Portal Imaging Technology:
`Past, Present, and Future
`
`made, more development needs to be directed towards
`making portal imaging convenient and reliable. Image
`quality must be improved further, to improve the robust-
`ness of image registration techniques and more thought
`must be given to integrating and automating the various
`steps in the image registration process. Otherwise, too
`much time will have to be devoted to these tasks.
`Finally, and most importantly, users will have to decide
`what is the best way of using EPlDs clinically. Much
`development is required before the full potential of this
`exciting technology can be realized.
`Copyright © 1995 by W. B. Saunders Company
`
`Peter Munro
`
`Many different electronic portal imaging devices (EP|Ds)
`have been developed to improve geometric accuracy in
`radiation therapy. This article describes the two types of
`EPlDs that have become available commercia|ly—the
`television camera-based EPID and the matrix ion cham-
`ber EP|D—as well as describing the amorphous silicon
`array, a device that may become available in the future
`for portal imaging. in addition, the various image regis-
`tration techniques that identify geometric errors from
`the portal images are described. These include interac-
`tive techniques, landmark-based techniques, contrast-
`based techniques, and hybrid techniques. Although great
`improvements in portal imaging technology have been
`
`he goal of radiation therapy is to deliver a
`prescribed radiation dose to the target volume
`accurately while sparing the surrounding normal and
`critical tissues. Experimental and clinical evidence
`shows that small changes in dose of 7% to 1. % can
`reduce local tumor control signiIicanLly,"3 or increase
`the rate of normal tissue complications.4 As a result,
`recommendations by the International Commission
`on Radiation Units (ICRU) suggest that the accu-
`racy in dose delivery be i‘5‘.’/o.'Z'3 Such accuracy can be
`achieved only if field placement is precise during the
`entire course of radiation treatment, so that the
`treatment beams irradiate only the prescribed re-
`gions.
`Unfortunately, the geometric accuracy of radia-
`tion treatments can be compromised resulting in the
`irradiation of regions other than those prescribed. A
`number of studies“? have shown that discrepancies
`in field placement occur frequently, especially for
`complicated treatment setups. Furthermore, these
`geometric discrepancies can also influence the out-
`come of treatment.8"3‘ *5 Fortunately, frequent moni-
`toring of patient positioning can reduce the fre-
`quency of discrepancies in field placement.7‘9-“H9 As a
`result, several studies have suggested that patient
`positioning be checked daily.19’20
`
` (cid:14) (cid:14)
`
`Trends in radiation treatments are increasing the
`need for accurate patient positioning.
`\/Vith the
`integration of computed tomography (CT) and, on
`occasion, magnetic resonance imaging (MRI) data
`into the treatment planning process, with the devel-
`opment of affordable three-dimensional treatment
`planning work stations, and with the advent of
`programmable multi—leaf collimators for field shap-
`ing, radiation treatment portals have become more
`highly tailored with smaller margins around the
`target volumes. In addition, new treatment tech-
`niques such as dynamic beam modulation are increas-
`ing the need for routine monitoring of patient
`positioning even further. Because of these trends in
`radiation therapy, much effort has been devoted to
`developing more convenient methods to image the
`patient during radiation treatment,
`the process
`known as portal imaging. This article examines the
`histoiy of portal imaging developments, describes
`some of the devices (known as electronic portal
`imaging devices or EPIDSJ that are currently avail-
`able commercially, discusses one of the imaging
`devices that may become available in the future, and
`examines some of the image processing and image
`registration techniques that are essential for EPIDs
`to be useful in the clinic.
`
`From t/zzDe,tnzrtnzmt_r cf/l/[radical Bin};/yxirx and Oncology, I/nioerxity Ly"
`I/Vestern Ontario, and London Regional Cancer Centre, London, Ontario,
`Canada.
`Addrm reprint requert: to Peter Munro, PhD, London Regional Cancer
`Centre, 790 Cammissinrmry Ro1zd,EzLrt, Lnmiun, 0m.‘z1r1'oN6/l 4L6, Canada.
`Supported by the Ontario Cancer Treatment r1ndRe5mrch Foundotzon.
`Copyright © 1995 by WB. Saundm Company
`I053-4296/ 95/ 0502~0005$05.00/ 0
`
`History
`
`There is a long history of imaging with high—energy
`radiation beams. As early as 1904, radiographs of
`human hands, mice, purses, and other objects had
`been made using radium as the radiation source.“
`However, because of the high energy of the radium
`gamma rays, the images suffered from low contrast,
`
`Ser/Lilian in Rachofiorz Oncology, V015, Nu 2 (April), 1995' M) 115-133
`
`1 1 5
`
`Page 1 of 19
`
`Elekta Exhibit 1006
`
`

`
`Peter Kl/Izmra
`
`which made them unsuited for the diagnostic applica-
`tions that were of interest at the time. One of the
`
`l958, several artielcsm” described the use of a
`television-roentgen (TVR) system for monitoring
`the position of patients during pendulum therapy.
`The patients lay on the treatment couch while the
`gantry-mounted TVR system (consisting of a X—ray
`image intensifier and a TV camera) rotated around
`the patient during the treatment with a 200-kV(p)
`x—ray beam. The video signal was sent to a monitor
`located in the control room of the treatment Ina-
`
`this TVR
`chine. The major limitation was that
`system had a field of View of only 5.0 inches (2.5
`inches at the patient). Independently, a somewhat
`similar system was being developed by Andrews et
`al” using a TV camera—based device called a ‘johns
`Hopkins screen intensifier”3‘ as the image receptor.
`It consisted of a fluorescent screen viewed by an
`image orthicon TV camera using a complicated
`series of relay mirrors and lenses known as a Schmidt
`optical system. These efforts resulted in some of the
`earliest megavoltage images, using a 2-MeV Van de
`Graaff generator as the X-ray source. Not surpris-
`ingly, the authors found that normal tissue contrast
`was insullicient and they had to use air or mercury
`contrast materials to visual anatomic structures. In
`
`
`
`first references to imaging in radiation therapy was
`by Nielsen and Jensen” who, in l942, described a
`rotation therapy technique for treating cancer of the
`esophagus. In the treatment, the patient sat upright
`in a rotating chair while a stationary x—ray beam was
`directed horizontally towards the patient. The thera-
`peutic radiation (l80 kV(p) x-ray beam) exiting the
`patient hit a fluorescent screen that was viewed by an
`observer who was looking through a lead glass
`window. Not only did the observer view the treat-
`ment in real time, but corrections to the position of
`the beam were made remotely during the treatment.
`Thus,
`this treatment may have been one of the
`earliest examples of dynamic conformal radiation
`therapy! By l95l, a similar treatment for cancer of
`the esophagus had been developed by Hare et 21123
`using a Q—MeV Van de Graafl‘ generator, They
`described the use of film radiography to ensure
`accurate positioning of the patient before commence-
`ment of the rotation therapy. They showed a great
`deal of sophistication in their portal lilrn activities,
`investigating the utility of introducing air into the
`bladder and rectum as a contrast agent as well as
`using double-exposure radiographs to visualize
`anatomy outside of the treatment field. Surprisingly,
`during the 1950s most of the scientific interest in
`imaging with megavoltage beams dealt with diagnos-
`tic applications. Several articles by Tuddenham et
`al2”*v25 promoted the use of megavoltage x—ray beams
`for chest radiography to optimize exposure in both
`the mediastinum and the lung regions and to mini-
`mize the visibility of overlying “osseous” structures.
`In 1960, Perryman et al26 described cobalt 60 radiog-
`raphy. The technique used Kodak type AA industrial
`film (Eastman Kodak, Rochester, NY) placed in a
`cassette where the standard intensiiying screens had
`been replaced by two 0.01-inch lead sheets. The only
`drawback was that Iilrn development took approxi-
`mately 30 minutes to complete. Not only were the
`radiographs used for treatment localization but also
`to give diagnostic information “regarding the extent
`and location of soft tissue lesions.” Finally, in 1962,
`Springer et al” suggested an improvement to cobalt
`60 radiography where two fluorescent screens were
`placed between the lead sheets and the film. This
`modification reduced the exposure time and, it was
`claimed, increased the contrast of the final radio-
`graphs.
`About this time, non-film imaging methods started
`to be introduced into radiation therapy clinics. In
`
`1962, another article“ described an imaging system
`that used an imaging device similar to the johns
`Hopkins screen intensifier. The field of view at the
`detector was 25 cm in diameter and the system was
`attached to a 30—MV betatron. Again, only high-
`contrast objects such as tungsten, gold, and lead
`markers produced sufficient contrast to be visualized
`by the system. The investigators showed great inge-
`nuity in introducing high—contrast, high-density ob-
`jects into various body orifices such as the esophagus.
`However, the limited image quality and the need for
`contrast agents reduced the utility of all of these
`imaging systems and they never came into wide
`spread clinical use.
`important developments for
`One of the most
`portal imaging was First described by Swain and
`Steckcl in L96633 and was refined further by Marks
`and Haus and their colleagues?/"35 The method used
`a slow, wide-latitude film that was placed in card-
`board film holders and was exposed for the entire
`duration of the treatment. Unlike portal film tech-
`niques common at the time, the patient was not
`moved between the exposure of the film and the
`treatment and the cardboard film~ho1ders were much
`more comfortable to lie on than the film-screen
`
`cassettes that were typically modified for therapy
`verification, Not only was much less effort required to
`treat the patient but, in addition, a record of the
`
`Page 2 of 19
`
`

`
`Portal Imaging Technology
`
`117
`
`entirc treatment was available from the film. Be-
`cause of the efibrts of Marks and Haus, a film
`(eventually known as Kodak XV-2 film), which was
`compatible with the 90—second lilm processors37 that
`had been introduced into hospitals in 1965, became
`available for therapy verification. This was a major
`development for portal imaging. Not only did Marks
`and Haus and their colleagues develop a convenient
`method ofverifying patient positioning during radia-
`tion treatment but they also showed in a series of
`studies7’” the importance of routine verification in
`reducing errors in patient positioning and in improv-
`ing local control. As a result, use of portal films
`became routine and eventually generated the de-
`mand for even more convenient methods of therapy
`verification that has since lead to the development of
`EPIDS.
`
`Commercially Available Imaging
`Devices
`
`Many different devices have been examined since the
`early 1980s as alternatives to film. These devices can
`be divided into two categories: scanning systems,
`where the radiation detector subtends only a small
`
`fraction of the radiation beam and must be scanned
`
`underneath the patient to form the image, and area
`systems, where the detector subtends the entire
`radiation beam. Examples of these devices include
`scanning diode arrays,33‘”‘0 scanning scintillator ar-
`rays,“ storage pl1osphors,”‘2”“ coded aperture ar-
`rays,“ matrix ion chamber systems,‘*5”*7 and TV
`camera—based systerr1s.”‘E*54 lnsullicient space is avail-
`able to discuss all of these systems and readers are
`referred to two comprehensive reviews of portal
`imaging devices for further details.55=55 The following
`discussion concentrates on the matrix ion chamber
`
`and the TV camera-based EPIDS, which are both
`available commercially, as well as the amorphous
`silicon array,57‘72 which is an imaging device that may
`become available commercially in the future.
`
`Matrix Ion Chamber
`
`A schematic of the matrix ion chamber device, which
`was originally developed by Meertens and van Ilerk
`and their colleagues‘*59”‘7 is shown in Fig l. The device
`consists of two sets of electrodes that are oriented
`
`perpendicularly to each other separated by a 0.8—mm
`gap, which is filled with a fluid (2,2,4-trimethylpen-
`tane) that is ionized when the device is irradiated.
`
`
`
`Processor
`2)] +256 KMemory
`
`Accelerator
`Sync Pulse
`
`Analog
`Inputs.
`Mulllplexar
`Control
`4.:
`
`Dedicated
`Control
`Proi:e5sor
`
`125 lol
`Multiplexer
`
`Control
`Electronics
`
`User
`
`Contro Unit
`
`Megavoltage”Camera" Cassette
`
`the matrix ion chamber. The device consists ofa set of 256 signal electrodes and a set of
`Figure 1. Schematic diagram 0
`256 high-voltage electrodes oriented perpendicularly to each other and separated by a 0.8—mm gap filled with a fluid called
`2,2,4 trimethylpentane. VVhen irradiated, the fluid is ionized and generates signals in the signal electrodes. The active area
`of the device is 32.5 X 32.5 cm and its overall dimensions are approximately 60 X 60 X 5 cm including the readout
`electronics, which are immediately adjacent to the ion chambers. The device is controlled by a computer located in the
`control area of the linear accelerator. (Reprinted with permission.46)
`
`Page 3 of 19
`
`

`
`Peter Fl/Iunro
`
`The electrode spacing is 1.27 mm and, since each set
`of electrodes consists of 256 electrodes, the active
`area of the matrix ion chamber array is 32.5 cm on a
`side. One set of electrodes is connected to 256
`electrometers and the other set of electrodes is
`
`
`
`connected to a high-voltage supply that can apply a
`300-V potential to each electrode individually. The
`matrix ion chamber array is read out by applying a
`high voltage to each of the high—voltage electrodes in
`succession (for approximately 20 milliseconds) and
`measuring the signal generated in each of the 256
`signal electrodes. This procedure takes 5.5 seconds to
`read out an image. In addition, a fast (lower resolu-
`tion) scanning mode is available that scans the array
`in 1.5 seconds by applying the high voltage for a
`l0—millisecond period to two high—voltage electrodes
`at a time. The fast acquisition mode is useful for
`acquiring double—exposure images.
`The most obvious advantage of the matrix ion
`chamber is its compact size, which makes the device
`a convenient replacement for film cassettes. Another
`advantage is geometric reliability—images acquired
`with this EPID have no geometric distortions. Fur-
`thermore, unlike other scanning EPIDS, the matrix
`ion chamber has no moving parts, reducing the
`likelihood ofmeehanical problems.
`The major limitation of most EPIDS that use a
`scanning radiation detector, such as the matrix ion
`chamber, is quantum utilization. Ideally, an image
`receptor should use all of the available radiation
`efficiently (even for megavoltage imaging) because
`this will improve image quality. Clearly, this is not
`the case for the matrix ion chamber, where only one
`high-voltage electrode (out of 256) is active at any
`one time. However, the physics of signal generation
`in the 2,2,4 trimethylpentane improves the quantum
`utilization of the matrix ion chamber considerably.
`The signal measured by the matrix ion chamber
`depends on the rate of formation and the rate of
`recombination of the ion pairs that are generated in
`the ionizing fluid. Even when no high voltage is
`applied to the electrodes, the rate of recombination
`of the ion pairs generated in the 2,2,4 trimethylpen-
`tane is relatively slow. Therefore, the concentration
`of ion pairs can increase over a period of time until an
`equilibrium is reached between ion-pair formation,
`which depends on the dose rate at the matrix ion
`chamber, and ion-pair recombination, which de-
`pends on the probability of ions encountering each
`other and is proportional to the square of the ion-pair
`concentration. The signal measured by any electrode
`of the matrix ion chamber does not depend greatly
`
`on the dose rate during the 20-millisecond period
`when the high voltage is applied but on the previous
`irradiation history of the electrode. Calculations
`have shown that after 0.5 second, a latent image has
`been form ed over the entire irradiated region of the
`matrix ion chamber and that irradiating for a longer
`time will not increase the size of the signal, ie, will not
`improve image quality. These observations have both
`positive and negative implications. The measured
`signal is six to seven times greater than would be
`expected if no charge integration occurred in the
`2,2,4 trimethylpentane. However, the effective pe-
`riod of the charge integration (~05 second) is still
`short compared with the total image acquisition time
`of 5.5 seconds. Therefore, a large fraction of the
`radiation that interacts with the matrix ion chamber
`
`does not generate any measurable signal. For this
`reason,
`the matrix ion chamber requires higher
`doses to generate images than other portal imaging
`devices.
`
`An example of an image acquired with the matrix
`ion chamber EPID is shown in Fig 2. One ofthe most
`noticeable characteristics of this image is how exten-
`sively the raw signals have to be processed before
`yielding a usable image.
`(The artifacts in Fig 2
`represent an extreme case because these images
`were acquired using a prototype device.) Because
`spurious signals are generated in the electrometers
`and ion chambers, and because the sensitivities of
`each ion chamber can vary,
`the device must be
`calibrated routinely. In addition, because the matrix
`ion chamber is a scanning EPID, it is more suscep-
`tible to artifacts if the dose rate of the accelerator
`
`changes during image acquisition. Thus, the radia-
`tion beam has to stabilize for some period (typically
`1.0 second) after startup before image acquisition
`can begin.
`
`TV Camera—Based EPIDs
`
`The design of TV camera—based EPIDS, which is
`shown in Fig 3, uses technologies that have long been
`used for other applications. The x—ray detector con—
`sists of a metal plate to which is attached a gadolin-
`ium oxysulphide (Gd2O2S) screen similar to the
`screens used in diagnostic radiology. VVhen irradi~
`ated, high—energy electrons generated in the metal
`plate and the Gd2OgS screen are converted into light.
`The light that diffuses through the screen and exits
`the rear surface of the x-ray detector is viewed by a
`TV camera using a 45° mirror. The video signal from
`the camera is digitized and the digitized image can
`
`Page 4 of 19
`
`

`
`Portal Imaging Technology
`
`1 19
`
`
`
`Figure 2. Image of an Alderson Rando phantom acquired using the matrix ion chamber EPID and an 8-MV x-ray beam:
`(A) before and (B) after corrections for variations in electrometer offset and sensitivity as well as variations in electrode
`shape and fluid thidmess. (Reprinted with permission."°)
`
`be viewed on a monitor located in the control area of
`the accelerator.
`
`This design has one major advantage. The x-ray
`detector subtends the entire area of the radiation
`beam so all of the radiation that exits from the
`
`patient has the potential to generate a signal in the
`EPID. However, the design suffers from one major
`limitation and this is the light collection efliciency of
`the optical chain. Figure 4 shows the problem sche-
`matically. Because the light is highly scattered within
`. the phosphor screen, the light is emitted from the
`rear of the screen in all directions with equal probabil-
`
`ity. Only those light photons that are emitted within
`a small cone subtended by the lens of the camera can
`generate a signal in the TV camera; typically only
`0.1% to 0.01% of the light emitted by the phosphor
`screen reaches the TV camera. This poor light
`collection efficiency can reduce image quality in two
`ways. Firstly, if an x-ray photon interacts in the x-ray
`detector and none of the light generated by this
`interaction reaches the TV camera, then no measur-
`able signal is produced. Seoondly,
`if only a small
`signal is produced in the TV camera then noise
`generated by the preamplifier and other electronics
`
`imaging Imlon
`. m.
`
`
`
`Figure 3. Schematic dia-
`gram of the TVcamera—baxed
`EP1Ds. These devices use a
`metal plate/phosphor screen
`as the x-ray detector. light
`emitted by the phosphor
`screen is viewed by a TV cam-
`era using a 45° mirror.
`
`Page 5 of 19
`
`

`
`I 20
`
`Peter ;’l'Iunra
`
`
`
`acce lance cone
`0 tholens
`
`Figure 4. Schematic diagram drawn approximately to
`scale showing the light collection efliciency of TV camera-
`based EPIDS. The narrow cone shows how only a small
`frauion of the light emitted from the phosphor screen
`reaches the TV camera.
`
`may be quite large compared with the small signal.
`As a result, much ofthc effort in the development of
`TV camcra—based EPIDS has been to improve the
`light collection of the optical chain, by increasing the
`light output of the phosphor scrccn,5'-73 using a large
`aperture lens that collects more of the light,“-55~7‘ or
`using TV cameras with high light-quantum efficien-
`cics.5'
`
`Care has to be taken when using these ap-
`proaches. \Vhile Munro et 2115‘ have shovm experimen-
`tally that using thick (-l00 mg/cm“) phosphor screens
`improves overall image quality, the current trend in
`the manufacture of TV camera—based EPTDs is
`
`towards the use of thinner (I50 mg/cm2) screens.
`This is because the thicker scrccns suffer from loss of
`
`spatial resolution and blemishes that generate visu-
`allydistracting flaws that have proven to be unaccept-
`able to users. Large aperture lenses suffer from
`decreased spatial resolution because of spherical
`aberrations (light rays reaching the edges of the lens
`do not focus to the same point as those reaching the
`center) and the spatial resolution of these lenses
`decreases from the center to the edge of the lens.
`Large aperture lenses also suffer from vignetting,
`which results in images that are brighter at the
`center of the lens than the edge. This change in
`image brightness complicates the display of portal
`
`images because level and windowing to display one
`part of the image will obscure other parts of the
`image that have different brightness levels. Thus,
`manufacturers have had to devise software or hard-
`
`ware correction schemes to compensate for lcns
`vignctting. Finally, large aperture lenses can gener-
`ate distortions, such as pin cushion or barrel distor-
`tion, which cause straight lines to appear curved in
`the image, especially at the edges of the Field of view.
`Care must be taken to ensure that these geometric
`distortions are negligible, otherwise the portal image
`will give a geometrically inaccurate representation of
`the patient's position.
`An example of an image acquired by the Infimed
`T.V. camera-based EPID (Infimed, Liverpool, NY)
`while using a special video digitizer (not available on
`the commercial system) is shown in Fig 5. The image
`was acquired using a 2-monitor unit irradiation and a
`6 MV x-ray beam. The TV camera. of the lnfimed
`EPID has a shading correction circuit that corrects
`for gradual changes in brightness caused by lens
`Vignetting and so the effect of lens vignetting is not
`visible in Fig 5A. No corrections have been applied to
`the image apart from changing the display contrast.
`Note that, although this EPID generates images that
`are free of artifacts such as lines or point blemishes,
`Fig 5 is not free from confounding structures. The
`tennis racket of the treatment couch is visible and
`can obscure some anatomic structures.
`
`Practical Considerations
`
`The above discussions have focussed on the factors
`
`that influence image quality. However, there are a
`number of other considerations that can also influ-
`
`ence the clinical utility of EPIDS. These include (1)
`the size of the mechanical assembly; (2) the field of
`view of the device; and, (3) the speed of the image
`acquisition. Table l summarizes some of the charac-
`teristics of the commercially available EPIDS.
`The aim of EPIDS is to improve the geometric
`accuracy of radiation treatments. Therefore, it is
`essential that the EPII)s do not make the setup ofthe
`patient more awkward, which could lead to an
`increase rather than decrease in field-placement
`errors. Photographs of EPIDS from different manu-
`facturers, whilc retracted and while deployed, are
`shown in Fig 6. Clearly, widely different designs have
`been used for the EPID assemblies. The mechanical
`
`assemblies are either retractable (Fig 6A,B,G,H),
`dcmountablc (Fig 6C,D), or partly dcmountable and
`partly retractable (Fig 6E,F). These differences
`
`Page 6 of 19
`
`

`
`Portal Imaging Technology
`
`121
`
`
`
`Figure 5. Image of the pelvic region of a patient acquired using the Infimed EPID, a 6-MV x-ray beam, and a 2-monitor
`unit irradiation: (A) before and (B) after window and leveling to optimize the display contrast. The TV camera of the
`Infimed EPID uses a shading correction circuit that corrects for lens vignctting during image acquisition.
`
`change how the devices are used. For instance: the
`gantry has to be set to 0° to attach and deploy the
`demountable EPID (Fig 6C,D); two of the EPID
`designs permit the detector-to-isocenter distance to
`be adjusted to acoomodate thicker patients or in-
`crease the efiective field ofvicw (Fig 6E,F,G,H); and,
`irradiation of the external electronics is a concern for
`
`one of the EPID designs (Fig 6G,H).
`One limitation of all currently available EPIDs is
`their small effective field of view. Whereas the sizes
`
`‘of the x—ray detectors are similar to that of standard
`(35 X 43 cm) film cassettes, the EPIDs are deployed
`at much larger isocenter-to-detector distances than
`film cassettes, thereby reducing their effective field
`of view. Having a large field of view is obviously
`important if one is imaging a large radiation field
`such as the mantle fields. However, Table I shows
`that one does not have to be treating with such large
`fields to reach the limits of the current EPIDs. Most
`
`EPIDs have fields ofview that cover an area approxi-
`mately 1/§ that ofa film cassette (which typically could
`be placed almost in Contact with the patient). It is not
`clear whether EPIDs with larger fields of View will
`become available soon because this would require
`that the current EPIDs be redesigned completely. It
`remains to be seen whether the advantages ofEPlDs,
`such as quantitative image analysis, can make up for
`their limited field ofview.
`
`Surprisingly, the characteristic that identifies each
`of the diflerent EPIDs most uniquely is not image
`quality, field of view, or physical size of the radiation
`detector, but the speed of image acquisition. More-
`over,
`it is this characteristic that has the largest
`influence on how these devices are used in the clinic.
`
`For instance, the Varian EPID (Varian Oncology
`Systems, Palo Alto, CA), which requires a long time
`to acquire an image,
`is unlikely to be used to
`intervene in patient positioning or to generate a
`movie of patient motion during treatment. In con-
`trast, the Siemens EPID (Siemens Medical Systems,
`Oncology Care Systems, Concord, CA) has been
`used extensively to generate movies of patient mo-
`tion during treatment,”-75 but requires that the user
`manually turn the radiation beam off‘when acquiring
`double-exposure images. Somewhere in between in
`acquisition rate is the Infimed EPID, which accumu-
`lates signals generated by a short (0.5 to l.0 second)
`irradiation and displays the image immediately after
`the irradiation has been completed, or generates a
`sequence of images every 0.5 to 1.0 second during a
`treatment. Thus,
`the Infimed EPID is very well
`suited to acquiring double-exposure images while
`minimizing the radiation exposure the patient. Al-
`though the products from the different manufactur-
`ers are becoming similar, distinctions remain be-
`cause of the different philosophies that initiated their
`
`Page 7 of 19
`
`

`
`122
`
`Peter 11/Iunm
`
`Table 1. The Characteristics of the Four Commercially Available EPIDS
`
`Supplier
`
`Iryimea.’/ GE
`
`Name
`
`Type
`Detector
`
`TV Camera
`
`Mechanical assembly
`
`Mounting
`
`Collision interlock
`
`Field ofview (cm)
`
`Detector to isocentre
`distance
`
`lsocentre accuracy
`Prototype descriptions
`
`Theraview/Target-
`View
`TV camera—based
`1.5 mm brass
`plate + 4-00
`mg/cm? Gd2O2S
`screen
`Plurnhicon or Chal-
`nicon
`Partly retractable and
`partly demountable
`Any accelerator (GE,
`Varian, Scandi-
`tronix)
`yes (customer must
`connect to accel-
`erator motion
`interlocks)
`Adjustable
`30.8 dia. Varian (iso-
`center)
`31.5 dia. Scanditronix
`31.7 dia. GE.
`40 X 40 (detector)
`30-60 cm (Varian)
`27-78 cm (Scanditro—
`nix)
`26-67 cm
`:5 m m
`Ref:')l
`
`Philips
`SR1 100
`
`Sicmem
`
`Varizzn
`
`Be amyiewph”
`
`PortalVision
`
`TV camera—based
`1.5 mm steel
`plate + 411
`mg/cm? GdgO2S
`screen
`CCD
`
`Demountable
`
`TV camera—based
`l—2mm plate + 160
`mg/cm?’ GdQOgS
`scrccn
`
`Newyicon
`
`Matrix ion chamber
`1.0 mm stainless steel
`plate + 0.8 mm
`2,2,4-trimethyl-
`pentane
`n/a
`
`Fully retractable
`
`Fully retractable
`
`Philips only
`
`Siemens only
`
`Any accelerator (cus-
`tomer must attach)
`
`Yes
`
`No (interlock prevents
`collision during
`deployment only)
`
`yes
`
`Fixed
`19 X 24- (isocenter)
`
`Fixed
`24 X 30 (isocenter)
`
`Adjustable
`25 X 25 (isocenter)
`
`30 X 38 (detector)
`60 cm
`
`35 X 4-4 (detector)
`39 cm
`
`325 X 32.5 (detector)
`5-80 cm
`
`ilmm
`Ref 19
`
`:2 mm
`Ref 53
`
`:5 mm
`Ref 46, 47
`
`
`
`NOTE. Some important considerations are how well the collision detection rnecliaiiisms ofthe (le\/ices wor