Effect of dataset size, image quality, and image type on deep learning-based automatic prostate segmentation in 3D ultrasound.

Three-dimensional (3D) transrectal ultrasound (TRUS) is utilized in prostate cancer diagnosis and treatment, necessitating time-consuming manual prostate segmentation. We have previously developed an automatic 3D prostate segmentation algorithm involving deep learning prediction on radially sampled 2D images followed by 3D reconstruction, trained on a large, clinically diverse dataset with variable image quality. As large clinical datasets are rare, widespread adoption of automatic segmentation could be facilitated with efficient 2D-based approaches and the development of an image quality grading method. The complete training dataset of 6761 2D images, resliced from 206 3D TRUS volumes acquired using end-fire and side-fire acquisition methods, was split to train two separate networks using either end-fire or side-fire images. Split datasets were reduced to 1000, 500, 250, and 100 2D images. For deep learning prediction, modified U-Net and U-Net++ architectures were implemented and compared using an unseen test dataset of 40 3D TRUS volumes. A 3D TRUS image quality grading scale with three factors (acquisition quality, artifact severity, and boundary visibility) was developed to assess the impact on segmentation performance. For the complete training dataset, U-Net and U-Net++ networks demonstrated equivalent performance, but when trained using split end-fire/side-fire datasets, U-Net++ significantly outperformed the U-Net. Compared to the complete training datasets, U-Net++ trained using reduced-size end-fire and side-fire datasets demonstrated equivalent performance down to 500 training images. For this dataset, image quality had no impact on segmentation performance for end-fire images but did have a significant effect for side-fire images, with boundary visibility having the largest impact. Our algorithm provided fast (<1.5 s) and accurate 3D segmentations across clinically diverse images, demonstrating generalizability and efficiency when employed on smaller datasets, supporting the potential for widespread use, even when data is scarce. The development of an image quality grading scale provides a quantitative tool for assessing segmentation performance.

Automatic prostate segmentation using deep learning on clinically diverse 3D transrectal ultrasound images.

PURPOSE: Needle-based procedures for diagnosing and treating prostate cancer, such as biopsy and brachytherapy, have incorporated three-dimensional (3D) transrectal ultrasound (TRUS) imaging to improve needle guidance. Using these images effectively typically requires the physician to manually segment the prostate to define the margins used for accurate registration, targeting, and other guidance techniques. However, manual prostate segmentation is a time-consuming and difficult intraoperative process, often occurring while the patient is under sedation (biopsy) or anesthetic (brachytherapy). Minimizing procedure time with a 3D TRUS prostate segmentation method could provide physicians with a quick and accurate prostate segmentation, and allow for an efficient workflow with improved patient throughput to enable faster patient access to care. The purpose of this study was to develop a supervised deep learning-based method to segment the prostate in 3D TRUS images from different facilities, generated using multiple acquisition methods and commercial ultrasound machine models to create a generalizable algorithm for needle-based prostate cancer procedures. METHODS: Our proposed method for 3D segmentation involved prediction on two-dimensional (2D) slices sampled radially around the approximate central axis of the prostate, followed by reconstruction into a 3D surface. A 2D U-Net was modified, trained, and validated using images from 84 end-fire and 122 side-fire 3D TRUS images acquired during clinical biopsies and brachytherapy procedures. Modifications to the expansion section of the standard U-Net included the addition of 50% dropouts and the use of transpose convolutions instead of standard upsampling followed by convolution to reduce overfitting and improve performance, respectively. Manual contours provided the annotations needed for the training, validation, and testing datasets, with the testing dataset consisting of 20 end-fire and 20 side-fire unseen 3D TRUS images. Since predicting with 2D images has the potential to lose spatial and structural information, comparisons to 3D reconstruction and optimized 3D networks including 3D V-Net, Dense V-Net, and High-resolution 3D-Net were performed following an investigation into different loss functions. An extended selection of absolute and signed error metrics were computed, including pixel map comparisons [dice similarity coefficient (DSC), recall, and precision], volume percent differences (VPD), mean surface distance (MSD), and Hausdorff distance (HD), to assess 3D segmentation accuracy. RESULTS: Overall, our proposed reconstructed modified U-Net performed with a median [first quartile, third quartile] absolute DSC, recall, precision, VPD, MSD, and HD of 94.1 [92.6, 94.9]%, 96.0 [93.1, 98.5]%, 93.2 [88.8, 95.4]%, 5.78 [2.49, 11.50]%, 0.89 [0.73, 1.09] mm, and 2.89 [2.37, 4.35] mm, respectively. When compared to the best-performing optimized 3D network (i.e., 3D V-Net with a Dice plus cross-entropy loss function), our proposed method performed with a significant improvement across nearly all metrics. A computation time <0.7 s per prostate was observed, which is a sufficiently short segmentation time for intraoperative implementation. CONCLUSIONS: Our proposed algorithm was able to provide a fast and accurate 3D segmentation across variable 3D TRUS prostate images, enabling a generalizable intraoperative solution for needle-based prostate cancer procedures. This method has the potential to decrease procedure times, supporting the increasing interest in needle-based 3D TRUS approaches.

Robust 2-D-3-D Registration Optimization for Motion Compensation During 3-D TRUS-Guided Biopsy Using Learned Prostate Motion Data.

In magnetic resonance (MR)-targeted, 3-D transrectal ultrasound (TRUS)-guided biopsy, prostate motion during the procedure increases the needle targeting error and limits the ability to accurately sample MR-suspicious tumor volumes. The robustness of the 2-D-3-D registration methods for prostate motion compensation is impacted by local optima in the search space. In this paper, we analyzed the prostate motion characteristics and investigated methods to incorporate such knowledge into the registration optimization framework to improve robustness against local optima. Rigid motion of the prostate was analyzed adopting a mixture-of-Gaussian (MoG) model using 3-D TRUS images acquired at bilateral sextant probe positions with a mechanically assisted biopsy system. The learned motion characteristics were incorporated into Powell's direction set method by devising multiple initial search positions and initial search directions. Experiments were performed on data sets acquired during clinical biopsy procedures, and registration error was evaluated using target registration error (TRE) and converged image similarity metric values after optimization. After incorporating the learned initialization positions and directions in Powell's method, 2-D-3-D registration to compensate for motion during prostate biopsy was performed with rms ± std TRE of 2.33 ± 1.09 mm with ~3 s mean execution time per registration. This was an improvement over 3.12 ± 1.70 mm observed in Powell's standard approach. For the data acquired under clinical protocols, the converged image similarity metric value improved in ≥8% of the registrations whereas it degraded only ≤1% of the registrations. The reported improvements in optimization indicate useful advancements in robustness to ensure smooth clinical integration of a registration solution for motion compensation that facilitates accurate sampling of the smallest clinically significant tumors.

Postediting prostate magnetic resonance imaging segmentation consistency and operator time using manual and computer-assisted segmentation: multiobserver study.

Prostate segmentation on T2w MRI is important for several diagnostic and therapeutic procedures for prostate cancer. Manual segmentation is time-consuming, labor-intensive, and subject to high interobserver variability. This study investigated the suitability of computer-assisted segmentation algorithms for clinical translation, based on measurements of interoperator variability and measurements of the editing time required to yield clinically acceptable segmentations. A multioperator pilot study was performed under three pre- and postediting conditions: manual, semiautomatic, and automatic segmentation. We recorded the required editing time for each segmentation and measured the editing magnitude based on five different spatial metrics. We recorded average editing times of 213, 328, and 393 s for manual, semiautomatic, and automatic segmentation respectively, while an average fully manual segmentation time of 564 s was recorded. The reduced measured postediting interoperator variability of semiautomatic and automatic segmentations compared to the manual approach indicates the potential of computer-assisted segmentation for generating a clinically acceptable segmentation faster with higher consistency. The lack of strong correlation between editing time and the values of typically used error metrics ([Formula: see text]) implies that the necessary postsegmentation editing time needs to be measured directly in order to evaluate an algorithm's suitability for clinical translation.

Comparison of prostate MRI-3D transrectal ultrasound fusion biopsy for first-time and repeat biopsy patients with previous atypical small acinar proliferation.

INTRODUCTION: This study evaluates the clinical benefit of magnetic resonance-transrectal ultrasound (MR-TRUS) fusion biopsy over systematic biopsy between first-time and repeat prostate biopsy patients with prior atypical small acinar proliferation (ASAP). MATERIALS: 100 patients were enrolled in a single-centre prospective cohort study: 50 for first biopsy, 50 for repeat biopsy with prior ASAP. Multiparameteric magnetic resonance imaging (MP-MRI) and standard 12-core ultrasound biopsy (Std-Bx) were performed on all patients. Targeted biopsy using MRI-TRUS fusion (Fn-Bx) was performed f suspicious lesions were identified on the pre-biopsy MP-MRI. Classification of clinically significant disease was assessed independently for the Std-Bx vs. Fn-Bx cores to compare the two approaches. RESULTS: Adenocarcinoma was detected in 49/100 patients (26 first biopsy, 23 ASAP biopsy), with 25 having significant disease (17 first, 8 ASAP). Fn-Bx demonstrated significantly higher per-core cancer detection rates, cancer involvement, and Gleason scores for first-time and ASAP patients. However, Fn-Bx was significantly more likely to detect significant cancer missed on Std-Bx for ASAP patients than first-time biopsy patients. The addition of Fn-Bx to Std-Bx for ASAP patients had a 166.7% relative risk reduction for missing Gleason ≥ 3 + 4 disease (number needed to image with MP-MRI=10 patients) compared to 6.3% for first biopsy (number to image=50 patients). Negative predictive value of MP-MRI for negative biopsy was 79% for first-time and 100% for ASAP patients, with median followup of 32.1 ± 15.5 months. CONCLUSIONS: MR-TRUS Fn-Bx has a greater clinical impact for repeat biopsy patients with prior ASAP than biopsy-naïve patients by detecting more significant cancers that are missed on Std-Bx.

Toward Prostate Cancer Contouring Guidelines on Magnetic Resonance Imaging: Dominant Lesion Gross and Clinical Target Volume Coverage Via Accurate Histology Fusion.

PURPOSE: Defining prostate cancer (PCa) lesion clinical target volumes (CTVs) for multiparametric magnetic resonance imaging (mpMRI) could support focal boosting or treatment to improve outcomes or lower morbidity, necessitating appropriate CTV margins for mpMRI-defined gross tumor volumes (GTVs). This study aimed to identify CTV margins yielding 95% coverage of PCa tumors for prospective cases with high likelihood. METHODS AND MATERIALS: Twenty-five men with biopsy-confirmed clinical stage T1 or T2 PCa underwent pre-prostatectomy mpMRI, yielding T2-weighted, dynamic contrast-enhanced, and apparent diffusion coefficient images. Digitized whole-mount histology was contoured and registered to mpMRI scans (error ≤2 mm). Four observers contoured lesion GTVs on each mpMRI scan. CTVs were defined by isotropic and anisotropic expansion from these GTVs and from multiparametric (unioned) GTVs from 2 to 3 scans. Histologic coverage (proportions of tumor area on co-registered histology inside the CTV, measured for Gleason scores [GSs] ≥6 and ≥7) and prostate sparing (proportions of prostate volume outside the CTV) were measured. Nonparametric histologic-coverage prediction intervals defined minimal margins yielding 95% coverage for prospective cases with 78% to 92% likelihood. RESULTS: On analysis of 72 true-positive tumor detections, 95% coverage margins were 9 to 11 mm (GS ≥ 6) and 8 to 10 mm (GS ≥ 7) for single-sequence GTVs and were 8 mm (GS ≥ 6) and 6 mm (GS ≥ 7) for 3-sequence GTVs, yielding CTVs that spared 47% to 81% of prostate tissue for the majority of tumors. Inclusion of T2-weighted contours increased sparing for multiparametric CTVs with 95% coverage margins for GS ≥6, and inclusion of dynamic contrast-enhanced contours increased sparing for GS ≥7. Anisotropic 95% coverage margins increased the sparing proportions to 71% to 86%. CONCLUSIONS: Multiparametric magnetic resonance imaging-defined GTVs expanded by appropriate margins may support focal boosting or treatment of PCa; however, these margins, accounting for interobserver and intertumoral variability, may preclude highly conformal CTVs. Multiparametric GTVs and anisotropic margins may reduce the required margins and improve prostate sparing.

Magnetic Resonance Imaging-Guided Transurethral Ultrasound Ablation of Prostate Tissue in Patients with Localized Prostate Cancer: A Prospective Phase 1 Clinical Trial.

BACKGROUND: Magnetic resonance imaging-guided transurethral ultrasound ablation (MRI-TULSA) is a novel minimally invasive technology for ablating prostate tissue, potentially offering good disease control of localized cancer and low morbidity. OBJECTIVE: To determine the clinical safety and feasibility of MRI-TULSA for whole-gland prostate ablation in a primary treatment setting of localized prostate cancer (PCa). DESIGN, SETTING, AND PARTICIPANTS: A single-arm prospective phase 1 study was performed at three tertiary referral centers in Canada, Germany, and the United States. Thirty patients (median age: 69 yr; interquartile range [IQR]: 67-71 yr) with biopsy-proven low-risk (80%) and intermediate-risk (20%) PCa were treated and followed for 12 mo. INTERVENTION: MRI-TULSA treatment was delivered with the therapeutic intent of conservative whole-gland ablation including 3-mm safety margins and 10% residual viable prostate expected around the capsule. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Primary end points were safety (adverse events) and feasibility (technical accuracy and precision of conformal thermal ablation). Exploratory outcomes included quality of life, prostate-specific antigen (PSA), and biopsy at 12 mo. RESULTS AND LIMITATIONS: Median treatment time was 36min (IQR: 26-44) and prostate volume was 44ml (IQR: 38-48). Spatial control of thermal ablation was ±1.3mm on MRI thermometry. Common Terminology Criteria for Adverse Events included hematuria (43% grade [G] 1; 6.7% G2), urinary tract infections (33% G2), acute urinary retention (10% G1; 17% G2), and epididymitis (3.3% G3). There were no rectal injuries. Median pretreatment International Prostate Symptom Score 8 (IQR: 5-13) returned to 6 (IQR: 4-10) at 3 mo (mean change: -2; 95% confidence interval [CI], -4 to 1). Median pretreatment International Index of Erectile Function 13 (IQR: 6-28) recovered to 13 (IQR: 5-25) at 12 mo (mean change: -1; 95% CI, -5 to 3). Median PSA decreased 87% at 1 mo and was stable at 0.8 ng/ml (IQR: 0.6-1.1) to 12 mo. Positive biopsies showed 61% reduction in total cancer length, clinically significant disease in 9 of 29 patients (31%; 95% CI, 15-51), and any disease in 16 of 29 patients (55%; 95% CI, 36-74). CONCLUSIONS: MRI-TULSA was feasible, safe, and technically precise for whole-gland prostate ablation in patients with localized PCa. Phase 1 data are sufficiently compelling to study MRI-TULSA further in a larger prospective trial with reduced safety margins. PATIENT SUMMARY: We used magnetic resonance imaging-guided transurethral ultrasound to heat and ablate the prostate in men with prostate cancer. We showed that the treatment can be targeted within a narrow range (1mm) and has a well-tolerated side effect profile. A larger study is under way. TRIAL REGISTRATION: NCT01686958, DRKS00005311.

Three-dimensional nonrigid landmark-based magnetic resonance to transrectal ultrasound registration for image-guided prostate biopsy.

Registration of three-dimensional (3-D) magnetic resonance (MR) to 3-D transrectal ultrasound (TRUS) prostate images is an important step in the planning and guidance of 3-D TRUS guided prostate biopsy. In order to accurately and efficiently perform the registration, a nonrigid landmark-based registration method is required to account for the different deformations of the prostate when using these two modalities. We describe a nonrigid landmark-based method for registration of 3-D TRUS to MR prostate images. The landmark-based registration method first makes use of an initial rigid registration of 3-D MR to 3-D TRUS images using six manually placed approximately corresponding landmarks in each image. Following manual initialization, the two prostate surfaces are segmented from 3-D MR and TRUS images and then nonrigidly registered using the following steps: (1) rotationally reslicing corresponding segmented prostate surfaces from both 3-D MR and TRUS images around a specified axis, (2) an approach to find point correspondences on the surfaces of the segmented surfaces, and (3) deformation of the surface of the prostate in the MR image to match the surface of the prostate in the 3-D TRUS image and the interior using a thin-plate spline algorithm. The registration accuracy was evaluated using 17 patient prostate MR and 3-D TRUS images by measuring the target registration error (TRE). Experimental results showed that the proposed method yielded an overall mean TRE of [Formula: see text] for the rigid registration and [Formula: see text] for the nonrigid registration, which is favorably comparable to a clinical requirement for an error of less than 2.5 mm. A landmark-based nonrigid 3-D MR-TRUS registration approach is proposed, which takes into account the correspondences on the prostate surface, inside the prostate, as well as the centroid of the prostate. Experimental results indicate that the proposed method yields clinically sufficient accuracy.

Statistical Biomechanical Surface Registration: Application to MR-TRUS Fusion for Prostate Interventions.

A common challenge when performing surface-based registration of images is ensuring that the surfaces accurately represent consistent anatomical boundaries. Image segmentation may be difficult in some regions due to either poor contrast, low slice resolution, or tissue ambiguities. To address this, we present a novel non-rigid surface registration method designed to register two partial surfaces, capable of ignoring regions where the anatomical boundary is unclear. Our probabilistic approach incorporates prior geometric information in the form of a statistical shape model (SSM), and physical knowledge in the form of a finite element model (FEM). We validate results in the context of prostate interventions by registering pre-operative magnetic resonance imaging (MRI) to 3D transrectal ultrasound (TRUS). We show that both the geometric and physical priors significantly decrease net target registration error (TRE), leading to TREs of 2.35 ± 0.81 mm and 2.81 ± 0.66 mm when applied to full and partial surfaces, respectively. We investigate robustness in response to errors in segmentation, varying levels of missing data, and adjusting the tunable parameters. Results demonstrate that the proposed surface registration method is an efficient, robust, and effective solution for fusing data from multiple modalities.

Computer-Aided Prostate Cancer Detection Using Ultrasound RF Time Series: In Vivo Feasibility Study.

UNLABELLED: This paper presents the results of a computer-aided intervention solution to demonstrate the application of RF time series for characterization of prostate cancer, in vivo. METHODS: We pre-process RF time series features extracted from 14 patients using hierarchical clustering to remove possible outliers. Then, we demonstrate that the mean central frequency and wavelet features extracted from a group of patients can be used to build a nonlinear classifier which can be applied successfully to differentiate between cancerous and normal tissue regions of an unseen patient. RESULTS: In a cross-validation strategy, we show an average area under receiver operating characteristic curve (AUC) of 0.93 and classification accuracy of 80%. To validate our results, we present a detailed ultrasound to histology registration framework. CONCLUSION: Ultrasound RF time series results in differentiation of cancerous and normal tissue with high AUC.

Ultrasound-Based Characterization of Prostate Cancer Using Joint Independent Component Analysis.

OBJECTIVE: This paper presents the results of a new approach for selection of RF time series features based on joint independent component analysis for in vivo characterization of prostate cancer. METHODS: We project three sets of RF time series features extracted from the spectrum, fractal dimension, and the wavelet transform of the ultrasound RF data on a space spanned by five joint independent components. Then, we demonstrate that the obtained mixing coefficients from a group of patients can be used to train a classifier, which can be applied to characterize cancerous regions of a test patient. RESULTS: In a leave-one-patient-out cross validation, an area under receiver operating characteristic curve of 0.93 and classification accuracy of 84% are achieved. CONCLUSION: Ultrasound RF time series can be used to accurately characterize prostate cancer, in vivo without the need for exhaustive search in the feature space. SIGNIFICANCE: We use joint independent component analysis for systematic fusion of multiple sets of RF time series features, within a machine learning framework, to characterize PCa in an in vivo study.

Unshielded asymmetric transmit-only and endorectal receive-only radiofrequency coil for (23) Na MRI of the prostate at 3 tesla.

BACKGROUND: To develop and optimize radiofrequency (RF) hardware for the detection of endogenous sodium ((23) Na) by 3.0 Tesla (T) MRI in the human prostate. METHODS: A transmit-only receive-only (TORO) RF system of resonators consisting of an unshielded, asymmetric, quadrature birdcage (transmit), and an endorectal (ER), linear, surface (receive) coil were developed and tested on a 3T MRI scanner. Two different ER receivers were constructed; a single-tuned ((23) Na) and a dual-tuned ((1) H/(23) Na). Both receivers were evaluated by the measurements of signal-to-noise ratio (SNR) and B1 homogeneity. For tissue sodium concentration (TSC) quantification, vials containing known sodium concentrations were incorporated into the ER. The system was used to measure the prostate TSC of three men (age 55 ± 5 years) with biopsy-proven prostate cancer. RESULTS: B1 field inhomogeneity of the asymmetric transmitter was estimated to be less than 5%. The mean SNR measured in a region of interest within the prostate using the single-tuned ER coil was 54.0 ± 4.6. The mean TSC in the central gland was 60.2 ± 5.7 mmol/L and in the peripheral gland was 70.5 ± 9.0 mmol/L. CONCLUSION: A TORO system was developed and optimized for (23) Na MRI of the human prostate which showed good sensitivity throughout the prostate for quantitative measurement of TSC.

Three-Dimensional Nonrigid MR-TRUS Registration Using Dual Optimization.

In this study, we proposed an efficient nonrigid magnetic resonance (MR) to transrectal ultrasound (TRUS) deformable registration method in order to improve the accuracy of targeting suspicious regions during a three dimensional (3-D) TRUS guided prostate biopsy. The proposed deformable registration approach employs the multi-channel modality independent neighborhood descriptor (MIND) as the local similarity feature across the two modalities of MR and TRUS, and a novel and efficient duality-based convex optimization-based algorithmic scheme was introduced to extract the deformations and align the two MIND descriptors. The registration accuracy was evaluated using 20 patient images by calculating the TRE using manually identified corresponding intrinsic fiducials in the whole gland and peripheral zone. Additional performance metrics [Dice similarity coefficient (DSC), mean absolute surface distance (MAD), and maximum absolute surface distance (MAXD)] were also calculated by comparing the MR and TRUS manually segmented prostate surfaces in the registered images. Experimental results showed that the proposed method yielded an overall median TRE of 1.76 mm. The results obtained in terms of DSC showed an average of 80.8±7.8% for the apex of the prostate, 92.0±3.4% for the mid-gland, 81.7±6.4% for the base and 85.7±4.7% for the whole gland. The surface distance calculations showed an overall average of 1.84±0.52 mm for MAD and 6.90±2.07 mm for MAXD.

Evaluation of MRI-TRUS fusion versus cognitive registration accuracy for MRI-targeted, TRUS-guided prostate biopsy.

OBJECTIVE: The purpose of this article is to compare transrectal ultrasound (TRUS) biopsy accuracies of operators with different levels of prostate MRI experience using cognitive registration versus MRI-TRUS fusion to assess the preferred method of TRUS prostate biopsy for MRI-identified lesions. SUBJECTS AND METHODS; One hundred patients from a prospective prostate MRI-TRUS fusion biopsy study were reviewed to identify all patients with clinically significant prostate adenocarcinoma (PCA) detected on MRI-targeted biopsy. Twenty-five PCA tumors were incorporated into a validated TRUS prostate biopsy simulator. Three prostate biopsy experts, each with different levels of experience in prostate MRI and MRI-TRUS fusion biopsy, performed a total of 225 simulated targeted biopsies on the MRI lesions as well as regional biopsy targets. Simulated biopsies performed using cognitive registration with 2D TRUS and 3D TRUS were compared with biopsies performed under MRI-TRUS fusion. RESULTS: Two-dimensional and 3D TRUS sampled only 48% and 45% of clinically significant PCA MRI lesions, respectively, compared with 100% with MRI-TRUS fusion. Lesion sampling accuracy did not statistically significantly vary according to operator experience or tumor volume. MRI-TRUS fusion-naïve operators showed consistent errors in targeting of the apex, midgland, and anterior targets, suggesting that there is biased error in cognitive registration. The MRI-TRUS fusion expert correctly targeted the prostate apex; however, his midgland and anterior mistargeting was similar to that of the less-experienced operators. CONCLUSION: MRI-targeted TRUS-guided prostate biopsy using cognitive registration appears to be inferior to MRI-TRUS fusion, with fewer than 50% of clinically significant PCA lesions successfully sampled. No statistically significant difference in biopsy accuracy was seen according to operator experience with prostate MRI or MRI-TRUS fusion.

Spatially varying accuracy and reproducibility of prostate segmentation in magnetic resonance images using manual and semiautomated methods.

PURPOSE: Three-dimensional (3D) prostate image segmentation is useful for cancer diagnosis and therapy guidance, but can be time-consuming to perform manually and involves varying levels of difficulty and interoperator variability within the prostatic base, midgland (MG), and apex. In this study, the authors measured accuracy and interobserver variability in the segmentation of the prostate on T2-weighted endorectal magnetic resonance (MR) imaging within the whole gland (WG), and separately within the apex, midgland, and base regions. METHODS: The authors collected MR images from 42 prostate cancer patients. Prostate border delineation was performed manually by one observer on all images and by two other observers on a subset of ten images. The authors used complementary boundary-, region-, and volume-based metrics [mean absolute distance (MAD), Dice similarity coefficient (DSC), recall rate, precision rate, and volume difference (ΔV)] to elucidate the different types of segmentation errors that they observed. Evaluation for expert manual and semiautomatic segmentation approaches was carried out. Compared to manual segmentation, the authors' semiautomatic approach reduces the necessary user interaction by only requiring an indication of the anteroposterior orientation of the prostate and the selection of prostate center points on the apex, base, and midgland slices. Based on these inputs, the algorithm identifies candidate prostate boundary points using learned boundary appearance characteristics and performs regularization based on learned prostate shape information. RESULTS: The semiautomated algorithm required an average of 30 s of user interaction time (measured for nine operators) for each 3D prostate segmentation. The authors compared the segmentations from this method to manual segmentations in a single-operator (mean whole gland MAD = 2.0 mm, DSC = 82%, recall = 77%, precision = 88%, and ΔV = - 4.6 cm(3)) and multioperator study (mean whole gland MAD = 2.2 mm, DSC = 77%, recall = 72%, precision = 86%, and ΔV = - 4.0 cm(3)). These results compared favorably with observed differences between manual segmentations and a simultaneous truth and performance level estimation reference for this data set (whole gland differences as high as MAD = 3.1 mm, DSC = 78%, recall = 66%, precision = 77%, and ΔV = 15.5 cm(3)). The authors found that overall, midgland segmentation was more accurate and repeatable than the segmentation of the apex and base, with the base posing the greatest challenge. CONCLUSIONS: The main conclusions of this study were that (1) the semiautomated approach reduced interobserver segmentation variability; (2) the segmentation accuracy of the semiautomated approach, as well as the accuracies of recently published methods from other groups, were within the range of observed expert variability in manual prostate segmentation; and (3) further efforts in the development of computer-assisted segmentation would be most productive if focused on improvement of segmentation accuracy and reduction of variability within the prostatic apex and base.

Evaluating the utility of intraprocedural 3D TRUS image information in guiding registration for displacement compensation during prostate biopsy.

PURPOSE: In targeted 3D transrectal ultrasound (TRUS)-guided biopsy, patient and prostate movement during the procedure can cause target misalignments that hinder accurate sampling of preplanned suspicious tissue locations. Multiple solutions have been proposed for displacement compensation via registration of intraprocedural TRUS images to a baseline 3D TRUS image acquired at the beginning of the biopsy procedure. While 2D TRUS images are widely used for intraprocedural guidance, some solutions utilize richer intraprocedural images such as bi- or multiplanar TRUS or 3D TRUS, acquired by specialized probes. In this work, the impact of such richer intraprocedural imaging on displacement compensation accuracy was measured to evaluate the tradeoff between cost and complexity of intraprocedural imaging versus improved displacement compensation. METHODS: Baseline and intraprocedural 3D TRUS images were acquired from 29 patients at standard sextant-template biopsy locations. Planes extracted from 3D TRUS images acquired at sextant positions were used to simulate 2D and 3D intraprocedural information available in different potential clinically relevant scenarios for co-registration with the baseline 3D TRUS image. In practice, intraprocedural 3D information can be acquired either via the use of specialized ultrasound probes (e.g., multiplanar or 3D probes) or via axial rotation of a tracked 2D TRUS probe. Registration accuracy was evaluated by calculating the target registration error (TRE) using manually identified homologous intrinsic fiducial markers (microcalcifications). The TRE was analyzed separately at the base, mid-gland and apex regions of the prostate. RESULTS: The results indicate that TRE improved gradually as the number of intraprocedural imaging planes used in registration was increased, implying that 3D TRUS information assisted the registration algorithm to robustly converge to more accurate solutions. The acquisition of a partial volume up to the angle of rotation supported more accurate displacement compensation than acquiring biplane configurations. Additional intraprocedural 3D TRUS image information was more beneficial to registration accuracy in the base and apex regions as compared with the mid-gland region. CONCLUSIONS: While the majority of the registrations using 2D TRUS images provided a clinically desired level of accuracy, intraprocedural 3D imaging helped improve the overall registration accuracy and robustness, especially in the base and apex regions of the prostate. These results are helpful for devising image-based registration methods for displacement compensation when designing 3D TRUS-guided biopsy systems.

Magnetic resonance imaging-targeted, 3D transrectal ultrasound-guided fusion biopsy for prostate cancer: Quantifying the impact of needle delivery error on diagnosis.

PURPOSE: Magnetic resonance imaging (MRI)-targeted, 3D transrectal ultrasound (TRUS)-guided "fusion" prostate biopsy intends to reduce the ∼23% false negative rate of clinical two-dimensional TRUS-guided sextant biopsy. Although it has been reported to double the positive yield, MRI-targeted biopsies continue to yield false negatives. Therefore, the authors propose to investigate how biopsy system needle delivery error affects the probability of sampling each tumor, by accounting for uncertainties due to guidance system error, image registration error, and irregular tumor shapes. METHODS: T2-weighted, dynamic contrast-enhanced T1-weighted, and diffusion-weighted prostate MRI and 3D TRUS images were obtained from 49 patients. A radiologist and radiology resident contoured 81 suspicious regions, yielding 3D tumor surfaces that were registered to the 3D TRUS images using an iterative closest point prostate surface-based method to yield 3D binary images of the suspicious regions in the TRUS context. The probabilityP of obtaining a sample of tumor tissue in one biopsy core was calculated by integrating a 3D Gaussian distribution over each suspicious region domain. Next, the authors performed an exhaustive search to determine the maximum root mean squared error (RMSE, in mm) of a biopsy system that gives P ≥ 95% for each tumor sample, and then repeated this procedure for equal-volume spheres corresponding to each tumor sample. Finally, the authors investigated the effect of probe-axis-direction error on measured tumor burden by studying the relationship between the error and estimated percentage of core involvement. RESULTS: Given a 3.5 mm RMSE for contemporary fusion biopsy systems,P ≥ 95% for 21 out of 81 tumors. The authors determined that for a biopsy system with 3.5 mm RMSE, one cannot expect to sample tumors of approximately 1 cm(3) or smaller with 95% probability with only one biopsy core. The predicted maximum RMSE giving P ≥ 95% for each tumor was consistently greater when using spherical tumor shapes as opposed to no shape assumption. However, an assumption of spherical tumor shape for RMSE = 3.5 mm led to a mean overestimation of tumor sampling probabilities of 3%, implying that assuming spherical tumor shape may be reasonable for many prostate tumors. The authors also determined that a biopsy system would need to have a RMS needle delivery error of no more than 1.6 mm in order to sample 95% of tumors with one core. The authors' experiments also indicated that the effect of axial-direction error on the measured tumor burden was mitigated by the 18 mm core length at 3.5 mm RMSE. CONCLUSIONS: For biopsy systems with RMSE ≥ 3.5 mm, more than one biopsy core must be taken from the majority of tumors to achieveP ≥ 95%. These observations support the authors' perspective that some tumors of clinically significant sizes may require more than one biopsy attempt in order to be sampled during the first biopsy session. This motivates the authors' ongoing development of an approach to optimize biopsy plans with the aim of achieving a desired probability of obtaining a sample from each tumor, while minimizing the number of biopsies. Optimized planning of within-tumor targets for MRI-3D TRUS fusion biopsy could support earlier diagnosis of prostate cancer while it remains localized to the gland and curable.

3D ultrasound imaging in image-guided intervention.

Ultrasound imaging is used extensively in diagnosis and image-guidance for interventions of human diseases. However, conventional 2D ultrasound suffers from limitations since it can only provide 2D images of 3-dimensional structures in the body. Thus, measurement of organ size is variable, and guidance of interventions is limited, as the physician is required to mentally reconstruct the 3-dimensional anatomy using 2D views. Over the past 20 years, a number of 3-dimensional ultrasound imaging approaches have been developed. We have developed an approach that is based on a mechanical mechanism to move any conventional ultrasound transducer while 2D images are collected rapidly and reconstructed into a 3D image. In this presentation, 3D ultrasound imaging approaches will be described for use in image-guided interventions.

2D-3D rigid registration to compensate for prostate motion during 3D TRUS-guided biopsy.

PURPOSE: Three-dimensional (3D) transrectal ultrasound (TRUS)-guided systems have been developed to improve targeting accuracy during prostate biopsy. However, prostate motion during the procedure is a potential source of error that can cause target misalignments. The authors present an image-based registration technique to compensate for prostate motion by registering the live two-dimensional (2D) TRUS images acquired during the biopsy procedure to a preacquired 3D TRUS image. The registration must be performed both accurately and quickly in order to be useful during the clinical procedure. METHODS: The authors implemented an intensity-based 2D-3D rigid registration algorithm optimizing the normalized cross-correlation (NCC) metric using Powell's method. The 2D TRUS images acquired during the procedure prior to biopsy gun firing were registered to the baseline 3D TRUS image acquired at the beginning of the procedure. The accuracy was measured by calculating the target registration error (TRE) using manually identified fiducials within the prostate; these fiducials were used for validation only and were not provided as inputs to the registration algorithm. They also evaluated the accuracy when the registrations were performed continuously throughout the biopsy by acquiring and registering live 2D TRUS images every second. This measured the improvement in accuracy resulting from performing the registration, continuously compensating for motion during the procedure. To further validate the method using a more challenging data set, registrations were performed using 3D TRUS images acquired by intentionally exerting different levels of ultrasound probe pressures in order to measure the performance of our algorithm when the prostate tissue was intentionally deformed. In this data set, biopsy scenarios were simulated by extracting 2D frames from the 3D TRUS images and registering them to the baseline 3D image. A graphics processing unit (GPU)-based implementation was used to improve the registration speed. They also studied the correlation between NCC and TREs. RESULTS: The root-mean-square (RMS) TRE of registrations performed prior to biopsy gun firing was found to be 1.87 ± 0.81 mm. This was an improvement over 4.75 ± 2.62 mm before registration. When the registrations were performed every second during the biopsy, the RMS TRE was reduced to 1.63 ± 0.51 mm. For 3D data sets acquired under different probe pressures, the RMS TRE was found to be 3.18 ± 1.6 mm. This was an improvement from 6.89 ± 4.1 mm before registration. With the GPU based implementation, the registrations were performed with a mean time of 1.1 s. The TRE showed a weak correlation with the similarity metric. However, the authors measured a generally convex shape of the metric around the ground truth, which may explain the rapid convergence of their algorithm to accurate results. CONCLUSIONS: Registration to compensate for prostate motion during 3D TRUS-guided biopsy can be performed with a measured accuracy of less than 2 mm and a speed of 1.1 s, which is an important step toward improving the targeting accuracy of a 3D TRUS-guided biopsy system.