Fast treatment plan modification with an over-relaxed Cimmino algorithm

Technical Note: A novel leaf sequencing optimization algorithm which considers previous underdose and overdose events for MLC tracking radiotherapy

PURPOSE

Multileaf collimator (MLC) tracking radiotherapy is complex as the beam pattern needs to be modified due to the planned intensity modulation as well as the real-time target motion. The target motion cannot be planned; therefore, the modified beam pattern differs from the original plan and the MLC sequence needs to be recomputed online. Current MLC tracking algorithms use a greedy heuristic in that they optimize for a given time, but ignore past errors. To overcome this problem, the authors have developed and improved an algorithm that minimizes large underdose and overdose regions. Additionally, previous underdose and overdose events are taken into account to avoid regions with high quantity of dose events.

METHODS

The authors improved the existing MLC motion control algorithm by introducing a cumulative underdose/overdose map. This map represents the actual projection of the planned tumor shape and logs occurring dose events at each specific regions. These events have an impact on the dose cost calculation and reduce recurrence of dose events at each region. The authors studied the improvement of the new temporal optimization algorithm in terms of the L1-norm minimization of the sum of overdose and underdose compared to not accounting for previous dose events. For evaluation, the authors simulated the delivery of 5 conformal and 14 intensity-modulated radiotherapy (IMRT)-plans with 7 3D patient measured tumor motion traces.

RESULTS

Simulations with conformal shapes showed an improvement of L1-norm up to 8.5% after 100 MLC modification steps. Experiments showed comparable improvements with the same type of treatment plans.

CONCLUSIONS

A novel leaf sequencing optimization algorithm which considers previous dose events for MLC tracking radiotherapy has been developed and investigated. Reductions in underdose/overdose are observed for conformal and IMRT delivery.

A method for generating large datasets of organ geometries for radiotherapy treatment planning studies

Background

With the rapidly increasing application of adaptive radiotherapy, large datasets of organ geometries based on the patient’s anatomy are desired to support clinical application or research work, such as image segmentation, re-planning, and organ deformation analysis. Sometimes only limited datasets are available in clinical practice. In this study, we propose a new method to generate large datasets of organ geometries to be utilized in adaptive radiotherapy.

Methods

Given a training dataset of organ shapes derived from daily cone-beam CT, we align them into a common coordinate frame and select one of the training surfaces as reference surface. A statistical shape model of organs was constructed, based on the establishment of point correspondence between surfaces and non-uniform rational B-spline (NURBS) representation. A principal component analysis is performed on the sampled surface points to capture the major variation modes of each organ.

Results

A set of principal components and their respective coefficients, which represent organ surface deformation, were obtained, and a statistical analysis of the coefficients was performed. New sets of statistically equivalent coefficients can be constructed and assigned to the principal components, resulting in a larger geometry dataset for the patient’s organs.

Conclusions

These generated organ geometries are realistic and statistically representative.

Adaptive replanning to account for lumpectomy cavity change in sequential boost after whole-breast irradiation

PURPOSE

To evaluate the efficiency of standard image-guided radiation therapy (IGRT) to account for lumpectomy cavity (LC) variation during whole-breast irradiation (WBI) and propose an adaptive strategy to improve dosimetry if IGRT fails to address the interfraction LC variations.

METHODS AND MATERIALS

Daily diagnostic-quality CT data acquired during IGRT in the boost stage using an in-room CT for 19 breast cancer patients treated with sequential boost after WBI in the prone position were retrospectively analyzed. Contours of the LC, treated breast, ipsilateral lung, and heart were generated by populating contours from planning CTs to boost fraction CTs using an auto-segmentation tool with manual editing. Three plans were generated on each fraction CT: (1) a repositioning plan by applying the original boost plan with the shift determined by IGRT; (2) an adaptive plan by modifying the original plan according to a fraction CT; and (3) a reoptimization plan by a full-scale optimization.

RESULTS

Significant variations were observed in LC. The change in LC volume at the first boost fraction ranged from a 70% decrease to a 50% increase of that on the planning CT. The adaptive and reoptimization plans were comparable. Compared with the repositioning plans, the adaptive plans led to an improvement in target coverage for an increased LC case (1 of 19, 7.5% increase in planning target volume evaluation volume V95%), and breast tissue sparing for an LC decrease larger than 35% (3 of 19, 7.5% decrease in breast evaluation volume V50%; P=.008).

CONCLUSION

Significant changes in LC shape and volume at the time of boost that deviate from the original plan for WBI with sequential boost can be addressed by adaptive replanning at the first boost fraction.

Characterization and management of interfractional anatomic changes for pancreatic cancer radiotherapy

PURPOSE

To quantitatively characterize interfractional anatomic variations in pancreatic cancer radiotherapy (RT) and to study dosimetric advantages for using an online adaptive replanning scheme to account for these variations.

METHODS AND MATERIALS

Targets and organs at risk (OAR) were delineated by autosegmentation based on daily computed tomography (CT) images acquired using a respiration-gated in-room CT during daily image-guided RT (IGRT) for 10 pancreatic cancer patients. Various parameters, including the maximum overlap ratio (MOR) between the volumes based on planning and daily CTs for a structure, while the overlapping volumes were maximized, were used to quantify the interfractional organ deformation with the intrafractional variations largely excluded. An online adaptive RT (ART) was applied to these daily CTs. To evaluate the dosimetric benefits of ART, the dose distributions from the online ART were compared to those from the repositioning in the current standard IGRT practice.

RESULTS

The interfractional anatomic variations, particularly the organ deformation, are significant during pancreas irradiation. For the patients studied, the average MORs of all daily CTs were 80.2%, 61.7%, and 72.2% for pancreatic head, duodenum, and stomach, respectively. The online ART leads to improved dosimetric plan with better target coverage and/or OAR sparing than IGRT repositioning. For the patients studied, the mean V(50.4 Gy) (volume covered by 50.4 Gy) for the duodenum was reduced from 43.4% for IGRT to 15.6% for the online ART scheme.

CONCLUSIONS

The online adaptive RT scheme can effectively account for the significant interfractional anatomic variations observed in pancreas irradiation. The dosimetric advantages with the online ART may enable safe dose escalation in radiation therapy for pancreatic cancer.

Ultra-Fast Digital Tomosynthesis Reconstruction Using General-Purpose GPU Programming for Image-Guided Radiation Therapy

The purpose of this work is to demonstrate an ultra-fast reconstruction technique for digital tomosynthesis (DTS) imaging based on the algorithm proposed by Feldkamp, Davis, and Kress (FDK) using standard general-purpose graphics processing unit (GPGPU) programming interface. To this end, the FDK-based DTS algorithm was programmed “in-house” with C language with utilization of 1) GPU and 2) central processing unit (CPU) cards. The GPU card consisted of 480 processing cores (2 × 240 dual chip) with 1,242 MHz processing clock speed and 1,792 MB memory space. In terms of CPU hardware, we used 2.68 GHz clock speed, 12.0 GB DDR3 RAM, on a 64-bit OS. The performance of proposed algorithm was tested on twenty-five patient cases (5 lung, 5 liver, 10 prostate, and 5 head-and-neck) scanned either with a full-fan or half-fan mode on our cone-beam computed tomography (CBCT) system. For the full-fan scans, the projections from 157.5°–202.5° (45°-scan) were used to reconstruct coronal DTS slices, whereas for the half-fan scans, the projections from both 157.5°–202.5° and 337.5°–22.5° (2 × 45°-scan) were used to reconstruct larger FOV coronal DTS slices. For this study, we chose 45°-scan angle that contained ~80 projections for the full-fan and ~160 projections with 2 × 45°-scan angle for the half-fan mode, each with 1024 × 768 pixels with 32-bit precision. Absolute pixel value differences, profiles, and contrast-to-noise ratio (CNR) calculations were performed to compare and evaluate the images reconstructed using GPU- and CPU-based implementations. The time dependence on the reconstruction volume was also tested with (512 × 512) × 16, 32, 64, 128, and 256 slices. In the end, the GPU-based implementation achieved, at most, 1.3 and 2.5 seconds to complete full reconstruction of 512 × 512 × 256 volume, for the full-fan and half-fan modes, respectively. In turn, this meant that our implementation can process > 13 projections-per-second (pps) and > 18 pps for the full-fan and half-fan modes, respectively. Since commercial CBCT system nominally acquires 11 pps (with 1 gantry-revolution-per-minute), our GPU-based implementation is sufficient to handle the incoming projections data as they are acquired and reconstruct the entire volume immediately after completing the scan. In addition, on increasing the number of slices (hence volume) to be reconstructed from 16 to 256, only minimal increases in reconstruction time were observed for the GPU-based implementation where from 0.73 to 1.27 seconds and 1.42 to 2.47 seconds increase were observed for the full-fan and half-fan modes, respectively. This resulted in speed improvement of up to 87 times compared with the CPU-based implementation (for 256 slices case), with visually identical images and small pixel-value discrepancies (< 6.3%), and CNR differences (< 2.3%). With this achievement, we have shown that time allocation for DTS image reconstruction is virtually eliminated and that clinical implementation of this approach has become quite appealing. In addition, with the speed achievement, further image processing and real-time applications that was prohibited prior due to time restrictions can now be tempered with.

A GPU-based finite-size pencil beam algorithm with 3D-density correction for radiotherapy dose calculation

Targeting at the development of an accurate and efficient dose calculation engine for online adaptive radiotherapy, we have implemented a finite-size pencil beam (FSPB) algorithm with a 3D-density correction method on graphics processing unit (GPU). This new GPU-based dose engine is built on our previously published ultrafast FSPB computational framework (Gu et al 2009 Phys. Med. Biol. 54 6287-97). Dosimetric evaluations against Monte Carlo dose calculations are conducted on ten IMRT treatment plans (five head-and-neck cases and five lung cases). For all cases, there is improvement with the 3D-density correction over the conventional FSPB algorithm and for most cases the improvement is significant. Regarding the efficiency, because of the appropriate arrangement of memory access and the usage of GPU intrinsic functions, the dose calculation for an IMRT plan can be accomplished well within 1 s (except for one case) with this new GPU-based FSPB algorithm. Compared to the previous GPU-based FSPB algorithm without 3D-density correction, this new algorithm, though slightly sacrificing the computational efficiency (∼5-15% lower), has significantly improved the dose calculation accuracy, making it more suitable for online IMRT replanning.

On-line adaptive radiation therapy: feasibility and clinical study

The purpose of this paper is to evaluate the feasibility and clinical dosimetric benefit of an on-line, that is, with the patient in the treatment position, Adaptive Radiation Therapy (ART) system for prostate cancer treatment based on daily cone-beam CT imaging and fast volumetric reoptimization of treatment plans. A fast intensity-modulated radiotherapy (IMRT) plan reoptimization algorithm is implemented and evaluated with clinical cases. The quality of these adapted plans is compared to the corresponding new plans generated by an experienced planner using a commercial treatment planning system and also evaluated by an in-house developed tool estimating achievable dose-volume histograms (DVHs) based on a database of existing treatment plans. In addition, a clinical implementation scheme for ART is designed and evaluated using clinical cases for its dosimetric qualities and efficiency.

Implementation and evaluation of various demons deformable image registration algorithms on a GPU

Online adaptive radiation therapy (ART) promises the ability to deliver an optimal treatment in response to daily patient anatomic variation. A major technical barrier for the clinical implementation of online ART is the requirement of rapid image segmentation. Deformable image registration (DIR) has been used as an automated segmentation method to transfer tumor/organ contours from the planning image to daily images. However, the current computational time of DIR is insufficient for online ART. In this work, this issue is addressed by using computer graphics processing units (GPUs). A gray-scale-based DIR algorithm called demons and five of its variants were implemented on GPUs using the compute unified device architecture (CUDA) programming environment. The spatial accuracy of these algorithms was evaluated over five sets of pulmonary 4D CT images with an average size of 256 x 256 x 100 and more than 1100 expert-determined landmark point pairs each. For all the testing scenarios presented in this paper, the GPU-based DIR computation required around 7 to 11 s to yield an average 3D error ranging from 1.5 to 1.8 mm. It is interesting to find out that the original passive force demons algorithms outperform subsequently proposed variants based on the combination of accuracy, efficiency and ease of implementation.

GPU-based ultra-fast dose calculation using a finite size pencil beam model

Online adaptive radiation therapy (ART) is an attractive concept that promises the ability to deliver an optimal treatment in response to the inter-fraction variability in patient anatomy. However, it has yet to be realized due to technical limitations. Fast dose deposit coefficient calculation is a critical component of the online planning process that is required for plan optimization of intensity-modulated radiation therapy (IMRT). Computer graphics processing units (GPUs) are well suited to provide the requisite fast performance for the data-parallel nature of dose calculation. In this work, we develop a dose calculation engine based on a finite-size pencil beam (FSPB) algorithm and a GPU parallel computing framework. The developed framework can accommodate any FSPB model. We test our implementation in the case of a water phantom and the case of a prostate cancer patient with varying beamlet and voxel sizes. All testing scenarios achieved speedup ranging from 200 to 400 times when using a NVIDIA Tesla C1060 card in comparison with a 2.27 GHz Intel Xeon CPU. The computational time for calculating dose deposition coefficients for a nine-field prostate IMRT plan with this new framework is less than 1 s. This indicates that the GPU-based FSPB algorithm is well suited for online re-planning for adaptive radiotherapy.

Objective assessment of deformable image registration in radiotherapy: a multi-institution study

The looming potential of deformable alignment tools to play an integral role in adaptive radiotherapy suggests a need for objective assessment of these complex algorithms. Previous studies in this area are based on the ability of alignment to reproduce analytically generated deformations applied to sample image data, or use of contours or bifurcations as ground truth for evaluation of alignment accuracy. In this study, a deformable phantom was embedded with 48 small plastic markers, placed in regions varying from high contrast to roughly uniform regional intensity, and small to large regional discontinuities in movement. CT volumes of this phantom were acquired at different deformation states. After manual localization of marker coordinates, images were edited to remove the markers. The resulting image volumes were sent to five collaborating institutions, each of which has developed previously published deformable alignment tools routinely in use. Alignments were done, and applied to the list of reference coordinates at the inhale state. The transformed coordinates were compared to the actual marker locations at exhale. A total of eight alignment techniques were tested from the six institutions. All algorithms performed generally well, as compared to previous publications. Average errors in predicted location ranged from 1.5 to 3.9 mm, depending on technique. No algorithm was uniformly accurate across all regions of the phantom, with maximum errors ranging from 5.1 to 15.4 mm. Larger errors were seen in regions near significant shape changes, as well as areas with uniform contrast but large local motion discontinuity. Although reasonable accuracy was achieved overall, the variation of error in different regions suggests caution in globally accepting the results from deformable alignment.

Dosimetric evaluation of daily rigid and nonrigid geometric correction strategies during on-line image-guided radiation therapy (IGRT) of prostate cancer

The purpose of this study is to evaluate a geometric image guidance strategy that simultaneously correct for various inter-fractional rigid and nonrigid geometric uncertainties in an on-line environment, using field shape corrections (called the "MU-MLC" technique). The effectiveness of this strategy was compared with two other simpler on-line image guidance strategies that are more commonly used in the clinic. To this end, five prostate cancer patients, with at least 15 treatment CT studies each, were analyzed. The prescription dose was set to the maximum dose that did not violate the rectum and bladder dose-volume constraints, and hence, was unique to each patient. Deformable image registration and dose-tracking was performed on each CT image to obtain the cumulative treatment dose distributions. From this, maximum, minimum, and mean dose, as well as generalized equivalent uniform dose (gEUD) were calculated for each image guidance strategy. As expected, some dosimetric differences in the clinical target volume (CTV) were observed between the three image guidance strategies investigated. For example, up to +/-2% discrepancy in prostate minimum dose were observed among the techniques. Of them, only the "MU -MLC" technique did not reduce the prostate minimum dose for all patients (i.e., > or = 100%). However, the differences were clinically not significant to indicate the preference of one strategy over another, when using a uniform 5 mm margin size. For the organ-at-risks (OARs), the large rectum sparing effect (< or =5.7 Gy, gEUD) and bladder overdosing effect (< or = 16 Gy, gEUD) were observed. This was likely due to the use of bladder contrast during CT simulation studies which was not done during the treatment CT studies. Therefore, ultimately, strategies to maintain relatively constant rectum and bladder volumes, throughout the treatment course, are required to minimize this effect. In conclusion, the results here suggest that simple translational corrections based on three-dimensional (3D) images is adequate to maintain target coverage, for margin sizes at least as large as 5 mm. In addition, due to large fluctuations in OAR volumes, innovative image guidance strategies are needed to minimize dose and maintain consistent sparing during the whole course of radiation therapy.

Direct aperture deformation: An interfraction image guidance strategy

A new scheme, called direct aperture deformation (DAD), for online correction of interfraction geometric uncertainties under volumetric imaging guidance is presented. Using deformable image registration, the three-dimensional geometric transformation matrix can be derived that associates the planning image set and the images acquired on the day of treatment. Rather than replanning or moving the patient, we use the deformation matrix to morph the treatment apertures as a potential online correction method. A proof-of-principle study using an intensity-modulated radiation therapy plan for a prostate cancer patient was conducted. The method, procedure, and algorithm of DAD are described. The dose-volume histograms from the original plan, reoptimized plan, and rigid-body translation plan are compared with the ones from the DAD plan. The study showed the feasibility of the DAD as a general method for both target dislocation and deformation. As compared with using couch translation to move the patient, DAD is capable of correcting both target dislocation and deformations. As compared with reoptimization, online correction using the DAD scheme could be completed within a few minutes rather than tens of minutes and the speed gain would be at a very small cost of plan quality.

Dose-guided radiation therapy with megavoltage cone-beam CT

Recent advances in fractionated external beam radiation therapy have increased our ability to deliver radiation doses that conform more tightly to the tumour volume. The steeper dose gradients delivered in these treatments make it increasingly important to set precisely the positions of the patient and the internal organs. For this reason, considerable research now focuses on methods using three-dimensional images of the patient on the treatment table to adapt either the patient position or the treatment plan, to account for variable organ locations. In this article, we briefly review the different adaptive methods being explored and discuss a proposed dose-guided radiation therapy strategy that adapts the treatment for future fractions to compensate for dosimetric errors from past fractions. The main component of this strategy is a procedure to reconstruct the dose delivered to the patient based on treatment-time portal images and pre-treatment megavoltage cone-beam computed tomography (MV CBCT) images of the patient. We describe the work to date performed to develop our dose reconstruction procedure, including the implementation of a MV CBCT system for clinical use, experiments performed to calibrate MV CBCT for electron density and to use the calibrated MV CBCT for dose calculations, and the dosimetric calibration of the portal imager. We also present an example of a reconstructed patient dose using a preliminary reconstruction program and discuss the technical challenges that remain to full implementation of dose reconstruction and dose-guided therapy.

Deformable registration of the planning image (kVCT) and the daily images (MVCT) for adaptive radiation therapy

The incorporation of daily images into the radiotherapy process leads to adaptive radiation therapy (ART), in which the treatment is evaluated periodically and the plan is adaptively modified for the remaining course of radiotherapy. Deformable registration between the planning image and the daily images is a key component of ART. In this paper, we report our researches on deformable registration between the planning kVCT and the daily MVCT image sets. The method is based on a fast intensity-based free-form deformable registration technique. Considering the noise and contrast resolution differences between the kVCT and the MVCT, an 'edge-preserving smoothing' is applied to the MVCT image prior to the deformable registration process. We retrospectively studied daily MVCT images from commercial TomoTherapy machines from different clinical centers. The data set includes five head-neck cases, one pelvis case, two lung cases and one prostate case. Each case has one kVCT image and 20-40 MVCT images. We registered the MVCT images with their corresponding kVCT image. The similarity measures and visual inspections of contour matches by physicians validated this technique. The applications of deformable registration in ART, including 'deformable dose accumulation', 'automatic re-contouring' and 'tumour growth/regression evaluation' throughout the course of radiotherapy are also studied.

Fast voxel and polygon ray-tracing algorithms in intensity modulated radiation therapy treatment planning

We present work on combining three algorithms to improve ray-tracing efficiency in radiation therapy dose computation. The three algorithms include: An improved point-in-polygon algorithm, incremental voxel ray tracing algorithm, and stereographic projection of beamlets for voxel truncation. The point-in-polygon and incremental voxel ray-tracing algorithms have been used in computer graphics and nuclear medicine applications while the stereographic projection algorithm was developed by our group. These algorithms demonstrate significant improvements over the current standard algorithms in peer reviewed literature, i.e., the polygon and voxel ray-tracing algorithms of Siddon for voxel classification (point-in-polygon testing) and dose computation, respectively, and radius testing for voxel truncation. The presented polygon ray-tracing technique was tested on 10 intensity modulated radiation therapy (IMRT) treatment planning cases that required the classification of between 0.58 and 2.0 million voxels on a 2.5 mm isotropic dose grid into 1-4 targets and 5-14 structures represented as extruded polygons (a.k.a. Siddon prisms). Incremental voxel ray tracing and voxel truncation employing virtual stereographic projection was tested on the same IMRT treatment planning cases where voxel dose was required for 230-2400 beamlets using a finite-size pencil-beam algorithm. Between a 100 and 360 fold cpu time improvement over Siddon's method was observed for the polygon ray-tracing algorithm to perform classification of voxels for target and structure membership. Between a 2.6 and 3.1 fold reduction in cpu time over current algorithms was found for the implementation of incremental ray tracing. Additionally, voxel truncation via stereographic projection was observed to be 11-25 times faster than the radial-testing beamlet extent approach and was further improved 1.7-2.0 fold through point-classification using the method of translation over the cross product technique.

History of tomotherapy

Tomotherapy is the delivery of intensity modulated radiation therapy using rotational delivery of a fan beam in the manner of a CT scanner. In helical tomotherapy the couch and gantry are in continuous motion akin to a helical CT scanner. Helical tomotherapy is inherently capable of acquiring CT images of the patient in treatment position and using this information for image guidance. This review documents technological advancements of the field concentrating on the conceptual beginnings through to its first clinical implementation. The history of helical tomotherapy is also a story of technology migration from academic research to a university-industrial partnership, and finally to commercialization and widespread clinical use.

Deformable registration for image-guided radiation therapy

Ths paper examines several applications of deformable registration algorithms in the field of image-guided radiotherapy. The first part focuses on the description of input and output of deformable registration algorithms, with a brief review of conventional and most current methods. The typical applications of deformable registration are then reviewed on the basis of four practical examples. The first two sets of examples deal with the fusion of images obtained from the same patient (inter-fraction registration), with time intervals of several days between each image. The other two examples deal with the fusion of images obtained in immediate sequence (intra-fraction registration); in this case, the focus is the displacement during image acquisition or patient treatment (mainly due to respiratory movement), with time intervals in the order of magnitude of tenths of seconds. Finally, the registration of images of different patients (inter-patient registration) is also discussed. In conclusion, deformable registration has become a fundamental tool for image analysis in radiotherapy. Although extensive validation of the numerous existing methods is required before extending its clinical use, deformable registration is expected to become a standard methodology in the treatment planning systems in the near future.

L-curve analysis of radiotherapy optimization problems

We attempt to select an optimal value of regularization parameter in the optimization problems for intensity-modulated radiotherapy which are solved using a variational regularization technique. We apply to inverse treatment planning the L-curve method which was developed to determine the regularization parameter in the discrete ill-posed problems. The L-curve method is based on finding the regularization parameter which minimizes the residual norm which is a measure of accuracy of fit and the solution norm which is a measure of smoothness of solution. The main idea of the L-curve method is to plot the smoothing norm as a function of the residual norm for all values of the regularization parameter. This characteristic curve has an L-shaped dependence and the optimal value of regularization parameter can be found at the "corner" of the L-curve. We plot the L-curves for the optimization problems which simulate prostate radiotherapy cancer treatment with intensity-modulated beams. Different numerical methods are applied to calculate the point of maximum curvature of the L-curves which is a criterion to locate the corner. We show that the point of maximum curvature can be located in a most robust way using a formula derived from the singular value decomposition analysis.

Dosimetric impact of image-guided 3D conformal radiation therapy of prostate cancer

The goal of this work is to quantify the impact of image-guided conformal radiation therapy (CRT) on the dose distribution by correcting patient setup uncertainty and inter-fraction tumour motion. This was a retrospective analysis that used five randomly selected prostate cancer patients that underwent approximately 15 computed tomography (CT) scans during their radiation treatment course. The beam arrangement from the treatment plan was imported into each repeat CT study and the dose distribution was recalculated for the new beam setups. Various setup scenarios were then compared to assess the impact of image guidance on radiation treatment precision. These included (1) daily alignment to skin markers, thus representing a conventional beam setup without image guidance, (2) alignment to bony anatomy for correction of daily patient setup error, thus representing on-line portal image guidance, and (3) alignment to the 'CTV of the day' for correction of inter-fraction tumour motion, thus representing on-line CT or ultrasound image guidance. Treatment scenarios (1) and (3) were repeated with a reduced CTV to PTV margin, where the former represents a treatment using small margins without daily image guidance. Daily realignment of the treatment beams to the prostate showed an average increase in minimum tumour dose of 1.5 Gy, in all cases where tumour 'geographic miss' without image guidance was apparent. However, normal tissue sparing did not improve unless the PTV margin was reduced. Daily realignment to the tumour combined with reducing the margin size by a factor of 2 resulted in an average escalation in tumour dose of 9.0 Gy for all five static plans. However, the prescription dose could be escalated by 13.8 Gy when accounting for changes in anatomy by accumulating daily doses using nonlinear image registration techniques. These results provide quantitative information on the effectiveness of image-guided radiation treatment of prostate cancer and demonstrate that the dosimetric impact is patient dependent.

Image-guided adaptive radiation therapy (IGART): Radiobiological and dose escalation considerations for localized carcinoma of the prostate

The goal of this work was to evaluate the efficacy of various image-guided adaptive radiation therapy (IGART) techniques to deliver and escalate dose to the prostate in the presence of geometric uncertainties. Five prostate patients with 15-16 treatment CT studies each were retrospectively analyzed. All patients were planned with an 18 MV, six-field conformal technique with a 10 mm margin size and an initial prescription of 70 Gy in 35 fractions. The adaptive strategy employed in this work for patient-specific dose escalation was to increase the prescription dose in 2 Gy-per-fraction increments until the rectum normal tissue complication probability (NTCP) reached a level equal to that of the nominal plan NTCP (i.e., iso-NTCP dose escalation). The various target localization techniques simulated were: (1) daily laser-guided alignment to skin tattoo marks that represents treatment without image-guidance, (2) alignment to bony landmarks with daily portal images, and (3) alignment to the clinical target volume (CTV) with daily CT images. Techniques (1) and (3) were resimulated with a reduced margin size of 5 mm to investigate further dose escalation. When delivering the original clinical prescription dose of 70 Gy in 35 fractions, the "CTV registration" technique yielded the highest tumor control probability (TCP) most frequently, followed by the "bone registration" and "tattoo registration" techniques. However, the differences in TCP among the three techniques were minor when the margin size was 10 mm (< or = 1.1 %). Reducing the margin size to 5 mm significantly degraded the TCP values of the "tattoo registration" technique in two of the five patients, where a large difference was found compared to the other techniques (< or = 11.8 %). The "CTV registration" technique, however, did maintain similar TCP values compared to their 10 mm margin counterpart. In terms of normal tissue sparing, the technique producing the lowest NTCP varied from patient to patient. Reducing the margin size seemed the only sure way to reduce the NTCP significantly, irrespective of the IGART technique employed. In escalating the dose with the iso-NTCP constraint, the largest average gain in dose was observed with the "tattoo registration" technique, followed by the "CTV registration" and "bone registration" techniques. This is attributed to the fact that in three of the five patients, the "tattoo registration" technique yielded the lowest NTCP, hence a greater window of opportunity to escalate the dose was possible with this technique. However, the variation among the five patients was also largest with the "tattoo registration" technique where, in the case of one patient, the required dose actually needed to be below the original prescription dose of 70 Gy to satisfy the iso-NTCP constraint. This was not the case with the "CTV registration" technique where positive and similar dose escalation was allowed on all five patients. Based on these data, an attractive dose escalation strategy may be to implement the "CTV registration" technique (for consistent dosimetric coverage) for daily target localization in combination with a margin reduction (for increased normal tissue sparing).

Dose sculpting with generalized equivalent uniform dose

With intensity-modulated radiotherapy (IMRT), a variety of user-defined dose distribution can be produced using inverse planning. The generalized equivalent uniform dose (gEUD) has been used in IMRT optimization as an alternative objective function to the conventional dose-volume-based criteria. The purpose of this study was to investigate the effectiveness of gEUD optimization to fine tune the dose distributions of IMRT plans. We analyzed the effect of gEUD-based optimization parameters on plan quality. The objective was to determine whether dose distribution to selected structures could be improved using gEUD optimization without adversely altering the doses delivered to other structures, as in sculpting. We hypothesized that by carefully defining gEUD parameters (EUD0 and n) based on the current dose distributions, the optimization system could be instructed to search for alternative solutions in the neighborhood, and we could maintain the dose distributions for structures already satisfactory and improve dose for structures that need enhancement. We started with an already acceptable IMRT plan optimized with any objective function. The dose distribution was analyzed first. For structures that dose should not be changed, a higher value of n was used and EUD0 was set slightly higher/lower than the EUD value at the current dose distribution for critical structures/targets. For structures that needed improvement in dose, a higher to medium value of n was used, and EUD0 was set to the EUD value or slightly lower/higher for the critical structure/target at the current dose distribution. We evaluated this method in one clinical case each of head and neck, lung and prostate cancer. Dose volume histograms, isodose distributions, and relevant tolerance doses for critical structures were used for the assessment. We found that by adjusting gEUD optimization parameters, the dose distribution could be improved with only a few iterations. A larger value of n could lead to faster convergence and a medium value of n could result in a search in a broader area. Such improvement could also be achieved by optimization based on other criteria, but the gEUD-based method has the advantage of efficiency and flexibility. Therefore, gEUD-based optimization can be used as a tool to improve IMRT plans by adjusting the planning parameters, thereby making dose sculpting feasible.

Four-dimensional radiotherapy planning for DMLC-based respiratory motion tracking

Four-dimensional (4D) radiotherapy is the explicit inclusion of the temporal changes in anatomy during the imaging, planning, and delivery of radiotherapy. Temporal anatomic changes can occur for many reasons, though the focus of the current investigation is respiration motion for lung tumors. The aim of this study was to develop 4D radiotherapy treatment-planning methodology for DMLC-based respiratory motion tracking. A 4D computed tomography (CT) scan consisting of a series of eight 3D CT image sets acquired at different respiratory phases was used for treatment planning. Deformable image registration was performed to map each CT set from the peak-inhale respiration phase to the CT image sets corresponding to subsequent respiration phases. Deformable registration allows the contours defined on the peak-inhale CT to be automatically transferred to the other respiratory phase CT image sets. Treatment planning was simultaneously performed on each of the eight 3D image sets via automated scripts in which the MLC-defined beam aperture conforms to the PTV (which in this case equaled the GTV due to CT scan length limitations) plus a penumbral margin at each respiratory phase. The dose distribution from each respiratory phase CT image set was mapped back to the peak-inhale CT image set for analysis. The treatment intent of 4D planning is that the radiation beam defined by the DMLC tracks the respiration-induced target motion based on a feedback loop including the respiration signal to a real-time MLC controller. Deformation with respiration was observed for the lung tumor and normal tissues. This deformation was verified by examining the mapping of high contrast objects, such as the lungs and cord, between image sets. For the test case, dosimetric reductions for the cord, heart, and lungs were found for 4D planning compared with 3D planning. 4D radiotherapy planning for DMLC-based respiratory motion tracking is feasible and may offer tumor dose escalation and/or a reduction in treatment-related complications. However, 4D planning requires new planning tools, such as deformable registration and automated treatment planning on multiple CT image sets.

Validation of contour-driven thin-plate splines for tracking fraction-to-fraction changes in anatomy and radiation therapy dose mapping

The goal of this study is to validate a deformable model using contour-driven thin-plate splines for application to radiation therapy dose mapping. Our testing includes a virtual spherical phantom as well as real computed tomography (CT) data from ten prostate cancer patients with radio-opaque markers surgically implanted into the prostate and seminal vesicles. In the spherical mathematical phantom, homologous control points generated automatically given input contour data in CT slice geometry were compared to homologous control point placement using analytical geometry as the ground truth. The dose delivered to specific voxels driven by both sets of homologous control points were compared to determine the accuracy of dose tracking via the deformable model. A 3D analytical spherically symmetric dose distribution with a dose gradient of approximately 10% per mm was used for this phantom. This test showed that the uncertainty in calculating the delivered dose to a tissue element depends on slice thickness and the variation in defining homologous landmarks, where dose agreement of 3-4% in high dose gradient regions was achieved. In the patient data, radio-opaque marker positions driven by the thin-plate spline algorithm were compared to the actual marker positions as identified in the CT scans. It is demonstrated that the deformable model is accurate (approximately 2.5 mm) to within the intra-observer contouring variability. This work shows that the algorithm is appropriate for describing changes in pelvic anatomy and for the dose mapping application with dose gradients characteristic of conformal and intensity modulated radiation therapy.

Technical note: A novel boundary condition using contact elements for finite element based deformable image registration

Deformable image registration is an important tool for image-guided radiotherapy. Physics-model-based deformable image registration using finite element analysis is one of the methods currently being investigated. The calculation accuracy of finite element analysis is dependent on given boundary conditions, which are usually based on the surface matching of the organ in two images. Such a surface matching, however, is hard to obtain from medical images. In this study, we developed a new boundary condition to circumvent the traditional difficulties. Finite element contact-impact analysis was employed to simulate the interaction between the organ of interest and the surrounding body. The displacement loading is not necessarily specified. The algorithm automatically deforms the organ model into the minimum internal energy state. The analysis was performed on CT images of the lung at two different breathing phases (exhalation and full inhalation). The result gave the displacement vector map inside the lung. Validation of the result showed satisfactory agreement in most parts of the lung. This approach is simple, operator independent and may provide improved accuracy of the prediction of organ deformation.

Inherent smoothness of intensity patterns for intensity modulated radiation therapy generated by simultaneous projection algorithms

The efficient delivery of intensity modulated radiation therapy (IMRT) depends on finding optimized beam intensity patterns that produce dose distributions, which meet given constraints for the tumour as well as any critical organs to be spared. Many optimization algorithms that are used for beamlet-based inverse planning are susceptible to large variations of neighbouring intensities. Accurately delivering an intensity pattern with a large number of extrema can prove impossible given the mechanical limitations of standard multileaf collimator (MLC) delivery systems. In this study, we apply Cimmino's simultaneous projection algorithm to the beamlet-based inverse planning problem, modelled mathematically as a system of linear inequalities. We show that using this method allows us to arrive at a smoother intensity pattern. Including nonlinear terms in the simultaneous projection algorithm to deal with dose-volume histogram (DVH) constraints does not compromise this property from our experimental observation. The smoothness properties are compared with those from other optimization algorithms which include simulated annealing and the gradient descent method. The simultaneous property of these algorithms is ideally suited to parallel computing technologies.