This manual gives specific details on the operation of the Dosimetry Check program. This program provides a feed back mechanism for checking the dosimetry of x-ray radiation fields that are to be applied to patients. This is accomplished by measuring each radiation field that is to be applied and then recomputing the dose distribution to the patient using the measured fields as input and the definition of each beam. This technique does not include electron therapy.

Dosimetry Check Tasks

The program may be invoked by running the DosimetryCheckTasks program shown below:

The user must either recreate the plan to be checked or down load it from the planning system, to include the CT scan set for the plan. The importing of the plan can be in either the Dicom RT or RTOG protocol, the RTOG protocol being replaced by Dicom RT. A process is run to import the plan, selected above.

If using an integrating device to measure the radiation fields, the data for the fields must be imported, normalized, and possibly processed to convert dose back to fluence. Separate programs are provided for reading in the integrated beam fields from various devices and may be selected from the above program. The fields are normalized to the center of a 10x10 for known monitor units in order to be able to compute the absolute dose. For an EPID and most devices, there is a necessary deconvolution to convert the integrated image back to fluence. The result of the process is written to files in the patient directory. Dosimetry Check will automatically pick up those files and match them to the respective beams if the beam names are embedded in the file names of the fluence files produced by this process.

This application rests on top of System 2100, a General Purpose Radiological Display System. Functions are here added to System 2100 under the program name of DosimetryCheck. For all functions regarding to image display, contouring, and image fusion, refer to the manual for System 2100.

Warning on intended use:

This system is not intended to be used as a radiation therapy treatment planning system, but rather as a system for providing a redundant dose calculation using the feed back mechanism of basing the calculation on measured radiation fields.

Indications for Use

510(k) Number: K010225

Device Name: Dosimetry Check

Indications for Use:

The product is to be used by radiation oncologist, dosimetrist, and radiation therapy physicist to check the correctness of the x-ray treatment fields from high energy treatment machines that are planned to be or have been applied to a patient. This product is to be used in addition to the treatment planning system to provide a means for additional and redundant verification that the plan is in fact successfully accomplished. This product is not a treatment planning system and is not to be used as one. This product only checks the applied dose based on the measurement of each x-ray field and a theoretical calculation. This product does not provide any quality assurance that the fields are in fact correctly applied to and correctly aligned with the patient anatomy as planned. In addition, the product may be used to display the above dose on other fused image sets which could provide additional clinical information to the radiation oncologist regarding the treatment.

Broad Considerations in the Use of this Program

Inherent in any activity carried out by human beings in particular is the possibility of making errors. Radiation therapy presents unique problems in this regard to the application of ionizing radiation to patients. This radiation cannot be seen, heard, smelled, felt, or tasted, as the five senses do not record its presence. By contrast, a nurse who administers a drug can visually read the label from the bottle from which the drug is taken. The amount of drug that is measured out can be visually seen at the time it is administered to the patient.

Radiation on the other hand must be measured at some other time by a physicist with complex equipment and complex calibration processes. The plan of radiation is often computed on a computer where all aspects of the treatment delivery system must be modeled mathematically by the planning software. This modeling may include the insertion of objects into the beam such as wedges and shielding blocks. Intensity modulation may be designed by the treatment planning system and accomplished by a compensator manufactured for the purpose or by the treatment machine during treatment with moving multi-leaf collimators. This lengthy and complex process has multiple opportunities for errors to be committed by persons and machines.

Radiation therapy, like other medical specialties that actively act upon the patient, has zero tolerance for error. It is generally not acceptable to injure, kill, or paralyze a patient. Yet injuries, death, and paralysis can result if mistakes are made in the delivery of the treatment. Quality control measures must be instituted in any treatment center to assure proper treatment of patients as planned. The standard procedures for quality control (see references 1, 2, 3, and 4 below) generally call for the checking of each component of the treatment planning process. It is assumed, and hoped, that when all the components are correct that the end result is correct. Yet without a feedback mechanism for the entire treatment planning and delivery process, failure to detect a problem with any component or underlying concept will most likely go unnoticed. A recent example of such a system failure would be the lost of the Mars lander caused by failure to convert English units to metric units in some component of the system. No one noticed this mistake until they flew the rocket. The use of this program will allow for “flying” the plan before treating the patient.

General Quality Control Principles

General principles have been discovered in the engineering and manufacturing fields to assure quality. They include having and following documented procedures that include redundancy and feedback mechanisms. The firing of the rocket above provided the ultimate feedback as to whether the satellite will hit the target. Medical practitioners need to be cognitive of quality control methods that are employed in other industries. In the airplane industry there is a principle that the failure of a single component should not bring down the airplane. Redundancy is employed to reduce failures.

Let us examine this concept for a moment. If the failure rate of a component is 0.01 or 1 percent, and the failure rate of a redundant component is also 0.01, then the failure rate of the two components employed together so that both must fail before the system can fail is the product of the individual components, or 0.01 x 0.01 = 0.0001 or 0.01 percent, a hundredfold improvement in reliability. Therefore if a component cannot be made better, simply add redundancy.

Feedback mechanisms are also powerful tools for achieving reliable systems. Consider the float in the flush toilet. The level of water shuts off the incoming flow by lifting the float. If a simple timer were used instead and if the water pressure were to be higher on some particular day the reservoir tank might overflow before the timer shuts off the incoming water. The overflow tube in the reservoir tank also provides redundant protection against such a failure. Consequently I have yet to experience the failure of the reservoir tank overflowing in my life time to date.

Human Fallibility

Human beings perform the task of treatment planning. They must operate complex programs on computers and may also have to use calculators to compute or check the results. Human beings are fallible. No person can perform even a routine procedure repetitively without eventually making a mistake. For the past two years I have been giving fluid subcutaneously to our cat twice a week. During the two years I have stuck myself twice with the needle. My failure rate with the needle is 2 out of 208 or about 1 percent. It seems unlikely that any individual can go to work everyday over a 40 or so year career and operate computers and calculators and never make a mistake. It seems unlikely that every patient can be treated without a failure if there is no quality control process.

Typical Possible Errors

Treatment aids such as blocks or compensators may have to be made. A milling machine might make a compensator for example. Things can go wrong in the manufacturing process. The tools might not be calibrated correctly. The physicist might make an error measuring the attenuation factor. The compensator might be mounted backwards or switched with the wrong field or patient. The wrong file might be down loaded to a multi-leaf collimator system for field shaping or intensity modulation. There was an incident where a compensator was made for a lung field. The central axis went through lung but the monitor units were not corrected for lung transmission. Evidently no one noticed that the result would be the raising of doses to structures around the perimeter of the lung, such as the cord, to what the lung would get without a transmission correction, rather then the intended lowering of the dose to the lung to correspond to the dose outside the lung. This can result in the cord getting 10 to 15 percent higher dose than anticipated. I have heard of another incident where a physicist assumed the wedge factor was not accounted for in an external beam program when in fact it was, and so ended up applying the wedge factor twice.

Undoubtedly anyone who works in radiation therapy has been involved in or heard about some kind of mistake. Because of the nature of mistakes and the possible consequences to the patient, not to mention law suits, mistakes are probably not publicized and we may not be aware of the full extent of this problem in radiation therapy. Our concern here is to develop a procedure that will minimize the chance of any mistake not being caught before treatment.

Feedback

There remains very little immediate feedback in radiation therapy that the treatment is delivered as planned. The only feedback mechanisms commonly employed to verify the dose consist of making a surface measurement on the patient’s skin. This surface measurement can be related to a predicted dose value. However, the measurement at one or a few points does not demonstrate how the effects of all the treatment beams are adding up nor show the dose to the target volume or critical structures. Errors can still exist at other positions within the radiation field that will not be detected by a single point measurement, such as performing a point measurement on the central ray but the wedge is reversed in position. Making measurements inside the patient is generally limited to a few points if there is a cavity available and is an invasive procedure.

This product provides a feedback loop for the dose and should prove to be a powerful tool for assuring the correct treatment of patients. By calculating the dose to the patient based upon the measured fields, any mistake in the treatment plan, monitor unit calculation, or devices placed within the beam should be evident if substantial. By substantial we would hope to flag any error in dose greater than five percent. This also means that if a wedge were applied backwards, or the wrong wedge used, or some error made in intensity modulation such as down loading the wrong file to the accelerator, that the resultant dose to the patient would be evident from examining the dose distribution computed from the measured fields applied to the patient.

Advantage of This Method

An advantage of this approach is that the dose will be presented against the same anatomy that the plan was done on. The review of the dose would not be abstract as compared to, say, measuring the dose in a cylindrical phantom for which the plan has been recomputed, not that such an approached doesn’t also have its merits. A further advantage here is that the confirming dose can be computed for any desired plane throughout the patient or shown in 3d perspective room views, and is not limited to the planes measured in a phantom. By downloading the dose distribution from the planning system, a direct comparison can be made to the planned dose.

Limitations

Warning on limitations:

At the same time one must bear in mind that only the dosimetry is being checked. Nothing is done in this process to check that the fields are in the right place on the patient. Further, we are also assuming that the treatments will be given with the same parameters that the field measurements were made with, i.e. that a mistake is not made after this dosimetry check is completed. Nor is the correct energy selection being directly checked.

A careful review and evaluation should be done in implementing procedures so that it is understood and documented what is checked and what is not. Nonetheless, this verification of the dose distribution based on measurement eliminates many possible sources of error by providing a much needed feedback loop.

Possible False Positives: Common Errors

A possible source of failure to flag an incorrect treatment would be the occurrence of the same error in the planning and treatment process and this checking system. Care must be taken that the same mistake cannot be made twice or that complimentary errors are not made. For example, suppose a wedge or compensator is inserted in a reversed position from that intended. If the measured field used as input to this system were to be then incorrectly reversed as well as to make the device to appear to be in the correct position, the error in the device position would not be detected. If the treatment machine calibration were incorrectly specified to both the planning system and this checking system, an actual error would go undetected. Errors with the orientation of the measured field can be avoided by having a procedure with a positive lock between the measured field and the collimator. Other common errors would normally be found during routine acceptance testing of treatment planning systems and quality control of the measured beam data.

However, it is desirable to reduce as much as possible any overlapping dependency between the checking system, the treatment planning system, and treatment machine. The use of generic beam data as provided here and or the separation of the beam data used by this system from the treatment planning system will allow for independence between those two systems.

References

1. Physical Aspects of Quality Assurance in Radiation Therapy, Report Number 13, 0-88318-457-5, 1984 Radiation Therapy Committee Task Group #24, with contribution from Task Group #22, Medical Physics Publishing, 4513 Vernon Blvd., Madison, WI 53705-4964.

2. “Medical Accelerator Safety Considerations: Report of AAPM Radiation Therapy Committee Task Group No. 35, James A. Purdy, et. al., Medical Physics Vol. 4, No. 4, July/August 1993, pages 1261-1275.

3. “Comprehensive QA for Radiation Oncology: Report of AAPM Radiation Therapy Committee Task Group 40”, Gerald J. Kutcher, et. al., Medical Physics Vol. 21, No. 4, April 1994, pages 581-618.

4. “American Association of Physicist in Medicine Radiation Therapy Committee Task Group 53: Quality Assurance for Clinical Radiotherapy Treatment Planning”, Benedick Fraass, et. al., Medical Physics Vol. 25, No. 10, Oct. 1998, pages 1773-1829.

Measurement of the X-ray Fields

As a general rule, we believe that the persons who treat the patient should also make the exposures for the measurement of the x-ray fields. The reason is for increased assurance that the field exposure will be made identical to that used in the treatment, i.e., that the same devices will be used in the beam in the same positions. Below we discuss the possibility of making such a measurement during treatment if suitable technology is available.

Geometry for making a measured picture of the x-ray field.

There are many technologies available for measuring the radiation field. X-ray film is tedious and many departments are getting rid of their x-ray processors.

EPID

Electronic portal imaging devices (EPID) and ion chamber or diode arrays are suitable replacement devices. Using these devices is covered in a separate document: “EPID Functions for Dosimetry Check”.

X-ray film

Kodak EC film in ready pack, known as EDR2 (Extended Dose Range), should have a dynamic range out to 650 monitor units with the film placed at 100 cm (we have measured density of 3.0 at 650 monitor units at 100 cm). Verification film in ready pack will be limited to less than 200 monitor units at 100 cm (we have measured a density of 3.0 at 200 mu). In the case of such limited dynamic range, this software can handle a truncated treatment because the user can enter the monitor units used to irradiate the film and the monitor units that the field otherwise is to get, with the measured dose scaled according to the ratio. Digital radiography systems applied to radiation therapy have the potential of also offering a wide dynamic range. Electronic devices such as phosphorescent screen, a 45 degree mirror, and a TV camera may be sufficient, as well as a diode or ion chamber arrays. Film will work and may be the most readily available device to use.

Measured Field Orientation

It will be important that there be no confusion as to the orientation of the imaging device relative to the collimator. An electronic imaging device that is mounted to the collimator or gantry would completely eliminate any possible orientation errors with distinction made as whether or not the imaging devices rotates with the collimator.

With film in ready pack, one can pin prick the film using a pattern that clearly marks the positive X and Y axis, such as two marks just outside the field to mark +X and one to mark +Y, with a second mark to mark the other end of the X or Y axis. Leave at least a 1 cm margin between the marks and the field image.

The imaging surface must be perpendicular to the central ray. Closer distances than 100 cm may be desirable to fit a large field within the film size available. This software can correct for a changed distance of measurement by applying the inverse square law to calibrations made at a different distance.

The location of the central ray must be known. The image of the field edges can also serve to locate the central ray with the location tool provided in Field Dose below. However, an aperture block would obscure the field edges.

Warning measured field and collimator:

There should be a positive locking system between the measured field and the collimator, so that the location of the central ray is known, and so that there is no confusion as to the rotation of the measured field relative to the collimator.

An electron shield will have to be above the imaging surface to provide just sufficient material to shield contamination electrons from fogging the film. Using water density materials will cause a small problem in that scatter produced in the bolus will round off the edges of the measured radiation field. We have noted that this will cause higher isodose levels, such as the 95% level, to contract somewhat.

With x-ray film, we have gotten best results using 3mm of copper for the bolus material to shield out contamination electrons. We have not seen any degradation of the measured field edge with 3mm of copper. At 6x one can use 1 mm copper.

We do not want a phantom under the imaging surface because the measured field is to approximate an in air measurement. A phantom under the measured surface would contribute full back scatter and we are assuming that there is no back scatter in the measured field. Otherwise a deconvolution process is needed.

It is assumed that a dry run will be made without the patient. However, making a measurement of the field during treatment is not precluded if the technology exist to do that, that is, if there is a suitable device that can also be treated through. However, in that case, careful consideration needs to be given to the total attenuation the imaging device presents between the source of x-rays and the patient and the effect this might have on one treatment. If the monitor units were increased accordingly, care must be taken that the monitor units are readjusted when the imaging system is not in the beam for the rest of the radiation course.

Relative Monitor Units

The imaging system must not only record the variation of the x-ray intensity over the area of the beam, but must accurately measure the applied relative monitor units. This can be accomplished with film by running a calibration curve to convert film response to relative monitor units at the same time. Using a step wedge this can be accomplished with the single exposure of an extra film. Another variation would be to put a parallel plate ion chamber or diode above or below the imaging surface. It will have to be possible to determine the point of measurement if different from the central axis, and the chamber and leads must not introduce large artifacts onto the rest of the image of the field. Please review the below Field Dose section for a definition of relative monitor units.

Making a Calibration Curve for Film

We would recommend that you simply place the film on the treatment couch at 100 cm and cover it by the dmax thickness using a solid water slab. Use a 10x10 cm field size (assuming you have normalized your scatter collimator factor to a 10x10, otherwise correct the monitor units you use). We also recommend that you use at least 10 steps in the calibration curve from the minimum exposure to the maximum density that can be digitized by the film scanner. For example, for therapy verification film you might consider running the curve up to a maximum of 60 monitor units and shoot all the treatment fields with 50 monitor units (you can probably go higher if your film scanner can handle the darker densities).

If you expose the same film multiple times, you must consider the leakage from the accelerator. If you fit six 10x10 field exposures on a single 14x17 inch film, each field will receive the leakage from five other exposures. We would suggest that you cut each film into three strips 14 by 5.7 inches. Place each strip in an empty paper ready pack, fold the open edge over and fix with a paper clip. The 14 inch length is long enough to run through the film processor. And you can make two 10x10 cm field exposures on each strip. But, the other part of the film still must be covered by bolus to protect the film from exposure to secondary electrons. If the leakage is more than 0.5%, we would recommend you use more film with only one exposure per strip (you can also tape a leader onto smaller pieces of film for running through the film processor). The calibration curve is simply monitor units versus the measured value at the center of each 10x10 cm field. An unexposed area would serve for 0 monitor units.

Rescale, Step Strip, or Dosimeter

Using multiple exposures every time the verification system is used to create a calibration curve will be time consuming. You can either make only one curve and rescale the curve for each use, rely on a dosimeter measurement to renormalize the treatment field films, or calibrate a step wedge to generate a calibration curve for each use of the system. The issue here is that there will be large variations in film response due to variations in the film processor and possibly from one film batch to another and we need to correct for that variation.

Rescale the Calibration Curve

To correct for changes in film processing, it may only be necessary to rescale the calibration curve. When making film images of each field for a patient case, expose one extra calibration film, typically 10x10 cm, to a known monitor unit. Use the existing calibration curve to determine the relative monitor units for the calibration film. Type in the monitor units used on the provided popup tool (see Field Dose in next chapter) to rescale the calibration curve. The rescaled curve is saved and then used separately and instead of the original curve.

Dosimeter

If using a dosimeter, you could use a parallel plate ion chamber imbedded in plastic and place it either immediately above or below the film. The software has options for you to simply type in the reading, the temperature and pressure if an ion chamber, and the distance. You need to calibrate the dosimeter in the same geometry used by making an exposure with the 10x10 cm field size for known monitor units. The film calibration curve will correct for the dose on a relative basis, with the dosimeter supplying an absolute value.

Making a Step Strip

The sensitometer meters used in radiology are too fast for our films and also have too much variation between exposures. Stated stability is typically +-0.02 log exposure, which means about +- 5% in exposure. You can instead make a step strip out of cerrobend or lead. The linear attenuation factor will very with energy. Using a value between 4 and 18 MV x-rays (-0.420/mm) we have computed suggested thicknesses for a step wedge made out of cerrobend:

approximate % transmission	Thickness cm
0	extend thickest part beyond field edge.
4.3	7.5
10	5.5
20	3.8
30	2.9
40	2.2
50	1.6
60	1.2
70	0.85
80	0.53
90	0.25
100.0	0.0

A band saw can be used to cut out each step. You may want to cut the sides on diverging rays.

To calibrate the step wedge, you must still run a calibration curve. Then place the step wedge on top of the film. Decide on the field size and orientation that will always be used (still assuming the film is at 100 cm), i.e. mark where the central axis is to go on top of the step wedge. Expose to slightly less than the largest monitor units you used when you ran the calibration curve. Digitize this film with the calibration curve. Next use the software to create a step wedge file. The calibration curve will be used to assign an equivalent monitor unit to each step.

Then when treatment films are exposed, make a film using the step wedge for the same distance, orientation, field size, and monitor units used to create the step wedge file. The software will then associate the measured value for each step with the above equivalent monitor units to generate a calibration curve that can be applied to the treatment field films.

Digitizing Film

The field image must be converted to digital form. Some consideration must be made to the dynamic range of the film scanner device, if digitizing film, as to the maximum density that can be measured. You must use less monitor units so as not to exceed the dynamic range of the film scanner. The program has an option to allow you to scale the measured values up by the ratio of monitor units used to expose the film to that for the treatment field if the monitor unit prescription exceeds the dynamic range, which will be the case if using therapy verification film.

The zero crossing of the response of the film digitizer must also be considered. Some signal should be seen for an unexposed area of the film, otherwise it may be that low doses are being truncated to zero.

Another consideration is the number of bits that can be digitized. Eight bits per pixel will only permit a number from 0 to 255 to represent the range of monitor units. Since field prescriptions can top off around 250 to 300 mu, we would barely have the resolution of 1 mu. With the necessary dynamic overrange needed to insure that the range is not exceeded, the resultant resolution will probably be even less. If the number of bits limits accuracy, consider adding a dosimeter reading to the field measurement as mentioned above.

Even if the device has 12 or 16 bit analogue to digital converters, the image file format must also be considered. The Tiff format for example only supports 8 bits per color, so the image pixel data would normally have to be truncated to 8 bits. Usually the same number is stored for each of red, green, and blue or a single byte (8 bits) is used for gray scale. However, in one case using software with the 12 bit Vidar film scanner, we noted that 12 bits were in fact being written to the Tiff file with the 12 bits distributed over the red, green, and blue color bytes. This is not part of the Tiff standard as gray scale can only go to 8 bits, but we have nonetheless modified our Tiff reader to decode a 12 bit number from the red, green, and blue fields. We are not able to read non-standard 16 bit Tiff pixels. We will attempt to deal with this issue as the need arises if other encoding schemes are encountered.

The PNG format can accommodate a 16 bit gray scale pixel, has provision for specifying a pixel size, and so is to be preferred over using TIFF.

The Dicom standard permits a data size of 12 or 16 bits per pixel and so is preferred for this application.

If the accuracy in terms of the absolute dose in relative monitor units cannot be achieved with only digitized film images, then a parallel plate ion chamber or a small diode (or mosfet) should be used either in the bolus stack immediately above the measuring surface, or might be placed immediately below. The point measurement is used to renormalize the dose array. The calibration curve is still used before the final renormalization. We would like to have the relative monitor units measured to within 1 or 2 percent to have confidence in the result. This is achievable with film relying upon a step wedge or a point measurement.

Pixel Size

Anther issue with the Tiff and PNG formats is that the pixel size is an optional parameter, not a required one (actually the image size is specified from which the pixel size can be computed). Yet we need the pixel size of the digitized image. If the pixel size is not in the file, a message will be popped up and the user will have to determine the pixel size. This can be accomplished with the provided tool that locates the field image in beam’s eye coordinates.

Dose Recomputation Procedure

The CT scans, beam positions, outlined regions of interest, and a 3d dose matrix can be downloaded in Dicom RT or RTOG format. If the skin boundary is included in the download, it is only necessary to select that volume as being the skin boundary. The Dicom RT protocol has a mechanism for specifying that. A pixel value to density conversion still must be done with this software, as the conversion curve is not part of the Dicom RT or RTOG protocols.

If the plan cannot be downloaded, the user must read into this software the CT scan set used for the treatment plan. A skin boundary must be generated and CT pixel number to density specified. Then the user must locate the isocenter of each beam. The plan will have to be referred to with at least two orthogonal (perpendicular) views of the image data to locate isocenter. A transverse scan will show where isocenter is in the plane of the scan, but a sagittal or coronal reconstruction may be needed to show where isocenter is along the axis parallel to the long axis of the body perpendicular to the transverse plane. The beam location is easily adjusted with this software along three axes. Subsequent beams will start with the isocenter of the prior beam. The user must also select the correct energy. The user must set the gantry, collimator, and couch angles for each beam.

Lastly, in either case, each beam must be associated with the respective measured field. This can happen automatically or the user can select the measured fields for each beam. Selecting two or more fields will result in them being added together. When this process is completed the dose distribution can be computed.

Comparison of Results

For the check of the dosimetry to be effective, the result of the dose computation must be compared to the treatment plan. The user should plot the same planes presented in the plan for comparison. The hot spot (maximum dose) should also be noted as this may indicate a conditioned that has gone unnoticed. The normalization point (usually isocenter) should be computed for comparison with the plan. The normalization point and hot spot can serve for an objective comparison. The dose distribution comparison will be more subjective in nature, but one can look at the target volume coverage and dose to critical structures.

Warning on quality control plan failure:

Treatment of the patient must NOT continue until any large difference in dose greater than five percent to significant volumes is understood and corrected. The treatment plan must not be changed based solely upon the result of this check, rather the cause of any such difference must be resolved and understood and a corrected treatment plan produced.

Gamma Method

The gamma analysis method provides a means to compare two dose distributions taking into account both dose at a point and distance to where the dose is the same. See the reference in Medical Physics by Daniel Low, et. al., Vol 25(5) May 1998, pp. 656-661, "A technique for the quantitative evaluation of dose distributions." The reason here is that in a low gradient area, the isodose curves from two distributions for the same dose can be far apart whereas the difference in dose is actually small. And in a high gradient area, the there can be a large change in dose at a point but only a very short distance to where the dose is the same.

Comparison Tools

Several comparison tools are available for the case when the plan is downloaded from the treatment planning system and includes a 3d dose matrix, and are described in a separate section. The tools include, besides the above gamma method, generating a difference dose volume histogram, showing the computed and downloaded dose on the same images, showing a plot of the dose difference on an image, and showing a 3d perspective room view of the dose and dose difference. Using this feature, one could also use this program to show the downloaded dose distribution on other fused image sets.

Algorithmic Differences

One can expect small differences in dose due to differences in the algorithm between this system and the planning system. Some knowledge of those differences may be needed. The pencil beam algorithm used here is not going to do well in internal interface areas and differences can be expected if compared to an algorithm that does account for interface effects such as Monte Carlo or a fully implemented point dose superposition algorithm. We can also expect differences at field and block edges in penumbra regions. Within the beam no closer than one cm to an edge, we should expect any two systems to agree within at least five percent, hopefully three percent. However, one should not be surprised at dose differences greater than five percent in regions of electronic disequilibrium which includes beam edges.

Another potential source of small differences might be the skin boundary. Here shape interpolation is performed (unless turned off) between contours that are further than several millimeters apart.

Systematic Procedures for Comparing Plans

The user should develop a systematic procedure for comparing the doses from the two systems, noting the dose that covers the target area, the dose to critical structures, and the maximum hot spot doses. Significant differences should be understood and resolved. Comparing the dose at a central point provides an objective measure, as well as the difference dose volume histogram and gamma volume histogram.

Possible Failure of This Checking System

Warning, program only tests dosimetry:

This system only tests the dosimetry, not the actual application of the fields to the patient.

The actual treatment to the patient will depend upon the proper placement of the fields on the patient and that the fields are not changed or different for actual treatment.

With the dosimetry check performed here, there are two modes of failure, false positive and false negative. False negative is showing an error when none exist and a false positive shows no error when one in fact exist.

A negative result alerts the user and investigative action can be taken to resolve the problem.

A false positive should be less likely but consideration should be given to the possibilities. A false positive would result if the same error applied to both the checking system and the original planning system. A common error might result from using the same beam data. However, one would expect normal quality control procedures for the planning system to screen out any such error in the beam data. Further, there is the possibility for using generic data provided with the checking system that would make the checking system more independent of the planning system. But even the generic data may require some editing by the physicist to properly define how the treatment machine is calibrated. For example, is the distance to the calibration point 100 cm + dmax or just 100 cm?

Another source of a false positive error would be if the fields are set up correctly for the dosimetry check but incorrectly for the actual treatment. But this last consideration applies equally to the daily treatment of the patient in that there must be consistency and some control for guaranteeing consistency.

Warning, input field parameters:

The dosimetry check program relies upon the user properly defining the isocenter and angles for each field, and for properly specifying the energy of each beam. The measured field must be correctly associated with each beam.

Special attention needs to be paid to the actual field measurement and the means to determine the orientation of the measured field relative to the collimator. The measurement device, be it film or otherwise, must be consistently placed in the same orientation relative to the collimator. If the measurement device attaches to the collimator this is assured. It not a procedure must be established, such as always putting the collimator in its unrotated position and always having the film or measurement device in the same position on the treatment couch. Further, some markers may be needed so that the beam’s eye view x and y axes are clearly discernible as well as which side of the film the x-ray source was on. With film in ready pack, a pin prick outside the field would serve this purpose. But be aware also that the image of any marker will also be interpreted by the checking system as a change in dose.

Warning, understand what part is checked:

The user must take care to clearly understand what part of the treatment process the dosimetry checking system is confirming and what parts of the process the system cannot provide a check.

Future enhancements of this system might include spatially calibrated images of the patient that can be compared to the rendered skin surface for verification of the patient’s position relative to isocenter. Similar measurements of the treatment machine gantry and collimator may be possible.

Fused Image Sets

Warning Fused Image Sets:

Clearly, if the checking system signals a difference and error in dose the system cannot be used to show the dose on a fused image set.

The dose distribution transferred to the same plane in a fused image set should be the same in both identical planes. The tools provided with image fusion provide means to verify that the fused solution is correct. Only if the fusion solution is correct and the dosimetry check agrees with the plan can the dose shown on a fused image set be relied upon.