Volume 35, Issue 6, June 2008
Index of content:
- Imaging Scientific Session In Memoriam of Professor Bruce Hasegawa: Room 332
- Multi‐Modality and Ultrasound Imaging
TU‐C‐332‐01: A Comparison of Initial Area Under Curve (IAUC) Obtained From DCE‐CT and ‐MR Imaging in Patients with Cervical Cancer35(2008); http://dx.doi.org/10.1118/1.2962488View Description Hide Description
Purpose: To compare the IAUC60 (initial area under curve taken up to 60 seconds) obtained from DCE‐CT and ‐MR imaging in patients with cervical cancer.Method and Materials: A group of 40 patients with cervical cancer received a DCE‐MRI scan followed by a DCE‐CT scan at the time of staging. A radiologist observer identified and contoured the tumours on CT and MR images. At least one slice was identified as the matching slice in the CT and MR images for each patient. IAUC60 obtained from tumour region was normalized by IAUC60 from muscle region for DCE‐CT and ‐MR data. Correlation study and Bland‐Altman analysis were performed to assess the relationship between the normalized IAUC60 obtained from the two imaging modalities. Regression analysis was also applied to assess the relationship between the normalized IAUC60 and the normalized transfer constant (rK trans) for DCE‐CT data. Results: The regression analysis between the normalized IAUC60 and the normalized transfer constant (rK trans) for DCE‐CT data resulted in a significant strong correlation (R = 0.98, P<0.0005). A significant correlation (R = 0.75, P<0.0005) was found in the correlation analysis of the normalized IAUC60 between DCE‐CT and ‐MR imaging. The Bland‐Altman plot analysis of the normalized IAUC60 resulted in the 95% limit of agreement ranging from −2.68 to 4.75 and mean difference of 1.03. Since the average of the normalized IAUC60 measurements from two modalities ranged from 1.81 to 13.73, the degree of agreement was considered to be acceptable for the use of the two modalities interchangeably. Conclusion: The comparison of the normalized IAUC60 showed that both DCE‐CT and ‐MR imaging modalities may be used interchangeably in assessing cervical cancers. The normalized IAUC60 may be considered as a reliable quantitative surrogate of the normalized transfer constant for both modalities.
35(2008); http://dx.doi.org/10.1118/1.2962492View Description Hide Description
Purpose: To develop an automatic and accurate technique for concomitant segmentations and registration of liveranatomy using SPECT and CTimages for unsealed sourceradiotherapy.Method and Materials: The link between segmentation and registration is given by the using the level set of a liver segmentation into the registration process. In the combined approach, the liver is automatically segmented from the CTimage by evolving an initial seed with a level set until it locks to the liver's border as observed in the CTimages. The time‐crossing map of the level set is then used to match gradients in the SPECTimage to the level set by using a data structure containing the signed distance values at a small band of neighboring pixels. Results: The technique was applied to three cases of metastatic liver disease treated with unsealed source therapy. Results indicated that the speed map of the level set plays an importance role in obtainng an accurate registration and produce a segmentation that is superior in registration time and accuracy over manual segmentation or the standard registration approach using mutual information. Accuracy measured with the convergence analysis method was of less than 0.5 mm rotation and 1 degree rotation. Conclusion: With the proposed combined segmentation‐registration technique, the uncertainty of soft‐tissue target localization could be greatly reduced ensuring accurate therapy assessment to be precisely delivered as planned. The combined all‐in‐one approach is automated and provides excellent accuracy over manual segmentation and mutual information approaches.
TU‐C‐332‐03: Automatic Definition of Radiation Targets Using Textural Characteristics of Both Co‐Registered PET and CT Images35(2008); http://dx.doi.org/10.1118/1.2962495View Description Hide Description
Purpose: To automatically segment the radiation target for treatment of head and neck cancer (HNC) from FDG‐PET/CT images using a textural classifier and to compare the automated results with contours defined by expert observers. Method and Materials: 27 image features, including textural features from Spatial Gray‐Level Dependence Matrices and Neighborhood Gray‐Tone‐Difference Matrices, as well as statistical and structural features were calculated for 476 head and neck regions of interest (ROIs) in PET/CT images of 20 patients with HNC and 20 patients with lungcancer. A voxel based automated segmentation method using a Decision Tree (DT) based K nearest neighbors (KNN) classifier was developed based on the features in these ROIs. PET/CT images of another 10 head and neck patients who had all primary tumors and positive nodes manually segmented by three radiation oncologists were used to evaluate the method. Features were calculated for each voxel from corresponding PET and CTimages within a window centered on the voxel. All voxels of head and neck soft tissues from the below the eye to the apex of the lung were automatically segmented. Results: The specificity was 95% ± 2% when all “true negative” voxels were considered to be all soft tissue voxels excluding the ROIs considered abnormal by one or more of three radiation oncologists. Sensitivity was 84% ± 19% when “true positive” was considered the intersection of at least two physicians' abnormal ROIs and sensitivity was 90% ± 16% when all “true positive” was the intersection of the abnormal ROIs of all three physicians. Conclusion: This work suggests that an automated segmentation method based on texture classification of FDG‐PET/CT images has potential to provide accurate delineation of HNC. This could potentially lead to reduction in inter‐observer variability in target delineation and improved accuracy of treatment delivery.
TU‐C‐332‐04: Pilot Patient Studies Using a Dedicated Dual‐Modality SPECT‐CT System for Breast Imaging35(2008); http://dx.doi.org/10.1118/1.2962498View Description Hide Description
Purpose: Acknowledging the limitations/discomfort of mammography has inspired the development of a dedicated SPECT‐CT system to detect breast cancer, monitor therapeutic responses, and improve patient comfort. This system provides semi‐quantitative 3D functional/anatomical imaging of a pendant, uncompressed breast. Fused images can potentially provide more valuable clinical information than independent systems alone. Method and Materials: The SPECT subsystem permits fully‐3D complex acquisition trajectories around the breast, avoiding physical hindrances, overcoming distortions due to inadequate sampling, and allowing lesion detection on the chest wall. The CT subsystem, restricted to circular rotation, uses a quasi‐monochromatic, cone‐beam x‐ray source, which allows for reduced radiationdose and increased contrast between similar soft tissue attenuation coefficients. With no breast compression and an open, common field‐of‐view geometry system, the patient lies prone on a customized patient bed while the hybrid device non‐invasively acquires 3D data underneath. A preliminary investigation on the clinical performance of the hybrid system was done by imaging women with biopsy confirmed breast cancerResults:SPECT patient images can clearly visualize the tracer uptake by the tumor and view into the chest wall. Physical system constraints limit chest wall visualization in the CT patient images and thus patient positioning is under modification. Eliminating overlapping tissues through 3D imaging, the CTimages improve lesion isolation versus 2D imaging modalities. Complementary functional and anatomical image information helps localize suspicious areas for subsequent analysis. Conclusion: Implementation of the world's first dedicated SPECT‐CT system promises greatly improved visualization of the 3D breast volume. Complementary information from functional and anatomical imaging can guide lesion localization for subsequent analysis. Conflict of Interest: MPT is an inventor of this technology, and is named as an inventor on the patent for this technology applied for by Duke. If this technology becomes commercially successful, he and Duke could benefit financially.
TU‐C‐332‐05: Simulation of Ultrasound Two‐Dimensional Array Transducers Using a Frequency Domain Model35(2008); http://dx.doi.org/10.1118/1.2962509View Description Hide Description
Purpose:Ultrasound imaging with two‐dimensional (2D) arrays has garnered broad interest from scanner manufacturers and researchers for real time three‐dimensional (3D) imaging. Previously we described a frequency domain B‐mode imaging model applicable for linear and phased array transducers. In this study, we extend this model to incorporate 2D array transducers.Method and Materials: The pressure field for a 64×64 square array with element dimension of 0.15 mm and center‐to‐center spacing of 0.2 mm was calculated by applying the paraxial approximation to solve the 2D Rayleigh integral. We assume a rigid baffle, no apodization, a 2.5 MHz center frequency, and a speed of sound of 1540 m/s. A single transmit focus at 30 mm and dynamic receive focus with an F‐number of 2 was utilized. The 2D array model is compared with the widely used ultrasound simulation program FIELD II, which utilizes an approximate form of the time domain impulse response function. Results: Discrepancies between waveforms computed using our model and FIELD II are less than 4%, regardless of the steering angle for distances greater than 2 cm, yet computation times are on the order of 1/35 of those using FIELD II. Modern beam‐forming techniques such as apodization, dynamic aperture, dynamic receive focusing and 3D beam steering can also be simulated. The simulated beam patterns and point spread function images allow evaluation of beam properties for specific transducer parameters. Simulations of B‐mode images provide vivid demonstrations of the ability of 2D arrays with specific imaging parameters to detect lesions of a given backscatter contrast and size. Conclusion: The frequency domain approach provides an effective and feasible tool to model transmitted and pulse‐echo fields as well as B‐mode images for 2D array transducers.
35(2008); http://dx.doi.org/10.1118/1.2962523View Description Hide Description
Purpose: To develop an accurate patient‐specific PET attenuation coefficients map to be used in hybrid MRI‐PET systems for braintumorimaging. The attenuation maps are obtained by warping a general (atlas) CT dataset to the patient‐specific MRI dataset using a deformable registration model.Method and Materials: Patient MR images and the atlas CTimages are registered using a B‐Spline deformable model and the Mattes formulation of the mutual information metric as registration criterion. The registration establishes a voxel‐to‐voxel correspondence that maps each voxel in the CT atlas voxels to the MRI dataset, creating an artificial, individualized CT scan of the patient's anatomy as observed in the MRI dataset. To evaluate the accuracy of the deformable‐based attenuation correction, ten clinical braintumor cases are studied with MR‐CT image sets. For each case an artificial CT is computed by warping the atlas to the MRI datasets. This artificial CT is compared to the true patient's CT in terms of geometrical accuracy of the deformation module as well as a voxel‐to‐voxel comparison of HU units. Results: In all cases, mapping form the atlas CT to the individual MR was achieved with great geometrical accuracy as visually judged using the qualitative visual inspection tools. The mean distance between the artificial and true CT's external contour and bony anatomy was 1 mm and 1.5 mm, respectively. In terms of HU unit comparison, the mean voxel‐to‐voxel difference was less than 5%. Conclusion: Attenuation correction for hybrid MRI‐PET scanners can be easily achieved by individualizing an atlas CT to the MRI dataset using the BSpline deformable model, with no user interaction required. The method provides clinical accuracy while eliminating the need for an additional CT scan for PET attenuation correction.
TU‐C‐332‐07: Accuracy and Reproducibility of Tumor Position During Prolonged and Multi‐Modality Animal Imaging Studies35(2008); http://dx.doi.org/10.1118/1.2962526View Description Hide Description
Purpose: Dedicated small‐animal imaging devices are being used more frequently for translational molecular imaging studies. However, few studies have investigated the magnitude of animal motion during extended dynamic imaging studies or the precision of animal repositioning in multi‐modality and/or serial imaging protocols. The objective of this work was to determine the positional accuracy and precision with which tumors in situ can be reliably and reproducibly imaged on dedicated small‐animal imaging equipment. Method and Materials: A custom rodent animal cradle with a stereotactic template served to define a coordinate system and to facilitate rigid‐body image registration. Attached to the template were fiduciary markers containing PET tracer and MRI and CTcontrast media for visualization on the respective scanners. To quantify animal tumor motion during imaging protocols, “gold standard” point markers were inserted into tumors grown on the hind limb of nude rats. Three types of imaging examination were performed with the animals continuously anesthetized and immobilized: (i) single‐modality imaging (microPET and MRI) in which the animals remained in the same scanner continuously for 2 hours, (ii) multi‐modality imaging studies in which the animals were transported from a microPET to an MR scanner located in another building, and (iii) serial microPET scans in which the animals were removed from the scanner, then re‐positioned and scanned. Results: The animal tumor moved by less than 0.2–0.3 mm over two‐hour microPET or MR imaging sessions. Transporting the animal between instruments introduced additional error of ∼0.2 mm. In serial animal imaging studies, in which the animal was returned to its cage and subsequently re‐positioned, reproducibility within ∼0.8 mm could be obtained. Conclusion: To our knowledge, this is the first study systematically and rigorously evaluating the accuracy and precision with which tumors can be repeatedly imaged in small‐animal imaging devices.
35(2008); http://dx.doi.org/10.1118/1.2962529View Description Hide Description
Purpose:Ultrasound attenuation in liver differs from normal values when diffuse disease is present. Additionally, clinical images suggest that many masses in the liver have attenuation that varies from that of background tissue. In this presentation we compare four attenuation measurement algorithms to determine strengths and weaknesses of each for determining global and local values of attenuation in liver.Method and Materials: Full frames of RF echo data were acquired from test phantoms using a Siemens Antares scanner equipped with the Axis Direct research interface. Signals were acquired using linear array transducers operating at a 6 MHz frequency. Data were analyzed offline using 4 different algorithms: “Video Signal Analysis” (VSA), a centroid frequency shift method (FS), Diffraction Corrected Spectral Cross‐correlation (DCSC), and a conventional Reference Phantom Method (RPM). Both global estimates of ultrasound attenuation and low resolution attenuation coefficient images of uniform and “inclusion” phantoms were obtained. In a similar manner, attenuation is being measured in the livers of patients (under an IRB approved protocol) who are undergoing needle biopsies under ultrasound guidance. Results: The RPM provides accurate attenuation estimates, with acceptable variance, for uniform regions in phantoms. All methods are subject to attenuationimage artifacts when the backscatter is not uniform, as shown with the inclusion phantoms. However, the DCSC appears to be the least susceptible to backscatter variations. Initial attenuation results in liver were obtained from a patient with a hemangioma, a tumor that exhibited lower attenuation and higher backscatter than background liver. Because of elevated backscatter, the DCSC method performed best. Additional liver samples are being processed. Conclusion: Modern machines provide RF data that can be used to measure acoustical properties of tissues. When measuring attenuation, the RPM performed best in uniform regions of phantoms, but the DCSC technique appeared to be least susceptible to backscatter variations.
TU‐C‐332‐09: Effects of Mild Temperature Hyperthermia On Rat HT29 Xenograft Hypoxia Measured with a Dual‐Radiolabel Hypoxia Marker35(2008); http://dx.doi.org/10.1118/1.2962531View Description Hide Description
Purpose: Quantitative measurements to directly determine the tumor hypoxia response following a mild temperature hyperthermia (MTH) treatment were conducted in a rat HT29 colorectal xenograft model. Method and Materials: A hypoxia marker iodoazomycin galactopyranoside (IAZGP) was labeled with two radioisotopes of iodine and . The two distinct IAZGP tracers were injected into HT29 tumor‐bearing nude rats 4‐hour before and immediately after 41.5°C, 45‐munite hyperthermia treatment respectively. The animals were sacrificed 3‐hour post hyperthermia, tumors resected, frozen and cryo‐sectioned for digital autoradiography on phosphorimaging plate. Novel methods were developed to acquire and analyze the dual‐isotope digital autoradiographic images, to unfold the pixel contribution of tracers administered before and after hyperthermia, thus providing quantitative information of the hypoxia change at the microscopic (50‐micron) level. Results: The results showed that, immediately following MTH treatment, there was a significant reduction in hypoxia tracer binding, indicating a reduction in the tumor hypoxic fraction, and that re‐oxygenation had taken place in this rat HT29 xenograft model. Pixel‐by‐pixel analysis of the data revealed a decline in hypoxia tracer uptake after hyperthermia in most regions, but with the concomitant emergence of some new regions of hypoxia identified by increased tracer uptake post treatment. In the body‐temperature control group, the overall hypoxic fraction remained almost constant, with some hypoxic tracer redistribution (putative acute hypoxia) observed. In conclusion, the pretreatment hypoxic fraction changed from between 18–42% to post‐hyperthermia values of between 7%–20% (spread among 5 animals). Conclusion: This study provided evidence for reoxygenation immediately following MTH treatment in the rat HT29 xenograft, with a preponderance of microscopic regional decreased radiotracer uptake. However, a few areas did exhibit increased hypoxia specific tracer uptake indicating the possible emergence of new hypoxia.
TU‐C‐332‐10: Evaluation of Combined Effects of Target Size, Background Activity, and Respiratory Motion On 3D and 4D PET/CT Images35(2008); http://dx.doi.org/10.1118/1.2962534View Description Hide Description
Purpose: In recent years, quantitative analysis of gated (4D) PET/CT images has been introduced for diagnosis, staging, and prediction of tumor response where internal organ motion is significant. However, the best methodology for applying 4D information to radiotherapy target definition is not currently well established. In order to accurately determine moving target volume, we have investigated the combined effects of target size, respiratory motion, target‐to‐background activity concentration ratio (TBR) on ungated (3D) and 4D PETimages as well as gating methods. Method and Materials: Using a GE Discovery PET/CT scanner, a 3D‐PET scan corrected with a 3D attenuation map from 3D‐CT scan and a 4D‐PET scan corrected with matching attenuation maps from 4D‐CT were performed using spherical targets (0.5–26.5 mL) filled with in a NEMA IEC body phantom at different TBRs (infinite, 8, and 4). To simulate respiratory motion, the phantoms were driven sinusoidally in the superior‐inferior direction with amplitudes of 0, 1, and 2 cm and a period of 4.5 s. Recovery coefficients were determined on PETimages. In addition, gating methods using different numbers of gating bins (1–20 bins) were evaluated by determining imagenoise and temporal resolution. Results: Signal loss in 3D‐PET images was measured from both the partial volume effect, due to the limited PET resolution, as well as respiratory motion. The results show that signal loss depends on both the amplitude and shape of respiratory motion. However, 4D‐PET successfully recovers most of the loss induced by respiratory motion. The 5‐bin gating method gives the best temporal resolution with acceptable imagenoise.Conclusion: The results based on the 4D scan protocols can be used to improve the accuracy of gross tumor volume definition in the lung and abdomen.