Jerry T. Wong

Real-time tumor tracking using implanted positron emission markers: Concept and simulation study

The delivery accuracy of radiation therapy for pulmonary and abdominal tumors suffers from tumor motion due to respiration. Respiratory gating should be applied to avoid the use of a large target volume margin that results in a substantial dose to the surrounding normal tissue. Precise respiratory gating requires the exact spatial position of the tumor to be determined in real time during treatment. Usually, fiducial markers are implanted inside or next to the tumor to provide both accurate patient setup and real-time tumor tracking. However, current tumor tracking systems require either substantial x-ray exposure to the patient or large fiducial markers that limit the value of their application for pulmonary tumors. We propose a real-time tumor tracking system using implanted positron emission markers (PeTrack). Each marker will be labeled with low activity positron emitting isotopes, such as 124I, 74As, or 84Rb. These isotopes have half-lives comparable to the duration of radiation therapy (from a few days to a few weeks). The size of the proposed PeTrack marker will be 0.5-0.8 mm, which is approximately one-half the size of markers currently employed in other techniques. By detecting annihilation gammas using position-sensitive detectors, multiple positron emission markers can be tracked in real time. A multimarker localization algorithm was developed using an Expectation-Maximization clustering technique. A Monte Carlo simulation model was developed for the PeTrack system. Patient dose, detector sensitivity, and scatter fraction were evaluated. Depending on the isotope, the lifetime dose from a 3.7 MBq PeTrack marker was determined to be 0.7-5.0 Gy at 10 mm from the marker. At the center of the field of view (FOV), the sensitivity of the PeTrack system was 240-320 counts/s per 1 MBq marker activity within a 30 cm thick patient. The sensitivity was reduced by 45% when the marker was near the edge of the FOV. The scatter fraction ranged from 12% (124I, 74As) to 16% (84Rb). In addition, four markers (labeled with 124I) inside a 30 cm diameter water phantom were simulated to evaluate the feasibility of the multimarker localization algorithm. Localization was considered successful if a marker was localized to within 2 mm from its true location. The success rate of marker localization was found to depend on the number of annihilation events used and the error in the initial estimate of the marker position. By detecting 250 positron annihilation events from 4 markers (average of 62 events per marker), the marker success rates for initial errors of +/-5, +/-10, and +/-15 mm were 99.9%, 99.6%, and 92.4%, respectively. Moreover, the average localization error was 0.55 (+/-0.27) mm, which was independent of initial error. The computing time for localizing four markers was less than 20 ms (Pentium 4, 2.8 GHz processor, 512 MB memory). In conclusion, preliminary results demonstrate that the PeTrack technique can potentially provide real-time tumor tracking with low doses associated with the markers activity. Furthermore, the small size of PeTrack markers is expected to facilitate implantation and reduce patient risk.

Feasibility of real time dual‐energy imaging based on a flat panel detector for coronary artery calcium quantification

The feasibility of a real-time dual-energy imaging technique with dynamic filtration using a flat panel detector for quantifying coronary arterial calcium was evaluated. In this technique, the x-ray beam was switched at 15 Hz between 60 kVp and 120 kVp with the 120 kVp beam having an additional 0.8 mm silver filter. The performance of the dynamic filtration technique was compared with a static filtration technique (4 mm Al+0.2 mm Cu for both beams). The ability to quantify calcium mass was evaluated using calcified arterial vessel phantoms with 20-230 mg of hydroxylapatite. The vessel phantoms were imaged over a Lucite phantom and then an anthropomorphic chest phantom. The total thickness of Lucite phantom ranges from 13.5-26.5 cm to simulate patient thickness of 16-32 cm. The calcium mass was measured using a densitometric technique. The effective dose to patient was estimated from the measured entrance exposure. The effects of patient thickness on contrast-to-noise ratio (CNR), effective dose, and the precision of calcium mass quantification (i.e., the frame to frame variability) were studied. The effects of misregistration artifacts were also measured by shifting the vessel phantoms manually between low- and high-energy images. The results show that, with the same detector signal level, the dynamic filtration technique produced 70% higher calcium contrast-to-noise ratio with only 4% increase in patient dose as compared to the static filtration technique. At the same time, x-ray tube loading increased by 30% with dynamic filtration. The minimum detectability of calcium with anatomical background was measured to be 34 mg of hydroxyapatite. The precision in calcium mass measurement, determined from 16 repeated dual-energy images, ranges from 13 mg to 41 mg when the patient thickness increased from 16 to 32 cm. The CNR was found to decrease with the patient thickness linearly at a rate of (-7%/cm). The anatomic background produced measurement root-mean-square (RMS) errors of 13 mg and 18 mg when the vessel phantoms were imaged over a uniform (over the rib) and nonuniform (across the edge of rib) bone background, respectively. Misregistration artifacts due to motions of up to 1.0 mm between the low- and high-energy images introduce RMS error of less than 4.3 mg, which is much smaller than the frame to frame variability and the measurement error due to anatomic background. The effective dose ranged from 1.1 to 6.6 microSv for each dual-energy image, depending on patient thickness. The study shows that real-time dual-energy imaging can potentially be used as a low dose technique for quantifying coronary arterial calcium.

Estimation of regional myocardial mass at risk based on distal arterial lumen volume and length using 3D micro-CT images ☆

The determination of regional myocardial mass at risk distal to a coronary occlusion provides valuable prognostic information for a patient with coronary artery disease. The coronary arterial system follows a design rule which allows for the use of arterial branch length and lumen volume to estimate regional myocardial mass at risk. Image processing techniques, such as segmentation, skeletonization and arterial network tracking, are presented for extracting anatomical details of the coronary arterial system using micro-computed tomography (micro-CT). Moreover, a method of assigning tissue voxels to their corresponding arterial branches is presented to determine the dependent myocardial region. The proposed micro-CT technique was utilized to investigate the relationship between the sum of the distal coronary arterial branch lengths and volumes to the dependent regional myocardial mass using a polymer cast of a porcine heart. The correlations of the logarithm of the total distal arterial lengths (L) to the logarithm of the regional myocardial mass (M) for the left anterior descending (LAD), left circumflex (LCX) and right coronary (RCA) arteries were log(L)=0.73log(M)+0.09 (R=0.78), log(L)=0.82log(M)+0.05 (R=0.77) and log(L)=0.85log(M)+0.05 (R=0.87), respectively. The correlation of the logarithm of the total distal arterial lumen volumes (V) to the logarithm of the regional myocardial mass for the LAD, LCX and RCA were log(V)=0.93log(M)-1.65 (R=0.81), log(V)=1.02log(M)-1.79 (R=0.78) and log(V)=1.17log(M)-2.10 (R=0.82), respectively. These morphological relations did not change appreciably for diameter truncations of 600-1400microm. The results indicate that the image processing procedures successfully extracted information from a large 3D dataset of the coronary arterial tree to provide prognostic indications in the form of arterial tree parameters and anatomical area at risk.

Determination of fractional flow reserve (FFR) based on scaling laws: a simulation study.

Fractional flow reserve (FFR) provides an objective physiological evaluation of stenosis severity. A technique that can measure FFR using only angiographic images would be a valuable tool in the cardiac catheterization laboratory. To perform this, the diseased blood flow can be measured with a first pass distribution analysis and the theoretical normal blood flow can be estimated from the total coronary arterial volume based on scaling laws. A computer simulation of the coronary arterial network was used to gain a better understanding of how hemodynamic conditions and coronary artery disease can affect blood flow, arterial volume and FFR estimation. Changes in coronary arterial flow and volume due to coronary stenosis, aortic pressure and venous pressure were examined to evaluate the potential use of flow and volume for FFR determination. This study showed that FFR can be estimated using arterial volume and a scaling coefficient corrected for aortic pressure. However, variations in venous pressure were found to introduce some error in FFR estimation. A relative form of FFR was introduced and was found to cancel out the influence of pressure on coronary flow, arterial volume and FFR estimation. The use of coronary flow and arterial volume for FFR determination appears promising.

Regional blood flow analysis and its relationship with arterial branch lengths and lumen volume in the coronary arterial tree

The limitations of visually assessing coronary artery disease are well known. These limitations are particularly important in intermediate coronary lesions (30-70% diameter stenosis) where it is difficult to determine whether a particular lesion is the cause of ischaemia. Therefore, a functional measure of stenosis severity is needed. The purpose of this study is to determine whether the expected maximum coronary blood flow in an arterial tree is predictable from its sum of arterial branch lengths or lumen volume. Using a computer model of a porcine coronary artery tree, an analysis of blood flow distribution was conducted through a network of millions of vessels that included the entire coronary artery tree down to the first capillary branch. The flow simulation results show that there is a linear relationship between coronary blood flow and the sum of its arterial branch lengths. This relationship holds over the entire arterial tree. The flow simulation results also indicate that there is a 3/4 power relation between coronary blood flow (Q) and the sum of its arterial lumen volume (V). Moreover, there is a linear relationship between normalized Q and normalized V raised to a power of 3/4 over the entire arterial tree. These results indicate that measured arterial branch lengths or lumen volumes can be used to predict the expected maximum blood flow in an arterial tree. This theoretical maximum blood flow, in conjunction with an angiographically measured blood flow, can potentially be used to calculate fractional flow reserve based entirely on angiographic data.

Dynamic dual-energy chest radiography: a potential tool for lung tissue motion monitoring and kinetic study

Dual-energy chest radiography has the potential to provide better diagnosis of lung disease by removing the bone signal from the image. Dynamic dual-energy radiography is now possible with the introduction of digital flat-panel detectors. The purpose of this study is to evaluate the feasibility of using dynamic dual-energy chest radiography for functional lung imaging and tumor motion assessment. The dual-energy system used in this study can acquire up to 15 frames of dual-energy images per second. A swine animal model was mechanically ventilated and imaged using the dual-energy system. Sequences of soft-tissue images were obtained using dual-energy subtraction. Time subtracted soft-tissue images were shown to be able to provide information on regional ventilation. Motion tracking of a lung anatomic feature (a branch of pulmonary artery) was performed based on an image cross-correlation algorithm. The tracking precision was found to be better than 1 mm. An adaptive correlation model was established between the above tracked motion and an external surrogate signal (temperature within the tracheal tube). This model is used to predict lung feature motion using the continuous surrogate signal and low frame rate dual-energy images (0.1-3.0 frames per second). The average RMS error of the prediction was (1.1 ± 0.3) mm. The dynamic dual energy was shown to be potentially useful for lung functional imaging such as regional ventilation and kinetic studies. It can also be used for lung tumor motion assessment and prediction during radiation therapy.

Quantitative coronary angiography using image recovery techniques for background estimation in unsubtracted images

Densitometry measurements have been performed previously using subtracted images. However, digital subtraction angiography (DSA) in coronary angiography is highly susceptible to misregistration artifacts due to the temporal separation of background and target images. Misregistration artifacts due to respiration and patient motion occur frequently, and organ motion is unavoidable. Quantitative densitometric techniques would be more clinically feasible if they could be implemented using unsubtracted images. The goal of this study is to evaluate image recovery techniques for densitometry measurements using unsubtracted images. A humanoid phantom and eight swine (25-35 kg) were used to evaluate the accuracy and precision of the following image recovery techniques: Local averaging (LA), morphological filtering (MF), linear interpolation (LI), and curvature-driven diffusion image inpainting (CDD). Images of iodinated vessel phantoms placed over the heart of the humanoid phantom or swine were acquired. In addition, coronary angiograms were obtained after power injections of a nonionic iodinated contrast solution in an in vivo swine study. Background signals were estimated and removed with LA, MF, LI, and CDD. Iodine masses in the vessel phantoms were quantified and compared to known amounts. Moreover, the total iodine in left anterior descending arteries was measured and compared with DSA measurements. In the humanoid phantom study, the average root mean square errors associated with quantifying iodine mass using LA and MF were approximately 6% and 9%, respectively. The corresponding average root mean square errors associated with quantifying iodine mass using LI and CDD were both approximately 3%. In the in vivo swine study, the root mean square errors associated with quantifying iodine in the vessel phantoms with LA and MF were approximately 5% and 12%, respectively. The corresponding average root mean square errors using LI and CDD were both 3%. The standard deviations in the differences between measured iodine mass in left anterior descending arteries using DSA and LA, MF, LI, or CDD were calculated. The standard deviations in the DSA-LA and DSA-MF differences (both approximately 21 mg) were approximately a factor of 3 greater than that of the DSA-LI and DSA-CDD differences (both approximately 7 mg). Local averaging and morphological filtering were considered inadequate for use in quantitative densitometry. Linear interpolation and curvature-driven diffusion image inpainting were found to be effective techniques for use with densitometry in quantifying iodine mass in vitro and in vivo. They can be used with unsubtracted images to estimate background anatomical signals and obtain accurate densitometry results. The high level of accuracy and precision in quantification associated with using LI and CDD suggests the potential of these techniques in applications where background mask images are difficult to obtain, such as lumen volume and blood flow quantification using coronary arteriography.

SU‐E‐J‐102: The Impact of the Number of Subjects for Atlas‐Based Automatic Segmentation

PURPOSE To determine the impact of atlas size on the performance of atlas-based automatic segmentation (ABAS) in delineation of organs at risk for adaptive radiation therapy. METHODS A total of 25 patients who had undergone intensity modulated radiation therapy for various head and neck cancers were retrospectively selected for inclusion in a library to be used for ABAS with the MIM VISTA software package (MIM Software, Cleveland OH). Treatment planning computed tomography (CT) scans and subsequent organ at risk (OAR) contours generated as part of the treatment planning process for these patients were added to the library. This library of 25 patients was then successively pruned to generate 5 atlases with 25, 20, 15, 10, and 5 patient subjects respectively. Atlas based segmentation was performed on 10 retrospectively selected treatment planning CT scans to automatically generate right and left parotid glands and brainstem contours. These planning CT scans belonged to a unique set of 10 patient subjects different from the ones used for generating the atlases. One physician (JW), who was blinded to the ABAS results, manually delineated gold-standard contours for the right and left parotid glands and brainstem. Dice similarity coefficients were calculated and analyzed as a function of atlas subject size. RESULTS For the sites selected in this study, the performance of ABAS was relatively insensitive to atlas size. Furthermore, some patient subjects were repeatedly selected implying that the adoption of a single standard patient for ABAS may be of benefit. CONCLUSIONS Our preliminary results indicate that the performance of the atlas based segmentation module in MIM VISTA Version 5.2 for the organs studied here may be relatively insensitive to the atlas size.

SU‐E‐T‐381: Dosimetric Analysis of Patients Treated with Accelerated Partial Breast Irradiation Using the Mammosite® and SAVI Applicators

Purpose: To present dosimetric data for patients undergoing Accelerated Partial Breast Irradiation (APBI) using the Mammosite® and Strut‐Adjusted Volume Implant (SAVI) applicators. Methods: We have treated 35 patients with Accelerated Partial Breast Irradiation (APBI) using high dose rate brachytherapy. These patients were treated per guidelines specified in the NSABP B‐39/RTOG 0413 protocol. The patients undergoing APBI have been treated with Mammosite® applicator (N=20) and SAVI applicator (N=15). The Mammosite® applicator used for all 20 patients was a single channel applicator. Single or multiple dwell positions were used as warranted by the type of balloon used and other clinical factors. The SAVI Applicator is a single‐entry, multi‐catheter device available in different sizes with 7, 9 and 11 catheters. The use of multiple catheters facilitates dose sculpting to improve treatment plan quality as well as reduce the skindose. Each treatment plan was evaluated for conformance of the dose to the PTV using commonly used dosimetric parameters. These parameters included balloon/cavity volume, PTV volume, V90, V100, V150 and V200. Results: The median V90 for the treatment plans delivered using the Mammosite® applicator was 99.7% (96.7%–99.8%) whereas it was 98.8% (96%–100.0%) for the patients treated with the SAVI applicator. Other data including the V100, V150, V200 and skindose will be presented. The impact/ importance of the data presented is that it provides useful information about the range of dosimetric parameters to be expected in treatment planning of APBI cases using these applicators. Conclusions: : Our results indicate that both the Mammosite® and SAVI applicators allow treatment planning and delivery of APBI per the guidelines specified in the NSABP B‐39/RTOG 0413 protocol.

SU‐GG‐T‐109: Comprehensive RapidArc™ Treatment Planning and Quality Assurance for Head and Neck Cancers

Purpose: To assess the dosimetric quality of a two‐arc RapidArc™ plans for the treatment of head and neck cancers and present corresponding quality assurance (QA) results. Method and Materials: Fifty three patients (male=32,female=21, = 61.9 years (range: 25–87 years) treated for nasopharynx, oropharynx, base of tongue and laryngealcancers were included in this study. Treatment doses ranged from 12Gy to 70.4Gy with many cases requiring irradiation of the cervical nodes as well as the primary site. RapidArc™ plans were generated using Varian Eclipse™ 8.6 and consisted primarily of two 358‐degree arcs delivered counterclockwise and clockwise, respectively with a ±5‐degree couch rotation for one of the arcs. Based on the treatment site and target location, there were instances where a 358‐degree arc and a partial arc with 90‐degree couch rotation were used. QA plans were generated and delivered to a solid‐water phantom with a Mapcheck™ detector array centrally mounted between the solid‐water slabs. Results: The RapidArc™ treatment plans were evaluated based on RTOG conformality index (CI), RTOG homogeneity index (HI), monitor units (MUs) and beam‐on time. The primary targets (PTVs) had mean±SD CI and HI values of 0.93±0.04 and 1.10±0.03, respectively. The average number of MUs was 565 (range: 350 – 2144) and beam‐on times ranged from 3 to 5 min. The average %PASS for the plan QA was 99.3% (range: 97.9% – 100%) using the 3% /3mm plan evaluation criteria in Mapcheck™. Conclusion: This dosimetric analysis indicates that a two‐arc RapidArc™ plan provides highly conformai dose distributions to head and neck treatments. QA results have shown that dose calculation and measurement are in good agreement for these plans. Short beam‐on times and few MUs have reduced the overall treatment time (setup + beam‐on times) by ∼40% per patient making RapidArc™ a more efficient delivery technique than multifield IMRT.