Vaishali Karnik

Assessment of image registration accuracy in three-dimensional transrectal ultrasound guided prostate biopsy

PURPOSE Prostate biopsy, performed using two-dimensional (2D) transrectal ultrasound (TRUS) guidance, is the clinical standard for a definitive diagnosis of prostate cancer. Histological analysis of the biopsies can reveal cancerous, noncancerous, or suspicious, possibly precancerous, tissue. During subsequent biopsy sessions, noncancerous regions should be avoided, and suspicious regions should be precisely rebiopsied, requiring accurate needle guidance. It is challenging to precisely guide a needle using 2D TRUS due to the limited anatomic information provided, and a three-dimensional (3D) record of biopsy locations for use in subsequent biopsy procedures cannot be collected. Our tracked, 3D TRUS-guided prostate biopsy system provides additional anatomic context and permits a 3D record of biopsies. However, targets determined based on a previous biopsy procedure must be transformed during the procedure to compensate for intraprocedure prostate shifting due to patient motion and prostate deformation due to transducer probe pressure. Thus, registration is a critically important step required to determine these transformations so that correspondence is maintained between the prebiopsied image and the real-time image. Registration must not only be performed accurately, but also quickly, since correction for prostate motion and deformation must be carried out during the biopsy procedure. The authors evaluated the accuracy, variability, and speed of several surface-based and image-based intrasession 3D-to-3D TRUS image registration techniques, for both rigid and nonrigid cases, to find the required transformations. METHODS Our surface-based rigid and nonrigid registrations of the prostate were performed using the iterative-closest-point algorithm and a thin-plate spline algorithm, respectively. For image-based rigid registration, the authors used a block matching approach, and for nonrigid registration, the authors define the moving image deformation using a regular, 3D grid of B-spline control points. The authors measured the target registration error (TRE) as the postregistration misalignment of 60 manually marked, corresponding intrinsic fiducials. The authors also measured the fiducial localization error (FLE), the effect of segmentation variability, and the effect of fiducial distance from the transducer probe tip. Lastly, the authors performed 3D principal component analysis (PCA) on the x, y, and z components of the TREs to examine the 95% confidence ellipsoids describing the errors for each registration method. RESULTS Using surface-based registration, the authors found mean TREs of 2.13 +/- 0.80 and 2.09 +/- 0.77 mm for rigid and nonrigid techniques, respectively. Using image-based rigid and non-rigid registration, the authors found mean TREs of 1.74 +/- 0.84 and 1.50 +/- 0.83 mm, respectively. Our FLE was 0.21 mm and did not dominate the overall TRE. However, segmentation variability contributed substantially approximately50%) to the TRE of the surface-based techniques. PCA showed that the 95% confidence ellipsoid encompassing fiducial distances between the source and target registra- tion images was reduced from 3.05 to 0.14 cm3, and 0.05 cm3 for the surface-based and image-based techniques, respectively. The run times for both registration methods were comparable at less than 60 s. CONCLUSIONS Our results compare favorably with a clinical need for a TRE of less than 2.5 mm, and suggest that image-based registration is superior to surface-based registration for 3D TRUS-guided prostate biopsies, since it does not require segmentation.

Evaluation of intersession 3D-TRUS to 3D-TRUS image registration for repeat prostate biopsies.

PURPOSE 3D-TRUS-guided prostate biopsy permits a 3D record of biopsy cores, supporting the planning of targets to resample or avoid during repeat biopsy sessions. Image registration is required in order to map biopsy targets planned on a previous sessions 3D-TRUS image into the context of the current session. The authors evaluated the performance of surface- and intensity-based rigid and nonrigid registration algorithms for this task using a clinically motivated success criterion of a maximum 2.5 mm target registration error (TRE). METHODS The authors collected two 3D-TRUS images for each of 13 patients, where each image was collected in a separate biopsy session, and the sessions were 1 week apart. The authors tested the iterative closest point and thin-plate spline surface-based registration methods, and the block matching and B-spline intensity-based methods. Manually marked intrinsic fiducials (calcifications) were used to calculate a TRE for each of the tested methods. In addition, error ellipsoids, anisotropy, and variability due to image segmentation were analyzed. All analysis was performed separately for the peripheral zone since this area harbors up to 80% of all prostate cancer. RESULTS Only the intensity-based nonrigid registration method met the success criterion for both the whole gland and the peripheral zone. Segmentation was a substantial contributor to registration error variability for the surface-based methods, and the surface-based methods resulted in greater error volumes and anisotropy. CONCLUSIONS Intensity-based rigid registration is clinically sufficient to register regions outside the peripheral zone, but nonrigid registration is required in order to register the peripheral zone with clinically needed accuracy. The clinical advantage of using nonrigid registration is questionable since the difference between the RMS TREs for rigid and nonrigid intensity-based registration could be considered to be small (0.3 mm) and is statistically significant. If the added clinical value in performing a nonrigid registration is insufficient given the additional time required for this computation, rigid registration alone may be suitable.

Fusion of MRI to 3D TRUS for mechanically-assisted targeted prostate biopsy: system design and initial clinical experience

This paper presents a mechanically-assisted 3D transrectal ultrasound (TRUS) biopsy system with MRI-3D TRUS fusion. The 3D TRUS system employs a 4 degree-of-freedom linkage for real-time TRUS probe tracking. MRI-TRUS fusion is achieved using a surface-based nonlinear registration incorporating thin-plate splines to provide real-time overlays of suspicious MRI lesions on 3D TRUS for intrabiopsy targeted needle guidance. Clinical use of the system is demonstrated on a prospective cohort study of 25 patients with clinical findings concerning for prostate adenocarcinoma (PCa). The MRI-3D TRUS registration accuracy is quantified and compared with alternative algorithms for optimal performance. Results of the clinical study demonstrated a significantly higher rate of positive biopsy cores and a higher Gleason score cancer grading for targeted biopsies using the MRI-3D TRUS fusion as compared to the standard 12-core sextant biopsy distribution. Lesion targeted biopsy cores that were positive for PCa contained a significantly higher percentage of tumor within each biopsy sample compared to the sextant cores and in some patients resulted in identifying higher risk disease.

Perfusion of Renal Allografts with Verapamil Improves Graft Function

The effect of adding a calcium channel antagonist to kidney allograft perfusate solution was assessed. All renal transplants in which both kidneys from the same donor used for transplantation were studied between November, 2003 and August, 2005 (n=46). The first renal allograft was perfused on the backtable with 1 L of histidine-tryptophan-ketoglurate solution and the second with 1 L of histidine-tryptophan-ketoglurate with 5 mg/L of verapamil. Both organs were transplanted in the usual manner. Baseline demographic parameters were similar between first and second kidney recipients other than BMI and cold ischemic time. At 6 and 12 months, renal function was significantly improved in the verapamil versus control cohort (creatinine clearance 73.8±23.5 mL/min vs. 55.8±17.0 mL/min, P<0.05 and 87.5±28.4 mL/min vs. 59.7±21.3 mL/min, P<0.05 respectively). Additionally, rates of hypotension during graft reperfusion and other adverse reactions were similar in both groups. In conclusion, verapamil supplemented perfusate significantly improved renal function posttransplantion.

SU‐E‐U‐04: MRI‐Targeted, 3D TRUS‐Guided Prostate Biopsy: Measurement of Inter‐Modality Prostate Deformation and Rigid Registration Algorithm Performance

Purpose: To measure the target registration error (TRE) of two approaches for the rigid registration of magnetic resonance(MR) prostate images to 3D transrectal ultrasound(TRUS)images, in order to improve biopsy planning and guidance for prostate cancer diagnosis. Methods: Six T2‐weighted MRimages and six 3D‐TRUS images from six different subjects were acquired during an MRI‐targeted, 3D TRUS‐guided biopsy procedure. All image pairs had identifiable calcifications and/or cysts which were used as fiducials for registration validation. A total of 78 fiducials were identified and all images were segmented using a semi‐automated method. The MRimages were designated as the source images, and the 3D‐TRUS images were designated as the target images. The source images were rigidly registered to the corresponding target images using 3 different methods. The first method was a visual alignment (VA) performed interactively by the operator, the second was a surface‐based iterative‐closest‐point (ICP) method, and the third was least‐squares best fit (LSF) transformation based on a set of identified homologous landmark pairs. We calculated the RMS TREs for each method and compared the results using a two‐tailed paired t‐test for each combination. Results: The RMS TREs for VA, ICP and LSF were 5.07mm, 5.87mm and 3.10mm, respectively (p 2.5mm). This is likely due to the substantial shape distortion of the prostate, caused mainly by transducer pressure in the 3D‐TRUS images, and the pressure induced by the endorectal coil in the MRimages. These preliminary results indicate that, at most, an ideal rigid registration algorithm based on the operators initialization (i.e. VA) would able to reduce the TRE by 1.97mm. The remaining error of 3.10mm would require a non‐rigid registration to correct for the deformation.

Co-registration Framework for Histology-registration-based Validation of Fused Multimodality Prostate Cancer Imaging

The correlation of multiple in vivo imaging modalities with graded whole-mount histology ideally requires co-registration with appropriately quantified accuracy. This work presents a framework for the co-registration of multiple in vivo prostate images with histology as a composition of pair wise registrations. This work also proposes a methodology for the measurement of the accuracy of these registrations based on a composition of target registration errors calculated from homologous landmarks in registered pairs of images. Preliminary results from the implementation of components of the framework are reported.

Mechanically Assisted 3D Prostate Ultrasound Imaging and Biopsy Needle-Guidance System

Prostate biopsy procedures are currently limited to using 2D transrectal ultrasound (TRUS) imaging to guide the biopsy needle. Being limited to 2D causes ambiguity in needle guidance and provides an insufficient record to allow guidance to the same suspicious locations or avoid regions that are negative during previous biopsy sessions. We have developed a mechanically assisted 3D ultrasound imaging and needle tracking system, which supports a commercially available TRUS probe and integrated needle guide for prostate biopsy. The mechanical device is fixed to a cart and the mechanical tracking linkage allows its joints to be manually manipulated while fully supporting the weight of the ultrasound probe. The computer interface is provided in order to track the needle trajectory and display its path on a corresponding 3D TRUS image, allowing the physician to aim the needle-guide at predefined targets within the prostate. The system has been designed for use with several end-fired transducers that can be rotated about the longitudinal axis of the probe in order to generate 3D image for 3D navigation. Using the system, 3D TRUS prostate images can be generated in approximately 10 seconds. The system reduces most of the user variability from conventional hand-held probes, which make them unsuitable for precision biopsy, while preserving some of the user familiarity and procedural workflow. In this paper, we describe the 3D TRUS guided biopsy system and report on the initial clinical use of this system for prostate biopsy.

TH‐C‐304A‐06: Assessment of Registration Accuracy in 3D Transrectal Ultrasound Images of Prostates

Purpose: To evaluate the accuracy of our rigid and non‐rigid registration methods to align three‐dimensional (3D) transrectal ultrasound(TRUS)images of the prostate in order to improve repeat biopsy planning and guidance for prostate cancer diagnosis. Methods and Materials: Three 3D prostate ultrasoundimages were obtained from each of the 16 patients during 3D prostate ultrasound‐guided biopsy procedures. These 3D images had identifiable calcifications and cysts, which were used as fiducials for the registration validation. For each prostate, images at two different time points (to and t1) during the procedure were selected and the 3D coordinates of the identified fiducials were recorded. A total of 62 fiducials were identified. All images were then segmented manually and semi‐automatically. Images at to were registered to the corresponding images at t1 using rigid and non‐rigid registration. Rigid and non‐rigid registration was performed using boundary‐based iterative closest point and thin‐plate spline algorithms, respectively. We evaluated the registration methods by determining the target registration error (TRE), effect of segmentation variability on registration variability, the TRE dependence on fiducial distance from the transducer tip, and the fiducial localization error (FLE). Results: Tests showed that the FLE was 1.3mm. The mean TRE for manual and semi‐automatic segmentation was 2.2mm and 2.3mm respectively in both rigid and non‐rigid cases. There was little correlation between the TRE and the distance of the fiducials from the ultrasound transducer tip, and segmentation variability resulted in a 0.57mm standard deviation of the TRE. Conclusions: Our results suggest that both rigid and non‐rigid registration techniques for 3D TRUSimages during a biopsy procedure are accurate to within 2.3mm, which is less than the 5mm radius of the smallest tumors considered clinically significant. Future tests will evaluate registration accuracy for multimodal cases, specifically magnetic resonance to ultrasoundimages.