Development of a semi-automatic alignment tool for accelerated localization of the prostate

Shortening image registration time using a deep neural network for patient positional verification in radiotherapy

We sought to accelerate 2D/3D image registration computation time using image synthesis with a deep neural network (DNN) to generate digitally reconstructed radiographic (DRR) images from X-ray flat panel detector (FPD) images. And we explored the feasibility of using our DNN in the patient setup verification application. Images of the prostate and of the head and neck (H&N) regions were acquired by two oblique X-ray fluoroscopic units and the treatment planning CT. DNN was designed to generate DRR images from the FPD image data. We evaluated the quality of the synthesized DRR images to compare the ground-truth DRR images using the peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM). Image registration accuracy and computation time were evaluated by comparing the 2D-3D image registration algorithm using DRR and FPD image data with DRR and synthesized DRR images. Mean PSNR values were 23.4 ± 3.7 dB and 24.1 ± 3.9 dB for the pelvic and H&N regions, respectively. Mean SSIM values for both cases were also similar (= 0.90). Image registration accuracy was degraded by a mean of 0.43 mm and 0.30°, it was clinically acceptable. Computation time was accelerated by a factor of 0.69. Our DNN successfully generated DRR images from FPD image data, and improved 2D-3D image registration computation time up to 37% in average.

Optimizing 3DCT image registration for interfractional changes in carbon-ion prostate radiotherapy

To perform setup procedures including both positional and dosimetric information, we developed a CT-CT rigid image registration algorithm utilizing water equivalent pathlength (WEPL)-based image registration and compared the resulting dose distribution with those of two other algorithms, intensity-based image registration and target-based image registration, in prostate cancer radiotherapy using the carbon-ion pencil beam scanning technique. We used the data of the carbon ion therapy planning CT and the four-weekly treatment CTs of 19 prostate cancer cases. Three CT-CT registration algorithms were used to register the treatment CTs to the planning CT. Intensity-based image registration uses CT voxel intensity information. Target-based image registration uses target position on the treatment CTs to register it to that on the planning CT. WEPL-based image registration registers the treatment CTs to the planning CT using WEPL values. Initial dose distributions were calculated using the planning CT with the lateral beam angles. The treatment plan parameters were optimized to administer the prescribed dose to the PTV on the planning CT. Weekly dose distributions using the three different algorithms were calculated by applying the treatment plan parameters to the weekly CT data. Dosimetry, including the dose received by 95% of the clinical target volume (CTV-D95), rectal volumes receiving > 20 Gy (RBE) (V20), > 30 Gy (RBE) (V30), and > 40 Gy (RBE) (V40), were calculated. Statistical significance was assessed using the Wilcoxon signed-rank test. Interfractional CTV displacement over all patients was 6.0 ± 2.7 mm (19.3 mm maximum standard amount). WEPL differences between the planning CT and the treatment CT were 1.2 ± 0.6 mm-H₂O (< 3.9 mm-H₂O), 1.7 ± 0.9 mm-H₂O (< 5.7 mm-H₂O) and 1.5 ± 0.7 mm-H₂O (< 3.6 mm-H₂O maxima) with the intensity-based image registration, target-based image registration, and WEPL-based image registration, respectively. For CTV coverage, the D95 values on the planning CT were > 95% of the prescribed dose in all cases. The mean CTV-D95 values were 95.8 ± 11.5% and 98.8 ± 1.7% with the intensity-based image registration and target-based image registration, respectively. The WEPL-based image registration was CTV-D95 to 99.0 ± 0.4% and rectal Dmax to 51.9 ± 1.9 Gy (RBE) compared to 49.4 ± 9.1 Gy (RBE) with intensity-based image registration and 52.2 ± 1.8 Gy (RBE) with target-based image registration. The WEPL-based image registration algorithm improved the target coverage from the other algorithms and reduced rectal dose from the target-based image registration, even though the magnitude of the interfractional variation was increased.

Development of CBCT-based prostate setup correction strategies and impact of rectal distension

Background

Cone-beam computed tomography (CBCT) image-guided radiotherapy (IGRT) systems are widely used tools to verify and correct the target position before each fraction, allowing to maximize treatment accuracy and precision. In this study, we evaluate automatic three-dimensional intensity-based rigid registration (RR) methods for prostate setup correction using CBCT scans and study the impact of rectal distension on registration quality.

Methods

We retrospectively analyzed 115 CBCT scans of 10 prostate patients. CT-to-CBCT registration was performed using (a) global RR, (b) bony RR, or (c) bony RR refined by a local prostate RR using the CT clinical target volume (CTV) expanded with 1-to-20-mm varying margins. After propagation of the manual CT contours, automatic CBCT contours were generated. For evaluation, a radiation oncologist manually delineated the CTV on the CBCT scans. The propagated and manual CBCT contours were compared using the Dice similarity and a measure based on the bidirectional local distance (BLD). We also conducted a blind visual assessment of the quality of the propagated segmentations. Moreover, we automatically quantified rectal distension between the CT and CBCT scans without using the manual CBCT contours and we investigated its correlation with the registration failures. To improve the registration quality, the air in the rectum was replaced with soft tissue using a filter. The results with and without filtering were compared.

Results

The statistical analysis of the Dice coefficients and the BLD values resulted in highly significant differences (p<10⁻⁶) for the 5-mm and 8-mm local RRs vs the global, bony and 1-mm local RRs. The 8-mm local RR provided the best compromise between accuracy and robustness (Dice median of 0.814 and 97% of success with filtering the air in the rectum). We observed that all failures were due to high rectal distension. Moreover, the visual assessment confirmed the superiority of the 8-mm local RR over the bony RR.

Conclusion

The most successful CT-to-CBCT RR method proved to be the 8-mm local RR. We have shown the correlation between its registration failures and rectal distension. Furthermore, we have provided a simple (easily applicable in routine) and automatic method to quantify rectal distension and to predict registration failure using only the manual CT contours.

A novel approach for establishing benchmark CBCT/CT deformable image registrations in prostate cancer radiotherapy

Deformable image registration (DIR) is an integral component for adaptive radiation therapy. However, accurate registration between daily cone-beam computed tomography (CBCT) and treatment planning CT is challenging, due to significant daily variations in rectal and bladder fillings as well as the increased noise levels in CBCT images. Another significant challenge is the lack of 'ground-truth' registrations in the clinical setting, which is necessary for quantitative evaluation of various registration algorithms. The aim of this study is to establish benchmark registrations of clinical patient data. Three pairs of CT/CBCT datasets were chosen for this institutional review board approved retrospective study. On each image, in order to reduce the contouring uncertainty, ten independent sets of organs were manually delineated by five physicians. The mean contour set for each image was derived from the ten contours. A set of distinctive points (round natural calcifications and three implanted prostate fiducial markers) were also manually identified. The mean contours and point features were then incorporated as constraints into a B-spline based DIR algorithm. Further, a rigidity penalty was imposed on the femurs and pelvic bones to preserve their rigidity. A piecewise-rigid registration approach was adapted to account for the differences in femur pose and the sliding motion between bones. For each registration, the magnitude of the spatial Jacobian (|JAC|) was calculated to quantify the tissue compression and expansion. Deformation grids and finite-element-model-based unbalanced energy maps were also reviewed visually to evaluate the physical soundness of the resultant deformations. Organ DICE indices (indicating the degree of overlap between registered organs) and residual misalignments of the fiducial landmarks were quantified. Manual organ delineation on CBCT images varied significantly among physicians with overall mean DICE index of only 0.7 among redundant contours. Seminal vesicle contours were found to have the lowest correlation amongst physicians (DICE = 0.5). After DIR, the organ surfaces between CBCT and planning CT were in good alignment with mean DICE indices of 0.9 for prostate, rectum, and bladder, and 0.8 for seminal vesicles. The Jacobian magnitudes |JAC| in the prostate, rectum, and seminal vesicles were in the range of 0.4-1.5, indicating mild compression/expansion. The bladder volume differences were larger between CBCT and CT images with mean |JAC| values of 2.2, 0.7, and 1.0 for three respective patients. Bone deformation was negligible (|JAC| = ∼ 1.0). The difference between corresponding landmark points between CBCT and CT was less than 1.0 mm after DIR. We have presented a novel method of establishing benchmark DIR accuracy between CT and CBCT images in the pelvic region. The method incorporates manually delineated organ surfaces and landmark points as well as pixel similarity in the optimization, while ensuring bone rigidity and avoiding excessive deformation in soft tissue organs. Redundant contouring is necessary to reduce the overall registration uncertainty.

Target Localization and Toxicity in Dose-Escalated Prostate Radiotherapy with Image-Guided Approach using Daily Planar Kilovoltage Imaging

Dose escalation with intensity-modulated radiation therapy (IMRT) for carcinoma of the prostate has augmented the need for accurate prostate localization prior to dose delivery. Daily planar kilovoltage (kV) imaging is a low-dose image-guidance technique that is prevalent among radiation oncologists. However, clinical outcomes evaluating the benefit of daily kV imaging are lacking. The purpose of this study was to report our clinical experience, including prostate motion and gastrointestinal (GI) and genitourinary (GU) toxicities, using this modality. A retrospective analysis of 100 patients treated consecutively between December 2005 and March 2008 with definitive external beam IMRT for T1c-T4 disease were included in this analysis. Prescription doses ranged from 74–78 Gy (median, 76) in 2 Gy fractions and were delivered following daily prostate localization using on-board kV imaging (OBI) to localize gold seed fiducial markers within the prostate. Acute and late toxicities were graded as per the NCI CTCAEv3.0. The median follow-up was 22 months. The magnitude and direction of prostate displacement and daily shifts in three axes are reported. Of note, 9.1% and 12.9% of prostate displacements were ≥ 5 mm in the anterior-posterior and superior-inferior directions, respectively. Acute grade 2 GI and GU events occurred in 11% and 39% of patients, respectively, however no grade 3 or higher acute GI or GU events were observed. Regarding late toxicity, 2% and 17% of patients developed grade 2 toxicities, and similarly no grade 3 or higher events had occurred by last follow-up. Thus, kV imaging detected a substantial amount of inter-fractional displacement and may help reduce toxicity profiles, especially high grade events, by improving the accuracy of dose delivery.

Comparison of daily prostate positions during conformal radiation therapy of prostate cancer using an integrated CT-linear accelerator system: in-room CT image versus digitally reconstructed radiograph

PURPOSE

The purpose of this study is to quantify the magnitudes of the position shifts of internal structures together with the correlation between the day-to-day positioning of the prostate and the bony anatomy using an integrated CT-linear accelerator system for external beam radiation therapy for prostate cancer.

MATERIALS AND METHODS

A total of 1176 pretreatment in-room CT images and their digitally reconstructed radio-graph (DRR) pairs from 33 patients were acquired over the course of the study. The differences between the isocenter of the prostate on in-room CT and the isocenter of the bony anatomy on DRR were analyzed. The agreement between positions in each direction was compared using Bland-Altman limits of agreement.

RESULTS

The 95% limits of agreement in lateral (LR), superoinferior (SI), and anteroposterior (AP) directions were -2.98 to 2.49 mm, -4.69 to 5.75 mm, and -8.23 to 7.30 mm, respectively. The isocenter was localized to within 3.0 mm on in-room CT images and DRR 99.0% in LR, 85.1% in SI, and 85.9% in AP.

CONCLUSIONS

Considerable differences between in-room CT images and DRR exist. These data demonstrate that there is a significantly greater shift in the SI and AP directions than in the lateral direction for the entire patient group. Applications such as our image guide system will, with routine clinical use, continue to improve the precision of external beam radiation therapy for prostate cancer.

Prostate localization on daily cone-beam computed tomography images: accuracy assessment of similarity metrics

PURPOSE

To evaluate different similarity metrics (SM) using natural calcifications and observation-based measures to determine the most accurate prostate and seminal vesicle localization on daily cone-beam CT (CBCT) images.

METHODS AND MATERIALS

CBCT images of 29 patients were retrospectively analyzed; 14 patients with prostate calcifications (calcification data set) and 15 patients without calcifications (no-calcification data set). Three groups of test registrations were performed. Test 1: 70 CT/CBCT pairs from calcification dataset were registered using 17 SMs (6,580 registrations) and compared using the calcification mismatch error as an endpoint. Test 2: Using the four best SMs from Test 1, 75 CT/CBCT pairs in the no-calcification data set were registered (300 registrations). Accuracy of contour overlays was ranked visually. Test 3: For the best SM from Tests 1 and 2, accuracy was estimated using 356 CT/CBCT registrations. Additionally, target expansion margins were investigated for generating registration regions of interest.

RESULTS

Test 1-Incremental sign correlation (ISC), gradient correlation (GC), gradient difference (GD), and normalized cross correlation (NCC) showed the smallest errors (mu +/- sigma: 1.6 +/- 0.9 approximately 2.9 +/- 2.1 mm). Test 2-Two of the three reviewers ranked GC higher. Test 3-Using GC, 96% of registrations showed <3-mm error when calcifications were filtered. Errors were left/right: 0.1 +/- 0.5mm, anterior/posterior: 0.8 +/- 1.0mm, and superior/inferior: 0.5 +/- 1.1 mm. The existence of calcifications increased the success rate to 97%. Expansion margins of 4-10 mm were equally successful.

CONCLUSION

Gradient-based SMs were most accurate. Estimated error was found to be <3 mm (1.1 mm SD) in 96% of the registrations. Results suggest that the contour expansion margin should be no less than 4 mm.

Evaluation of Image Enhancement Method on Target Registration Using Cone Beam CT in Radiation Therapy

An intensity based six-degree image registration algorithm between cone-beam CT (CBCT) and planning CT has been developed for image-guided radiation therapy (IGRT). CT images of an anthropomorphic chest phantom were acquired using conventional CT scanner and corresponding CBCT was reconstructed based on projection images acquired by an on-board imager (OBI). Both sets of images were initially registered to each other using attached fudicial markers to achieve a golden standard registration. Starting from this point, an offset was applied to one set of images, and the matching result was found by a gray-value based registration method. Finally, The registration error was evaluated by comparing the detected shifts with the known shift. Three window-level (WL) combinations commonly used for image enhancement were examined to investigate the effect of anatomical information of Bony only (B), Bone+Tissue (BT), and Bone+Tissue+Air (BTA) on the accuracy and robustness of gray-value based registration algorithm. Extensive tests were performed in searching for the attraction range of registration algorithm. The widest attraction range was achieved with the WL combination of BTA. The average attraction ranges of this combination were 73.3 mm and 81.6 degree in the translation and rotation dimensions, respectively, and the average registration errors were 0.15 mm and 0.32 degree. The WL combination of BT shows the secondary largest attraction ranges. The WL combination of B shows limited convergence property and its attraction range was the smallest among the three examined combinations (on average 33.3 mm and 25.0 degree). If two sets of 3D images in original size (512 × 512) were used, registration could be accomplished within 10~20 minutes by current algorithm, which is only acceptable for off-line reviewing purpose. As the size of image set reduced by a factor of 2~4, the registration time would be 2~4 minutes which is feasible for on-line target localization.

Different styles of image-guided radiotherapy

To account for geometric uncertainties during radiotherapy, safety margins are applied. In many cases, these margins overlap organs at risk, thereby limiting dose escalation. The aim of image-guided radiotherapy is to improve the accuracy by imaging tumors and critical structures on the machine just before irradiation. The availability of high-quality imaging systems and automatic image registration on the machine leads to many new clinical applications, such as high-precision hypofractionated treatments of brain metastases and solitary long tumors with online tumor position corrections. In this review, the prerequisites for image guidance in terms of planning, image acquisition, and processing are first described. Then, the various methods of correction are discussed such as table shifts and rotation and direct adaptation of machine parameters. Then, online, offline, and intrafraction correction strategies are discussed. Finally, some imaging dose issues are discussed showing that kilovoltage cone-beam computed tomography guidance has a net positive impact on the integral dose; the gain caused by margin reduction is larger than the image dose. We can conclude that image-guided radiotherapy is very much a clinical reality and that the development of optimal clinical protocols should currently be the focus of research.

eNAL: An Extension of the NAL Setup Correction Protocol for Effective Use of Weekly Follow-up Measurements

PURPOSE

The no action level (NAL) protocol reduces systematic displacements relative to the planning CT scan by using the mean displacement of the first few treatment fractions as a setup correction in all subsequent fractions. This approach may become nonoptimal in case of time trends or transitions in the systematic displacement of a patient. Here, the extended NAL (eNAL) protocol is introduced to cope with this problem.

METHODS AND MATERIALS

The initial setup correction of eNAL is the same as in NAL. However, in eNAL, additional weekly follow-up measurements are performed. The setup correction is updated after each follow-up measurement based on linear regression of the available measured displacements to track and correct systematic time-dependent changes. We investigated the performance of eNAL with Monte Carlo simulations for populations without systematic displacement changes over time, with large gradual changes (time trends), and with large sudden changes (transitions). Weekly follow-up measurements were simulated for 35 treatment fractions. We compared the outcome of eNAL with NAL and optimized shrinking action level (SAL) protocol with weekly measurements.

RESULTS

Without time-dependent changes, eNAL, SAL, and NAL performed comparably, but SAL required the largest imaging workload. For time trends and transitions, eNAL performed superiorly to the other protocols and reduced systematic displacements to the same magnitude as in case of no time-dependent changes (SD approximately 1 mm).

CONCLUSION

Extended NAL can reduce systematic displacements to a minor level irrespective of the precise nature of the systematic time-dependent changes that may occur in a population.

Imaging and alignment for image-guided radiation therapy

Image-guided radiation therapy is an exciting new area that focuses heavily on the potential benefit of advanced imaging and image registration to improve precision, thus limiting morbidity and potentially allowing for safe delivery of increased dose. This review explores the issues surrounding the use of imaging and image registration for treatment planning and verification, with emphasis on the underlying patient model and alignment algorithms.

Influence of the initial rectal distension on posterior margins in primary and postoperative radiotherapy for prostate cancer

BACKGROUND AND PURPOSE

The aim of the study was to define the effect of different rectum fillings in the planning CT study on the posterior clinical target volume (CTV) displacements (PD) in primary and postoperative radiotherapy (RT) for prostate cancer.

MATERIALS AND METHODS

Fifty patients underwent CT scans in supine position with a full bladder and an empty bladder before RT and at several points in time during the treatment. PD were determined depending on the initial rectum volume (RV), average cross-sectional rectal area (CSA), and the rectal diameter at the level of the bladder neck (RD).

RESULTS

Posterior CTV motion was not found to be minimal with a particularly small initial rectum filling. Steeply increasing PD resulted for patients with RV>120cm(3), CSA>12cm(2), and RD>4.5cm. While below these critical values a posterior margin of 6mm/9mm allowed to cover 80%/90% of displacements, 18mm/24mm were needed for patients with larger rectum fillings. No correlation of increasing rectum distension with increasing PD was found at the apex level. PD could not be reduced by voiding the bladder.

CONCLUSIONS

Defining the posterior margin in prostate RT, the initial rectum distension and the superior-inferior CTV level has to be considered. Patients with large initial rectum fillings have preferentially the need for repeated planning CT scans or image-guided RT.

Quantification des mouvements prostatiques lors de l'irradiation prostatique

Decrease treatment uncertainties is one of the most important challenge in radiation oncology. Numerous techniques are available to quantify prostate motion and visualise prostate location day after day before each irradiation: CT-scan, cone-beam-CT-Scan, ultrason, prostatic markers... The knowledge of prostate motion is necessary to define the minimal margin around the target volume needed to avoid mispositioning during treatment session. Different kind of prostate movement have been studied and are reported in the present work: namely, those having a large amplitude extending through out the whole treatment period on one hand; and those with a shorter amplitude happening during treatment session one the other hand. The long lasting movement are mostly anterior-posterior (3 mm standard deviation), secondary in cranial-caudal (1-2 mm standard deviation) and lateral directions (0.5-1 mm standard deviation). They are mostly due to the rectal state of filling and mildly due to bladder filling or inferior limbs position. On the other hand, the shorter movement that occurs during the treatment session is mostly variation of position around a steady point represented by the apex. Ones again, the rectal filling state is the principle cause. This way, during the 20 minutes of a treatment session, including the positioning of the patient, a movement of less than 3 mm could be expected when the rectum is empty. Ideally, real time imaging tools should allow an accurate localisation of the prostate and the adaptation of the dosimetry before each treatment session in a time envelope not exceeding 20 minutes.

In-room CT techniques for image-guided radiation therapy

Accurate patient setup and target localization are essential to advanced radiation therapy treatment. Significant improvement has been made recently with the development of image-guided radiation therapy, in which image guidance facilitates short treatment course and high dose per fraction radiotherapy, aiming at improving tumor control and quality of life. Many imaging modalities are being investigated, including x-ray computed tomography (CT), ultrasound imaging, positron emission tomography, magnetic resonant imaging, magnetic resonant spectroscopic imaging, and kV/MV imaging with flat panel detectors. These developments provide unique imaging techniques and methods for patient setup and target localization. Some of them are different; some are complementary. This paper reviews the currently available kV x-ray CT systems used in the radiation treatment room, with a focus on the CT-on-rails systems, which are diagnostic CT scanners moving on rails installed in the treatment room. We will describe the system hardware including configurations, specifications, operation principles, and functionality. We will review software development for image fusion, structure recognition, deformation correction, target localization, and alignment. Issues related to the clinical implementation of in-room CT techniques in routine procedures are discussed, including acceptance testing and quality assurance. Clinical applications of the in-room CT systems for patient setup, target localization, and adaptive therapy are also reviewed for advanced radiotherapy treatments.

Comparison of three radiotherapy treatment planning protocols of definitive external-beam radiation for localized prostate cancer

BACKGROUND

Three radiotherapy treatment planning (RTTP) protocols for definitive external-beam radiation for localized prostate cancer, designed and clinically applied at Kyoto University, were compared.

METHODS

Treatment plans were created according to three different RTTP protocols (old three-dimensional conformal radiotherapy [3D-CRT], new 3D-CRT, and intensity-modulated radiotherapy [IMRT]) on computed tomography (CT) data sets of five patients with localized prostate cancer. The dynamic-arc conformal technique was used in the 3D-CRT protocols. Differences in dose distribution were evaluated and compared based on dose-volume histogram (DVH) analyses.

RESULTS

The coverage of the clinical target volume (= prostate alone) was comparable among the three RTTP protocols. However, the average values for the percent volume that received at least 95% of the prescription dose (V95), the percent of the prescription dose covering 95% of the volume (D95), and the conformity index of the planning target volume (PTV) were 99%, 97%, and 0.88 for the IMRT; 93.9%, 94.5%, and 0.76 for the new 3D-CRT; and 59.6%, 82.9%, and 0.6 for the old 3D-CRT protocol, respectively. Inhomogeneity of doses to the PTV was larger with the IMRT protocol than with the new 3D-CRT protocol. Doses to both the rectal wall and bladder wall were almost comparable with the new 3D-CRT and IMRT protocols, but were lower with the old 3D-CRT protocol, due to the lowest prescription dose and incomplete dose coverage of the PTV.

CONCLUSION

The old 3D-CRT protocol could not achieve the goals for the PTV set in the IMRT protocol. The new 3D-CRT and IMRT protocols were generally comparable in terms of both the PTV coverage and normal tissue-sparing, although the IMRT protocol achieved the most conformal dose distribution to the PTV, in return for a larger, but acceptable, dose inhomogeneity.

Radiothérapie guidée par tomodensitométrie associée à l'accélérateur linéaire dans la salle de traitement

Target localization has become increasingly important in the advent of IMRT, as treatment margins are reduced and target doses are increased with high-dose gradients outside this target volume. The in-room CT on rails-LINAC system allows CT imaging while the patient remains immobilized in the treatment position just prior to treatment. The anatomic inter- and intra-fractional variations can be therefore quantified during a course of treatment. The position of the tumour can be checked and corrected before the fraction. In case of modification of tumour shape, a re-planning of the treatment is also feasible. However, several issues remain: the integration with routine clinical treatment due to a lack of software tools, the frequency of imaging, and the cost-efficiency ratio. The clinical experience is yet very limited but CT-image-guided radiotherapy appears promising for prostate, brain and spinal tumours.

Update on Radiation Therapy in Prostate Cancer

Higher doses of radiation result in improved clinical control of prostate cancer,and the recent advances in prostate cancer radiotherapy are designed to escalate dose while minimizing toxicity. To achieve this goal, tighter treatment margins are needed, which require more accurate delineation of the prostate target and normal tissue at the time of treatment planning and before actual daily treatments. Modem radiation therapy techniques can deposit conformal dose virtually anywhere in the body; however, this precise therapy is of no value if it is not accurately hitting the target. Whether dose escalation is achieved by external beam techniques (eg, IMRT, protons) or brachytherapy, these ba-sic planning and delivery considerations are essentially the same. Future directions in prostate radiation therapy will use even higher radiation doses,alternative fractionation patterns, intraprostatic targets (eg, prostate tumor seen on MRI), and improved patient selection regarding which patients will benefit the most from these advanced techniques.

Prostate contouring uncertainty in megavoltage computed tomography images acquired with a helical tomotherapy unit during image-guided radiation therapy

PURPOSE

To evaluate the image-guidance capabilities of megavoltage computed tomography (MVCT), this article compares the interobserver and intraobserver contouring uncertainty in kilovoltage computed tomography (KVCT) used for radiotherapy planning with MVCT acquired with helical tomotherapy.

METHODS AND MATERIALS

Five prostate-cancer patients were evaluated. Each patient underwent a KVCT and an MVCT study, a total of 10 CT studies. For interobserver variability analysis, four radiation oncologists, one physicist, and two radiation therapists (seven observers in total) contoured the prostate and seminal vesicles (SV) in the 10 studies. The intraobserver variability was assessed by asking all observers to repeat the contouring of 1 patient's KVCT and MVCT studies. Quantitative analysis of contour variations was performed by use of volumes and radial distances.

RESULTS

The interobserver and intraobserver contouring uncertainty was larger in MVCT compared with KVCT. Observers consistently segmented larger volumes on MVCT where the ratio of average prostate and SV volumes was 1.1 and 1.2, respectively. On average (interobserver and intraobserver), the local delineation variability, in terms of standard deviations [Deltasigma = radical(sigma2MVCT-sigma2KVCT)], increased by 0.32 cm from KVCT to MVCT.

CONCLUSIONS

Although MVCT was inferior to KVCT for prostate delineation, the application of MVCT in prostate radiotherapy remains useful.

Geometric and dosimetric evaluations of an online image-guidance strategy for 3D-CRT of prostate cancer

PURPOSE

To evaluate an online image-guidance strategy for conformal treatment of prostate cancer and to estimate margin-reduction benefits.

METHODS AND MATERIALS

Twenty-eight patients with at least 16 helical computed tomography scans were each used in this study. Two prostate soft-tissue registration methods, including sagittal rotation, were evaluated. Setup errors and rigid organ motion were corrected online; non-rigid and intrafraction motion were included in offline analysis. Various clinical target volume-planning target volume (CTV-PTV) margins were applied. Geometrical evaluations included analyses of isocenter shifts and rotations and overlap index. Dosimetric evaluations included minimum dose and equivalent uniform dose (EUD) for prostate and gEUD for rectum.

RESULTS

Average isocenter shift and rotation were (dX,dY,dZ,theta) = (0.0 +/- 0.7,-1.1 +/- 4.0,-0.1 +/- 2.5,0.7 degrees +/- 2.0 degrees ) mm. Prostate motion in anterior-posterior (AP) direction was significantly higher than superior-inferior and left-right (LR) directions. This observation was confirmed by isocenter shift in perspectives AP (1.8 +/- 1.8 mm) and RL (3.7 +/- 3.0 mm). Organ motion degrades target coverage and reduces doses to rectum. If 2% dose reduction on prostate D(99) is allowed for 90% patients, then minimum 3 mm margins are necessary with ideal image registration.

CONCLUSIONS

Significant margin reduction can be achieved through online image guidance. Certain margins are still required for nonrigid and intrafraction motion. To further reduce margin, a strategy that combines online geometric intervention and offline dose replanning is necessary.

From IMRT to IGRT: Frontierland or Neverland?

The recent enthusiasm for real-time image guidance in radiotherapy (IGRT) is in part due to the commercial availability of advanced on-line imaging technologies. Perhaps more important than its potential to improve conventional radiotherapy, IGRT may lead to a paradigm shift in facilitating hypo-fractionated or single-dose treatment. However, there are uncertainty regarding features and approaches of competing IGRT systems and as to whether a sub-set of the features of an ideal IGRT system would suffice for specific disease sites and clinical applications. Clinical studies are necessary for the quantification of benefit needed for evidence-based medicine.

Automatic prostate localization on cone-beam CT scans for high precision image-guided radiotherapy

PURPOSE

Previously, we developed an automatic three-dimensional gray-value registration (GR) method for fast prostate localization that could be used during online or offline image-guided radiotherapy. The method was tested on conventional computed tomography (CT) scans. In this study, the performance of the algorithm to localize the prostate on cone-beam CT (CBCT) scans acquired on the treatment machine was evaluated.

METHODS AND MATERIALS

Five to 17 CBCT scans of 32 prostate cancer patients (332 scans in total) were used. For 18 patients (190 CBCT scans), the CBCT scans were acquired with a collimated field of view (FOV) (craniocaudal). This procedure improved the image quality considerably. The prostate (i.e., prostate plus seminal vesicles) in each CBCT scan was registered to the prostate in the planning CT scan by automatic 3D gray-value registration (normal GR) starting from a registration on the bony anatomy. When these failed, registrations were repeated with a fixed rotation point locked at the prostate apex (fixed apex GR). Registrations were visually assessed in 3D by one observer with the help of an expansion (by 3.6 mm) of the delineated prostate contours of the planning CT scan. The percentage of successfully registered cases was determined from the combined normal and fixed apex GR assessment results. The error in gray-value registration for both registration methods was determined from the position of one clearly defined calcification in the prostate gland (9 patients, 71 successful registrations).

RESULTS

The percentage of successfully registered CBCT scans that were acquired with a collimated FOV was about 10% higher than for CBCT scans that were acquired with an uncollimated FOV. For CBCT scans that were acquired with a collimated FOV, the percentage of successfully registered cases improved from 65%, when only normal GR was applied, to 83% when the results of normal and fixed apex GR were combined. Gray-value registration mainly failed (or registrations were difficult to assess) because of streaks in the CBCT scans caused by moving gas pockets in the rectum during CBCT image acquisition (i.e., intrafraction motion). The error in gray-value registration along the left-right, craniocaudal, and anteroposterior axes was 1.0, 2.4, and 2.3 mm (1 SD) for normal GR, and 1.0, 2.0, and 1.7 mm (1 SD) for fixed apex GR. The systematic and random components of these SDs contributed approximately equally to these SDs, for both registration methods.

CONCLUSIONS

The feasibility of automatic prostate localization on CBCT scans acquired on the treatment machine using an adaptation of the previously developed three-dimensional gray-value registration algorithm, has been validated in this study. Collimating the FOV during CBCT image acquisition improved the CBCT image quality considerably. Artifacts in the CBCT images caused by large moving gas pockets during CBCT image acquisition were the main cause for unsuccessful registration. From this study, we can conclude that CBCT scans are suitable for online and offline position verification of the prostate, as long as the amount of nonstationary gas is limited.

Target localization for post-prostatectomy patients using CT and ultrasound image guidance

We conducted a study comparing B‐mode acquisition and targeting (BAT) ultrasound alignments based on CT data in the postoperative setting. CT scans were obtained with a Primatom CT‐on‐rails on nine patients. Two CT scans were obtained each week, while setup error was minimized by BAT ultrasounds. For the first three patients, a direct comparison was performed. For the next six patients, a template based on the shifts from the week 1 CT during treatment was used for subsequent setup. Comparison of isocenter shifts between the BAT ultrasound and CT was made by the difference, absolute difference, and improvement (using CT alignments as the reference technique). A total of 90 image comparisons were made. The average interfraction motion was 3.2 mm in the lateral, 3.0 mm in the longitudinal, and 5.1 mm in the AP direction. The results suggest that the CT‐based ultrasound templates can improve the localization of the prostate bed when the initial displacements are greater than 4 mm. For initial displacements smaller than 4 mm, the technique neither improved nor worsened target localization. However, ultrasound alignments performed without the use of a template deteriorated patient positioning for two out of three patients, demonstrating that the use of a CT template was beneficial even at small initial displacements.

PACS numbers: 87.53.‐j, 87.53.Kn, 87.53.Xd

Prostate position relative to pelvic bony anatomy based on intraprostatic gold markers and electronic portal imaging

PURPOSE

To describe the relative positions and motions of the prostate, pelvic bony anatomy, and intraprostatic gold fiducial markers during daily electronic portal localization of the prostate.

METHODS AND MATERIALS

Twenty prostate cancer patients were treated supine with definitive external radiotherapy according to an on-line target localization protocol using three or four intraprostatic gold fiducial markers and an electronic portal imaging device. Daily pretherapy and through-treatment electronic portal images (EPIs) were obtained for each of four treatment fields. The patients' pelvic bony anatomy, intraprostatic gold markers, and a best visual match to the target (i.e., prostate) were identified on simulation digitally reconstructed radiographs and during daily treatment setup and delivery. These data provided quantitative inter- and intrafractional analysis of prostate motion, its position relative to the bony anatomy, and the individual intraprostatic fiducial markers. Treatment planning margins, with and without on-line localization, were subsequently compared.

RESULTS

A total of 22,266 data points were obtained from daily pretherapy and through-treatment EPIs. The pretherapy three-dimensional (3D) average displacement of the fiducial markers, as a surrogate for the prostate, was 5.6 mm, which improved to 2.8 mm after use of the localization protocol. The bony anatomy 3D average displacement was 4.4 mm both before and after localization to the prostate (p = 0.46). Along the superior-inferior (SI), anterior-posterior (AP), and right-left (RL) axes, the average prostate displacement improved from 2.5, 3.7, and 1.9 mm, respectively, before localization to 1.4, 1.6, and 1.1 mm after (all p < 0.001). The pretherapy to through-treatment position of the bony landmarks worsened from 1.7 to 2.5 mm (p < 0.001) in the SI axis, remained statistically unchanged at 2.8 mm (p = 0.39) in the AP axis, and improved from 2.0 to 1.2 mm in the RL axis (p < 0.001). There was no significant intrafractional displacement of prostate position or bony anatomic landmarks. An intermarker distance was identified for all fiducial markers, and 96 were followed daily. Seventy-nine percent had a standard deviation of <1 mm, and 96% were <1.5 mm. Margins were 5.1, 7.3, and 5.0 mm in the SI, AP, and RL axes, respectively, before localization and 2.7, 2.9, and 2.8 mm after localization.

CONCLUSIONS

Significant interfractional motion exists for patients' prostate and pelvic bony anatomy. However, these move independently, so the pelvic bony anatomy should not be used as a surrogate for prostate motion. Fiducial markers are stable within the prostate and allow significant margin reduction when used for on-line localization of the prostate.

An automatic CT-guided adaptive radiation therapy technique by online modification of multileaf collimator leaf positions for prostate cancer

PURPOSE

To propose and evaluate online adaptive radiation therapy (ART) using in-room computed tomography (CT) imaging that detects changes in the target position and shape of the prostate and seminal vesicles (SVs) and then automatically modifies the multileaf collimator (MLC) leaf pairs in a slice-by-slice fashion.

METHODS AND MATERIALS

For intensity-modulated radiation therapy (IMRT) using a coplanar beam arrangement, each MLC leaf pair projects onto a specific anatomic slice. The proposed strategy assumes that shape deformation is a function of only the superior-inferior (SI) position. That is, there is no shape change within a CT slice, but each slice can be displaced in the anteroposterior (AP) or right-left (RL) direction relative to adjacent slices. First, global shifts (in SI, AP, and RL directions) were calculated by three-dimensional (3D) registration of the bulk of the prostate in the treatment planning CT images with the daily CT images taken immediately before treatment. Local shifts in the AP direction were then found using slice-by-slice registration, in which the CT slices were individually registered. The translational shift within a slice could then be projected to a translational shift in the position of the corresponding MLC leaf pair for each treatment segment for each gantry angle. Global shifts in the SI direction were accounted for by moving the open portal superiorly or inferiorly by an integral number of leaf pairs. The proposed slice-by-slice registration technique was tested by using daily CT images from 46 CT image sets (23 each from 2 patients) taken before the standard delivery of IMRT for prostate cancer. A dosimetric evaluation was carried out by using an 8-field IMRT plan.

RESULTS

The shifts and shape change of the prostate and SVs could be separated into 3D global shifts in the RL, AP, and SI directions, plus local shifts in the AP direction, which were different for each CT slice. The MLC leaf positions were successfully modified to compensate for these global shifts and local shape variations. The ART method improved geometric coverage of the prostate and SVs compared with the couch-shift method, particularly for the superior part of the prostate and all the SVs, for which the interfraction shape change was the largest. The dosimetric comparison showed that the ART method covered the target better and reduced the rectal dose more than a simple couch-translation method.

CONCLUSIONS

ART corrected for interfraction changes in the position and shape of the prostate and SVs and gave dose distributions that were considerably closer to the planned dose distributions than could be achieved with simple alignment strategies that neglect shape change. The ART proposed in this investigation requires neither contouring of the daily CT images nor extensive calculations; therefore, it may prove to be an effective and clinically practical solution to the problem of interfraction shape changes.

Application of the No Action Level (NAL) protocol to correct for prostate motion based on electronic portal imaging of implanted markers

PURPOSE

To evaluate the efficacy of the No Action Level (NAL) off-line correction protocol in the reduction of systematic prostate displacements as determined from electronic portal images (EPI) using implanted markers.

METHODS AND MATERIALS

Four platinum markers, two near the apex and two near the base of the prostate, were implanted for localization purposes in patients who received fractionated high dose rate brachytherapy. During the following course of 25 fractions of external beam radiotherapy, the position of each marker relative to the corresponding position in digitally reconstructed radiographs (DRRs) was measured in EPI in 15 patients for on average 17 fractions per patient. These marker positions yield the composite displacements due to both setup error and internal prostate motion, relative to the planning computed tomography scan. As the NAL protocol is highly effective in reducing systematic errors (recurring each fraction) due to setup inaccuracy alone, we investigated its efficacy in reducing systematic composite displacements. The analysis was performed for the center of mass (COM) of the four markers, as well as for the cranial and caudal markers separately. Furthermore, the impact of prostate rotation on the achieved positioning accuracy was determined.

RESULTS

In case of no setup corrections, the standard deviations of the systematic composite displacements of the COM were 3-4 mm in the craniocaudal and anterior-posterior directions, and 2 mm in the left-right direction. The corresponding SDs of the random displacements (interfraction fluctuations) were 2-3 mm in each direction. When applying a NAL protocol based on three initial treatment fractions, the SDs of the systematic COM displacements were reduced to 1-2 mm. Displacements at the cranial end of the prostate were slightly larger than at the caudal end, and quantitative analysis showed this originates from left-right axis rotations about the prostate apex. Further analysis revealed that significant time trends are present in these prostate rotations. No significant trends were observed for the prostate translations.

CONCLUSIONS

The NAL protocol based on marker positions in EPI halved the composite systematic displacements using only three imaged fractions per patient, and thus allowed for a significant reduction of planning margins. Although large rotations of the prostate, and time trends therein, were observed, the net impact on the measured displacements and on the accuracy obtained with NAL was small.

Model-based segmentation of medical imagery by matching distributions

The segmentation of deformable objects from three-dimensional (3-D) images is an important and challenging problem, especially in the context of medical imagery. We present a new segmentation algorithm based on matching probability distributions of photometric variables that incorporates learned shape and appearance models for the objects of interest. The main innovation over similar approaches is that there is no need to compute a pixelwise correspondence between the model and the image. This allows for a fast, principled algorithm. We present promising results on difficult imagery for 3-D computed tomography images of the male pelvis for the purpose of image-guided radiotherapy of the prostate.

Automatic localization of the prostate for on-line or off-line image-guided radiotherapy

PURPOSE

With higher radiation dose, higher cure rates have been reported in prostate cancer patients. The extra margin needed to account for prostate motion, however, limits the level of dose escalation, because of the presence of surrounding organs at risk. Knowledge of the precise position of the prostate would allow significant reduction of the treatment field. Better localization of the prostate at the time of treatment is therefore needed, e.g. using a cone-beam computed tomography (CT) system integrated with the linear accelerator. Localization of the prostate relies upon manual delineation of contours in successive axial CT slices or interactive alignment and is fairly time-consuming. A faster method is required for on-line or off-line image-guided radiotherapy, because of prostate motion, for patient throughput and efficiency. Therefore, we developed an automatic method to localize the prostate, based on 3D gray value registration.

METHODS AND MATERIALS

A study was performed on conventional repeat CT scans of 19 prostate cancer patients to develop the methodology to localize the prostate. For each patient, 8-13 repeat CT scans were made during the course of treatment. First, the planning CT scan and the repeat CT scan were registered onto the rigid bony structures. Then, the delineated prostate in the planning CT scan was enlarged by an optimum margin of 5 mm to define a region of interest in the planning CT scan that contained enough gray value information for registration. Subsequently, this region was automatically registered to a repeat CT scan using 3D gray value registration to localize the prostate. The performance of automatic prostate localization was compared to prostate localization using contours. Therefore, a reference set was generated by registering the delineated contours of the prostates in all scans of all patients. Gray value registrations that showed large differences with respect to contour registrations were detected with a chi(2) analysis and were removed from the data set before further analysis.

RESULTS

Comparing gray value registration to contour registration, we found a success rate of 91%. The accuracy for rotations around the left-right, cranial-caudal, and anterior-posterior axis was 2.4 degrees, 1.6 degrees, and 1.3 degrees (1 SD), respectively, and for translations along these axes 0.7, 1.3, and 1.2 mm (1 SD), respectively. A large part of the error is attributed to uncertainty in the reference contour set. Automatic prostate localization takes about 45 seconds on a 1.7 GHz Pentium IV personal computer.

CONCLUSIONS

This newly developed method localizes the prostate quickly, accurately, and with a good success rate, although visual inspection is still needed to detect outliers. With this approach, it will be possible to correct on-line or off-line for prostate movement. Combined with the conformity of intensity-modulated dose distributions, this method might permit dose escalation beyond that of current conformal approaches, because margins can be safely reduced.

Evaluation of a contour-alignment technique for CT-guided prostate radiotherapy: an intra- and interobserver study

PURPOSE

The recent introduction of integrated CT/linear accelerator systems may mean that daily CT localization can become a reality in the clinic, possibly allowing further dose escalation to the prostate while limiting unwanted doses to the rectum and bladder. However, the implementation of CT localization is currently impeded by the lack of precise and robust techniques to align the treatment plan with the daily CT images. The purpose of this study was to evaluate a manual alignment technique, in which the gross target volume contours are overlaid on the daily CT images and then shifted to match the structures visible in the images.

METHODS AND MATERIALS

A total of 28 CT image sets were taken before the standard delivery of intensity-modulated radiotherapy for prostate cancer for 2 patients. Seven observers (four radiation oncologists and three medical physicists) manually shifted the gross target volume contours from the treatment plan to best match the daily CT images. One observer repeated the process 1 week later to evaluate intraobserver variations. The experiment was then repeated, but the CT images from the original treatment plan were used as a reference to reduce interobserver uncertainty when aligning the contours. The shifts in prostate position found by different observers, both with and without reference data, were evaluated using a factorial analysis of variance to determine the standard errors of measurement for the intra- and interobserver uncertainty (SEM(intra) and SEM(inter), respectively). The differences in the SEM for the two groups of observers (radiation oncologists and medical physicists), the two alignment techniques (with and without reference information), and the two patients were evaluated using the t test at 90% confidence levels.

RESULTS

With no reference information, the SEM(inter) using one patient data set (Patient 1) was 0.8 mm, 2.0 mm, and 2.2 mm in the right-left (RL), anterior-posterior (AP), and superior-inferior (SI) directions, respectively. The use of the treatment plan as a reference reduced the SEM(inter) to 0.7 mm, 1.0 mm, and 1.6 mm in the RL, AP, and SI directions, respectively. In Patient 2, localization of the prostate was more difficult; the best SEM(inter) achieved with this patient was 0.8 mm, 1.9 mm, and 2.0 mm in the RL, AP, and SI directions, respectively. The SEM(intra) values with Patient 1 were also slightly better than with Patient 2. When reference data were used, the SEM(intra) value was 0.5 mm, 0.7 mm, and 0.5 mm for Patient 1 and 0.6 mm, 1.0 mm, and 0.7 mm for Patient 2 in the RL, AP, and SI directions, respectively. Despite the larger than expected interobserver variation reported here, the SEM(inter) was smaller than the typical day-to-day variation in prostate position. The contour alignment technique may still be useful to aid daily prostate localization or in a correction scheme to minimize the effect of target positional error.

CONCLUSION

The interobserver uncertainties associated with aligning the gross target volume contours with daily CT images were sufficiently small that this method may be used for daily CT localization of the prostate. The use of a reference image is important to improve the consistency among different users in this technique.

Intensity-modulated Stereotactic Radiotherapy of Paraspinal Tumors: A Preliminary Report

OBJECTIVE

Radioresistant paraspinal tumors may benefit from conformal treatment techniques such as intensity-modulated radiotherapy (IMRT). Local tumor control and long-term palliation for both primary and metastatic tumors may be achieved with IMRT while reducing the risk of spinal cord toxicity associated with conventional radiotherapy techniques. In this article, we report our initial clinical experience in treating 16 paraspinal tumors with IMRT in which the planning target volume was 2 mm or greater from the spinal cord.

METHODS

IMRT was administered by using a linear accelerator mounted with a multileaf collimator. Two immobilization body frames developed at Memorial Sloan-Kettering Cancer Center were used for patients with and without spinal implants. During a 30-month period, 16 patients underwent IMRT for metastatic and primary tumors. Eleven patients were treated for symptomatic recurrences after undergoing surgery and prior external beam radiotherapy, and one patient was treated after undergoing radiotherapy for a metastatic pancreatic gastrinoma with overlapping ports to the spine. Four patients with primary tumors were treated after primary resection that resulted in positive histological margins. Twelve patients were symptomatic with pain, functional radiculopathy, or both. Tumoral doses were determined on the basis of the relative radiosensitivity of tumors. Patients with metastatic tumors were administered a median tumoral dose of 20 Gy in four to five fractions and a spinal cord maximum dose of 6.0 Gy in addition to the full tolerance dose administered in previous radiation treatments. The primary tumors were delivered a median dose of 70 Gy in 33 to 37 fractions and a spinal cord maximum dose of 16 Gy. The median tumoral volume was 7.8 cm3.

RESULTS

Of the 15 patients who underwent radiographic follow-up, 13 demonstrated either no interval growth or a reduction in tumor size in a median follow-up period of 12 months (range, 2–23 mo). Two patients, one with a thoracic chondrosarcoma and one with a chordoma, showed tumor progression 1 year after undergoing IMRT. Pain symptoms improved in 11 of 11 patients, and 4 of 4 patients had significant improvement in their functionally significant radiculopathy and/or plexopathy. Pain relief was durable in all patients except the two with tumor progression. No patient showed signs or symptoms of radiation-induced myelopathy, radiculopathy, or plexopathy, including 12 patients with a median follow-up of 18 months.

CONCLUSION

IMRT was effective for treating pain and improving functional radiculopathy in patients with metastatic and primary tumors. Although long-term tumor control is not established in this study, high-dose tumoral irradiation can be performed without causing radiation myelopathy in more than 1 year of follow-up.

Daily target localization for prostate patients based on 3D image correlation

There are several localization techniques that have been used for prostate treatment. Recently, the potential use of a variety of CT-based equipment in the treatment room has been discussed. The goal of our study was to develop an automated procedure for daily treatment table shift calculation based on two CT data sets: simulation CT data and localization CT data. The method suggested in this study is a 3D image cross-correlation of small regions of interest (ROI) within the two data sets. The relative position of the two ROIs with respect to each other is determined by the maximum value of the normalized cross-correlation function, calculated for all possible relative locations of the two ROIs. After the best match is found the shifts are given by the vector connecting the treatment isocentre and the planning isocentre (both determined by the radio opaque fiducial markers on the patient's skin). The results have been compared with shifts calculated through manual fusion. The shift differences, averaged over 17 statistically independent shift calculations, are less then 1 mm in the lateral and longitudinal directions, and about 1 mm in the AP direction. The impact of image noise on the performance of the algorithm has been tested. The results show that the algorithm accurately adjusts for target positional changes even with Gaussian noise levels as high as 20% inserted.

Automatic registration of the prostate for computed-tomography-guided radiotherapy

The recent development of integrated computed tomography (CT)/linear accelerator (LINAC) combinations, where the CT scanner and the LINAC use the same patient couch, and of kilovoltage cone-beam CT systems attached to the LINAC gantry, means that suitable hardware is now available for CT-guided localization of the prostate. Clinical implementation is, however, currently impeded by the lack of robust and accurate software tools to compare the position of the prostate in the CT images used for the treatment plan with its position in the daily CT images. Manual registration of the planning CT images with the daily CT images can be slow and can introduce significant inter-user variations. We have developed an automatic registration technique that is not adversely influenced by changes in prostate shape, size or orientation, presence of rectal gas, or bladder filling. The cost function used in the registration is the mean absolute difference in CT numbers voxel-by-voxel between the daily CT image and the planning CT image for a volume extracted from the planning CT images using the original physician-drawn gross tumor volume contours. To enhance soft tissue contrast in the prostate region and to reduce the impact of rectal gas calcifications and bone on the registration, voxels with CT numbers that represent gas or bone are filtered out from the calculation. The results of the automatic registration agreed with the mean results of seven human observers, with standard deviations of 0.5 mm, 0.5 mm, and 1.0 mm in the left/right (RL), anterior/posterior (AP), and superior/inferior (SI) directions, respectively, for a patient that was relatively easy to localize. Agreement (one standard deviation) for a patient that was difficult to localize was 0.6 mm, 1.4 mm, and 1.9 mm in the RL, AP, and SI directions, respectively. These results are better than the interuser uncertainties reported for a manual alignment technique and are close to the reported intrauser uncertainties. The results are independent of the shape of contours in the original treatment plan, reducing the impact of interobserver variations in contouring the prostate. The algorithm is fast and reliable, allowing the entire CT localization process to take place in 5-9 minutes. In 120 CT image sets from seven patients, the failure rate was found to be less than 1%. The use of this algorithm will facilitate the clinical implementation of CT-guided localization of the prostate.

Evaluation of mechanical precision and alignment uncertainties for an integrated CT/LINAC system

A new integrated CT/LINAC combination, in which the CT scanner is inside the radiation therapy treatment room and the same patient couch is used for CT scanning and treatment (after a 180-degree couch rotation), should allow for accurate correction of interfractional setup errors. The purpose of this study was to evaluate the sources of uncertainties, and to measure the overall precision of this system. The following sources of uncertainty were identified: (1) the patient couch position on the LINAC side after a rotation, (2) the patient couch position on the CT side after a rotation, (3) the patient couch position as indicated by its digital readout, (4) the difference in couch sag between the CT and LINAC positions, (5) the precision of the CT coordinates, (6) the identification of fiducial markers from CT images, (7) the alignment of contours with structures in the CT images, and (8) the alignment with setup lasers. The largest single uncertainties (one standard deviation or 1 SD) were found in couch position on the CT side after a rotation (0.5 mm in the RL direction) and the alignment of contours with the CT images (0.4 mm in the SI direction). All other sources of uncertainty are less than 0.3 mm (1 SD). The overall precision of two setup protocols was investigated in a controlled phantom study. A protocol that relies heavily on the mechanical integrity of the system, and assumes a fixed relationship between the LINAC isocenter and the CT images, gave a predicted precision (1 SD) of 0.6, 0.7, and 0.6 mm in the SI, RL and AP directions, respectively. The second protocol reduces reliance on the mechanical precision of the total system, particularly the patient couch, by using radio-opaque fiducial markers to transfer the isocenter information from the LINAC side to the CT images. This protocol gave a slightly improved predicted precision of 0.5, 0.4, and 0.4 mm in the SI, RL and AP directions, respectively. The distribution of phantom position after CT-based correction confirmed these results. Knowledge of the individual sources of uncertainty will allow alternative setup protocols to be evaluated in the future without the need for significant additional measurements.