Proc. SPIE. 9785, Medical Imaging 2016: Computer-Aided Diagnosis
KEYWORDS: Image fusion, Principal component analysis, Tumor growth modeling, Cancer, Data modeling, Tissues, Ultrasonography, Magnetic resonance imaging, Receivers, Image registration, Prostate cancer, In vivo imaging
Recently, multi-parametric Magnetic Resonance Imaging (mp-MRI) has been used to improve the sensitivity of detecting high-risk prostate cancer (PCa). Prior to biopsy, primary and secondary cancer lesions are identified on mp-MRI. The lesions are then targeted using TRUS guidance. In this paper, for the first time, we present a fused mp-MRI-temporal-ultrasound framework for characterization of PCa, in vivo. Cancer classification results obtained using temporal ultrasound are fused with those achieved using consolidated mp-MRI maps determined by multiple observers. We verify the outcome of our study using histopathology following deformable registration of ultrasound and histology images. Fusion of temporal ultrasound and mp-MRI for characterization of the PCa results in an area under the receiver operating characteristic curve (AUC) of 0.86 for cancerous regions with Gleason scores (GSs)≥3+3, and AUC of 0.89 for those with GSs≥3+4.
Temporal ultrasound has been shown to have high classification accuracy in differentiating cancer from benign tissue. In this paper, we extend the temporal ultrasound method to classify lower grade Prostate Cancer (PCa) from all other grades. We use a group of nine patients with mostly lower grade PCa, where cancerous regions are also limited. A critical challenge is to train a classifier with limited aggressive cancerous tissue compared to low grade cancerous tissue. To resolve the problem of imbalanced data, we use Synthetic Minority Oversampling Technique (SMOTE) to generate synthetic samples for the minority class. We calculate spectral features of temporal ultrasound data and perform feature selection using Random Forests. In leave-one-patient-out cross-validation strategy, an area under receiver operating characteristic curve (AUC) of 0.74 is achieved with overall sensitivity and specificity of 70%. Using an unsupervised learning approach prior to proposed method improves sensitivity and AUC to 80% and 0.79. This work represents promising results to classify lower and higher grade PCa with limited cancerous training samples, using temporal ultrasound.
In this paper, we present a registration pipeline to compensate for prostate motion and deformation during targeted freehand prostate biopsies. We perform 2D-3D registration by reconstructing a thin-volume around the real-time 2D ultrasound imaging plane. Constrained Sum of Squared Differences (SSD) and gradient descent optimization are used to rigidly align the moving volume to the fixed thin-volume. Subsequently, B-spline de- formable registration is performed to compensate for remaining non-linear deformations. SSD and zero-bounded Limited memory Broyden Fletcher Goldfarb Shannon (LBFGS) optimizer are used to find the optimum B-spline parameters. Registration results are validated on five prostate biopsy patients. Initial experiments suggest thin- volume-to-volume registration to be more effective than slice-to-volume registration. Also, a minimum consistent 2 mm improvement of Target Registration Error (TRE) is achieved following the deformable registration.
We present a parallel implementation of a statistical shape model registration to 3D ultrasound images of the
lumbar vertebrae (L2-L4). Covariance Matrix Adaptation Evolution Strategy optimization technique, along
with Linear Correlation of Linear Combination similarity metric have been used, to improve the robustness and
capture range of the registration approach. Instantiation and ultrasound simulation have been implemented on
a graphics processing unit for a faster registration. Phantom studies show a mean target registration error of 3.2
mm, while 80% of all the cases yield target registration error of below 3.5 mm.