weighted imaging, T2-weighted, and postcontrast T1-weighted imaging after MR-guided, high intensity, focused ultrasound treatment of uterine leiomyomata: preliminary results

PURPOSE

To investigate the comparison between diffusion-weighted imaging (DWI), T2-weighted imaging, (T2WI) and contrast T1-weighted imaging (cT1WI) in uterine leiomyoma following treatment by magnetic resonance imaging-guided, high intensity focused ultrasound surgery (MRg-HIFUS).

METHODS

Twenty one patients (45 +/- 5 yrs) with clinical symptoms of uterine leiomyoma (fibroids) were treated by MRg-HIFUS using an integrated 1.5T MRI-HIFUS system. MRI parameters consisted of DWI, T2WI, and T1-weighted fast spoiled gradient echo before and after contrast. The post-MRg-HIFUS treatment volume in the fibroid was assessed by cT1WI and DWI. Trace apparent diffusion coefficient maps were constructed for quantitative analysis. The regions of the treated uterine tissue were defined by a semisupervised segmentation method called the "eigenimage filter," using both cT1WI and DWI. Signal-to-noise ratios were determined for the T2WI pretreatment images. Segmented regions were tested by a similarity index for congruence. Descriptive, regression, and Bland-Altman statistics were calculated.

RESULTS

All the patients exhibited heterogeneously increased DWI signal intensity localized in the treated fibroid regions and were colocalized with the cT1WI defined area. The mean pretreatment T2WI signal intensity ratios were T2WI/muscle = 1.8 +/- 0.7 and T2WI/myometrium = 0.7 +/- 0.4. The congruence between the regions was significant, with a similarity of 84% and a difference of 8% between the regions. Regression analyses of the cT1WI and DWI segmented treatment volume were found to be significantly correlated (r2 = 0.94, p < 0.05) with the linear equation, (cT1WI) = 1.1 (DWI)-0.66. There is good agreement between the regions defined by cT1WI and DWI in most of the cases as shown from the Bland-Altman plots.

CONCLUSIONS

Diffusion-weighted imaging exhibited excellent agreement, congruence, and correlation with the cT1WI-defined region of treatment in uterine fibroid. Therefore, DWI could be useful as an adjunct for assessing treatment of uterine fibroids by MRg-HIFUS.

Proton, diffusion-weighted imaging, and sodium (23Na) MRI of uterine leiomyomata after MR-guided high-intensity focused ultrasound: a preliminary study

PURPOSE

To determine the feasibility of using combined proton (1H), diffusion-weighted imaging (DWI), and sodium (23Na) magnetic resonance imaging (MRI) to monitor the treatment of uterine leiomyomata (fibroids).

MATERIALS AND METHODS

Eight patients with uterine leiomyomata were enrolled and treated using MRI-guided high-intensity frequency ultrasound surgery (MRg-HIFUS). MRI scans collected at baseline and posttreatment consisted of T2-, T1-, and 1H DWI, as well as posttreatment 23Na MRI. The 23Na and 1H MRi were coregistered using a replacement phantom method. Regions of interest in treated and untreated uterine leiomyoma tissue were drawn on 1H MRI and DWI, wherein the tissue apparent diffusion coefficient of water (ADC) and absolute sodium concentrations were measured.

RESULTS

Regions of treated uterine tissue were clearly identified on both DWI and 23Na images. The sodium concentrations in normal myometrium tissue were 35.8+/-2.1 mmol (mM), in the fundus; 43.4+/-3.8 mM, and in the bladder; 65.3+/-0.8 mM with ADC in normal myometrium of 2.2+/-0.3x10(-3) mm2/sec. Sodium concentration in untreated leiomyomata were 28+/-5 mM, and were significantly elevated (41.6+/-7.6 mM, P<0.05) after treatment. Apparent diffusion coefficient values in the treated leiomyomata (1.30+/-0.38x10(-3) mm2/sec) were decreased compared to areas of untreated leiomyomata (1.75+/--4048micro-4050micro36x10(-3) mm2/sec; P=0.04).

CONCLUSION

Multiparametric imaging permits identification of uterine leiomyomata, revealing altered 23Na MRI and DWI levels following noninvasive treatment that provides a mechanism to explore the molecular and metabolic pathways after treatment.

Reliability of tumor volume estimation from MR images in patients with malignant glioma. Results from the American College of Radiology Imaging Network (ACRIN) 6662 Trial

Reliable assessment of tumor growth in malignant glioma poses a common problem both clinically and when studying novel therapeutic agents. We aimed to evaluate two software-systems in their ability to estimate volume change of tumor and/or edema on magnetic resonance (MR) images of malignant gliomas. Twenty patients with malignant glioma were included from different sites. Serial post-operative MR images were assessed with two software systems representative of the two fundamental segmentation methods, single-image fuzzy analysis (3DVIEWNIX-TV) and multi-spectral-image analysis (Eigentool), and with a manual method by 16 independent readers (eight MR-certified technologists, four neuroradiology fellows, four neuroradiologists). Enhancing tumor volume and tumor volume plus edema were assessed independently by each reader. Intraclass correlation coefficients (ICCs), variance components, and prediction intervals were estimated. There were no significant differences in the average tumor volume change over time between the software systems (p > 0.05). Both software systems were much more reliable and yielded smaller prediction intervals than manual measurements. No significant differences were observed between the volume changes determined by fellows/neuroradiologists or technologists.Semi-automated software systems are reliable tools to serve as outcome parameters in clinical studies and the basis for therapeutic decision-making for malignant gliomas, whereas manual measurements are less reliable and should not be the basis for clinical or research outcome studies.

Uterine fibroids: diffusion-weighted MR imaging for monitoring therapy with focused ultrasound surgery--preliminary study

PURPOSE

To prospectively determine the feasibility of using diffusion-weighted (DW) imaging and apparent diffusion coefficient (ADC) mapping before (baseline) and after treatment and at 6-month follow-up to monitor magnetic resonance (MR) image-guided focused ultrasound surgical ablation of uterine fibroids.

MATERIALS AND METHODS

Informed consent was obtained from patients before treatment with our study protocol, as approved by the institutional review board, and the study complied with the Health Insurance Portability and Accountability Act. Fourteen patients (mean age, 46 years +/- 5 [standard deviation]) who underwent DW imaging were enrolled in this study, and 12 of 14 completed the inclusive MR examination with DW imaging at 6-month follow-up. Treatment was performed by one radiologist with a modified MR image-guided focused ultrasound surgical system coupled with a 1.5-T MR imager. Pre- and posttreatment and 6-month follow-up MR images were obtained by using phase-sensitive T1-weighted fast spoiled gradient-recalled acquisition, T1-weighted contrast material-enhanced, and DW imaging sequences. Total treatment time was 1-3 hours. Trace ADC maps were constructed for quantitative analysis. Regions of interest localized to areas of hyperintensity on DW images were drawn on postcontrast images, and quantitative statistics were obtained from treated and nontreated uterine tissue before and after treatment and at 6-month follow-up. Statistical analysis was performed with analysis of variance. Differences with P < .05 were considered statistically significant.

RESULTS

T1-weighted contrast-enhancing fibroids selected for treatment had no hyperintense or hypointense signal intensity changes on the DW images or ADC maps before treatment. Considerably increased signal intensity changes that were localized within the treated areas were noted on DW images. Mean baseline ADC value in fibroids was 1504 mm(-6)/sec2 +/- 290. Posttreatment ADC values for nontreated fibroid tissue (1685 mm(-6)/sec2 +/- 468) differed from posttreatment ADC values for fibroid tissue (1078 mm(-6)/sec2 +/- 293) (P = .001). A significant difference (P < .001) between ADC values for treated (1905 mm(-6)/sec2 +/- 446) and nontreated (1437 mm(-6)/sec2 +/- 270) fibroid tissue at 6-month follow-up was observed.

CONCLUSION

DW imaging and ADC mapping are feasible for identification of ablated tissue after focused ultrasound treatment of uterine fibroids.

Multiparametric and multinuclear magnetic resonance imaging of human breast cancer: current applications

The exploration of novel imaging methods that have the potential to improve specificity for the identification of malignancy is still critically needed in breast imaging. Changes in physiologic alterations of soft tissue water associated with breast cancer can be visualized by magnetic resonance (MR) imaging. However, it is unlikely that a single MR parameter can characterize the complexity of breast tissue. Techniques such as multiparametric MR imaging, proton magnetic resonance spectroscopic (MRSI) imaging, and 23Na sodium MR imaging when used in combination provide a comprehensive data set with potentially more power to diagnose breast disease than any single measure alone. A combination of MR, MRSI, and 23Na sodium MR parameters may be examined in a single MR imaging examination, potentially resulting in improved specificity for radiologic evaluation of malignancy.

Fast method for brain image segmentation: Application to proton magnetic resonance spectroscopic imaging

The interpretation of brain metabolite concentrations measured by quantitative proton magnetic resonance spectroscopic imaging (MRSI) is assisted by knowledge of the percentage of gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) within each MRSI voxel. Usually, this information is determined from T(1)-weighted magnetic resonance images (MRI) that have a much higher spatial resolution than the MRSI data. While this approach works well, it is time-consuming. In this article, a rapid data acquisition and analysis procedure for image segmentation is described, which is based on collection of several, thick slice, fast spin echo images (FSE) of different contrast. Tissue segmentation is performed with linear "Eigenimage" filtering and normalization. The method was compared to standard segmentation techniques using high-resolution 3D T(1)-weighted MRI in five subjects. Excellent correlation between the two techniques was obtained, with voxel-wise regression analysis giving GM: R2 = 0.893 +/- 0.098, WM: R2 = 0.892 +/- 0.089, ln(CSF): R2 = 0.831 +/- 0.082). Test-retest analysis in one individual yielded an excellent agreement of measurements with R2 higher than 0.926 in all three tissue classes. Application of FSE/EI segmentation to a sample proton MRSI dataset yielded results similar to prior publications. It is concluded that FSE imaging in conjunction with Eigenimage analysis is a rapid and reliable way of segmenting brain tissue for application to proton MRSI.

Application of independent component analysis to magnetic resonance imaging for enhancing the contrast of gray and white matter

An application of independent component analysis (ICA) was attempted to develop a method of processing magnetic resonance (MR) images to extract physiologically independent components representing tissue relaxation times and achieve improved visualization of normal and pathologic structures. Anatomical T1-weighted, T2-weighted and proton density images were obtained from 10 normal subjects, 3 patients with brain tumors and 1 patient with multiple sclerosis. The data sets were analyzed using ICA based on the learning rule of Bell and Sejnowski after prewhitening operations. The three independent components obtained from the three original data sets corresponded to (1) short T1 components representing myelin of white matter and lipids, (2) relatively short T1 components representing gray matter and (3) long T2 components representing free water. The involvement of gray or white matter in brain tumor cases and the demyelination in the case of multiple sclerosis were enhanced and visualized in independent component images. ICA can potentially achieve separation of tissues with different relaxation characteristics and generate new contrast images of gray and white matter. With the proper choice of contrast for the original images, ICA may be useful not only for extracting subtle or hidden changes but also for preprocessing transformation before clustering and segmenting the structure of the human brain.

Combined dynamic contrast enhanced breast MR and proton spectroscopic imaging: A feasibility study

PURPOSE

To investigate the feasibility of combined dynamic contrast enhanced (DCE) and magnetic resonance spectroscopy (MRS) in evaluating breast lesions.

METHODS

Nine patients with positive mammograms scheduled for either biopsy or mastectomy were examined on a 1.5-T MR scanner. DCE was performed with administration of gadolinium-DTPA contrast using a two-dimensional spoiled gradient recall echo (SPGR) sequence. Proton spectroscopy (TR/TE = 2000/272 msec) was performed using PRESS single slice (10 mm). Lesion time intensity curves were classified as persistent (type 1), plateau (type 2), or washout (type 3) pattern enhancement. Choline (Cho) signal-to-noise ratios (SNRs) and enhancement patterns were compared between benign and malignant lesions as determined by histopathology.

RESULTS

Five patients had breast carcinoma and four had benign lesions. Type 1 enhancement was found in two benign cases, type 2 enhancement in two of four benign and four of five malignant lesions, and one malignant case exhibited a type 3 pattern. Choline SNR was significantly different (P < 0.003) between benign and malignant lesions (2.0 +/- 0.3 vs. 5.7 +/- 1.4; P < 0.003). Choline SNR was less than 4.0 in all of the benign lesions, including the two lesions with type 2 enhancement.

CONCLUSION

Proton MRS appears to be a promising technique for classification of breast lesions when DCE results are equivocal. A combination of DCE and MRS is feasible, and may have improved specificity compared to either modality alone.

Benign and Malignant Breast Lesions: Diagnosis with Multiparametric MR Imaging

PURPOSE

To both develop and use a tissue signature method for the identification and classification of breast lesions and healthy breast tissue with magnetic resonance (MR) imaging.

MATERIALS AND METHODS

Thirty-six patients underwent breast MR imaging (T1- and T2-weighted imaging and three-dimensional T1-weighted imaging with and without contrast material enhancement), followed by biopsy or mastectomy and histopathologic analysis. Tissue cluster analysis was performed by using the iterative self-organizing data technique to identify glandular, adipose, and lesion tissue signature vectors. Glandular and lesion tissue vectors were characterized by angular separation from the reference adipose tissue vector. Differences in angular separation of histologically proved benign and malignant lesion groups were evaluated with an independent t test. The usefulness of the angular separation model for distinguishing benign from malignant lesions was evaluated with nonparametric receiver operating characteristic curve analysis.

RESULTS

The model enabled successful identification and characterization of breast lesion tissue clusters in all patients; 18 lesions were benign, and 18 were malignant. Angular separation +/- SD was 17.8 degrees +/- 6.1 degrees between adipose tissue and malignant lesions and 29.0 degrees +/- 11.2 degrees between adipose tissue and benign lesions. Angular separations of benign lesions and malignant lesions were significantly different (P =.002), with a specificity of 78% and sensitivity of 89% at a cutoff value of 21 degrees. Significant differences in angular separation from adipose tissue also were found between glandular tissue and lesion tissue (P <.001) and, in glandular tissue, between patients with benign lesions and those with malignant lesions (P =.04). The area under the receiver operating characteristic curve was 0.84.

CONCLUSION

Multispectral analysis of conventional breast MR images based on the iterative self-organizing data model and on measurement of angular separation between tissue signature vectors may enable automated lesion identification and classification.

MRI tissue characterization of experimental cerebral ischemia in rat

PURPOSE

To extend the ISODATA image segmentation method to characterize tissue damage in stroke, by generating an MRI score for each tissue that corresponds to its histological damage.

MATERIALS AND METHODS

After preprocessing and segmentation (using ISODATA clustering), the proposed method scores tissue regions between 1 and 100. Score 1 is assigned to normal brain matter (white or gray matter), and score 100 to cerebrospinal fluid (CSF). Lesion zones are assigned a score based on their relative levels of similarities to normal brain matter and CSF. To evaluate the method, 15 rats were imaged by a 7T MRI system at one of three time points (acute, subacute, chronic) after MCA occlusion. Then they were killed and their brains were sliced and prepared for histological studies. MRI of two or three slices of each rat brain (using two DWI (b = 400, b = 800), one PDWI, one T2WI, and one T1WI) was performed, and an MRI score between 1 and 100 was determined for each region. Segmented regions were mapped onto the histology images and scored on a scale of 1-10 by an experienced pathologist. The MRI scores were validated by comparison with histology scores. To this end, correlation coefficients between the two scores (MRI and histology) were determined.

RESULTS

Experimental results showed excellent correlations between MRI and histology scores at different time points. Depending on the reference tissue (gray matter or white matter) used in the standardization, the correlation coefficients ranged from 0.73 (P < 0.0001) to 0.78 (P < 0.0001) using the entire dataset, including acute, subacute, and chronic time points. This suggests that the proposed multiparametric approach accurately identified and characterized ischemic tissue in a rat model of cerebral ischemia at different stages of stroke evolution.

CONCLUSION

The proposed approach scores tissue regions and characterizes them using unsupervised clustering and multiparametric image analysis techniques. The method can be used for a variety of applications in the field of computer-aided diagnosis and treatment, including evaluation of response to treatment. For example, volume changes for different zones of the lesion over time (e.g., tissue recovery) can be evaluated.

Proton magnetic resonance spectroscopic imaging of human breast cancer: A preliminary study

PURPOSE

To investigate the diagnostic value of proton magnetic resonance spectroscopic imaging (MRSI) in patients with breast lesions.

MATERIALS AND METHODS

Eighteen patients underwent breast MRSI and MRI at 1.5 T. Contrast-enhanced MR was used to identify the lesion, after which single-slice MRSI (TR/TE = 2000/272 msec, 10-mm slice thickness) was performed. Water, lipid, and choline (Cho) images were reconstructed from MRSI data. The area of the Cho was measured in the lesion and expressed relative to the background noise level (signal-to-noise ratio (SNR)), measured between 7.0 and 9.0 ppm. Cho SNRs were compared between benign and malignant lesions as determined by histopathology.

RESULTS

Three cases were considered technical failures on MRSI. Of the remaining 15 cases, on histopathology, eight were classified as malignant carcinoma and seven were benign. The Cho SNR from malignant tissue was significantly elevated compared to benign tissue (6.2 +/- 2.1 vs. 2.4 +/- 0.7, P < 0.0008).

CONCLUSIONS

MRSI measurements of Cho are feasible in the human breast, and the SNR for Cho was significantly different between benign and malignant lesions. The potential advantages of MRSI over SV spectroscopy include the ability to assess multiple lesions as well as tissue with normal MRI appearance, as well as to perhaps gauge lesion borders and infiltration into surrounding tissue.

A model for multiparametric mri tissue characterization in experimental cerebral ischemia with histological validation in rat: part 1

BACKGROUND AND PURPOSE

After stroke, brain tissue undergoes time-dependent heterogeneous histopathological change. These tissue alterations have MRI characteristics that allow segmentation of ischemic from nonischemic tissue. Moreover, MRI segmentation generates different zones within the lesion that may reflect heterogeneity of tissue damage.

METHODS

A vector tissue signature model is presented that uses multiparametric MRI for segmentation and characterization of tissue. An objective (unsupervised) computer segmentation algorithm was incorporated into this model with the use of a modified version of the Iterative Self-Organizing Data Analysis Technique (ISODATA). The ability of the model to characterize ischemic tissue after permanent middle cerebral ischemia occlusion in the rat was tested. Multiparametric ISODATA measurements of the ischemic tissue were compared with quantitative histological characterization of the tissue from 4 hours to 1 week after stroke.

RESULTS

The ISODATA segmentation of tissue identified a gradation of cerebral tissue damage at all time points after stroke. The histological scoring of ischemic tissue from 4 hours to 1 week after stroke on all the animals was significantly correlated with ISODATA segmentation (r=0.78, P<0.001; n=20) when a multiparametric (T2-, T1-, diffusion-weighted imaging) data set was used, less correlated (r=0.70, P<0.01; n=20) when a T2- and T1-weighted data set was used, and not correlated (r=-0.12, P>0.47; n=20) when only a diffusion-weighted imaging data set was used.

CONCLUSIONS

Our data indicate that an integrated set of MRI parameters can distinguish and stage ischemic tissue damage in an objective manner.

Multiparametric MRI tissue characterization in clinical stroke with correlation to clinical outcome: part 2

BACKGROUND AND PURPOSE

Multiparametric MRI generates different zones within the lesion that may reflect heterogeneity of tissue damage in cerebral ischemia. This study presents the application of a novel model of tissue characterization based on an angular separation between tissues obtained with the use of an objective (unsupervised) computer segmentation algorithm implementing a modified version of the Iterative Self-Organizing Data Analysis Technique (ISODATA). We test the utility of this model to identify ischemic tissue in clinical stroke.

METHODS

MR parameters diffusion-, T2-, and T1-weighted imaging (DWI, T2WI, and T1WI, respectively) were obtained from 10 patients at 3 time points (30 studies) after stroke: acute (</=12 hours), subacute (3 to 5 days), and chronic (3 months). The National Institutes of Health Stroke Scale (NIHSS) was measured, and volumes were obtained from the ISODATA, DWI, and T2WI maps on patients at each time point.

RESULTS

The acute (</=12 hours) multiparametric ISODATA volume was significantly correlated with the acute (</=12 hours) DWI (r=0.96, P<0.05; n=10) and chronic (3 months) T2WI volume (r=0.69, P<0.05; n=10). The ISODATA-defined tissue regions exhibited MR indices consistent with ischemic and/or infarcted tissue at each time point. The acute (</=12 hours) multiparametric ISODATA volumes were significantly correlated (r=0.82, P<0.009; n=10) with the final NIHSS score. In comparison, the acute (</=12 hours) DWI volumes were less correlated (r=0.77, P<0.05; n=10) and T2WI volume (</=12h) exhibited a marginal correlation (r=0.66, P<0.05; n=10) with the final NIHSS score.

CONCLUSIONS

The integrated ISODATA approach to tissue segmentation and classification discriminated abnormal from normal tissue at each time point. The ISODATA volume was significantly correlated with the current MR standards used in the clinical setting and the 3-month clinical status of the patient.

Boundary-based warping of brain MR images

The goal of this work was to develop a warping technique for mapping a brain image to another image or atlas data, with minimum user interaction and independent of gray level information. We have developed and tested three different methods for warping magnetic resonance (MR) brain images. We utilize a deformable contour to extract and warp the boundaries of the two images. A mesh-grid coordinate system is constructed for each brain, by applying a distance transformation to the resulting contours, and scaling. In the first method (MGC), the first image is mapped to the second image based on a one-to-one mapping between different layers defined by the mesh-grid. In the second method (IDW), the corresponding pixels in the two images are found using the above mesh-grid system and a local inverse-distance weights interpolation. In the third proposed method (TSB), a subset of grid points is used for finding the parameters of a spline transformation, which defines the global warping. The warping methods were applied to clinical MR consisting of diffusion-weighted and T2-weighted images of the human brain. The IDW and TSB methods were superior in ranking of diagnostic quality of the warped MR images to the MGC (P < 0.01) as defined by a neuroradiologist. The deformable contour warping produced excellent diagnostic quality for the diffusion-weighted images coregistered and warped to T2 weighted images. J. Magn. Reson. Imaging 2000;12:417-429.

Unsupervised segmentation of multiparameter MRI in experimental cerebral ischemia with comparison to T2, diffusion, and ADC MRI parameters and histopathological validation

This study presents histological validation of an objective (unsupervised) computer segmentation algorithm, the iterative self-organizing data analysis technique (ISODATA), for analysis of multiparameter magnetic resonance imaging (MRI) data in experimental focal cerebral ischemia. T2-, T1-, and diffusion (DWI) weighted coronal images were acquired from 4 to 168 hours after stroke on separate groups of animals. Animals were killed immediately after MRI for histological analysis. MR images were coregistered/warped to histology. MRI lesion areas were defined using DWI, apparent diffusion coefficient (ADC) maps, T2-weighted images, and ISODATA. The last techniques clearly discriminated between ischemia-altered and morphologically intact tissue. ISODATA areas were congruent and significantly correlated (r = 0.99, P < 0.05) with histologically defined lesions. In contrast, DWI, ADC, and T2 lesion areas showed no significant correlation with histologically evaluated lesions until subacute time points. These data indicate that multiparameter ISODATA methodology can accurately detect and identify ischemic cell damage early and late after ischemia, with ISODATA outperforming ADC, DWI, and T2-weighted images in identification of ischemic lesions from 4 to 168 hours after stroke.