Skip to main content

Über dieses Buch

Prostatic adenocarcinoma (CAP) is the second most common malignancy with an estimated 190,000 new cases in the USA in 2010 (Source: American Cancer Society), and is the most frequently diagnosed cancer among men. If CAP is caught early, men have a high, five-year survival rate. Unfortunately there is no standardized ima- based screening protocol for early detection of CAP (unlike for breast cancers). In the USA high levels of prostate-specific antigen (PSA) warrant a trans-rectal ultrasound (TRUS) biopsy to enable histologic confirmation of presence or absence of CAP. With recent rapid developments in multi-parametric radiological imaging te- niques (spectroscopy, dynamic contrast enhanced MR imaging, PET, RF ultrasound), some of these functional and metabolic imaging modalities are allowing for definition of high resolution, multi-modal signatures for prostate cancer in vivo. Distinct com- tational and technological challenges for multi-modal data registration and classifi- tion still remain in leveraging this multi-parametric data for directing therapy and optimizing biopsy. Additionally, with the recent advent of whole slide digital sc- ners, digitized histopathology has become amenable to computerized image analysis. While it is known that outcome of prostate cancer (prognosis) is highly correlated with Gleason grade, pathologists often have difficulty in distinguishing between interme- ate Gleason grades from histopathology. Development of computerized image analysis methods for automated Gleason grading and predicting outcome on histopathology have to confront the significant computational challenges associated with working these very large digitized images.



Prostate Cancer MR Imaging

With a total of 192,280 new cases predicted for 2009, prostate cancer (PC) now accounts for 25% of all new male cancers diagnosed in the United States [1]. Furthermore, in their lifetime, one in six men will be clinically diagnosed with having PC, although many more men are found to have histological evidence of PC at autopsy [2,3,4]. Presently, approximately 1 in 10 men will die of PC [5,6]. The ever-aging population and wider spread use of the blood prostate-specific antigen (PSA) test [7,8], as well as the tendency to apply lower cut-off levels for this test [9], will further increase the diagnosis of this disease [10].
Jurgen J. Fütterer

Computer Aided Detection of Prostate Cancer Using T2, DWI and DCE MRI: Methods and Clinical Applications

One in 10 men will be diagnosed with prostate cancer during their life. PSA screening in combination with MR is likely to save lifes at low biopsy and overtreatment rates. Computer Aided Diagnosis for prostate MR will become mandatory in a high volume screening application. This paper presents an overview including our recent work in this area. It includes screening MR setup, quantitative imaging features, prostate segmentation, and pattern recognition.
Henkjan Huisman, Pieter Vos, Geert Litjens, Thomas Hambrock, Jelle Barentsz

Prostate Cancer Segmentation Using Multispectral Random Walks

Several studies have shown the advantages of multispectral magnetic resonance imaging (MRI) as a noninvasive imaging technique for prostate cancer localization. However, a large proportion of these studies are with human readers. There is a significant inter and intra-observer variability for human readers, and it is substantially difficult for humans to analyze the large dataset of multispectral MRI. To solve these problems a few studies were proposed for fully automated cancer localization in the past. However, fully automated methods are highly sensitive to parameter selection and often may not produce desirable segmentation results. In this paper, we present a semi-supervised segmentation algorithm by extending a graph based semi-supervised random walker algorithm to perform prostate cancer segmentation with multispectral MRI. Unlike classical random walker which can be applied only to dataset of single type of MRI, we develop a new method that can be applied to multispectral images. We prove the effectiveness of the proposed method by presenting the qualitative and quantitative results of multispectral MRI datasets acquired from 10 biopsy-confirmed cancer patients. Our results demonstrate that the multispectral MRI noticeably increases the sensitivity and jakkard measures of prostate cancer localization compared to single MR images; 0.71 sensitivity and 0.56 jakkard for multispectral images compared to 0.51 sensitivity and 0.44 jakkard for single MR image based segmentation.
Yusuf Artan, Masoom A. Haider, Imam Samil Yetik

Automatic MRI Atlas-Based External Beam Radiation Therapy Treatment Planning for Prostate Cancer

Prostate radiation therapy dose planning currently requires computed tomography (CT) scans as they contain electron density information needed for patient dose calculations. However magnetic resonance imaging (MRI) images have significantly superior soft-tissue contrast for segmenting organs of interest and determining the target volume for treatment. This paper describes work on the development of an alternative treatment workflow enabling both organ delineation and dose planning to be performed using MRI alone. This is achieved by atlas based segmentation and the generation of pseudo-CT scans from MRI. Planning and dosimetry results for three prostate cancer patients from Calvary Mater Newcastle Hospital (Australia) are presented supporting the feasibility of this workflow. Good DSC scores were found for the atlas based segmentation of the prostate (mean 0.84) and bones (mean 0.89). The agreement between MRI/pseudo-CT and CT planning was quantified by dose differences and distance to agreement in corresponding voxels. Dose differences were found to be less than 2%. Chi values indicate that the planning CT and pseudo-CT dose distributions are equivalent.
Jason Dowling, Jonathan Lambert, Joel Parker, Peter B. Greer, Jurgen Fripp, James Denham, Sébastien Ourselin, Olivier Salvado

An Efficient Inverse-Consistent Diffeomorphic Image Registration Method for Prostate Adaptive Radiotherapy

Deformable image registration is a key enabling technology for adaptive radiation therapy (ART) as it can facilitate structure segmentation as well as dose tracking and accumulation. In this work, we develop an efficient inverse-consistent diffeomorphic registration method applying the log-Euclidean formulation of diffeomorphisms. Unlike existing log-Euclidean deformable registration approaches, the proposed method deforms two images towards each other in a completely symmetric fashion during the registration optimization, which leads to higher efficiency and better accuracy in recovering large deformations. The method is applied for the automatic segmentation of daily CT images in prostate ART. To address difficulties caused by large bladder and rectum content change, we propose further improvements and combine deformable registration with model-based image segmentation. Validation results on real clinical data showed that the proposed method gives highly accurate segmentation of interested structures.
Xiao Han, Lyndon S. Hibbard, Virgil Willcut

Atlas Based Segmentation and Mapping of Organs at Risk from Planning CT for the Development of Voxel-Wise Predictive Models of Toxicity in Prostate Radiotherapy

The prediction of toxicity is crucial to managing prostate cancer radiotherapy (RT). This prediction is classically organ wise and based on the dose volume histograms (DVH) computed during the planning step, and using for example the mathematical Lyman Normal Tissue Complication Probability (NTCP) model. However, these models lack spatial accuracy, do not take into account deformations and may be inappropiate to explain toxicity events related with the distribution of the delivered dose. Producing voxel wise statistical models of toxicity might help to explain the risks linked to the dose spatial distribution but is challenging due to the difficulties lying on the mapping of organs and dose in a common template. In this paper we investigate the use of atlas based methods to perform the non-rigid mapping and segmentation of the individuals’ organs at risk (OAR) from CT scans. To build a labeled atlas, 19 CT scans were selected from a population of patients treated for prostate cancer by radiotherapy. The prostate and the OAR (Rectum, Bladder, Bones) were then manually delineated by an expert and constituted the training data. After a number of affine and non rigid registration iterations, an average image (template) representing the whole population was obtained. The amount of consensus between labels was used to generate probabilistic maps for each organ. We validated the accuracy of the approach by segmenting the organs using the training data in a leave one out scheme. The agreement between the volumes after deformable registration and the manually segmented organs was on average above 60% for the organs at risk. The proposed methodology provides a way to map the organs from a whole population on a single template and sets the stage to perform further voxel wise analysis. With this method new and accurate predictive models of toxicity will be built.
Oscar Acosta, Jason Dowling, Guillaume Cazoulat, Antoine Simon, Olivier Salvado, Renaud de Crevoisier, Pascal Haigron

Realtime TRUS/MRI Fusion Targeted-Biopsy for Prostate Cancer: A Clinical Demonstration of Increased Positive Biopsy Rates

In this paper, a system for fusion of realtime transrectal ultrasound (TRUS) with pre-acquired 3D images of the prostate is presented with a clinical demonstration on a cohort of 101 patients with suspicion of prostate cancer. Electromagnetically tracked biopsy guides for endocavity ultrasound transducers were calibrated and used to fuse MRI-based suspicious lesion locations with ultrasound image coordinates. The prostate shape is segmented from MRI in a semi-automated fashion via a model-based approach, and intraoperative image registration is performed between MR and ultrasound image space to superimpose target fiducials markers on the ultrasound image. In order to align both modalities, a surface model is automatically extracted from 2D swept TRUS images using a partial active shape model, utilizing image features and prior statistics. An automatic prostate motion compensation algorithm can be triggered as needed. The results were used to display live TRUS images fused with spatially corresponding realtime multiplanar reconstructions (MPRs) of the MR image volume. In this study, all patients were scanned with 3T MRI and TRUS for biopsy. Clinical results show significant improvement of target visualization and of positive detection rates during TRUS-guided biopsies. It also demonstrates the feasibility of realtime MR/TRUS image fusion for out-of-gantry procedures.
Samuel Kadoury, Pingkun Yan, Sheng Xu, Neil Glossop, Peter Choyke, Baris Turkbey, Peter Pinto, Bradford J Wood, Jochen Kruecker

HistoCAD: Machine Facilitated Quantitative Histoimaging with Computer Assisted Diagnosis

Prostatic adenocarcinoma (CAP) is the most common malignancy in American men. In 2010 there will be an estimated 217,730 new cases and 32,050 deaths from CAP in the US. The diagnosis of prostatic adenocarcinoma is made exclusively from the histological evaluation of prostate tissue. The sampling protocols used to obtain 18 gauge (1.5 mm diameter) needle cores are standard sampling templates consisting of 6-12 cores performed in the context of an elevated serum value for prostate specific antigen (PSA). In this context, the prior probability of cancer is somewhat increased. However, even in this screened population, the efficiency of finding cancer is low at only approximately 20%. Histopathologists are faced with the task of reviewing the 5-10 million cores of tissue resulting from approximately 1,000,000 biopsy procedures yearly, parsing all the benign scenes from the worrisome scenes, and deciding which of the worrisome images are cancer.
John E. Tomaszewski

Registration of In Vivo Prostate Magnetic Resonance Images to Digital Histopathology Images

Early and accurate diagnosis of prostate cancer enables minimally invasive therapies to cure the cancer with less morbidity. The purpose of this work is to non-rigidly register in vivo pre-prostatectomy prostate medical images to regionally-graded histopathology images from post-prostatectomy specimens, seeking a relationship between the multi parametric imaging and cancer distribution and aggressiveness. Our approach uses image-based registration in combination with a magnetically tracked probe to orient the physical slicing of the specimen to be parallel to the in vivo imaging planes, yielding a tractable 2D registration problem. We measured a target registration error of 0.85 mm, a mean slicing plane marking error of 0.7 mm, and a mean slicing error of 0.6 mm; these results compare favourably with our 2.2 mm diagnostic MR image thickness. Qualitative evaluation of in vivo imaging-histopathology fusion reveals excellent anatomic concordance between MR and digital histopathology.
A. D. Ward, C. Crukley, C. McKenzie, J. Montreuil, E. Gibson, J. A. Gomez, M. Moussa, G. Bauman, A. Fenster

High-Throughput Prostate Cancer Gland Detection, Segmentation, and Classification from Digitized Needle Core Biopsies

We present a high-throughput computer-aided system for the segmentation and classification of glands in high resolution digitized images of needle core biopsy samples of the prostate. It will allow for rapid and accurate identification of suspicious regions on these samples. The system includes the following three modules: 1) a hierarchical frequency weighted mean shift normalized cut (HNCut) for initial detection of glands; 2) a geodesic active contour (GAC) model for gland segmentation; and 3) a diffeomorphic based similarity (DBS) feature extraction for classification of glands as benign or cancerous. HNCut is a minimally supervised color based detection scheme that combines the frequency weighted mean shift and normalized cuts algorithms to detect the lumen region of candidate glands. A GAC model, initialized using the results of HNCut, uses a color gradient based edge detection function for accurate gland segmentation. Lastly, DBS features are a set of morphometric features derived from the nonlinear dimensionality reduction of a dissimilarity metric between shape models. The system integrates these modules to enable the rapid detection, segmentation, and classification of glands on prostate biopsy images. Across 23 H & E stained prostate studies of whole-slides, 105 regions of interests (ROIs) were selected for the evaluation of segmentation and classification. The segmentation results were evaluated on 10 ROIs and compared to manual segmentation in terms of mean distance (2.6 ±0.2 pixels), overlap (62±0.07%), sensitivity (85±0.01%), specificity (94±0.003%) and positive predictive value (68±0.08%). Over 105 ROIs, the classification accuracy for glands automatically segmented was (82.5 ±9.10%) while the accuracy for glands manually segmented was (82.89 ±3.97%); no statistically significant differences were identified between the classification results.
Jun Xu, Rachel Sparks, Andrew Janowczyk, John E. Tomaszewski, Michael D. Feldman, Anant Madabhushi

Automated Analysis of PIN-4 Stained Prostate Needle Biopsies

Prostate Needle biopsies are stained with the PIN-4 marker cocktail to help the pathologist distinguish between HGPIN and adenocarcinoma. The correct interpretation of multiple IHC markers can be challenging. Therefore we propose the use of computer aided diagnosis algorithms for the identification and classification of glands in a whole slide image of prostate needle biopsy. The paper presents the different issues related to the automated analysis of prostate needle biopsies and the approach taken by BioImagene in its first generation algorithms.
Bikash Sabata, Boris Babenko, Robert Monroe, Chukka Srinivas

Augmented Reality Image Guidance in Minimally Invasive Prostatectomy

This paper presents our work aimed at providing augmented reality (AR) guidance of robot-assisted laparoscopic surgery (RALP) using the da Vinci system. There is a good clinical case for guidance due to the significant rate of complications and steep learning curve for this procedure. Patients who were due to undergo robotic prostatectomy for organ-confined prostate cancer underwent preoperative 3T MRI scans of the pelvis. These were segmented and reconstructed to form 3D images of pelvic anatomy. The reconstructed image was successfully overlaid onto screenshots of the recorded surgery post-procedure. Surgeons who perform minimally-invasive prostatectomy took part in a user-needs analysis to determine the potential benefits of an image guidance system after viewing the overlaid images. All surgeons stated that the development would be useful at key stages of the surgery and could help to improve the learning curve of the procedure and improve functional and oncological outcomes. Establishing the clinical need in this way is a vital early step in development of an AR guidance system. We have also identified relevant anatomy from preoperative MRI. Further work will be aimed at automated registration to account for tissue deformation during the procedure, using a combination of transrectal ultrasound and stereoendoscopic video.
Daniel Cohen, Erik Mayer, Dongbin Chen, Ann Anstee, Justin Vale, Guang-Zhong Yang, Ara Darzi, Philip ’Eddie’ Edwards

Texture Guided Active Appearance Model Propagation for Prostate Segmentation

Fusion of Magnetic Resonance Imaging (MRI) and Trans Rectal Ultra Sound (TRUS) images during TRUS guided prostate biopsy improves localization of the malignant tissues. Segmented prostate in TRUS and MRI improve registration accuracy and reduce computational cost of the procedure. However, accurate segmentation of the prostate in TRUS images can be a challenging task due to low signal to noise ratio, heterogeneous intensity distribution inside the prostate, and imaging artifacts like speckle noise and shadow. We propose to use texture features from approximation coefficients of Haar wavelet transform for propagation of a shape and appearance based statistical model to segment the prostate in a multi-resolution framework. A parametric model of the propagating contour is derived from Principal Component Analysis of prior shape and texture informations of the prostate from the training data. The parameters are then modified with prior knowledge of the optimization space to achieve optimal prostate segmentation. The proposed method achieves a mean Dice Similarity Coefficient value of 0.95±0.01, and mean segmentation time of 0.72±0.05 seconds when validated on 25 TRUS images, grabbed from video sequences, in a leave-one-out validation framework. Our proposed model performs computationally efficient accurate prostate segmentation in presence of intensity heterogeneity and imaging artifacts.
Soumya Ghose, Arnau Oliver, Robert Martí, Xavier Lladó, Jordi Freixenet, Joan C. Vilanova, Fabrice Meriaudeau

Novel Stochastic Framework for Accurate Segmentation of Prostate in Dynamic Contrast Enhanced MRI

Prostate segmentation is an essential step in developing any non-invasive Computer-Assisted Diagnostic (CAD) system for the early diagnosis of prostate cancer using Dynamic Contrast Enhancement Magnetic Resonance Images (DCE-MRI). In this paper we propose a novel approach for segmenting the prostate region from DCE-MRI based on using a graph cut framework to optimize a new energy function consists of three descriptors: (i) 1 st -order visual appearance descriptors of the DCE-MRI; (ii) a spatially invariant 2 nd -order homogeneity descriptor, and (iii) a prostate shape descriptor. The shape prior is learned from a subset of co-aligned training images. The visual appearances are described with marginal gray level distributions obtained by separating their mixture over the image. The spatial interactions between the prostate pixels are modeled by a 2 nd -order translation and rotation invariant Markov-Gibbs random field of object / background labels with analytically estimated potentials. Experiments with prostate DCE-MR images confirm robustness and accuracy of the proposed approach.
Ahmad Firjany, Ahmed Elnakib, Ayman El-Baz, Georgy Gimel’farb, Mohamed Abo El-Ghar, Adel Elmagharby

Boundary Delineation in Prostate Imaging Using Active Contour Segmentation Method with Interactively Defined Object Regions

Active contour methods are often methods of choice for demanding segmentation problems, yet segmentation of medical images with complex intensity patterns still remains a challenge for these methods. This paper proposes a method to incorporate interactively specified foreground/background regions into the active model framework while keeping the user interaction to the minimum. To achieve that, the proposed functional to be minimized includes a term to encourage active contour to separate the points close to the specified foreground region from the points close to the specified background region in terms of geodesic distance. The experiments on multi-modal prostate images demonstrate that the proposed method not only can achieve robust and accurate results, but also provides an efficient way to interactively improve the results.
Yan Zhang, Bogdan J. Matuszewski, Aymeric Histace, Frédéric Precioso, Judith Kilgallon, Christopher Moore


Weitere Informationen