Quality assurance procedures for the peacock system

Medical Dosimetry, Vol. 26, No. 1, pp. 83–90, 2001 Copyright © 2001 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/01/$–see front matter

PII: S0958-3947(00)00058-3

QUALITY ASSURANCE PROCEDURES FOR THE PEACOCK SYSTEM CHENG B. SAW, PH.D.,* KOMANDURI M. AYYANGAR, PH.D., WEINING ZHEN, M.D., ROBERT B. THOMPSON, M.D., and CHARLES A. ENKE, M.D. Department of Radiation Oncology, University of Nebraska Medical Center, Omaha, NE, USA (Accepted 2 December 2000)

Abstract—The Peacock system is the product of technological innovations that are changing the practice of radiotherapy. It uses dynamic beam modulation technique and inverse planning algorithm, both of which are new methodologies, to perform intensity-modulation radiation therapy (IMRT). The quality assurance (QA) procedure established by Task Group No. 40 did not adequately consider these emerging modalities. A review of literature indicates that published articles on QA procedures concentrate primarily on the verification of dose delivered to phantom during commissioning of the system and dose delivered to phantom before treating patients. Absolute dose measurements using ion chambers and relative dose measurements using film dosimetry have been used to verify delivered doses. QA on equipment performance and equipment safety is limited. This paper will discuss QA on equipment performance, equipment safety, and patient setup reproducibility. © 2001 American Association of Medical Dosimetrists. Key Words: Quality assurance, IMRT, Dosimetry, Photon beam.

INTRODUCTION

conventional radiotherapy equipment because it incorporates beam modulation and implements inverse planning algorithm. With these distinct differences, existing quality assurance (QA) procedures for conventional radiotherapy are not applicable. Because the Peacock system was not widely available at the time of publication, the comprehensive QA for the radiation oncology report of AAPM Task Group No. 40 did not address these distinct features of the Peacock system.5 The task group suggested that the QA procedures should strictly follow the guidelines established by the manufacturer. QA suggested by the manufacturer (NOMOS) has also been limited. Suggestions made are primarily for commissioning of the Peacock system. A section of the Beam Utility manual discussed commissioning test procedure.6 Beam data can be analyzed and validated using the verification plan mode software. A review of literature indicates that published QA procedures emphasize dose verification during commissioning of the system and dose delivered to phantom before treating the patients.2,7–12 It is understandable that accurate dose delivery is of paramount importance, because conventional methods of dose verification are not applicable. The aspects of equipment performance and equipment safety are emphasized to a lesser extent compared to dose verification.3,13 Dose verification during commissioning has been addressed by Saw et al.14 This paper will review QA for equipment performance, safety, and patient setup reproducibility. QA tasks for the Peacock System are listed in Table 1.

The Peacock system is a product of technological advancement that is expected to change the practice of radiotherapy.1 The components of the Peacock system are the CORVUS 3.0 planning system, which uses an inverse-planning algorithm to create the beam modulation instructions and the multileaf modulating collimator (MIMiC), which modulates the beam during dose delivery. The MIMiC is a computerized modulator comprised of 40 leaves divided into 2 banks. The controller directs the MIMiC’s leaves to open and close via pneumatic mechanism and hence modulates the radiation beam. In addition to the MIMiC, controller, and electrical and air supplies, additional hardware used to assist in the operation of the beam delivery are the radiotherapy table adapter (RTA) device and the CRANE. The CRANE is an indexing device that is used to precisely move the linear accelerator couch for the slice-by-slice method of dose delivery. The RTA device is used for the fixation of the Aquaplast mask and Talon system. The components of the Peacock beam delivery system attached to our Siemens PRIMUS linear accelerator are shown in Fig. 1. The hardware received Federal Food and Drug Administration (FDA) clearance in 1995, while the total integrated system received FDA clearance in 1996.2,3 The Peacock system became the first commercially available system to perform intensity modulation radiation therapy (IMRT) using multileaf modulating collimator and inverse planning algorithm.4 The Peacock system is drastically different from

QUALITY ASSURANCE FOR EQUIPMENT The beam delivery components of the Peacock system are attachments to our PRIMUS linear accelerator. It

Reprint requests to: *Cheng B. Saw, Ph.D., Department of Radiation Oncology/UNMC, 987521 Nebraska Medical Center, Omaha, NE 68198-7521. E-mail: [email protected] 83

84

Medical Dosimetry

Volume 26, Number 1, 2001

Fig. 1. Components of the beam delivery system of the Peacock system are: (a) MIMiC attached to the treatment head of a linear accelerator, (b) RTA device attached to the treatment couch and a linear accelerator for immobilization of head and neck, (c) CRANE II attached to the treatment couch of a linear accelerator, and (d) power and air supplies for functioning of the MIMiC.

takes advantage of the safety features of a linear accelerator. These safety features, when invoked, should also terminate the delivery of IMRT. The MIMiC system is actually independent of the linear accelerator. It only responds to gantry position sensed by the inclinometer inside the controller. Because the inclinometer is sensitive to gantry speed, the MIMiC will perform beam modulation if the gantry speed value is within the limit set in the controller. The MIMiC cannot sense whether or not the radiation is energized. The QA procedures for a linear accelerator have been established by Task Group No. 40.5 Safety features associated with linear accelerators should be checked on a daily basis. These daily safety checks, when failed,

should also halt the delivery of IMRT. In addition to the safety features, the alignment lasers should be checked daily. Lasers are used to align fiducial markers placed on fabricated masks or markers on the Talon system to the isocenter serving as an alignment point. The MIMiC must be properly attached to the linear accelerator and aligned to the beam axis. Alignment is achieved with pins fitted into designated slots drilled onto the treatment head. Alignment is checked daily and whenever the MIMiC is attached to the linear accelerator. Irradiating through retracted alternate leaves like a checkerboard can conveniently be used to check MIMiC alignment. The alignment is considered acceptable if a film at isocenter exposed from the laterals using 90° and

QA for Peacock System ● C. B. SAW et al.

Table 1. Quality assurance tasks 1. 2. 3. 4. 5. 6. 7. 8.

Daily Daily QA on linear accelerator Check laser alignment Attach MIMiC; follow MIMiC setup checklist Check MIMiC alignment with lasers Check low/high (20/60 psi) pressure on air supplier Attach RTA device, aligned and leveled Check integrity of MIMiC, cables, RTA device Detach MIMiC; follow MIMiC setup checklist in reverse order

1. 2. 3. 4.

Patient Verification Trial run to check MIMiC clearance Perform hybrid plan for absolute dose measurement in phantom Perform hybrid plan for relative dose measurement in phantom Perform independent dose verification if available

1. 2. 3. 4. 5. 6. 7.

Periodic - Weekly/Monthly/As needed Bleed oil and water from compressor Perform radiographic check on MIMiC alignment with beam axis Perform clinometer accuracy check Indexing couch width check for 1- and 2-cm treatment mode CRANE travel vs. readout check Reboot CORVUS to clear temporary files Archive patients from CORVUS for additional storage space

270° gantry angles show well-separated beamlets, as illustrated in Fig. 2. A less stringent test would be to irradiate and capture the image of the checker designed on an electronic portal imager. Such a test would identify incorrect field orientation but may not show misalignment of the MIMiC with beam axis. In addition to alignment, proper electrical and pneumatic connection must be checked. Loose pneumatic connections can result in continuous leakage of pressure, leading to leaf fault. Likewise, loose electronics would also lead to leaf

Fig. 2. Checkerboard pattern design using MIMiC leaves and the resulting exposed film from laterals.

85

and incompatible gantry angle faults. An improperly attached RTA device to the treatment couch would lead to difficulty in patient setup. A poorly bolted CRANE to the side of the treatment couch can cause shearing of threads on the device that holds the CRANE to the treatment couch rails. Lastly, the treatment couch step size should also be checked periodically for the 2 treatment modes. The method of determining the treatment couch step size has been described by Saw et al.14 Equipment performance must be monitored at regular intervals (daily, weekly, or monthly). The low/high pressure on the MIMiC power supply should be checked and set at the appropriate level. Oil and water in the compressor should be bled to reduce leaf errors during treatment. The integrity of the MIMiC should be checked for loose screws. In addition, the electrical and pneumatic cables should also be checked for integrity. Likewise, the integrity of the RTA device attached to treatment couch should be checked on regular intervals for alignment and leveling. The CRANE attached to the treatment couch should be checked for its integrity (i.e., loose screws). The moving components attached to the hand rail of the treatment couch and the general movement of the CRANE should also be evaluated. The accuracy and functionality of the LCD readout on the CRANE should be checked; the readout should correspond to the distance traveled by the treatment couch. PATIENT TREATMENT PROCEDURE The Peacock system incorporates evolving imaging technologies such as computed tomography (CT), magnetic resonance imaging (MRI), and nuclear medicine scans, advancement of computer technology, and beam modulation delivery systems. A patient undergoing IMRT will undergo a patient data collection using scanning modalities, treatment setup, target localization, planning, and treatment delivery in a very systematic fashion. These steps are reviewed for familiarity before addressing QA issues, to provide optimized patient care. Patient data, such as external patient contour and internal anatomical structure for IMRT, are obtained from the CT scan. Prior to undergoing a CT scan, the patient must be immobilized to ensure reproducibility of patient setup. Immobilization is performed in our simulator room. For head and neck treatment, the CT attachment is positioned on the simulation couch. The mask device is placed on top of the attachment after which a head holder is positioned. A mask is made after the patient is positioned properly. At our institution, the patient must be positioned such that the target should not be more than 10 cm from the isocenter. This requirement is important because the MIMiC has a field width of 20 cm. The inhomogeneity caused by abutment is further reduced if the isocenter is centrally positioned within the target. After the mask hardens, fiducial markers are placed on the mask at the position marked by the lateral

86

Medical Dosimetry

Volume 26, Number 1, 2001

Table 2. Quality assurance—MIMiC setup checklist No. 1. 2. 3.

4. 5. 6. 7.

8. 9. 10. 11. 12. 13.

14.

Checks

Mon

Tue

Wed

Thu

Fri

Sat

Sun

1. 2. 1. 1. 2. 3. 4. 5. 1. 1. 2. 3. 1. 2. 1. 2.

Set treatment couch rotation to zero Check laser alignment Remove treatment couch head piece Set field size to Jx ⫽ 5.0 cm and Jy ⫽ 21.0 cm Raise treatment couch close to isocenter Pull treatment couch longitudinally out and away from gantry Rotate collimator to 0° Rotate gantry to 180° Insert shield in wedged tray mount Insert MIMiC into tray mount Release 2 side fixation pins Tighten 4 fixation screws Attach controller with “NOMOS” upside down Connect cable controller to MIMiC Rotate collimator to 89.7° Set pendant to gantry rotation mode to avoid accidental rotation after attaching cables from gantry to MIMiC and controller 3. Attach power and air umbilical cables from MIMiC and controller to the linear accelerator 1. Turn electrical and air supplies power on 2. Switch interlock to Peacock treatment 1. Attach collimator rotation fixation device 1. Rotate gantry to 0° 1. Follow computer bootup instructions 2. Use touch screen and follow directions 1. Attach RTA device for head & neck treatment 2. Tighten the RTA device using hex wrench 1. Bring CRANE close to LEFT side of treatment couch 2. Slowly guide the CRANE to slide on groove on the attachment 3. Screw CRANE to the treatment couch base 4. Unscrew the CRANE to storage table 5. Press down the storage table and remove it 1. Check to make sure the counter balance weight is attached to linear accelerator Performed By:

lasers and anterior crosshair. After the immobilization devices have been made, the patient undergoes CT scans. In general, the CT scan is taken 5 cm above and 5 cm below the target, with a slice thickness of 3 mm. For the head and neck region, the scan starts from the top of the head slice to 5 cm below the target. This procedure permits beam entry from the superior. The CT scans are then exported to the CORVUS planning system. After the CT scans are imported into the CORVUS system, it is reviewed for acceptability to perform IMRT planning. After the CT is reviewed, patient information, such as attending physician, is entered. The next steps consist of defining tissue or external contour, setting region of dose calculation, and adjusting the window levels. Next, the target and critical structure are identified and contoured. Doses are then assigned to both targets and critical structures. Generally, the assigned doses should be less than the desired doses because the optimized plan from CORVUS generally gives doses to target and organ at risks that closely match the desired doses. A treatment unit is then chosen and the data are sent to the CORVUS dose engine for computation. After the computation, the plan is reviewed. If the plan is unacceptable, the dose assignment is modified until an

acceptable plan is generated. Once an acceptable plan is obtained, the beam delivery instructions are transferred to the MIMiC controller via a floppy disk. The beam data for the patient is also applied to a phantom for dose verification. The application of patient beam geometry to a phantom is referred to as a hybrid plan by the vendor. The hybrid plan allows a means of obtaining confidence that the system delivers a known dose to a phantom and delivers a defined dose to the patient. The dose to a phantom can be different from the dose to the patient. The hybrid plan is carried out to 2 phantoms; one measures the absolute dose using the ion chamber, and the other measures the relative dose using film dosimetry. ABSOLUTE DOSE MEASUREMENT IN PHANTOM Although absolute dose can be determined from film dosimetry, this procedure is currently not used because of the many uncertainties associated with film processing. Flat phantom is used for ion chamber measurement. At our institution, a solid water phantom of 30 ⫻ 30 cm, with a thickness of 12 cm, is used for

QA for Peacock System ● C. B. SAW et al.

87

Fig. 4. NOMOS verification cassette for film dosimetry.

Fig. 3. Ion chamber phantom for point dose measurement.

patients treated at the head and neck region. For large patient thickness, a 22-cm-thick solid water phantom is used. Dose measurement is performed to the region of high dose with minimal dose gradient, which is generally the center of the target volume. The planning system will compute doses such that a point in the patient plan corresponds to the point at the center of the ion chamber. After the plan is generated, the beam data are transferred via a floppy disk to the MIMiC controller. The phantom and ion chamber setup is shown in Fig. 3. No adjustment of monitor units is made if the dose measurement is within 5% of the dose predicted by the hybrid plan. Thus far, we have not made any adjustment or repeat the plan based on this criterion.

processors may not have stringent QA, leading to changes of optical density depending on the time of processing the film. In addition, the optical density may change from one box to another. As such, film dosimetry is currently considered inaccurate. The film is scanned to determine the relative optical density and converted to relative dose distribution. Currently, digital optical density can be obtained using a digital scanner such as Vidar XVR-12 or Vidar XVR-16. A sample of irradiated film and its relative dose distribution is shown in Fig. 6. Because the film is totally enclosed within the verification cassette, the spatial coordinates of the point on the film cannot be correlated precisely with the isocenter. This is of concern when the dose to a critical structure, such as the spinal cord, cannot be determined with certainty. The spatial information can be defined using a beamlet because the beamlets can be referenced to the isocenter, as illustrated in Fig. 6. However, one needs to identify the location of the intersection

RELATIVE DOSE MEASUREMENT IN PHANTOM In addition to point dose measurement, the relative dose distribution in phantom is obtained using film dosimetry. A vendor-supplied verification cassette made of high-impact polystyrene, as shown in Fig. 4, is used for film dosimetry. There are several slabs of polystyrene within the verification cassettes. Films cut to size are inserted between the slab. The positioning of the film at isocenter may show leakage radiation between the 2 banks of leaves, as illustrated in Fig. 5, and can impact on the measurements. One of the many difficulties of film dosimetry has been the variation of film processing. Film processors are now automated and hence the chemical flow, temperature, and mixture should be properly maintained. Older

Fig. 5. Radiation leakage between the 2 banks of leaves.

88

Medical Dosimetry

Volume 26, Number 1, 2001

Fig. 6. Irradiated film and relative dose distribution. Beamlets were used to define the location of critical structures.

of the beamlets using an accurate method so that the cold spot on the plan can be located precisely. Some software allows the superposition of measured isodose distributions from film onto planned isodose distributions. INDEPENDENT MONITOR UNIT CALCULATION Point dose and relative dose measurements are currently important methods of verifying the treatment delivery for IMRT. The physical measurements are laborious but crucial until other effective methods of validation of dose delivered become available. Several authors15,16 have attempted to verify the dose delivered for IMRT. A method of verifying dose delivered using the Peacock System is proposed by Ayyangar et al.17 PATIENT LOCALIZATION AND TREATMENT IMRT requires accurate and highly reproducible patient setup that is more stringent than conventional technique. As such, particular attention must be given to immobilization and alignment to the lasers, which would yield simple and easy reproducible patient setup. To ensure accurate patient setup, a left lateral film is taken for verification purpose. The portal image is taken with the 2-cm mode presenting a 4 ⫻ 20-cm field as illustrated in Fig. 7b. The 2-cm mode with open-leaves option is accessed under the main menu by selecting the MIMiC test followed by all leaves-open item menu. After taking the portal film, the patient data on the floppy disk from CORVUS can be downloaded to the controller of the MIMiC for patient treatment. The controller requests an approval number to validate the beam delivery instructions for the patient selected by the operator. The

CRANE is then attached to the treatment couch handrail. The LCD readouts of the CRANE are zeroed and the indexing is set according to the planning instruction printout. The treatment couch must be disengaged before performing any indexing using the CRANE, otherwise there will be torque applied between the CRANE and treatment couch locking system, causing incorrect setup and/or wrecking the CRANE. It is advisable to draw laser lines on the mask or immobilization device after each indexing, as depicted in Fig. 8. This procedure would ensure the reproducibility of the daily indexing. DISCUSSION The implementation of the Peacock system requires the addition of hardware to a linear accelerator. This addition of hardware creates limited clearance between the unit and the treatment couch. As such, extra care must be taken to ensure that collisions between components are avoided. The advances of linear accelerators, toward more computer controls and automatic setup, also increase the risk of collision between components. Institutions such as ours, which do not use the Peacock system for the entire day, must remove and re-install the MIMiC as needed. A checklist of instructions for the sequential assembly of hardware as listed in Table 2 is essential to ensure that the attachment is proper and complete to avoid collision and/or destruction of equipment. Attachment of the cables first followed by a collimator rotation would easily tear off the electrical and pneumatic cables. Although complete arc rotation of the MIMiC is the best option for a treatment plan, such an option may not be possible because of the limited clearance between the

QA for Peacock System ● C. B. SAW et al.

89

Fig. 7. Simulation and localization films to verify patient setup.

Peacock hardware and the patient and/or treatment couch. As such, a trial run should be performed for each patient plan to avoid equipment collision. In addition to the avoidance of collision, a trial run also ensures that the MU per degree is adjusted properly, hence avoiding rotational fault during treatment. Since the installation, interlock communications between the linear accelerator and the MIMiC system

Fig. 8. Laser markers of treatment slice on masks to check setup reproducibility.

have not been established at our institution. The use of the Peacock system as a separate system has assisted us in reducing errors. First, therapists are made aware that the MIMiC must be disassembled before bringing in another patient from the IMPAC system into the PRIMUS. As stated above, if the MIMiC is left on the machine and the next patient requires a collimator rotation, automatic setup procedure would tear the electrical and pneumatic connection apart. We have halted patient treatments for a variety of reasons. When the linear accelerator is halted, the MIMiC immediately closes its leaves and provides instructions on how to proceed with the treatment. This feature of the Peacock system has been helpful in assisting therapists and also reduces the concern of inadvertent irradiation. The beam modulation start after 15° arc is another useful features of the MIMiC system. Accidental programming in the fixed gantry mode would not cause irradiation to the patient except radiation leakage because the leaves are closed all the time. However, therapists should concentrate on listening for the popping sounds from the MIMiC, caused by moving leaves, after the gantry travels 10° 15°, otherwise there is error at the controller of the MIMiC system. Because the Peacock system uses a rotational mode for beam delivery, it is not possible to perform localization for the continuous mode. Generally, 2 orthogonal portals, the anterior and lateral portals, are sufficient for

90

Medical Dosimetry

localization in the rotational mode. For the Peacock system, the anterior portal is not useful because the radiograph is overshadowed by the hardware. Typical localization and verification of daily patient setup is carried out using only the lateral portal. However, the MIMiC opening of 21 ⫻ 5 cm presents (actually 20 ⫻ 3.3 cm) gives a limited portal field size. Because of the limited field size and lack of blocks, assessing MIMiC portal films is done differently, compared to conventional technique. Although there are many ways of assessing the lateral radiograph for daily reproducibility, we have chosen to use the geometrical features of beam superimposed onto patient anatomical structure as a means of evaluating patient setup compared to simulation radiograph or DRR. A radiopaque marker, which defines the center of the beam axis, should be in the center of the field. To avoid confusion, the marker is placed only on the beam entry side. The center is validated by equal distance from the inferior to superior edge of the field. A line is drawn through the center of the field parallel to the edge. This line should pass through the radiopaque marker, defining the position of the isocenter and the lasers. The identification of anatomical structure relative to this line would show whether the patient setup is rotated and positioned correctly. The relative position of the anatomical structure relative the field presents a method of validating the setup when compared to simulation film, as depicted in Fig. 7. QA for the CORVUS planning system is limited. The maintenance of the system is limited to purging temporary files, removal of raw patient data from other modalities pushed to the CORVUS workstation, and archiving of patients completing treatment at regular intervals. Rebooting the workstation computer will purge the temporary files. Documentation of approval number for each patient is important. Restored patient files would not have the same study number or approval status. The printout of plans from the CORVUS has defined parameters. These parameters such as the x- and y-jaw positions must be changed when collimator rotation is involved. Likewise, the collimator angle may need to be changed to avoid incorrect field size orientation. In our

Volume 26, Number 1, 2001

institution, we use 89.7°, because the MIMiC alignment is not exactly 90°. REFERENCES 1. Carol, M.P. Peacock: a system for planning and rotational delivery of intensity-modulated fields. Int. J. Imaging System. Technol. 6:56 – 61; 1995. 2. Tsai, J-S.; Wazer, D.E.; Ling, M.N.; et al. Dosimetric verification of the dynamic intensity-modulated radiation therapy of 92 patients. Int. J. Radiat. Oncol. Biol. Phys. 40:1213–30; 1998. 3. Sternick, E.S.; Sussman M.L. Safety management of IMRT. In: Sternick ES, editor. The Theory and Practice of Intensity Modulated Radiation Therapy. Madison: Advanced Medical Publishing; 1997: 145–59. 4. Curran, B. Conformal radiation therapy using a multileaf intensity modulating collimator. In: Sternick ES, editor. The Theory and Practice of Intensity Modulated Radiation Therapy. Madison: Advanced Medical Publishing; 1997: 75–90. 5. Kutcher, G.J.; Coia, L.; Gillin, M.; et al. Comprehensive QA for radiation oncology: report of AAPM Radiation Therapy Committee Task Group 40. Med. Phys. 21:581– 618; 1994. 6. CORVUS Beam Utilities 1.0 User Manual. NOMOS Corporation, Sewickley, PA. 7. Tsai, J-S.; Engler, M.J.; Ling, M.N.; et al. A non-invasive immobilization system and related quality assurance for dynamic intensity modulated radiation therapy of intracranial and head and neck disease. Int. J. Radiat. Oncol. Biol. Phys. 43:455– 67; 1999. 8. Low, D.A.; Mutic, S.; Dempsey, J.F.; et al. Quantitative dosimetric verification of an IMRT planning and delivery system. Radiother. Oncol. 49:305–16; 1998. 9. Low D.A.; Chao, K.S.C.; Mutic S.; Gerber R.L.; Perez C.A.; Purdy J.A. Quality assurance of serial tomotherapy for head and neck patient treatments. Int. J. Radiat. Oncol. Biol. Phys. 42:681–92; 1998. 10. Low, D.A.; Gerber, R.L.; Mutic, S.; Purdy, J.A. Phantoms for IMRT dose distribution measurement and treatment verification. Int. J. Radiat. Oncol. Biol. Phys. 40:1231–35; 1998. 11. Verellen D.; Linthout N.; Van de Berge D.; Bel A.; Storme G. Initial experience with intensity-modulated conformal radiation therapy for treatment of the head and neck region. Int. J. Radiat. Oncol. Biol. Phys. 39:99 –114; 1997. 12. Carol, M.P. IMRT: Where we are today. In: Sternick ES, editor. The Theory and Practice of Intensity Modulated Radiation Therapy. Madison: Advanced Medical Publishing; 1997: 17–36. 13. Grant W.H. III. Commissioning & quality assurance of an IMRT system. In: Sternick ES, editor. The Theory and Practice of Intensity Modulated Radiation Therapy. Madison: Advanced Medical Publishing; 1997: 121–26. 14. Saw, C.B.; Ayyangar, K.M., Thompson, R.B.; Zhen, W.; Enke, C.A. Commissioning of Peacock System for Intensity Modulated Radiation Therapy. Med. Dosim. In press. 15. Xing, L.; Chen, Y.; Luxton, G.; Boyer, A.L. Monitor unit calculation for an intensity modulated photon field by a simple scatter summation algorithm. Phys. Med. Biol. 45:N1–7; 2000. 16. Kung J.; Chen G. A modified Clarkson integration for IMRT. (Abstr.) Med. Phys. 26:1135; 1999. 17. Ayyangar, K.M.; Saw, C.B.; Shen, B.; Enke, C.A.; Nizin, P.S. Independent dose calculations for the Peacock System. Med. Dosim. In press.