Computerized design of target margins for treatment uncertainties in conformal radiotherapy

Evaluation of the accuracy of a six-degree-of-freedom robotic couch using optical surface and cone beam CT images of an SRS QA phantom

Purpose:

To assess the accuracy of the Varian PerfectPitch six-degree-of-freedom (6DOF) robotic couch by using a Varian SRS QA phantom.

Methods:

The stereotactic radiosurgery (SRS) phantom has five tungsten carbide BBs each with 7·5 mm in diameter arranged with the known geometry. Optical surface images and cone beam CT (CBCT) images of the phantom were taken at different pitch, roll and rotation angles. The pitch, roll, and rotation angles were varied from −3 to 3 degrees by inputs from the linac console. A total of 39 Vision RT images with different rotation angle combinations were collected, and the Vision RT software was used to determine the rotation angles and translational shifts from those images. Eight CBCT images at most allowed rotational angles were analysed by in-house software. The software took the coordinates of the voxel of the maximum CT number inside a 7·5-mm sphere surrounding one BB to be the measured position of this BB. Expected BB positions at different rotation angles were determined by multiplying measured BB positions at zero pitch and roll values by a rotation matrix. Applying the rotation matrix to 5 BB positions yielded 15 equations. A linear least square method was used for regression analysis to approximate the solutions of those equations.

Results:

Of the eight calculations from CBCT images, the maximum rotation angle differences (degree) were 0·10 for pitch, 0·15 for roll and 0·09 for yaw. The maximum translation differences were 0·3 mm in the left–right direction, 0·5 mm in the anterior–posterior direction and 0·4 mm in the superior–inferior direction.

Conclusions:

The uncertainties of the 6-DOF couch were examined with the methods of optical surface imaging and CBCT imaging of the SRS QA phantom. The rotational errors were less than 0·2 degree, and the isocentre shifts were less than 0·8 mm.

Application of an automatic, uncertainty model-guided, target-generating algorithm to lung stereotactic body radiotherapy

PURPOSE

This work evaluated a new radiotherapy target-generating framework (the αTarget algorithm) for creating internal target volumes for lung SBRT.

METHODS

Nineteen patients previously treated with definitive intent SBRT to the lung were identified from a clinical database. For each patient's 4DCT simulation scan, deformable image registration was used between phases of the scan in order to generate voxelized models of motion for 35 individual gross tumor volumes. These motion models were then used with a new implementation of a previously described target-generating algorithm to create new internal target volumes (αITVs). The resulting αITVs were analyzed with respect to their volume and the coverage they provided each tumor voxel per that voxel's motion model. The clinically used ITVs were similarly analyzed, and were then compared to the αITVs using paired Student's t-tests. In addition, isotropic margins were added to the αITVs in order to determine the largest margin magnitude that could be added without exceeding the volume of the clinical ITVs.

RESULTS

The αITVs increased the target coverage provided to each tumor's 5th-percentile-most-covered-voxel an average of 50.3% compared to the clinical ITVs (p < 0.0001). At the same time, the αITVs had volumes that were, on average, 31.4% smaller (p < 0.0001). The differences in volume were large enough that, on average, an extra 2 mm isotropic margin could be added to the αITV before it had a volume greater than the clinical ITV.

CONCLUSIONS

The αTarget algorithm can generate more effective lung SBRT internal target volumes that provide greater coverage with smaller volumes. In combination with numerous other advantages of the framework, this effectiveness makes the αTarget algorithm a powerful new method for advanced IGRT or adaptive radiotherapy techniques.

Generating amorphous target margins in radiation therapy to promote maximal target coverage with minimal target size

BACKGROUND AND SIGNIFICANCE

This work provides proof-of-principle for two versions of a heuristic approach that automatically creates amorphous radiation therapy planning target volume (PTV) margins considering local effects of tumor shape and motion to ensure adequate voxel coverage with while striving to minimize PTV size. The resulting target thereby promotes disease control while minimizing the risk of normal tissue toxicity.

METHODS

This work describes the mixed-PDF algorithm and the independent-PDF algorithm which generate amorphous margins around a radiation therapy target by incorporating user-defined models of target motion. Both algorithms were applied to example targets - one circular and one "cashew-shaped." Target motion was modeled by four probability density functions applied to the target quadrants. The spatially variant motion model illustrates the application of the algorithms even with tissue deformation. Performance of the margins was evaluated in silico with respect to voxelized target coverage and PTV size, and was compared to conventional techniques: a threshold-based probabilistic technique and an (an)isotropic expansion technique. To demonstrate the algorithm's clinical utility, a lung cancer patient was analyzed retrospectively. For this case, 4D CT measurements were combined with setup uncertainty to compare the PTV from the mixed-PDF algorithm with a PTV equivalent to the one used clinically.

RESULTS

For both targets, the mixed-PDF algorithm performed best, followed by the independent-PDF algorithm, the threshold algorithm, and lastly, the (an)isotropic algorithm. Superior coverage was always achieved by the amorphous margin algorithms for a given PTV size. Alternatively, the margin required for a particular level of coverage was always smaller (8-15%) when created with the amorphous algorithms. For the lung cancer patient, the mixed-PDF algorithm resulted in a PTV that was 13% smaller than the clinical PTV while still achieving ≥99.9% coverage.

CONCLUSIONS

The amorphous margin algorithms are better suited for the local effects of target shape and positional uncertainties than conventional margins. As a result, they provide superior target coverage with smaller PTVs, ensuring dose delivered to the target while decreasing the risk of normal tissue toxicity.

Accuracy evaluation of a six-degree-of-freedom couch using cone beam CT and IsoCal phantom with an in-house algorithm

PURPOSE

The accuracy of a six degree of freedom (6DoF) couch was evaluated using a novel method.

METHODS

Cone beam CT (CBCT) images of a 3D phantom (IsoCal) were acquired with different, known combinations of couch pitch and roll angles. Pitch and roll angles between the maximum allowable values of 357 and 3 degrees were tested in one degree increments. A total of 49 combinations were tested at 0 degrees of yaw (couch rotation angle). The 3D positions of 16 tungsten carbide ball bearings (BBs), each 4 mm in diameter and arranged in a known geometry within the IsoCal phantom, were determined in the 49 image sets with in-house software. The BB positions at different rotation angles were determined using a rotation matrix from the original BB positions at zero pitch and roll angles. A linear least squares fit method estimated the rotation angles and differences between detected and nominal rotation angles were calculated. This study was conducted for the case with and without extra weight on the couch. Couch walk shifts for the system were investigated using eight combinations of rotation, roll and pitch.

RESULTS

A total of 49 CBCT images with voxel sizes 0.5 × 0.5 × 1.0 mm³ were taken for the case without extra weight on the couch. The 16 BBs were determined to evaluate the isocenter translation and rotation differences between the calculated and nominal couch values. Among all 49 calculations, the maximum rotation angle differences were 0.10 degrees for pitch, 0.15 degrees for roll and 0.09 degrees for yaw. The corresponding mean and standard deviation values were 0.028 ± 0.032, -0.043 ± 0.058, and -0.009 ± 0.033 degrees. The maximum translation differences were 0.3 mm in the left-right direction, 0.5 mm in the anterior-posterior direction and 0.4 mm in the superior-inferior direction. The mean values and corresponding standard deviations were 0.07 ± 0.12, -0.05 ± 0.25, and -0.12±0.14 mm for the planes described above. With an 80 kg phantom on the couch, the maximum translation shift was 0.69 mm. The couch walk translation shifts were less than 0.1 mm and rotation shifts were less than 0.1 degree.

CONCLUSIONS

Errors of a new 6DoF couch were tested using CBCT images of a 3D phantom. The rotation errors were less than 0.3 degree and the translation errors were less than or equal to 0.8 mm in each direction. This level of accuracy is warranted for clinical radiotherapy utilization including stereotactic radiosurgery.

Population model of bladder motion and deformation based on dominant eigenmodes and mixed-effects models in prostate cancer radiotherapy

In radiotherapy for prostate cancer irradiation of neighboring organs at risk may lead to undesirable side-effects. Given this setting, the bladder presents the largest inter-fraction shape variations hampering the computation of the actual delivered dose vs. planned dose. This paper proposes a population model, based on longitudinal data, able to estimate the probability of bladder presence during treatment, using only the planning computed tomography (CT) scan as input information. As in previously-proposed principal component analysis (PCA) population-based models, we have used the data to obtain the dominant eigenmodes that describe bladder geometric variations between fractions. However, we have used a longitudinal analysis along each mode in order to properly characterize patient's variance from the total population variance. We have proposed is a mixed-effects (ME) model in order to separate intra- and inter-patient variability, in an effort to control confounding cohort effects. Other than using PCA, bladder shapes are represented by using spherical harmonics (SPHARM) that additionally enables data compression without information lost. Based on training data from repeated CT scans, the ME model was thus implemented following dimensionality reduction by means of SPHARM and PCA. We have evaluated the model in a leave-one-out cross validation framework on the training data but also using independent data. Probability maps (PMs) were thus generated with several draws from the learnt model as predicted regions where the bladder will likely move and deform. These PMs were compared with the actual regions using metrics based on mutual information distance and misestimated voxels. The prediction was also compared with two previous population PCA-based models. The proposed model was able to reduce the uncertainties in the estimation of the probable region of bladder motion and deformation. This model can thus be used for tailoring radiotherapy treatments.

The effects of the shape and size of the clinical target volume on the planning target volume margin

PURPOSE

To investigate the impact of clinical target volume (CTV) shape and size on CTV to planning target volume (PTV) margin expansion.

METHODS AND MATERIALS

Using numerical integration methods, margins accounting for random errors and systematic errors were calculated for CTVs of different shapes and sizes. We use k(r-95) and k(s-95) to represent the coefficients, for random errors and systematic errors, respectively, that ensure that every point of the CTV receives ≥95% of the prescribed dose.

RESULTS

The part of the margin accounting for random errors depends on CTV shape and size; generally, a convex part of a CTV would have a larger margin than a concave part. However, the part of the margin accounting for systematic errors is independent of CTV shape and size.

CONCLUSIONS

CTV shape and size should be considered when generating a PTV. For a complex CTV, the margins of the various parts of the CTV are different and related to local forms.

Three independent one-dimensional margins for single-fraction frameless stereotactic radiosurgery brain cases using CBCT

PURPOSE

Setting a proper margin is crucial for not only delivering the required radiation dose to a target volume, but also reducing the unnecessary radiation to the adjacent organs at risk. This study investigated the independent one-dimensional symmetric and asymmetric margins between the clinical target volume (CTV) and the planning target volume (PTV) for linac-based single-fraction frameless stereotactic radiosurgery (SRS).

METHODS

The authors assumed a Dirac delta function for the systematic error of a specific machine and a Gaussian function for the residual setup errors. Margin formulas were then derived in details to arrive at a suitable CTV-to-PTV margin for single-fraction frameless SRS. Such a margin ensured that the CTV would receive the prescribed dose in 95% of the patients. To validate our margin formalism, the authors retrospectively analyzed nine patients who were previously treated with noncoplanar conformal beams. Cone-beam computed tomography (CBCT) was used in the patient setup. The isocenter shifts between the CBCT and linac were measured for a Varian Trilogy linear accelerator for three months. For each plan, the authors shifted the isocenter of the plan in each direction by ±3 mm simultaneously to simulate the worst setup scenario. Subsequently, the asymptotic behavior of the CTV V80% for each patient was studied as the setup error approached the CTV-PTV margin.

RESULTS

The authors found that the proper margin for single-fraction frameless SRS cases with brain cancer was about 3 mm for the machine investigated in this study. The isocenter shifts between the CBCT and the linac remained almost constant over a period of three months for this specific machine. This confirmed our assumption that the machine systematic error distribution could be approximated as a delta function. This definition is especially relevant to a single-fraction treatment. The prescribed dose coverage for all the patients investigated was 96.1% ± 5.5% with an extreme 3-mm setup error in all three directions simultaneously. It was found that the effect of the setup error on dose coverage was tumor location dependent. It mostly affected the tumors located in the posterior part of the brain, resulting in a minimum coverage of approximately 72%. This was entirely due to the unique geometry of the posterior head.

CONCLUSIONS

Margin expansion formulas were derived for single-fraction frameless SRS such that the CTV would receive the prescribed dose in 95% of the patients treated for brain cancer. The margins defined in this study are machine-specific and account for nonzero mean systematic error. The margin for single-fraction SRS for a group of machines was also derived in this paper.

Feasibility study of real-time planning for stereotactic radiosurgery

PURPOSE

3D rotational setup errors in radiotherapy are often ignored by most clinics due to inability to correct or simulate them accurately and efficiently. There are two types of rotation-related problems in a clinical setting. One is to assess the affected dose distribution in real-time if correction is not applied and the other one is to correct the rotational setup errors prior to the initiation of the treatment. Here, the authors present the analytical solutions to both problems.

METHODS

(1) To assess the real-time dose distribution, eight stereotactic radiosurgery (SRS) cases were used as examples. For each plan, two new sets of beams with different table, gantry, and collimator angles were given in analytical forms as a function of patient rotational errors. The new beams simulate the rotational effects of the patient during the treatment setup. By using one arbitrary set of beams, SRS plans were recomputed with a series of different combinations of patient rotational errors, ranging from (-5°, -5°, -5°) to (5°, 5°, 5°) (roll, pitch, and yaw) with an increment of 1° and compared with those without rotational errors. For each set of rotational errors, its corresponding equivalent beams were computed using the analytical solutions and then used for dose calculation. (2) To correct for the rotational errors, two new sets of table, gantry, and collimator angles were derived analytically to validate the previously published derivation. However, in the derivation, a novel methodology was developed and two sets of table, gantry, and collimator angles were obtained in analytical forms. The solutions provide an alternative approach to rotational error correction by rotating the couch, gantry, and collimator rather than the patient.

RESULTS

For demonstration purpose, the above-derived new beams were implemented in a treatment planning system (TPS) to study the rotational effects on the SRS cases. For each case, the authors have generated ten additional plans that accounted for different rotations of the patient. They have found that rotations have an insignificant effect on the minimal, maximum, mean doses, and V80% of the planning target volume (PTV) when the rotations were relatively small. This was particularly true for the small and near-spherical targets. They, however, did change V95% significantly when the rotations approached 5°. The theory has been validated with clinical SRS cases and proven to be practical and viable. The preliminary results demonstrate that the rotational effects are patient-specific and depend on several important factors, such as the PTV size, the PTV location, and the beam configuration. The solutions given in this paper are of great potential values in clinical applications.

CONCLUSIONS

They have derived the analytical solutions to a new set of table, gantry, and collimator angles for a given treatment beam configuration as a function of patient rotational errors. One solution was used to assess the dosimetric effects of an imperfect patient setup and the other one was used to correct for the setup errors without rotating the patient. Compared to the widely adopted method of rotation effect assessment by importing the rotational CT images into TPS, the equivalent beam approach is simple and accurate. The analytical solutions to correcting for rotational setup errors prior to treatment were also derived. Based on the initial clinical investigations, they firmly believe that clinically viable real-time treatment planning and adaptive radiation therapy are feasible with this novel method.

Reduction of prostate intrafractional motion from shortening the treatment time

This study aims to quantify the reduction of the intrafractional motion when the prostate intensity modulated radiation therapy (IMRT) treatment time is shortened. Prostate intrafractional motion data recorded by the Calypso system for 105 patients was analyzed. Statistical distributions of the prostate displacements for the regular IMRT treatment and the first 1, 2, 3 and 5 min of the treatment were calculated and used for treatment margin estimation for all the selected patients. The treatment margins estimated for the first 1, 2, 3 and 5 min were compared with those for the regular IMRT treatment to quantify the reduction of the motion. If the treatment can be completed within 5 (3) min, the standard deviation of the prostate displacement could be reduced by up to 45% and the required treatment margins could be reduced to 1.2 (1.1), 0.9 (0.8), 2.2 (1.9), 1.9 (1.5), 1.9 (1.7) and 2.8 (2.4) mm from 1.5, 1.1, 2.8, 3.0, 2.4 and 3.9 mm in the left, right, superior, inferior, anterior and posterior directions, respectively. The same work was also performed for 19 of the 105 patients who exhibited the largest motion with 30% of their treatment time having 3D motion more than 3 mm. For this group of patients, the required margins change to 1.4 (1.2), 0.8 (0.8), 1.8 (1.6), 2.3 (1.8), 1.7 (1.5) and 3.4 (2.8) mm from 1.9, 1.2, 1.7, 3.7, 1.6 and 4.9 mm in the six directions when the treatment time is reduced to 5 (3) min. The intrafractional motion effects on prostate treatment are significantly smaller and the required margins can be therefore reduced when the treatment is shortened.

Correction of systematic set-up error in breast and head and neck irradiation through a no-action level (NAL) protocol

PURPOSE

To quantify systematic and random patient set-up errors in breast and head and neck conventional irradiation and to evaluate a no-action level (NAL) protocol for systematic set-up error off-line correction in head and neck cancer and breast cancer patients.

MATERIAL AND METHODS

Verification electronic portal images of orthogonal set-up fields were obtained daily for the initial four consecutive fractions for 20 patients treated for breast cancer and for 20 head and neck cancer patients. The calculated systematic error was used to shift the isocentre accordingly on the fifth treatment day. From then until the end of the treatment course, pair orthogonal portal images of set-up fields were obtained weekly. To assess the impact of the protocol, pre- and post-correction systematic errors were compared and PTV margins were estimated before and after correction using published margin recipes.

RESULTS

Population systematic set-up error decreased in the breast cancer patient group after the implementation of NAL protocol from 4.0 to 1.7 mm on the x-axis, from 4.7 to 2.1 mm on the y-axis and from 2.8 to 0.9 mm on the z axis. The percentage of patients with individual systematic set-up error reduction was 80%, 90% and 80% on the x-, y and z-axes respectively. Population systematic set-up error decreased also in the head and neck cancer patient group from 2.3 to 1.1 mm on the x-axis, from 1.6 to 1.4 mm on the y-axis and from 1.7 to 0.7 mm on the z-axis. The percentage of patients with individual systematic set-up error reduction was 70%, 65% and 85% on the x-, y- and z-axes respectively. Margin reduction achievable with NAL protocol implementation on the x-, y- and z-axes was 6.3, 7.2 and 4.8 mm for breast cancer patients and 3.3, 0.6 and 2.8 mm for head and neck cancer patients.

CONCLUSION

NAL off-line protocol is useful for systematic set-up error correction and PTV margin reduction in conventional breast and head and neck irradiation.

Optimizing principal component models for representing interfraction variation in lung cancer radiotherapy

PURPOSE

To optimize modeling of interfractional anatomical variation during active breath-hold radiotherapy in lung cancer using principal component analysis (PCA).

METHODS

In 12 patients analyzed, weekly CT sessions consisting of three repeat intrafraction scans were acquired with active breathing control at the end of normal inspiration. The gross tumor volume (GTV) and lungs were delineated and reviewed on the first week image by physicians and propagated to all other images using deformable image registration. PCA was used to model the target and lung variability during treatment. Four PCA models were generated for each specific patient: (1) Individual models for the GTV and each lung from one image per week (week to week, W2W); (2) a W2W composite model of all structures; (3) individual models using all images (weekly plus repeat intrafraction images, allscans); and (4) composite model with all images. Models were reconstructed retrospectively (using all available images acquired) and prospectively (using only data acquired up to a time point during treatment). Dominant modes representing at least 95% of the total variability were used to reconstruct the observed anatomy. Residual reconstruction error between the model-reconstructed and observed anatomy was calculated to compare the accuracy of the models.

RESULTS

An average of 3.4 and 4.9 modes was required for the allscans models, for the GTV and composite models, respectively. The W2W model required one less mode in 40% of the patients. For the retrospective composite W2W model, the average reconstruction error was 0.7 +/- 0.2 mm, which increased to 1.1 +/- 0.5 mm when the allscans model was used. Individual and composite models did not have significantly different errors (p = 0.15, paired t-test). The average reconstruction error for the prospective models of the GTV stabilized after four measurements at 1.2 +/- 0.5 mm and for the composite model after five measurements at 0.8 +/- 0.4 mm.

CONCLUSIONS

Retrospective PCA models were capable of reconstructing original GTV and lung shapes and positions within several millimeters with three to four dominant modes, on average. Prospective models achieved similar accuracy after four to five measurements.

Determination of target volumes in radiotherapy and the implications of technological advances: a literature review

This study assesses the influence of new techniques and technologies in radiotherapy on the derivation and applicability of the margins currently used for treatment planning. The validity of the continued use of the recommendations of International Commission on Radiation Units and Measurements (ICRU) and other recommendations as a result of the additional information derived from these emerging techniques is also reviewed. The ICRU formulations still remain fundamental in the derivation of target volumes in radiotherapy; however, revisions to these have been recommended through various experimental and modelling techniques leading to the publication of various margin recipes. These recipes are used for margin definitions in new radiotherapy techniques including intensity-modulated radiotherapy (IMRT). The use of image-guided radiotherapy (IGRT) techniques leads to the reduction in organ motion uncertainties and setup errors, allowing for the adjustment of margins and treatment plans as well as dose escalation. Clinical trials are still needed to validate most of the new techniques in radiotherapy, particularly in IGRT techniques leading to adaptive radiotherapy. It is recommended that well devised clinical trials should be conducted to investigate fully the efficacy of these new techniques, particularly in radiotherapy image guidance and adaptive radiotherapy. Such trials would validate any recommendations regarding the current clinical margins and impact on their continued clinical use.

The investigation on the location effect of external markers in respiratory-gated radiotherapy

PURPOSE

To investigate the effect of the marker placement on the correlation relationship between the motions of external markers and the internal target under different breathing patterns for several lung cancer patients.

METHOD AND MATERIAL

To monitor and record simultaneous motions of internal target and associated surrogate markers during respiratory gated radiotherapy, an infrared camera system synchronized with a medical simulator was installed in our institute. Multiple external markers were placed on the patients' chest wall with proper geometrical arrangement in closely monitoring the motion of skin near tumor. The motion signals of three breathing sessions (free breathing, breath-holding, and free breathing after breath-holding) were recorded and the quality of correlation between them was analyzed. For a single marker motion, its correlation with the internal target was analyzed using cross-covariance function. For the multiple markers, their correlation with the internal target was analyzed based on additive model.

RESULT

Seven patients undergoing radiotherapy with right upper or middle lobe lesions were enrolled in this study. Statistic analysis based on the internal-external motion signals shows that the effect of marker location on the quality of its correlation with the internal target is varied from patient to patient. There was no specific marker location where consistently demonstrated superior quality of correlation with the internal target motion over three breathing sessions for all patients. As the composite surrogate signal which was generated from the motions of multiple external markers was used to correlate the internal target motion, significant improvement of the quality of correlation was achieved.

CONCLUSION

The correlation of external marker to the internal target could be influenced by several factors such as patient population, marker locations, and breathing patterns, considerably. The quality of correlation and predictability to the internal target furnished by a single external marker is inferior to that of the composite signal generated from multiple external markers. The use of composite signal shows great potential in improving the predictability of internal target motion and presents an effective way to track tumor more accurately.

A method to calculate coverage probability from uncertainties in radiotherapy via a statistical shape model

In this paper we describe a technique that may be used to model the geometric uncertainties that accrue during the radiotherapy process. Using data from in-treatment cone beam CT scans, we simultaneously analyse non-uniform observer delineation variability and organ motion together with patient set-up errors via the creation of a point distribution model (PDM). We introduce a novel method of generating a coverage probability matrix, that may be used to determine treatment margins and calculate uncertainties in dose, from this statistical shape model. The technique does not assume rigid body motion and can extrapolate shape variability in a statistically meaningful manner. In order to construct the PDM, we generate corresponding surface points over a set of delineations. Correspondences are established at a set of points in parameter space on spherically parameterized and canonical aligned outlines. The method is demonstrated using rectal delineations from serially acquired in-treatment cone beam CT image volumes of a prostate patient (44 image volumes total), each delineated by a minimum of two observers (maximum six). Two PDMs are constructed, one with set-up errors included and one without. We test the normality assumptions of the PDMs and find the distributions to be Gaussian in nature. The rectal PDM variability is in general agreement with data in the literature. The two resultant coverage probability matrices show differences as expected.

How does performance of ultrasound tissue typing affect design of prostate IMRT dose-painting protocols?

PURPOSE

To investigate how the performance characteristics of ultrasound tissue typing (UTT) affect the design of a population-based prostate dose-painting protocol.

METHODS AND MATERIALS

The performance of UTT is evaluated using the receiver operating characteristic curve. As the imager's sensitivity increases, more tumors are detected, but the specificity worsens, causing more false-positive results. The UTT tumor map, obtained with a specific sensitivity and specificity setup, was used with the patient's CT image to guide intensity-modulated radiotherapy (IMRT) planning. The optimal escalation dose to the UTT positive region, as well as the safe dose to the negative background, was obtained by maximizing the uncomplicated control (i.e., a combination of tumor control probability and weighted normal tissue complication probability). For high- and low-risk tumors, IMRT plans guided by conventional ultrasound or UTT with a one-dimensional or two-dimensional spectrum analysis technique were compared with an IMRT plan in which the whole prostate was dose escalated.

RESULTS

For all imaging modalities, the specificity of 0.9 was chosen to reduce complications resulting from high false-positive results. If the primary tumors were low risk, the IMRT plans guided by all imaging modalities achieved high tumor control probability and reduced the normal tissue complication probability significantly compared with the plan with whole gland dose escalation. However, if the primary tumors were high risk, the accuracy of the imaging modality was critical to maintain the tumor control probability and normal tissue complication probability at acceptable levels.

CONCLUSION

The performance characteristics of an imager have important implications in dose painting and should be considered in the design of dose-painting protocols.

On probabilistically defined margins in radiation therapy

Margins about a target volume subject to external beam radiation therapy are designed to assure that the target volume of tissue to be sterilized by treatment is adequately covered by a lethal dose. Thus, margins are meant to guarantee that all potential variation in tumour position relative to beams allows the tumour to stay within the margin. Variation in tumour position can be broken into two types of dislocations, reducible and irreducible. Reducible variations in tumour position are those that can be accommodated with the use of modern image-guided techniques that derive parameters for compensating motions of patient bodies and/or motions of beams relative to patient bodies. Irreducible variations in tumour position are those random dislocations of a target that are related to errors intrinsic in the design and performance limitations of the software and hardware, as well as limitations of human perception and decision making. Thus, margins in the era of image-guided treatments will need to accommodate only random errors residual in patient setup accuracy (after image-guided setup corrections) and in the accuracy of systems designed to track moving and deforming tissues of the targeted regions of the patient's body. Therefore, construction of these margins will have to be based on purely statistical data. The characteristics of these data have to be determined through the central limit theorem and Gaussian properties of limiting error distributions. In this paper, we show how statistically determined margins are to be designed in the general case of correlated distributions of position errors in three-dimensional space. In particular, we show how the minimal margins for a given level of statistical confidence are found. Then, how they are to be used to determine geometrically minimal PTV that provides coverage of GTV at the assumed level of statistical confidence. Our results generalize earlier recommendations for statistical, central limit theorem-based recommendations for margin construction that were derived for uncorrelated distributions of errors (van Herk, Remeijer, Rasch and Lebesque 2000 Int. J. Radiat. Oncol. Biol. Phys. 47 1121-35; Stroom, De Boer, Huizenga and Visser 1999 Int. J. Radiat. Oncol. Biol. Phys. 43 905-19).

Spinal cord planning risk volumes for intensity-modulated radiation therapy of head-and-neck cancer

PURPOSE

To assess planning organ at risk volume (PRV) margins of the spinal cord in intensity-modulated radiotherapy (IMRT) of oropharyngeal cancers, by modeling the effect of geometric uncertainties to estimate the probability of the spinal cord receiving a particular dose.

METHODS AND MATERIALS

Five patients with oropharyngeal cancer were treated by IMRT with simultaneous doses of 66 Gy (gross disease) and 54 Gy (subclinical disease) in 30 fractions. Spinal cord doses were limited to 45 Gy. The probability, due to random and systematic patient positioning uncertainties (3-mm standard deviation), of the cord receiving a particular dose was determined. The effect of an on-line setup correction protocol was also modeled.

RESULTS

The mean probability of a maximum spinal cord dose of 45 Gy was 1%, with a 6-mm PRV margin. The mean probability of a maximum dose exceeding 40 Gy was 37% (range, 13-77%); this probability is reduced with a setup correction protocol.

CONCLUSION

A spinal cord PRV generated with a 6-mm margin leads to a 99% probability of maintaining the maximum spinal cord dose below 45 Gy. The application of an on-line setup correction protocol reduces the cord dose by approximately 5 Gy.

Robust treatment planning for intensity modulated radiotherapy of prostate cancer based on coverage probabilities

BACKGROUND AND PURPOSE

To evaluate an optimization approach where coverage probabilities are incorporated into the optimization of intensity modulated radiotherapy (IMRT) to overcome the problem of margin definition in the case of overlapping planning target volume and organs at risk.

PATIENTS AND METHODS

IMRT plans were generated for three optimization approaches: based on a planning CT plus margin (A), on prostate and rectum contours from five pre-treatment CT plus margin (B), and on coverage probabilities (C). For approach (C), the probability of organ occupation was computed for each voxel from five pre-treatment CTs and the population distribution of systematic setup error and it was used as local weight in the costfunctions. Monte Carlo simulations of treatment courses were used to compute the probability distribution of prostate and rectal wall equivalent uniform dose (EUD).

RESULTS

Treatment simulations showed best and most robust results for prostate and rectal wall EUD within the population for (C). For (A) the rectal wall EUD was on average about 1.5 Gy greater than in (C), while the prostate EUD was lower than those from (C) for most of the patients for (B) (especially for those with great organ motion).

CONCLUSIONS

The incorporation of coverage probabilities as local weights allows for dose escalation as well as improved rectal sparing and results in a safer and more robust IMRT treatment.

Intensity modulated radiation therapy for oropharyngeal cancer: the sensitivity of plan objectives and constraints to set-up uncertainty

The goal of this study was to assess the impact of set-up uncertainty on compliance with the objectives and constraints of an intensity modulated radiation therapy protocol for early stage cancer of the oropharynx. As the convolution approach to the quantitative study of set-up uncertainties cannot accommodate either surface contours or internal inhomogeneities, both of which are highly relevant to sites in the head and neck, we have employed the more resource intensive direct simulation method. The impact of both systematic (variable from 0 to 6 mm) and random (fixed at 2 mm) set-up uncertainties on compliance with the criteria of the RTOG H-0022 protocol has been examined for eight geometrically complex structures: CTV66 (gross tumour volume and palpable lymph nodes suspicious for metastases), CTV54 (lymph node groups or surgical neck levels at risk of subclinical metastases), glottic larynx, spinal cord, brainstem, mandible and left and right parotids. In a probability-based approach, both dose-volume histograms and equivalent uniform doses were used to describe the dose distributions achieved by plans for two patients, in the presence of set-up uncertainty. The equivalent uniform dose is defined to be that dose which, when delivered uniformly to the organ of interest, will lead to the same response as the non-uniform dose under consideration. For systematic set-up uncertainties greater than 2 mm and 5 mm respectively, coverage of the CTV66 and CTV54 could be significantly compromised. Directional sensitivity was observed in both cases. Most organs at risk (except the glottic larynx which did not comply under static conditions) continued to meet the dose constraints up to 4 mm systematic uncertainty for both plans. The exception was the contra lateral parotid gland, which this protocol is specifically designed to protect. Sensitivity to systematic set-up uncertainty of 2 mm was observed for this organ at risk in both clinical plans.

Target Definition in Prostate, Head, and Neck

Target definition is a major source of errors in both prostate and head and neck external-beam radiation treatment. Delineation errors remain constant during the course of radiation and therefore have a large impact on the dose to the tumor. Major sources of delineation variation are visibility of the target including its extensions, disagreement on the target extension, and interpretation or lack of delineation protocols. The visibility of the target can be greatly improved with the use of multimodality imaging. Both in the head and neck and the prostate, computed tomography (CT)-magnetic resonance imaging coregistration decreases the target volume and its variability. CT-positron emission tomography delineation is promising for delineation in head and neck cancer. Despite the better visibility, a different interpretation of the target extension remains a major source of error. The use of coregistration of CT with a second modality, together with improved guidelines for delineation and an online anatomical atlas, increases agreement between observers in prostate, lung, and nasopharynx tumors. Delineation errors should not be treated differently from other geometrical errors. Similar margin recipes for the correction of setup errors and organ motion should be adapted to incorporate the effect of delineation errors. A calculation of a 3-dimensional clinical target volume-planning target volume margin incorporating delineation errors for the head and neck is around 6.1 to 9.7 mm. Given the good local control of IMRT with smaller margins and smaller pathological specimens, it is likely that the delineated CTV frequently overestimates the actual volume.

Adequate margins for random setup uncertainties in head-and-neck IMRT

PURPOSE

To investigate the effect of random setup uncertainties on the highly conformal dose distributions produced by intensity-modulated radiotherapy (IMRT) for clinical head-and-neck cancer patients and to determine adequate margins to account for those uncertainties.

METHODS AND MATERIALS

We have implemented in our clinical treatment planning system the possibility of simulating normally distributed patient setup displacements, translations, and rotations. The planning CT data of 8 patients with Stage T1-T3N0M0 oropharyngeal cancer were used. The clinical target volumes of the primary tumor (CTV(primary)) and of the lymph nodes (CTV(elective)) were expanded by 0.0, 1.5, 3.0, and 5.0 mm in all directions, creating the planning target volumes (PTVs). We performed IMRT dose calculation using our class solution for each PTV margin, resulting in the conventional static plans. Then, the system recalculated the plan for each positioning displacement derived from a normal distribution with sigma = 2 mm and sigma = 4 mm (standard deviation) for translational deviations and sigma = 1 degrees for rotational deviations. The dose distributions of the 30 fractions were summed, resulting in the actual plan. The CTV dose coverage of the actual plans was compared with that of the static plans.

RESULTS

Random translational deviations of sigma = 2 mm and rotational deviations of sigma = 1 degrees did not affect the CTV(primary) volume receiving 95% of the prescribed dose (V(95)) regardless of the PTV margin used. A V(95) reduction of 3% and 1% for a 0.0-mm and 1.5-mm PTV margin, respectively, was observed for sigma = 4 mm. The V(95) of the CTV(elective) contralateral was approximately 1% and 5% lower than that of the static plan for sigma = 2 mm and sigma = 4 mm, respectively, and for PTV margins <5.0 mm. An additional reduction of 1% was observed when rotational deviations were included. The same effect was observed for the CTV(elective) ipsilateral but with smaller dose differences than those for the contralateral side. The effect of the random uncertainties on the mean dose to the parotid glands was not significant. The maximal dose to the spinal cord increased by a maximum of 3 Gy.

CONCLUSIONS

The margins to account for random setup uncertainties, in our clinical IMRT solution, should be 1.5 mm and 3.0 mm in the case of sigma = 2 mm and sigma = 4 mm, respectively, for the CTV(primary). Larger margins (5.0 mm), however, should be applied to the CTV(elective), if the goal of treatment is a V(95) value of at least 99%.

Dose correlation for thoracic motion in radiation therapy of breast cancer

This work investigates the dose correlation for deformed objects due to thoracic motion for radiotherapy treatment of breast cancer. An analytical model has been developed to reconstruct patient anatomy based on the assumption that the body will expand or compress proportionally during respiration. The patient geometry at any phase during a breathing pattern is reconstructed using the CT data taken at the inspiration and expiration phases and the breathing level which can be related to the measured chest wall motion. A correlation between the voxels in the inspiration (or expiration) geometry and the voxels in the reconstructed geometry at any phase of the breathing pattern is established so that the dose can be accumulated during a treatment. The method has been implemented for treatment planning dose calculation by interfacing with a Monte Carlo code. The patient geometry files for different phases of the breathing pattern are generated and the three-dimensional dose data are obtained from the Monte Carlo simulations. The final dose distribution is reconstructed from the dose data at different breathing phases based on patient's breathing pattern associated with chest wall movements.

Experience of ultrasound-based daily prostate localization

PURPOSE

The NOMOS (Sewickley, PA) B-mode Acquisition and Targeting System (BAT) ultrasound system provides a rapid means of correcting for interfraction prostate positional variation. In this investigation, we report our experience on the clinical issues relevant to the daily use of the BAT system and the analysis of combined setup error and organ motion for 3509 BAT alignment procedures in 147 consecutive patients treated with IMRT for prostate cancer.

METHODS AND MATERIALS

After setup to external skin marks, therapists performed the BAT ultrasound alignment procedure before each IMRT treatment. In this study, a single physician (A.C.) reviewed all BAT images and classified image quality and accuracy of image alignment by the therapist. On a scale of 1-3, near-perfect image quality or alignment was given a 1, fair image quality or misalignment > or = 5 mm (likely within the PTV) was given a 2, and unacceptable image quality or misalignment >5 mm (potential to violate the PTV) was given a value of 3. The distribution of shifts made was analyzed in each dimension and for all patients. The time required to perform the BAT alignment was also assessed in 17 patients.

RESULTS

Among the 3509 attempted BAT procedures, the image quality was judged to be poor or unacceptable in 5.1% (181). Of the remaining 3328 BAT images, with quality scores of 1-2, alignments were unacceptable (>5 mm misalignment as judged by the reviewing physician) in 3% (100). The mean shift in each direction, averaged over all patients, was 0.5-0.7 mm. Interfraction standard deviation (1 SD) of prostate position based on combined setup error and internal organ motion is 4.9 mm, 4.4 mm, and 2.8 mm in the anteroposterior (AP), superior-inferior (SI), and lateral (RL) dimensions, respectively. The distribution of the shifts was a near-random Gaussian-type in all three major axes, with greater variations in AP and SI directions. The percent of BAT procedures in which the shift was >5 mm was 28.6% in AP, 23% in SI, and 9% in RL directions. The average BAT procedure took extra 5 min out of a 20-min time slot in a typical eight-field IMRT treatment.

CONCLUSIONS

The quality of the daily ultrasound images was deemed acceptable in 95%. Major alignment errors by therapists were only 3%. The BAT system is clinically effective and feasible in a matter of 5 min. Although the accuracy of the BAT was not addressed in this investigation, we found a significant percentage of large shifts being made from the initial alignment position.

Can PET provide the 3D extent of tumor motion for individualized internal target volumes? A phantom study of the limitations of CT and the promise of PET

PURPOSE

To characterize the limitations of fast, spiral computed tomography (CT) when imaging a moving object and to investigate whether positron emission tomography (PET) can predict the internal target volume (ITV) and ultimately improve the planning target volume (PTV) for moving tumors.

METHODS AND MATERIALS

To mimic tumors, three fillable spheres were imaged while both stationary and during periodic motion using spiral CT and PET. CT- and PET-imaged volumes were defined quantitatively using voxel values. Ideal PTVs for each scenario were calculated. CT-based PTVs were generated using margins of 7.5, 10, and 15 mm to account for both organ motion and setup uncertainties. PET-based PTVs were derived with the assumption that motion was captured in the PET images and only a margin (7.5 mm) for setup errors was necessary. Comparisons between CT-based and PET-based PTVs with ideal PTVs were performed.

RESULTS

CT imaging of moving spheres resulted in significant distortions in the three-dimensional (3D) image-based representations, and did not, in general, result in images well representative of either moving or stationary spheres. PET images were similar to the ideal capsular shape encompassing the sphere and its motion. In all cases, CT-imaged volumes were larger than that for the stationary sphere (range of excess volume from 0.4 to 29 cm(3) for stationary volumes of 2.14 to 172 cm(3)), but smaller than that for the true motion volume. PET-imaged volumes were larger than the true motion volume (difference from ideal ranged from 3 to 94 cm(3) for motion volumes of 1.2 to 243 cm(3)) and much larger than the stationary volume. Using CT data, geographic miss of some part of the ideal PTV occurred for 0 of 24 cases, 11 of 24 cases, and 18 of 24 cases using a 15-mm, 10-mm, and 7.5-mm margin, respectively. Geographic miss did not occur in any case for the PET-based PTV. The amount of "normal tissue" included in CT-based PTVs was dramatically greater than that included in PET-based PTVs.

CONCLUSION

Fast CT imaging of a moving tumor can result in poor representation of the time-averaged position and shape of the tumor. PET imaging can provide a more accurate representation of the 3D volume encompassing motion of model tumors and has potential to provide patient-specific motion volumes for an individualized ITV.

[Probabilities of controlling tumors and complications (TCP/NTCP) after radiotherapy: methodologic, physical, and biological aspects]

Radiotherapy is aimed at getting the best possible therapeutic ratio (tumor local control versus morbidity). Physicists and radiation oncologists have to evaluate explicitly or implicitly the probability of induced complications to normal surrounding tissues. This is based on published data and clinician's experience. Quantitative methods have been introduced with different models in order to predict the impact of partial or global irradiation on a normal organ. These models correspond to the Tumor Control Probability (TCP) and Normal Tissue Complication Probability (NTCP). These biological models may be useful to evaluate the quality of a treatment planning or for the optimization process. The methodologies used and the clinical data are developed and discussed.

The effect of set-up uncertainties, contour changes, and tissue inhomogeneities on target dose-volume histograms

Understanding set-up uncertainty effects on dose distributions is an important clinical problem but difficult to model accurately due to their dependence on tissue inhomogeneities and changes in the surface contour (i.e., variant effects). The aims are: (1) to evaluate and quantify the invariant and variant effects of set-up uncertainties, contour changes and tissue inhomogeneities on target dose-volume histograms (DVHs); (2) to propose a method to interpolate (variant) DVHs. We present a lung cancer patient to estimate the significance of set-up uncertainties, contour changes and tissue inhomogeneities in target DVHs. Differential DVHs are calculated for 15 displacement errors (with respect to the isocenter) using (1) an invariant shift of the dose distribution at the isocenter, (2) a full variant calculation, and (3) a B-spline interpolation applied to sparsely sampled variant DVHs. The collapsed cone algorithm was used for all dose calculations. Dosimetric differences are quantified with the root mean square (RMS) deviation and the equivalent uniform dose (EUD). To determine set-up uncertainty effects, weighted mean EUDs, assuming normally distributed displacement errors, are used. The maximum absolute difference and RMS deviation in the integral DVHs' relative dose between (1) the invariant and calculated curves are 65.2% and 5.8% and (2) the interpolated and calculated curves are 16.9% and 2.5%. Similarly, the maximum absolute difference and RMS deviation in mean EUD as a function of the set-up uncertainty's standard deviation between (1) the invariant and calculated curves are 0.02 and 0.01 Gy; and (2) the interpolated and calculated curves are 0.01 and 0.006 Gy. Since a "worst-case" example is selected, we conclude that, in the majority of clinical cases, the variant effects of contour changes, tissue inhomogeneities and set-up uncertainties on EUD are negligible. Interpolation is a valid, efficient method to approximate DVHs.

Dose-volume specification: new challenges with intensity-modulated radiation therapy

It has long been recognized that the specification of volumes and doses is an important issue for radiation oncology. Although in any individual center, policies and procedures of treatment delivery may be well understood by staff, reporting of treatment techniques in the archival literature in an unambiguous manner has been found to be less than desirable in many instances. For clinical studies utilizing three-dimensional conformal radiation therapy (3D-CRT), and even more so, intensity-modulated radiation therapy (IMRT), the situation has become even more complex. 3D-CRT and IMRT are now recognized to be more sensitive to geometric uncertainties than conventional radiation therapy because of their ability to create sharper dose gradients around target volumes and organs at risk (OARs). This article reviews the current status of specifying target volumes and doses for 3D-CRT and IMRT, and discusses some of the pertinent issues regarding the use of recommendations in Reports 50 and 62 of the International Commission on Radiation Units and Measurements (ICRU) in this task. It is imperative that physician and physicist fully appreciate the need to account for clinical and spatial uncertainties in the planning and delivery of cancer patients' treatment, paying even more attention to these issues for those cases in which 3D-CRT and/or IMRT is used. A brief review of the reporting requirements for Radiation Therapy Oncology Group (RTOG) 3D-CRT and IMRT protocols is also presented.

Margins for translational and rotational uncertainties: a probability-based approach

PURPOSE

To define margins for systematic rotations and translations, based on known statistical distributions of these deviations.

METHODS AND MATERIALS

The confidence interval-based expansion method for translations, known as the "rolling ball algorithm," was extended to include rotations. This new method, which we call the Rotational and Translational Confidence Limit (RTCL) method, is exact for a point with arbitrary rotations and translations or for a finite shape with rotations only. The method was compared with two existing expansion methods: a rolling ball algorithm without rotations, and a convolution (blurring) method which included rotations. On the basis of these methods, planning target volumes (PTVs, expanded clinical target volumes [CTVs]) were constructed for a number of shapes (a sphere, a sphere with an extension, and three prostate cases), and evaluated in several ways by means of a Monte Carlo method. The accuracy of each method was measured by determining the probability of finding the CTV completely inside the PTV (P(CTVinPTV)), using parameters that yield a 90% probability for a sphere-shaped CTV without rotations. Furthermore, with the expansion parameters adjusted to give an equal P(CTVinPTV) for all methods, PTV volumes were compared.

RESULTS

With the expansion algorithm parameters chosen to yield P(CTVinPTV) = 90% for a sphere, an average P(CTVinPTV) of 84%, 57%, and 46% was obtained for the other shapes, using the RTCL method, coverage probability, and rolling ball, respectively. With the parameters adjusted to yield an equal P(CTVinPTV) for all methods, the PTV volume was on average 8% larger for the coverage probability method and 15% larger for the rolling ball algorithm compared to the RTCL method.

CONCLUSION

The RTCL method provides an accurate way to include the effects of systematic rotations in the margin. Compared to other algorithms, the method is less sensitive to the shape of the CTV, and, given a fixed probability of finding the CTV inside the PTV, a smaller PTV volume can be obtained.

[Conformal radiotherapy in prostate cancer: for whom and how?]

External radiotherapy is one of the modalities used to cure localized prostate carcinoma. Most of localized prostate carcinomas, specially those of the intermediate prognostic group, may benefit from escalated dose above 70 Gy at least as regard biochemical and clinical relapse free survival. 3D-CRT allows a reduction of the dose received by organs at risk and an increase of prostate dose over 70 Gy. It is on the way to become a standard. Intensity modulated radiation therapy increases dose homogeneity and reduces rectal dose. These methods necessitate rigorous procedures in reproducibility, delineation of volumes, dosimetry, daily treatment. They need also technological and human means. It is clear that localized prostate cancer is a good example for evaluation of these new radiotherapy modalities.

Inclusion of geometric uncertainties in treatment plan evaluation

PURPOSE

To correctly evaluate realistic treatment plans in terms of absorbed dose to the clinical target volume (CTV), equivalent uniform dose (EUD), and tumor control probability (TCP) in the presence of execution (random) and preparation (systematic) geometric errors.

MATERIALS AND METHODS

The dose matrix is blurred with all execution errors to estimate the total dose distribution of all fractions. To include preparation errors, the CTV is randomly displaced (and optionally rotated) many times with respect to its planned position while computing the dose, EUD, and TCP for the CTV using the blurred dose matrix. Probability distributions of these parameters are computed by combining the results with the probability of each particular preparation error. We verified the method by comparing it with an analytic solution. Next, idealized and realistic prostate plans were tested with varying margins and varying execution and preparation error levels.

RESULTS

Probability levels for the minimum dose, computed with the new method, are within 1% of the analytic solution. The impact of rotations depends strongly on the CTV shape. A margin of 10 mm between the CTV and planning target volume is adequate for three-field prostate treatments given the accuracy level in our department; i.e., the TCP in a population of patients, TCP(pop), is reduced by less than 1% due to geometric errors. When reducing the margin to 6 mm, the dose must be increased from 80 to 87 Gy to maintain the same TCP(pop). Only in regions with a high-dose gradient does such a margin reduction lead to a decrease in normal tissue dose for the same TCP(pop). Based on a rough correspondence of 84% minimum dose with 98% EUD, a margin recipe was defined. To give 90% of patients at least 98% EUD, the planning target volume margin must be approximately 2.5 Sigma + 0.7 sigma - 3 mm, where Sigma and sigma are the combined standard deviations of the preparation and execution errors. This recipe corresponds accurately with 1% TCP(pop) loss for prostate plans with clinically reasonable values of Sigma and sigma.

CONCLUSION

The new method computes in a few minutes the influence of geometric errors on the statistics of target dose and TCP(pop) in clinical treatment plans. Too small margins lead to a significant loss of TCP(pop) that is difficult to compensate for by dose escalation.

Margins for geometric uncertainty around organs at risk in radiotherapy

BACKGROUND AND PURPOSE

ICRU Report 62 suggests drawing margins around organs at risk (ORs) to produce planning organ at risk volumes (PRVs) to account for geometric uncertainty in the radiotherapy treatment process. This paper proposes an algorithm for drawing such margins, and compares the recommended margin widths with examples from clinical practice and discusses the limitations of the approach.

METHOD

The use of the PRV defined in this way is that, despite the geometric uncertainties, the dose calculated within the PRV by the treatment planning system can be used to represent the dose in the OR with a certain confidence level. A suitable level is where, in the majority of cases (90%), the dose-volume histogram of the PRV will not under-represent the high-dose components in the OR. In order to provide guidelines on how to do this in clinical practice, this paper distinguishes types of OR in terms of the tolerance doses relative to the prescription dose and suggests appropriate margins for serial-structure and parallel-structure ORs.

RESULTS

In some instances of large and parallel ORs, the clinician may judge that the complication risk in omitting a margin is acceptable. Otherwise, for all types of OR, systematic, treatment preparation uncertainties may be accommodated by an OR-->PRV margin width of 1.3Sigma. Here, Sigma is the standard deviation of the combined systematic (treatment preparation) uncertainties. In the case of serial ORs or small, parallel ORs, the effects of blurring caused by daily treatment execution errors (set-up and organ motion) should be taken into account. Near a region of high dose, blurring tends to shift the isodoses away from the unblurred edge as shown on the treatment planning system by an amount that may be represented by 0.5sigma. This margin may be used either to increase or to decrease the margin already calculated for systematic uncertainties, depending upon the size of the tolerance dose relative to the detailed planned dose distribution. Where the detailed distribution is unknown before the OR is delineated, then the overall margin for serial or small parallel ORs should be 1.3Sigma+0.5sigma. Examples are given where the application of this algorithm leads to margin widths around ORs similar to those in use clinically.

CONCLUSIONS

Using PRVs is appropriate both for forward and inverse planning. Dose-volume histograms of PRVs for serial- and parallel-structure ORs require careful interpretation. Nevertheless, use of the proposed algorithms for drawing margins around both serial and parallel ORs can alert the dosimetrist/radiation oncologist to the possibility of high-dose complications in individual treatment plans, which might be missed if no such margins were drawn.

Uncertainties in model-based outcome predictions for treatment planning

PURPOSE

Model-based treatment-plan-specific outcome predictions (such as normal tissue complication probability [NTCP] or the relative reduction in salivary function) are typically presented without reference to underlying uncertainties. We provide a method to assess the reliability of treatment-plan-specific dose-volume outcome model predictions.

METHODS AND MATERIALS

A practical method is proposed for evaluating model prediction based on the original input data together with bootstrap-based estimates of parameter uncertainties. The general framework is applicable to continuous variable predictions (e.g., prediction of long-term salivary function) and dichotomous variable predictions (e.g., tumor control probability [TCP] or NTCP). Using bootstrap resampling, a histogram of the likelihood of alternative parameter values is generated. For a given patient and treatment plan we generate a histogram of alternative model results by computing the model predicted outcome for each parameter set in the bootstrap list. Residual uncertainty ("noise") is accounted for by adding a random component to the computed outcome values. The residual noise distribution is estimated from the original fit between model predictions and patient data.

RESULTS

The method is demonstrated using a continuous-endpoint model to predict long-term salivary function for head-and-neck cancer patients. Histograms represent the probabilities for the level of posttreatment salivary function based on the input clinical data, the salivary function model, and the three-dimensional dose distribution. For some patients there is significant uncertainty in the prediction of xerostomia, whereas for other patients the predictions are expected to be more reliable. In contrast, TCP and NTCP endpoints are dichotomous, and parameter uncertainties should be folded directly into the estimated probabilities, thereby improving the accuracy of the estimates. Using bootstrap parameter estimates, competing treatment plans can be ranked based on the probability that one plan is superior to another. Thus, reliability of plan ranking could also be assessed.

CONCLUSIONS

A comprehensive framework for incorporating uncertainties into treatment-plan-specific outcome predictions is described. Uncertainty histograms for continuous variable endpoint models provide a straightforward method for visual review of the reliability of outcome predictions for each treatment plan.

The reproducibility of organ position using active breathing control (ABC) during liver radiotherapy

PURPOSE

To evaluate the intrafraction and interfraction reproducibility of liver immobilization using active breathing control (ABC).

METHODS AND MATERIALS

Patients with unresectable intrahepatic tumors who could comfortably hold their breath for at least 20 s were treated with focal liver radiation using ABC for liver immobilization. Fluoroscopy was used to measure any potential motion during ABC breath holds. Preceding each radiotherapy fraction, with the patient setup in the nominal treatment position using ABC, orthogonal radiographs were taken using room-mounted diagnostic X-ray tubes and a digital imager. The radiographs were compared to reference images using a 2D alignment tool. The treatment table was moved to produce acceptable setup, and repeat orthogonal verification images were obtained. The positions of the diaphragm and the liver (assessed by localization of implanted radiopaque intra-arterial microcoils) relative to the skeleton were subsequently analyzed. The intrafraction reproducibility (from repeat radiographs obtained within the time period of one fraction before treatment) and interfraction reproducibility (from comparisons of the first radiograph for each treatment with a reference radiograph) of the diaphragm and the hepatic microcoil positions relative to the skeleton with repeat breath holds using ABC were then measured. Caudal-cranial (CC), anterior-posterior (AP), and medial-lateral (ML) reproducibility of the hepatic microcoils relative to the skeleton were also determined from three-dimensional alignment of repeat CT scans obtained in the treatment position.

RESULTS

A total of 262 fractions of radiation were delivered using ABC breath holds in 8 patients. No motion of the diaphragm or hepatic microcoils was observed on fluoroscopy during ABC breath holds. From analyses of 158 sets of positioning radiographs, the average intrafraction CC reproducibility (sigma) of the diaphragm and hepatic microcoil position relative to the skeleton using ABC repeat breath holds was 2.5 mm (range 1.8-3.7 mm) and 2.3 mm (range 1.2-3.7 mm) respectively. However, based on 262 sets of positioning radiographs, the average interfraction CC reproducibility (sigma) of the diaphragm and hepatic microcoils was 4.4 mm (range 3.0-6.1 mm) and 4.3 mm (range 3.1-5.7 mm), indicating a change of diaphragm and microcoil position relative to the skeleton over the course of treatment with repeat breath holds at the same phase of the respiratory cycle. The average population absolute intrafraction CC offset in diaphragm and microcoil position relative to skeleton was 2.4 mm and 2.1 mm respectively; the average absolute interfraction CC offset was 5.2 mm. Analyses of repeat CT scans demonstrated that the average intrafraction excursion of the hepatic microcoils relative to the skeleton in the CC, AP, and ML directions was 1.9 mm, 0.6 mm, and 0.6 mm respectively and the average interfraction CC, AP, and ML excursion of the hepatic microcoils was 6.6 mm, 3.2 mm, and 3.3 mm respectively.

CONCLUSION

Radiotherapy using ABC for patients with intrahepatic cancer is feasible, with good intrafraction reproducibility of liver position using ABC. However, the interfraction reproducibility of organ position with ABC suggests the need for daily on-line imaging and repositioning if treatment margins smaller than those required for free breathing are a goal.

Radiation-induced pulmonary toxicity: a dose-volume histogram analysis in 201 patients with lung cancer

PURPOSE

To relate lung dose-volume histogram-based factors to symptomatic radiation pneumonitis (RP) in patients with lung cancer undergoing 3-dimensional (3D) radiotherapy planning.

METHODS AND MATERIALS

Between 1991 and 1999, 318 patients with lung cancer received external beam radiotherapy (RT) with 3D planning tools at Duke University Medical Center. One hundred seventeen patients were not evaluated for RP because of <6 months of follow-up, development of progressive intrathoracic disease making scoring of pulmonary symptoms difficult, or unretrievable 3D dosimetry data. Thus, 201 patients were analyzed for RP. Univariate and multivariate analyses were performed to test the association between RP and dosimetric factors (i.e., mean lung dose, volume of lung receiving >or=30 Gy, and normal tissue complication probability derived from the Lyman and Kutcher models) and clinical factors, including tobacco use, age, sex, chemotherapy exposure, tumor site, pre-RT forced expiratory volume in 1 s, weight loss, and performance status.

RESULTS

Thirty-nine patients (19%) developed RP. In the univariate analysis, all dosimetric factors (i.e., mean lung dose, volume of lung receiving >or=30 Gy, and normal tissue complication probability) were associated with RP (p range 0.006-0.003). Of the clinical factors, ongoing tobacco use at the time of referral for RT was associated with fewer cases of RP (p = 0.05). These factors were also independently associated with RP according to the multivariate analysis (p = 0.001). Models predictive for RP based on dosimetric factors only, or on a combination with the influence of tobacco use, had a concordance of 64% and 68%, respectively.

CONCLUSIONS

Dosimetric factors were the best predictors of symptomatic RP after external beam RT for lung cancer. Multivariate models that also include clinical variables were slightly more predictive.

[Physical and methodological aspects of multimodality imaging and principles of treatment planning in 3D conformal radiotherapy]

The recent evolutions of the imaging modalities, the dose calculation models, the linear accelerators and the portal imaging permit to improve the quality of the conformal radiation therapy treatment planning. With DICOM protocols, the acquired imaging data coming from different modalities are treated by performant image fusion algorithms and yield more precise target volumes and organs at risk. The transformation of the clinical target volumes (CTV) to planning target volumes (PTV) can be realised using advanced probabilistic techniques based on clinical experience. The treatment plans evaluation is based on the dose volume histograms. Their precision and clinical relevance are improved by the multi-modality imaging and the advanced dose calculation models. The introduction of the inverse planning systems permitting to realise modulated intensity radiation therapy generates highly conformal dose distributions. All the previously cited complex techniques require the application of rigorous quality assurance programs.

Set-up verification of cervix cancer patients treated with long treatment fields; implications of a non-rigid bony anatomy

BACKGROUND AND PURPOSE

For cervix cancer patients, treatment fields may extend up to vertebra L1. In clinical practice, set-up verification is based on measured displacements of the pelvic rim as visible in the caudal part of the treatment fields. The implications of this procedure for the positions of bony structures in the cranial part of the fields were investigated.

MATERIALS AND METHODS

Twelve patients had four repeat simulator sessions. Both during treatment simulation (the reference) and the repeat sessions, anterior radiographs were acquired covering the whole treatment field. The films were used to investigate differences between the cranial and the caudal parts of the treatment field in day-to-day bony anatomy displacements.

RESULTS

Both in the transversal and the longitudinal directions, these differences were significant (3.5 mm, 1 SD). Indications were found that large differences in the cranio-caudal direction may be correlated with (non-rigid) internal pelvic rim rotations around a lateral axis. In the longitudinal direction, the position of L1 correlated much better with the position of vertebra S1 than with the position of the pelvic rim, which is usually used for set-up verification.

CONCLUSIONS

Due to the non-rigid bony anatomy of the studied patients, the usual set-up verification and correction procedure can result in set-up errors of 10 mm and more for structures in the cranial part of the treatment field, even in the case of a perfect set-up of the pelvic rim. Possibly, other patient set-up and immobilization procedures may result in a better day-to-day reproducibility of the 3D bony anatomy shape. (Remaining) Differences in anatomy position changes between the caudal and cranial field ends may be accounted for by using non-uniform clinical target volume-to-planning target volume margins, or by an adapted patient set-up verification and correction protocol.

Modelling the dosimetric consequences of organ motion at CT imaging on radiotherapy treatment planning

Treatment planning algorithms usually assume that the correct or at least the mean organ position is derived from the CT imaging procedure, and that this position is reproduced throughout the treatment. In reality a mobile organ is unlikely to be in its exact mean position at the time of imaging, causing the treatment to be planned with an organ off-set from its assumed mean position. This introduces an extra 'CT uncertainty' into the treatment. A Monte Carlo (MC) model is used to simulate organ translations at imaging and evaluate the effect of this uncertainty (above the treatment delivery uncertainties) on the dose distribution. An underdose by 4 Gy in a 60 Gy treatment is calculated in the penumbral region of a single-field dose distribution as a result of the CT uncertainty. The effect is reduced to less then 0.5 Gy when the organ position at planning is derived as the average from multiple pretreatment CT scans. It is shown that a convolution method can be applied to predict the effect of CT uncertainty on the dose distribution for a patient population. Additionally, a variation kernel for a convolution method is derived that incorporates uncertainty at both imaging and treatment.

Extracranial radiosurgery: immobilizing liver motion in dogs using high-frequency jet ventilation and total intravenous anesthesia

PURPOSE

Extracranial radiosurgery requires control of organ motion. The purpose of this study is to quantitatively determine the extent of liver motion in anesthetized dogs with continuous i.v. propofol infusion with or without muscle relaxants and high-frequency jet ventilation.

METHODS AND MATERIALS

Five dogs were used in the experiment. Each dog was restrained while anesthetized in the supine position using an alpha cradle. Surgical metal clips were implanted around the liver periphery so that its motion could be visualized using a fluoroscopic imaging device in a conventional simulator. Initially, two orthogonal simulation films were taken to correlate locations of implanted clips. Two orthogonal views of fluoroscopic images for each anesthetized dog were recorded on a magnetic tape and analyzed from the post-imaging data. Liver motion was documented under the following three conditions: 1) ventilated with a conventional mechanical ventilator, 2) ventilated with a high-frequency jet ventilator, and 3) ventilated with a high-frequency jet ventilator and total muscle paralysis (with vecuronium injection). The maximum liver motion for each dog was analyzed in three orthogonal directions: the inferior-to-superior direction, the anterior-to-posterior direction, and the right-to-left direction.

RESULTS

When the anesthetized dogs were ventilated with a conventional mechanical ventilator, the average liver motions were 1.2 cm in the inferior-to-superior direction, 0.4 cm in the anterior-to-posterior direction, and 0.2 cm in the right-to-left direction, respectively. After the introduction of high-frequency jet ventilation, the average liver motions were reduced to 0.2 cm in the inferior-to-superior direction, 0.2 cm in the anterior-to-posterior direction, and 0.1 cm in the right-to-left direction. The maximum liver motion was dependent on ventilator settings. There was no additional measurable motion reduction with the addition of the muscle relaxant.

CONCLUSION

The liver motion in each anesthetized dog was controlled under 3.0 mm in all directions with the use of high-frequency jet ventilation. No detectable advantage was identified by the injection of muscle relaxant in terms of further reducing the liver motion. The preclinical animal study indicated that the use of high-frequency jet ventilation (HFJV) would be able to limit the liver motion to an extent acceptable for the application of extracranial radiosurgery in humans. Radiosurgery for localized liver tumors warrants further investigation.

Considerations for the implementation of target volume protocols in radiation therapy

PURPOSE

Uncertainties in patient repositioning and organ motion are accounted for by defining a planning target volume (PTV). We make recommendations on issues not explicitly discussed in existing protocols for PTV design.

METHODS

A quantity called "coverage" is defined to quantify how effectively a PTV encompasses the clinical target volume, and is applied to examine the impact of several factors. A stochastic simulation is used to determine the coverage required for a desirable balance between tumor control probability (TCP) and the irradiated volume. Using a sample anatomy, we assess the importance of the method used to add uncertainties, the shape of the uncertainty distribution, the effect of systematic uncertainties, and the use of nonuniform margins. Additionally, we examine the benefit of patient immobilization techniques.

RESULTS

Our example indicates that 95% coverage is a reasonable goal for treatment planning. Using this as a comparison value, our example indicates quadrature addition of uncertainties predicts smaller margins (7 mm) than linear addition (11 mm), Gaussian distribution of uncertainties (7 mm) require the same margin as a uniform distribution (7 mm), systematic uncertainties have a small effect on TCP below a threshold value (4 mm), and nonuniform margins allow only a slight reduction of irradiated volume.

CONCLUSION

We recommend that uncertainties should generally be added in quadrature, the exact shape of the uncertainty distribution is not critical, systematic uncertainties should be maintained below some threshold value, and nonuniform margins may be effective when uncertainties are anisotropic.

Physical aspects of a real-time tumor-tracking system for gated radiotherapy

PURPOSE

To reduce uncertainty due to setup error and organ motion during radiotherapy of tumors in or near the lung, by means of real-time tumor tracking and gating of a linear accelerator.

METHODS AND MATERIALS

The real-time tumor-tracking system consists of four sets of diagnostic X-ray television systems (two of which offer an unobstructed view of the patient at any time), an image processor unit, a gating control unit, and an image display unit. The system recognizes the position of a 2.0-mm gold marker in the human body 30 times per second using two X-ray television systems. The marker is inserted in or near the tumor using image guided implantation. The linear accelerator is gated to irradiate the tumor only when the marker is within a given tolerance from its planned coordinates relative to the isocenter. The accuracy of the system and the additional dose due to the diagnostic X-ray were examined in a phantom, and the geometric performance of the system was evaluated in 4 patients.

RESULTS

The phantom experiment demonstrated that the geometric accuracy of the tumor-tracking system is better than 1.5 mm for moving targets up to a speed of 40 mm/s. The dose due to the diagnostic X-ray monitoring ranged from 0.01% to 1% of the target dose for a 2.0-Gy irradiation of a chest phantom. In 4 patients with lung cancer, the range of the coordinates of the tumor marker during irradiation was 2.5-5.3 mm, which would have been 9.6-38.4 mm without tracking.

CONCLUSION

We successfully implemented and applied a tumor-tracking and gating system. The system significantly improves the accuracy of irradiation of targets in motion at the expense of an acceptable amount of diagnostic X-ray exposure.

Autoactivation of source dwell positions for HDR brachytherapy treatment planning

The most accurate classical dose optimization algorithms in HDR brachytherapy strongly depend on an appropriate selection of source dwell positions which fulfill user-defined geometrical boundary conditions which are relative to patient anatomy. Most anatomical situations, such as for prostate and head and neck tumors, are complex and can require geometries with 5-15 catheters with 48 possible dwell positions per catheter depending on the tumor volume. The manual selection of dwell positions using visual checks by trial and error is very time consuming. This can only be improved by the use of a technique which automatically recognizes and selects the optimum dwell positions for each catheter. We have developed an algorithm, termed an autoactivation algorithm, which improves implant planning by providing a facility for the necessary automatic recognition of HDR source dwell positions.

The width of margins in radiotherapy treatment plans

Publication of ICRU Reports 50 and 62 has highlighted the need to devise protocols for the process of drawing the planning target volume (PTV) around the clinical target volume (CTV). The margin surrounding the CTV should be wide enough to account for all geometric errors so that no part of the CTV accumulates a dose less than, for instance, 95% of that prescribed. One approach to the problem has been to draw a margin around the CTV delineated at the treatment preparation stage which is sufficiently wide that the mean position of the CTV will be encompassed in a specific percentage of cases, for example 90%. This accounts for the systematic errors. A further margin is then drawn to account for random set-up and organ-motion uncertainties during treatment. The width of this second margin has previously been shown to be 1.64(sigma - sigmap). Here sigma, a vector quantity, is the standard deviation which results from convolving the penumbra spread function of standard deviation sigmap with the Gaussian distributions of the daily positional uncertainties of organ motion and set-up error. However, it is shown in this paper that the calculation should take into account the beam configuration of the treatment plan. In a typical coplanar multibeam plan, usually in the transverse plane, any given edge of the target volume is normally defined by a single beam or two parallel and opposed beams. However, because of the presence of the other beams, the effect of the blurring of the edge-defining beam(s) is reduced, which changes the value of the required margin to beta (sigma - sigmap) where, for example, beta can be as low as 1.04 in the transverse plane of a three-beam plan. The width of the required margins is calculated for up to six beams and presented in a table. It is shown that, while the table was derived using an idealized plan of equally weighted plane beams irradiating a spherical target, it is also valid for non-uniform beam weightings, wedged-beam plans, target volumes of general shape and intensity-modulated radiotherapy (IMRT).