Deanna L. Langer

Prostate Tissue Composition and MR Measurements: Investigating the Relationships between ADC, T2, Ktrans, ve, and Corresponding Histologic Features

PURPOSE To investigate relationships between magnetic resonance (MR) imaging measurements and the underlying composition of normal and malignant prostate tissue. MATERIALS AND METHODS Twenty-four patients (median age, 63 years; age range, 44-72 years) gave informed consent to be examined for this research ethics board-approved study. Before undergoing prostatectomy, patients were examined with T2-weighted, diffusion-weighted, T2 mapping, and dynamic contrast material-enhanced MR imaging at 1.5 T. Maps of apparent diffusion coefficient (ADC), T2, volume transfer constant (K(trans)), and extravascular extracellular space (v(e)) were calculated. Whole-mount hematoxylin-eosin-stained sections were generated and digitized at histologic resolution. Percentage areas of tissue components (nuclei, cytoplasm, stroma, luminal space) were measured by using image segmentation. Corresponding regions on MR images and histologic specimens were defined by using anatomically defined segments in peripheral zone (PZ) and central gland tissue. Cancer and normal PZ regions were identified at histopathologic analysis. Each MR parameter-histologic tissue component pair was assessed by using linear mixed-effects models, and cancer versus normal PZ values were compared by using nonparametric tests. RESULTS ADC and T2 were inversely related to percentage area of nuclei and percentage area of cytoplasm and positively related to percentage area of luminal space (P < or = .01). These trends were reversed for K(trans) (P < .001). K(trans) had a significantly negative (P = .01) slope versus percentage area of stroma, and v(e) had a positive (P = .008) slope versus percentage area of stroma. The v(e) was inversely proportional to the percentage area of nuclei (P = .05). All MR imaging parameters (P < or = .05) and the percentage areas of all tissue components (P < or = .001) except stroma (P > .48) were significantly different between cancer and normal PZ tissue. CONCLUSION MR imaging-derived parameters measured in the prostate were significantly related to the proportion of specific histologic components that differ between normal and malignant PZ tissue. These relationships may help define imaging-related histologic prognostic parameters for prostate cancer.

Focal Laser Ablation for Prostate Cancer Followed by Radical Prostatectomy: Validation of Focal Therapy and Imaging Accuracy

An increased incidence of low-risk prostate cancer (PCa) has led investigators to develop focal therapy as a management option for PCa. We evaluated the effects of focal laser ablation (FLA) on PCa tissue and the accuracy of magnetic resonance imaging (MRI) in determining ablated lesion volume by comparing the whole-mount histology and MRI in four patients that underwent FLA followed by radical prostatectomy. Ablated areas were characterized by homogeneous coagulation necrosis. The MRI-calculated ablated volume correlated well with histopathology. We found that FLA creates confluent ablation with no evidence of viable cells in treated regions. Postablation MRI is able to determine the ablation accurately.

Supervised and unsupervised methods for prostate cancer segmentation with multispectral MRI

PURPOSE Magnetic resonance imaging (MRI) has been proposed as a promising alternative to transrectal ultrasound for the detection and localization of prostate cancer and fusing the information from multispectral MR images is currently an active research area. In this study, the goal is to develop automated methods that combine the pharmacokinetic parameters derived from dynamic contrast enhanced (DCE) MRI with quantitative T2 MRI and diffusion weighted imaging (DWI) in contrast to most of the studies which were performed with human readers. The main advantages of the automated methods are that the observer variability is removed and easily reproducible results can be efficiently obtained when the methods are applied to a test data. The goal is also to compare the performance of automated supervised and unsupervised methods for prostate cancer localization with multispectral MRI. METHODS The authors use multispectral MRI data from 20 patients with biopsy-confirmed prostate cancer patients, and the image set consists of parameters derived from T2, DWI, and DCE-MRI. The authors utilize large margin classifiers for prostate cancer segmentation and compare them to an unsupervised method the authors have previously developed. The authors also develop thresholding schemes to tune support vector machines (SVMs) and their probabilistic counterparts, relevance vector machines (RVMs), for an improved performance with respect to a selected criterion. Moreover, the authors apply a thresholding method to make the unsupervised fuzzy Markov random fields method fully automatic. RESULTS The authors have developed a supervised machine learning method that performs better than the previously developed unsupervised method and, additionally, have found that there is no significant difference between the SVM and RVM segmentation results. The results also show that the proposed methods for threshold selection can be used to tune the automated segmentation methods to optimize results for certain criteria such as accuracy or sensitivity. The test results of the automated algorithms indicate that using multispectral MRI improves prostate cancer segmentation performance when compared to single MR images, a result similar to the human reader studies that were performed before. CONCLUSIONS The automated methods presented here can help diagnose and detect prostate cancer, and improve segmentation results. For that purpose, multispectral MRI provides better information about cancer and normal regions in the prostate when compared to methods that use single MRI techniques; thus, the different MRI measurements provide complementary information in the automated methods. Moreover, the use of supervised algorithms in such automated methods remain a good alternative to the use of unsupervised algorithms.

Prostate Cancer Localization With Multispectral MRI Using Cost-Sensitive Support Vector Machines and Conditional Random Fields

Prostate cancer is a leading cause of cancer death for men in the United States. Fortunately, the survival rate for early diagnosed patients is relatively high. Therefore, in vivo imaging plays an important role for the detection and treatment of the disease. Accurate prostate cancer localization with noninvasive imaging can be used to guide biopsy, radiotheraphy, and surgery as well as to monitor disease progression. Magnetic resonance imaging (MRI) performed with an endorectal coil provides higher prostate cancer localization accuracy, when compared to transrectal ultrasound (TRUS). However, in general, a single type of MRI is not sufficient for reliable tumor localization. As an alternative, multispectral MRI, i.e., the use of multiple MRI-derived datasets, has emerged as a promising noninvasive imaging technique for the localization of prostate cancer; however almost all studies are with human readers. There is a significant inter and intraobserver variability for human readers, and it is substantially difficult for humans to analyze the large dataset of multispectral MRI. To solve these problems, this study presents an automated localization method using cost-sensitive support vector machines (SVMs) and shows that this method results in improved localization accuracy than classical SVM. Additionally, we develop a new segmentation method by combining conditional random fields (CRF) with a cost-sensitive framework and show that our method further improves cost-sensitive SVM results by incorporating spatial information. We test SVM, cost-sensitive SVM, and the proposed cost-sensitive CRF on multispectral MRI datasets acquired from 21 biopsy-confirmed cancer patients. Our results show that multispectral MRI helps to increase the accuracy of prostate cancer localization when compared to single MR images; and that using advanced methods such as cost-sensitive SVM as well as the proposed cost-sensitive CRF can boost the performance significantly when compared to SVM.

T2*-weighted and arterial spin labeling MRI of calf muscles in healthy volunteers and patients with chronic exertional compartment syndrome: preliminary experience.

OBJECTIVE The purpose of our study was to assess temporal changes with exercise in T2* and arterial spin labeling signals in patients with chronic exertional compartment syndrome of the anterior compartment of the lower leg and in control subjects using T2* mapping and arterial spin labeling MRI. SUBJECTS AND METHODS This prospective study was approved by the institutional research ethics board. Ten control subjects (five women and five men; mean age, 29.0 years) and nine patients with chronic exertional compartment syndrome (three women and six men; mean age, 33.7 years) gave informed written consent and underwent MRI of the calf muscles using an axial T2*-weighted multiecho gradient-recalled echo and a flow-sensitive alternating inversion recovery sequence with echo-planar imaging readouts before (baseline) and 3, 6, 9, 12, and 15 minutes after exercise. T2* and arterial spin labeling signal changes (DeltaT2* and DeltaASL, respectively) over time were calculated relative to the baseline examination. DeltaT2* and DeltaASL between patients and control subjects were compared using the Students t test. RESULTS In both patients and control subjects, DeltaT2* and DeltaASL showed a peak at 3 minutes after exercise, followed by a decrease over time. The maximum DeltaT2* was 26% and 29% for patients and control subjects, respectively. The maximum DeltaASL was 183% and 224% for patients and control subjects, respectively. After 15 minutes, arterial spin labeling signal returned to baseline; however, T2* remained elevated (8% in patients; 10% in control subjects). No statistically significant differences between patients and control subjects in postexercise DeltaT2* and DeltaASL were found (p = 0.21-0.98). CONCLUSION After calf muscle exercise, no statistically significant differences in T2* relaxation times or arterial spin labeling signal, indicative of differences in muscle oxygenation and perfusion status, were found between patients with chronic exertional compartment syndrome and control subjects.

Unsupervised segmentation of the prostate using MR images based on level set with a shape prior

Prostate cancer is the second leading cause of cancer death in American men. Current prostate MRI can benefit from automated tumor localization to help guide biopsy, radiotherapy and surgical planning. An important step of automated prostate cancer localization is the segmentation of the prostate. In this paper, we propose a fully automatic method for the segmentation of the prostate. We firstly apply a deformable ellipse model to find an ellipse that best fits the prostate shape. Then, this ellipse is used to initiate the level set and constrain the level set evolution with a shape penalty term. Finally, certain post processing methods are applied to refine the prostate boundaries. We apply the proposed method to real diffusion-weighted (DWI) MRI images data to test the performance. Our results show that accurate segmentation can be obtained with the proposed method compared to human readers.

Prostate cancer localization with multispectral MRI based on Relevance Vector Machines

Prostate cancer is one of the leading causes of cancer death for men. However, early detection before cancer spreads beyond the prostate can reduce the mortality. Therefore, invivo imaging techniques play an important role to localize the prostate cancer for treatment. Although Magnetic Resonance Imaging (MRI) has been proposed to localize prostate cancer, the studies on automated localization with multispectral MRI have been limited. In this study we propose combining the pharmacokinetic parameters derived from DCE MRI with T2 MRI and DWI. We also propose to use Relevance Vector Machines (RVM) for automatic prostate cancer localization, compare its performance to Support Vector Machines (SVM) and show that RVM can produce more accurate and more efficient segmentation results than SVM for automated prostate cancer localization with multispectral MRI.

Using relative contrast and iterative normalization for improved prostate cancer localization with multispectral MRI

In this paper, a new method that uses relative contrast is proposed for medical image segmentation problems. Generally, the absolute intensity values of different features are mapped into a comparable range with a normalization method, but the differences across patients are not considered. In order to utilize the patient-specific information from medical images, we use relative contrast between the normal and malignant tissues to perform training. The proposed relative contrast based method mimics the image segmentation procedure performed by human readers based on relative intensity values rather than absolute intensity values. The proposed method requires the knowledge of normal and malignant tissues since it is based on their relative intensities. This is known at the training stage, but unknown for the test data. Therefore, we present an iterative algorithm to estimate the relative contrast based on the current estimate of the class membership for the test data. Our experimental results show that the suggested algorithm outperforms the classical z-score normalization for prostate cancer localization with multispectral MR images.

Improved prostate cancer localization with spatially regularized dynamic contrast-enhanced magnetic resonance imaging

Imaging methods to localize prostate cancer with sufficient accuracy are extremely useful in guiding biopsy, radiotherapy and surgery as well as to monitor disease progression. Imaging prostate cancer with multispectral magnetic resonance imaging (MRI) has shown a superior performance when compared to classical imaging modality transrectal ultrasound (TRUS). An important component of multispectral MRI is dynamic contrast-enhanced magnetic resonance imaging (DCE MRI). However, parametric images based on DCE MRI suffer from low signal-to-noise ratio (SNR). In this study, we propose a kinetic parametric imaging method with DCE MRI to overcome this problem using spatial regularization for improved prostate cancer localization. We demonstrate that the proposed method outperforms pixel-wise parametric imaging method, and that the performance of resulting tumor localization has a considerable improvement. Both visual and quantitative evaluations based on a task-based approach focused on tumor localization are provided.