Characterization of a two-dimensional liquid-filled ion chamber detector array used forverification of the treatments in radiotherapyMiljenko Markovic, Sotirios Stathakis, Panayiotis Mavroidis, Ines-Ana Jurkovic, and Nikos Papanikolaou

Citation: Medical Physics 41, 051704 (2014); doi: 10.1118/1.4870439 View online: http://dx.doi.org/10.1118/1.4870439 View Table of Contents: http://scitation.aip.org/content/aapm/journal/medphys/41/5?ver=pdfcov Published by the American Association of Physicists in Medicine Articles you may be interested in Experimental analysis of general ion recombination in a liquid-filled ionization chamber in high-energy photonbeams Med. Phys. 40, 062104 (2013); 10.1118/1.4805109 Dosimetric characterization of a multileaf collimator for a new four-dimensional image-guided radiotherapysystem with a gimbaled x-ray head, MHI-TM2000a) Med. Phys. 37, 4684 (2010); 10.1118/1.3480510 An MLC-based linac QA procedure for the characterization of radiation isocenter and room lasers’ position Med. Phys. 33, 1780 (2006); 10.1118/1.2198171 The stability of liquid-filled matrix ionization chamber electronic portal imaging devices for dosimetry purposes Med. Phys. 31, 819 (2004); 10.1118/1.1668411 Validation of dynamic MLC-controller log files using a two-dimensional diode array Med. Phys. 30, 799 (2003); 10.1118/1.1567951

Characterization of a two-dimensional liquid-filled ion chamber detectorarray used for verification of the treatments in radiotherapy

Miljenko Markovic,a) Sotirios Stathakis, Panayiotis Mavroidis, Ines-Ana Jurkovic,and Nikos PapanikolaouDepartment of Radiation Oncology, University of Texas Health Sciences Center at San Antonio,San Antonio, Texas 78229

(Received 23 April 2013; revised 3 February 2014; accepted for publication 23 March 2014;published 14 April 2014)

Purpose: The purpose of the study is to investigate the characteristics of a two-dimensional (2D)liquid-filled ion chamber detector array, which is used for the verification of radiotherapy treatmentplans that use small field sizes of up to 10 × 10 cm.Methods: The device used in this study was Octavius 1000 SRS model (PTW, Freiburg, Germany).Its 2D array of detectors consists of 977 liquid-filled ion chambers arranged over an area of 11× 11 cm. The size of the detectors is 2.3 × 2.3 × 0.5 mm (volume of 0.003 cm3) and their spacing inthe inner area of 5.5 × 5.5 cm is 2.5 mm center-to-center, whereas in the outer area it is 5 mm center-to-center. The detector reproducibility, dose linearity, and sensitivity to positional changes of thecollimator were tested. Also, the output factors of field sizes ranging from 0.5 × 0.5 to 10 × 10 cm2

both for open and wedged fields have been measured and compared against those measured by apin-point ionization chamber, liquid filled microchamber, SRS diode, and EDR2 film.Results: Its short-term reproducibility was within 0.2% and its medium and long-term reproducibilitywas within 0.5% (verified with air ionization chamber absolute dose measurements), which is anexcellent result taking into account the daily fluctuation of the linear accelerator and the errors inthe device setup reproducibility. The dose linearity and dose rate dependence were measured in therange of 0.5–85 Gy and 0.5–10 Gy min−1, respectively, and were verified with air ionization chamberabsolute dose measurements was within 3%. The measurements of the sensitivity showed that the 2DArray could detect millimetric collimator positional changes. The measured output factors showedan agreement of better than 0.3% with the pinpoint chamber and microliquid filled chamber for thefield sizes between 3 × 3 and 10 × 10 cm2. For field sizes down to 1 × 1 cm2, the agreement withSRS diode and microliquid filled chamber is better than 2%. The measurements of open and wedge-modulated field profiles were compared to the film and ionization chamber in water measurements.Conclusions: The Octavius Detector 1000 SRS is an accurate, precise, and reliable detector, veryuseful for the daily performance of the patient specific quality assurance of radiotherapy treatmentplans. © 2014 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4870439]

Key words: Octavius 1000 SRS, 2D liquid-filled ion chamber detector array, quality assurance oftreatment plans, IMRT QA

1. INTRODUCTION

With the advent of stereotactic body radiation therapy(SBRT), stereotactic radiosurgery (SRS), and intensity modu-lation radiotherapy (IMRT), beamlets with transversal dimen-sions of less than 10 mm are common. The doses for suchmodalities are calculated by dividing the beams into beam-lets that have varying intensities. Since the dimensions of thebeamlets may be too small to establish electronic equilibriumwithin them, calculations based on corrections to broad-beamdata will not suffice.1 Furthermore, in SBRT higher doses infewer fractions are delivered, which increases the biologicaleffective dose.2 A reliable dose delivery is required in termsof both spatial and numerical accuracy in order to achievegood treatment results without damaging the healthy tissuesurrounding the target. Raising the standards of the deliver-able treatments has been possible due to advances in radiationtherapy. Complex fields and dose painting using multileaf col-limators (MLCs) and micromultileaf collimators (mMLCs)

have boosted the clinical implementation of intensity modu-lated treatments as well as the use of small fields for preciselytargeting the treatment volumes. A number of studies showedthat the use of MLCs has improved the conformal radiother-apy techniques3 and they have pointed out the clinical advan-tages of these planning and delivery techniques.4–6 Paskalevet al.7 concluded that linac-based treatments with very smallphoton beams are clinically feasible.

With the clinical implementation of more complex treat-ments there is a need for a more accurate verification of doseas well as of QA systems and procedures.8, 9 Conventionalsmall field measurements require very precise fixation of de-tector at center of the small field at the depth of interest. Asubmillimeter displacement of the measuring point due to me-chanical uncertainties of the detectors may yield a significanterror in the measured value.10 Kutcher et al.11, 12 and Fraasset al.13 provided comprehensive reports on the implementa-tion of standard quality assurance (QA) programs in radio-therapy treatment planning. Clinical routine IMRT QA and

051704-1 Med. Phys. 41 (5), May 2014 © 2014 Am. Assoc. Phys. Med. 051704-10094-2405/2014/41(5)/051704/14/$30.00

051704-2 Markovic et al.: 2D ion chamber detector array for plan verification 051704-2

verification require specific software tools and measurementdevices.14–17 Gonzalez-Catano et al.18 investigated a liquidfilled ionization chamber for high precision relative dosime-try and showed its ability to perform profile measurementsand penumbrae determination with excellent accuracy. Clin-ical application for the real time profile verification of highgradient beam profiles like those that are present in inten-sity modulated radiation therapy and radiosurgery would re-quire a detector with good repeatability, small pixel size, andhigh spatial resolution.19 The intrinsic characteristics of theliquid filled ionization chamber have to be taken into con-sideration when it comes to the accuracy of the measure-ments. The escape probability from initial recombination ofthe created charge in an ionized liquid, which gives the freeion yield, depends on both the collecting electric field andtemperature.20, 21 By keeping polarization voltage stable andmonitoring temperature is the only way of dealing with thesetwo initial recombination dependencies. With the rapid in-crease in application of rotational therapy, such as arc ther-apy, angular dependence of the detector has to be taken intoaccount. It has been shown that detector measurements un-derestimate the dose. Esch et al.22 investigated online qualityassurance of rotational radiotherapy using a two-dimensional(2D) array. They observed that the array measurements under-estimated the dose on the beam axis for the open field half arctreatment deliveries by 4% in 6 MV and 2% in 18 MV.

The present investigation evaluated a two-dimensionalliquid-filled ion chamber detector array for planar dose mea-surement of clinical radiation beams. The detector array hasa limited field size with high resolution designed to be suit-able for patient specific quality assurance of SBRT, SRS, andstereotactic fractionated radiotherapy (SRT) targets. We fo-cused on measurements of the field sizes up to 10 × 10 cm2,which are relevant for the clinical application of the detector.Additionally, the dosimetric properties and the performanceof the two-dimensional liquid-filled ion chamber detector ar-ray were evaluated.

2. MATERIAL AND METHODS

In this study, the properties and characteristics of the 2Dliquid-filled ion chamber detector array Octavius 1000 SRS(PTW, Freiburg, Germany) are investigated (Fig. 1, upperimage).

The 2D detector array consists of 977 liquid-filled ioniza-tion chambers that are arranged in a square plane and are ableto measure field sizes of up to 11 × 11 cm2. The size of eachdetector is 2.3 × 2.3 × 0.5 mm (volume = 2.65 mm3). Thespacing of the detector in the inner, high resolution area (itrefers to field sizes of up to 5.5 × 5.5 cm2) is 2.5 mm center-to-center, whereas the spacing of the detector in outer area is5 mm center-to-center as shown in the lower image of Fig. 1.

The linear dimensions of the 2D Octavius Detector 1000SRS (Fig. 1, upper image) are 30 × 42 × 2.2 cm (W × D × H)and its weight is approximately 5.4 kg. According to the man-ufacturer, the reference point of measurement is located 9 mmbelow the surface of the array although other investigations30

have indicated that the effective point of measurement is lo-

FIG. 1. (Upper) The two-dimensional liquid-filled ion chamber detector ar-ray Octavius Detector 1000 SRS. (Lower) Schematic illustration of the 977liquid-filled ion chambers of the device. The detector size is 2.3 × 2.3× 0.5 mm (volume = 0.003 cm3). The detector spacing in the inner area(maximum field size = 5.5 × 5.5 cm) is 2.5 mm center-to-center and in theouter area (maximum field size = 11 × 11 cm) is 5 mm center-to-center.

cated 9.5 mm (±0.2 mm) below the surface of the array. Thedevice is able to measure absorbed dose and absorbed doserate to water. The Octavius Detector 1000 SRS and its mul-tichannel array interface are calibrated by the manufacturer.The calibration consists of a relative output of each detectorcompared to the central one. An onsite factor correcting forthe quality of the beam and daily output variation of the linaccan be measured and used during measurements by the ac-companying acquisition and analysis software. In this study,the correction factor was measured before each measurementsession. The Octavius Detector 1000 SRS is controlled bythe VeriSoft software, which is used during measurementacquisition and analysis of the measurements. The software

Medical Physics, Vol. 41, No. 5, May 2014

051704-3 Markovic et al.: 2D ion chamber detector array for plan verification 051704-3

provides the user with evaluation tools, such as profile com-parison (horizontal, vertical, diagonal), planar isodose over-lay, and gamma index calculation.

A Varian Novalis Tx (Varian Medical System, Palo Alto,CA, Brainlab AG, Germany) linear accelerator was used in allthe measurements of this study. All the measurements wereperformed using 6 MV photon beams, with a constant pulserepetition frequency of 360 Hz. Although small fields can bedelivered with higher MV modalities, the vast majority of thesmall field stereotactic radiotherapy treatments are carried outat low energy. Therefore, the device has only been tested at6 MV.

2.A. Phantom setup

Figure 2 (upper) shows the setup of the Octavius Detector1000 SRS used throughout the investigation. During the mea-surements dose integration time was set to 900 s with an in-terval time of 0.2 s. Plastic Water Phantom material (CNMC,Nashville, TN) was used for the buildup and backscatter. Plas-tic Water slabs are plates of a 30 × 30 cm2 surface size, ofvarious thicknesses, with a density of 1.03 g/cm3.

In this setup, the thickness of the build-up slab was 4.1 cmand the thickness of the backscatter slab was 5 cm. The sourceto surface distance (SSD) for the measurement setup was95 cm. The reference point of measurement of the OctaviusDetector 1000 SRS was set to be at 100 cm. Figure 3 showsthe vertical position of the Octavius Detector 1000 SRS usedfor the measurements of the directional (angular) dependenceof the detector with the gantry angles that were used for thebeams delivery.

2.B. Reproducibility

The Octavius Detector 1000 SRS was assessed over a pe-riod of 30 days. The reproducibility of the output of the de-tector has been checked for periods of: (a) hours (short-term),(b) a few days (midterm), and (c) several weeks (long-term).All the reproducibility measurements were performed withthe detector setup shown in Fig. 2 (upper). A 5 × 5 cm2 fixedbeam and 100 MU were used in each measurement and thedose at the central chamber was recorded. Each measurementwas repeated multiple times and the corresponding averageand standard deviation were calculated. Each measurement inthe linac was accompanied by a PTW N31003 air ionizationchamber absolute dose measurement to remove linac outputvariation and evaluate only the device response. Stability be-tween different detectors in the same setup was evaluated.

2.C. Charge collection efficiency

Several methods have been suggested for performing col-lection efficiency determination.23–26 Since Octavius DetectorSRS 1000 has fixed voltage (1000 V), the collection efficiencymeasurements were carried out with PTW 31003 air ioniza-tion chamber and the central chamber of the Octavius De-tector using the method described by Tölli et al.27 The mea-surements were performed with the detector setup shown in

FIG. 2. (Upper) The standard setup that was used for the measurementsthat were acquired using Octavius Detector 1000 SRS. The detector arrayis placed between two slabs of Plastic Water. The source to surface distanceis 95 cm. The effective point of measurement of the detector is set to be at100 cm from the radiation source. (Middle) Experimental setup for the mea-surements using an EDR film. The 5 cm Plastic Water phantom is placed ontop of the EDR film. (Lower) Setup for the measurements using OctaviusDetector 1000 SRS. The detector array is placed on a slab of Plastic Waterwithout any build-up slab. The SSD is 99.1 cm. The detectors’ effective pointof measurement is set to be at 100 cm from the radiation source. The PlasticWater slab used for backscatter has a 5 cm thickness.

Fig. 2 (upper). For the constant pulse repetition frequency of360 Hz, the measurements were done at different distancesof the effective point of measurement of the detectors to theradiation source with the dose per pulse varying from 1.47× 10−4 to 5.1 × 10−4 Gy/pulse.

2.D. Leakage current stability

With the detector setup shown in Fig. 4 the leakage currentstability was studied. A 10 × 10 cm2 fixed beam was used todeliver 1 Gy in each measurement at the central chamber.

The measurements were recorded before and after the ir-radiation of 2 Gy. The measurements were repeated using thesame setup and 2 Gy irradiation allowing 8 h between the

Medical Physics, Vol. 41, No. 5, May 2014

051704-4 Markovic et al.: 2D ion chamber detector array for plan verification 051704-4

FIG. 3. (Left) The setup that was used for the angular dependence measurements that were acquired using the Octavius Detector 1000 SRS. The detector arrayis placed vertically between two slabs of Plastic Water. The source to surface distance is 95 cm. The effective point of measurement of the detector is set to be at100 cm from the radiation source. To avoid tabletop from the beam, gantry angles started at 90◦ and ended at 270◦. (Right) The setup that was used for the airion chamber measurements.

measurements and they were also verified along different daysfor a period of 11 days. Each measurement in the linac wasaccompanied by a PTW N31003 air ionization chamber abso-lute dose measurement to remove linac output variation andevaluate only the device response. Also, the leakage currentstability was verified for a period of 1 h with the detector setupshown in Fig. 2 (upper). A 10 × 10 cm2 fixed beam was usedto deliver 100 MUs and the charge collected was recorded atdifferent times during the 1-h period. The measurements wererepeated in different days using the same setup and time in-terval. The charge collected was recorded using a sample ofnine different detectors.

2.E. Linearity

The dose linearity of Octavius Detector 1000 SRS wastested using the same conditions as the ones shown in

FIG. 4. The setup that was used for the leakage current stability measure-ments of the Octavius Detector 1000 SRS. The detector array is placed be-tween slabs of Plastic Water. The source to surface distance is 100 cm. Theeffective point of measurement of the detector is set to be at 5 cm depth. Theair ion chamber is placed on central axis at a depth of 3.1 cm.

Fig. 2 (upper). A 5 × 5 cm2 field of 6 MV photons was usedto expose the detector. To determine the linearity with dose,varying the dose rate to the highest possible value was ac-complished within Boag28 conditions. All the measurementswere compared to those of an air ionization chamber PTWN310013.

2.F. Energy dependence

For this dose fidelity aspect, energy dependence was alsostudied considering the measurement in a standard field asa function of phantom depth, which compared to standarddosimetry techniques by measuring PDD and profiles at dif-ferent depths.

2.G. Output factors

The linear accelerator radiation output as a function of thefield size using the Octavius Detector 1000 SRS was inves-tigated. The output factor measurements were carried out bydelivering 100 MUs for square fields ranging from 0.5 × 0.5to 10 × 10 cm2 using the setup of Fig. 2 (upper). Due to a na-ture of the clinical application of the Octavius Detector 1000SRS, the measurements were limited to a maximum field sizeof 10 × 10 cm2. The dose output measurements acquired bythe Octavius Detector 1000 SRS were compared with mea-surements acquired using a PTW pinpoint ionization chamber(type 31014, PTW, Freiburg, Germany), PTW MicroLion ion-ization chamber, and PTW SRS diode. The pinpoint ioniza-tion, MicroLion, and SRS diode measurements were made inwater under the same setup conditions (SSD = 95 and 100 cmto the effective point of measurement). The Unidos Weblineuniversal dosimeter (PTW, Freiburg, Germany) was used inthe pinpoint, MicroLion, and SRS diode measurements.

2.H. Sensitivity

The Octavius Detector 1000 SRS has ionizing chambersequally spaced at 2.5 mm intervals in the inner area (for field

Medical Physics, Vol. 41, No. 5, May 2014

051704-5 Markovic et al.: 2D ion chamber detector array for plan verification 051704-5

sizes up to 5.5 × 5.5 cm) and at 5 mm intervals in outer area(for field sizes up to 11 × 11 cm). In the central part, thedetectors are adjacent to each other without spacing in be-tween. The effect of the signal perturbation caused by thephoton scattering in the close vicinity of the single detectorhas been studied by Martens et al.29 for the PTW LA48 linearion chamber array. The geometrical separation of detectorscould lead to loss of sensitivity to positional changes of thecollimation system. While LA48 has guard-ring liquid ioniza-tion chambers, the trend along these last 10 years is to realizethat for pixilated LICs the guard ring is unnecessary. Thus, inthe inner matrix of Octavius only a small gap between pixelsexists. Whenever the matrix is polarized and irradiated, thedielectric materials in the gap become superficially charged,thus reaching a steady electric field configuration in which allthe electric field lines from the drift electrode end in the met-allized layer of the Octavius matrix. Therefore, in the innermatrix there is no dead area while the possible lack of sen-sitivity refers to the outer part of the matrix in which elec-trodes are separated with metallized guard-ring electrodes.Poppe et al.30 investigated detector response function and theydetermined the lateral detector response function of a singlechamber of the array. The sensitivity of the detectors of theOctavius Detector 1000 SRS to positional changes of the col-limation system was investigated.

For testing purposes, the Octavius Detector 1000 SRS wasirradiated with three rectangular asymmetric fields (A, B, C)as shown in the left panel of Fig. 5. The center of the field Ahas been offset to −4.0 cm along the x direction, the centerof the field B has been offset to −2.5 cm, and the center ofthe field C has been set to 0.0 cm. The initial dimension ofeach field was 2 × 11 cm2. One of the collimator jaws (Y1in Fig. 4) was then gradually moved to close the field with anincrement step of 1 mm. The final size of each radiation fieldwas 0.5 × 11 cm2 as shown in the right panel of Fig. 5.

At each step a new acquisition was taken by delivering 100MUs. Using the fields A, B, and C as shown in Fig. 5, half ofthe detectors were able to be tested given that symmetry ex-ists between the left and right parts of the detector. All themeasurements for the sensitivity tests of the Octavius Detec-tor 1000 SRS were performed using the setup shown in Fig. 2(upper). The pinpoint ion chamber (PTW model 31014) wasused to repeat the sequence of measurements in a water phan-tom, under the same geometrical conditions placed at the po-sition of the central array chamber for the measured field. Thesensitivity of the detector was investigated by changing thebuildup. The detector was irradiated in the same manner withthree rectangular asymmetric fields (A, B, C) as shown in theleft panel of Fig. 5 and described above. In every set of mea-surements, the buildup was changed by adding 1 mm (thinnestavailable) of plastic water slab in the standard setup shown inFig. 2 (lower).

2.I. Directional (angular) dependence

With the detector setup shown in Fig. 3, the isotropical be-havior of the device was investigated. For this test, differentincidence angle beams in an adequate phantom were deliv-

ered in order to provide the tissue-equivalence and angulardependence of the detector. The dose distribution for eachfield was also calculated on the CT scan of the phantom setupwith treatment planning system (Phillips Pinnacle 9.2) usingsuperposition/convolution31 dose calculation algorithm. Mea-surements for different incidence angles with respect to thedevice were acquired and compared with an air ionizationchamber (PTW model 31003).

2.J. Clinical application

The Octavius Detector 1000 SRS was used to measure thedose maps of clinical open and wedge fields and its perfor-mance was analyzed. The field sizes that were used in themeasurements were 2 × 2, 4 × 4, 6 × 6, and 8 × 8 cm2.The results were then compared against the measurementswith the pinpoint ion chamber at 5 cm depth in water at95 cm SSD. Furthermore, measurements were acquired usingan EDR film (Kodak, Rochester, NY) placed between a two-slab Plastic Water phantom as shown in Fig. 2 (middle). Thenumber of monitor units delivered during the measurementswith the EDR film was the same as that delivered during themeasurements with the Octavius Detector 1000 SRS. The off-axis measurements were performed using Octavius Detector1000 SRS and the setup shown in Fig. 2 (upper). The fieldsizes used for the measurements were 0.5 × 0.5, 1 × 1, 1.5× 1.5, and 2 × 2 cm2. These measurements were comparedwith measurements acquired with SRS diode in water underthe same conditions (5 cm depth at 95 cm SSD).

3. RESULTS

3.A. Reproducibility

The output of the Octavius Detector 1000 SRS as a func-tion of the time period, during which the measurements weretaken, is shown in Fig. 6. The measurements were acquiredover a period of 30 days and they were collected in batchesof ten. For every batch, the average of the measurements wascalculated along with the respective standard deviation. Eachmeasurement in the linac for the reproducibility study was ac-companied by an air ionization chamber (PTW model 31003)absolute dose measurement.

Liquid-filled ionization chambers are electric field andtemperature dependant.20 By keeping the polarization voltageconstant we managed to minimize their effect on recombi-nation dependency. Daily changes of temperature were mea-sured and a correction due to those variations was applied byusing the correction values feature of the VeriSoft software,which was used during the measurement acquisition andanalysis.

The reproducibility of the measurements within ev-ery batch showed good consistency and stability with thehighest value of the standard deviation of the mean notexceeding 0.2%, which is an indicator of the very short-termreproducibility performance of the Octavius Detector 1000SRS. The standard deviation between the batches was alsocalculated and its value does not exceed 0.5%, which is an

Medical Physics, Vol. 41, No. 5, May 2014

051704-6 Markovic et al.: 2D ion chamber detector array for plan verification 051704-6

FIG. 5. (Left panel) Experimental setup for the investigation of the perturbation effect of the irradiated area of the Octavius Detector 1000 SRS and of thesingle detector sensitivity to positional changes of the collimation system. Three rectangular and asymmetric fields and different off-axis positions were chosenfor the irradiation. The initial field size was 2 × 11 cm2. For each field, the secondary collimator Y1 was moved in increments of 1 mm to close the radiationfield. (Right panel) The final size of each radiation field was 0.5 × 11 cm2. At each step a separate acquisition was taken.

Medical Physics, Vol. 41, No. 5, May 2014

051704-7 Markovic et al.: 2D ion chamber detector array for plan verification 051704-7

FIG. 6. (Left) Diagram presenting the results of the reproducibility test for the Octavius Detector 1000 SRS. Each measurement in the linac was accompaniedby an air ionization chamber absolute dose measurement in order to have complete knowledge of the device contribution. (Right) The percentage differencesbetween the absolute doses measured by an air ionization chamber and those of the Octavius Detector are also plotted.

excellent indicator of the reproducibility with the longer pe-riod of time. The standard deviation between the detectormeasurements and an air ionization chamber absolute dosemeasurements between the batches does not exceed 0.3%,with maximum daily percentage difference being 0.6%. Thesample of the nine different chambers was used to study theintercalibration stability. The intercomparison of the measure-ments showed an average interchamber deviation of 0.2%with a maximum deviation of 0.5%, which was noted on onechamber.

3.B. Charge collection efficiency

The charge collection efficiency measurements were per-formed with the detector setup shown in Fig. 2 (upper). Asinvestigated by Tölli et al.,27 the charge collection efficiencywas calculated using the two-dose-rate method with the doseper pulse varying from 1.47 × 10−4 to 5.1 × 10−4 Gy/pulse.The measurements showed that the maximum dose per pulseof 2 × 10−4 Gy/pulse yields a charge collection efficiencyhigher than 0.98. The results are shown in Fig. 7. This di-agram also illustrates the dose rate dependence of the Oc-tavius Detector since the charge collection efficiency versusdose per pulse (dose rate) is a more accurate indication of thisproperty.

FIG. 7. The Octavius Detector 1000 SRS charge collection efficiency curvehas been determined by measurements using two-dose rates. The percentageof collection efficiency in the measurements is plotted.

3.C. Leakage current stability

The measurements of pre- and postirradiation of 2 Gyshowed no significant change in the absolute dose measured.The maximum percentage difference against the measure-ments carried out with an ion ionization chamber (PTW31003) was 0.1% (Fig. 8, left). The daily changes in temper-ature were measured and a correction for the variations wasapplied by using the relevant tool of the VeriSoft software,which was used during the measurement acquisition and anal-ysis. Also, the leakage current stability was verified using asample of nine detectors for 1 h time period. The leakage cur-rent showed that the response is linear with a mean value of3.26 × 10−14 C/s and a standard deviation of 1.61 × 10−15 C/s(Fig. 8, right).

3.D. Linearity

The results of the linearity test are shown in Fig. 9. To de-termine the linearity with dose while varying the dose rate tothe highest possible value, measurements were made both un-der Boag conditions (pulses separated more than 8 ms) andbreaking Boag conditions. In the first case, the linac was setto a proper long period pulse repetition and the source to de-tector distance was varied. In the second condition, the pulserepetition was varied from the longest to the shortest value.All the measurements were compared to those of an air ion-ization chamber.

The measured results of the central axis (CAX) of the ion-ization chamber using Boag and breaking Boag conditionsshowed linear response with the dose rate. The measurementsthat were made under the Boag condition with an air ion-ization chamber were compared to the corresponding mea-surements made with the Octavius detector. The standard de-viation of the local percentage difference was 1.1 with themaximum percentage difference being 1.7%. On the otherhand, the measurements made under the breaking Boag con-ditions with an air ionization chamber were compared to thosemade with the Octavius detector and they showed that thestandard deviation for the local percentage difference was1.6% with the maximum percentage difference being 3.0%.

Medical Physics, Vol. 41, No. 5, May 2014

051704-8 Markovic et al.: 2D ion chamber detector array for plan verification 051704-8

FIG. 8. Octavius Detector 1000 SRS leakage current stability test. (Left) Percentage difference in absolute dose measured before and after 2 Gy irradiation.(Right) Stability of the charge collected during the period of 1 h.

The data analysis indicates that the clinical implementationof the linearity variations underestimates values for the mea-surements at the highest achievable dose rate.

3.E. Energy dependence

For dose fidelity aspect, energy dependence has been stud-ied. Measurements in a standard field as a function of phan-tom depth for different field sizes were acquired and com-pared to standard dosimetry techniques. A study showingPDD/profiles measured at different depths were conducted.Using the standard setup (Fig. 2), measurements for differentfield sizes (3 × 3, 5 × 5, 7 × 7 cm2) were made at differentdepths by adding different build-up thickness of solid waterphantom. Profiles were acquired at depth of 9 mm, dmax,50, 100, and 200 mm. The measured values were normal-ized to the value at the depth of 10 cm. The measured datahave been accompanied with measurements in water by a pin-point chamber and SRS diode under the same setup condi-tions. It is noted that the detector sensitivity increases withdepth. The detector measurements were compared with thosemade with the pinpoint chamber and they showed a maxi-mum percentage difference of 1.1% for the depths betweendmax and 200 mm, whereas the respective comparison withthe measurements by the SRS diode showed a difference of2.3%. The diode convolved profiles were compared with thearray and penumbra regions. In relation to the diode-measuredprofiles, the σ value was determined to be between 0.7 and1 mm for the measured field sizes between the depths of dmax

and 50 mm. For the purpose of this study, several diagrams ofPDD and profiles measured at different depths are shown inFig. 10. Using the method proposed by Looe et al.,32 the ef-fective point of measurement of the detector was determinedby shifting the TPR curve of the Octavius detector centralchamber to match the TPR curve of the Roos chamber. Basedon our measurements, the effective point of measurement wasfound to be located 9 mm below the surface of the array.

3.F. Output factors

The results of the output factors are shown in Fig. 11.Small field dosimetry is more demanding than dosimetry ofconventional field sizes.

Volume averaging effect and the lack of lateral chargeequilibrium start to play a non-negligible role and the ap-proximations of classical radiation physic, such as Bragg-Gray conditions tend to be valid to a lesser extent comparedto larger field dosimetry. Wuerfel1 described how to decidewhich detector to choose for the measurements of the smallfield sizes and how to perform the measurement. Positioningaccuracy that can lead to a relevant uncertainty is very im-portant in small fields and it has to be taken into account.Following Wuerfel’s1 recommendations, we compared theOctavius Detector 1000 SRS measurement with those of thepinpoint air ionization chamber (PTW 31014) for field sizesfrom 3 × 3 to 10 × 10 cm2. The agreement between theOctavius Detector and the pinpoint chamber measurementsis within ±0.3% (local relative deviation). A single chamber

FIG. 9. Octavius Detector 1000 SRS linearity test: the output values of the single channel. (Left) Measured under the Boag conditions. (Right) Measured underbreaking Boag conditions.

Medical Physics, Vol. 41, No. 5, May 2014

051704-9 Markovic et al.: 2D ion chamber detector array for plan verification 051704-9

FIG. 10. The measurements in a standard field as a function of phantom depth are compared with measurements by an air chamber and a SRS diode. (Upper)The diagrams show the PDD profiles measured for the field size 3 × 3, 5 × 5, 7 × 7 cm2. (Lower) Measured profiles at 5 cm depth for the field size 3 × 3,5 × 5 cm2 compared with diode convolved profiles. For the 3 × 3 cm2 field size, a zoom-in image of the penumbra region is also shown. The σ values determinedare 0.9 and 0.7, respectively.

FIG. 11. Comparison of the output factors between the Octavius Detector 1000 SRS, pinpoint chamber, SRS diode, and microLion chamber. The field sizesrange between 0.5 × 0.5 and 10 × 10 cm2. (Upper left) Field sizes 3 × 3–10 × 10 cm2. (Upper right) Field sizes 0.5 × 0.5–3 × 3 cm2. (Lower) The percentagedifferences are also plotted for all the examined field sizes.

Medical Physics, Vol. 41, No. 5, May 2014

051704-10 Markovic et al.: 2D ion chamber detector array for plan verification 051704-10

of the Octavius Detector 1000 SRS has an active volume of0.003 cm3 whereas the PTW PinPoint chamber has an activevolume of 0.015 cm3. For field sizes from 3 × 3 down to0.5 × 0.5 cm2, the Octavius Detector measurements werecompared with those of the PTW SRS diode (type 60018).The SRS diode has an active volume of 0.0003 cm3 and itwas cross-calibrated with the pinpoint chamber. For field sizesfrom 3 × 3 down to 1 × 1 cm2, the Octavius detector and theSRS diode measurements are within 2% (local relative de-viation). For field sizes smaller than 1 × 1 cm2, the maxi-mum percentage difference is 12% (local relative deviation).Relevant measurements were also made using the PTW Mi-croLion chamber, which has an active volume of 0.002 cm3.The MicroLion chamber was also cross-calibrated with thepinpoint chamber. For field sizes from 10 × 10 down to1.5 × 1.5 cm2, the maximum percentage difference betweenthe Octavius Detector and the MicroLion chamber is ±0.8%(local relative deviation) and for the field size of 1 × 1 cm2 itis less than 2% (local relative deviation).

Other studies33, 34 have shown that for field sizes smallerthan 3 × 3 cm2, significant differences can be observedamong different detectors regarding the measured output fac-tors. The major factors that influence the measurements werefound to be detector’s material and the volume effect of de-tectors as well as positioning accuracy. The method of de-termining the center of the radiation field using an add ondevice such as the Peacock System (NOMOS CorporationPennsylvania) allows a 0.01 mm shift but at the same timeit is time consuming and not practical for the regular clinicaluse. If measured with a diode detector, which is nonwater-equivalent, the increase of the importance of secondary elec-trons in small field leads to an overestimation of output fac-tors. If measured with an ionization chamber, the reason foran underestimation of output factors is the increase of lateralelectron disequilibria with an increase of the detectors mea-suring volume. Air cavities may further reduce the relativedose reading of detectors. Moreover, the relative dose readingof the detector is also reduced due to the fact that the detectoraverages the dose across its sensitive volume. The percent-age relative difference of the output factors between the Oc-tavius detector and the other examined detectors are shown inFig. 11.

3.G. Sensitivity

Figure 12 shows the output variation of the nine chambersof the Octavius Detector 1000 SRS (Fig. 5), as a function ofthe position of the secondary collimator (Y1).

In Field A, the starting point of the Y1 collimator was−5 cm and the end point −3.5 cm. Similarly, in Field B,the Y1 collimator moved from −3.5 to −2 cm and in FieldC, from −1 to 0.5 cm. In Field C, the first acquisition wastaken with the chamber at position −1 cm and the chamber atposition 1 cm was partially covered by the projection of thecollimators. At the positions −0.75, −0.5, −0.25, 0, 2.5, and0.5 cm, the chamber at 0.75 cm is in the field and is not cov-ered by the projections of the collimators. In Fig. 12 (upper),it is shown that the chambers in the field detect the millimetric

FIG. 12. The sensitivity of the detectors of the Octavius Detector 1000 SRSto positional changes of the collimation system and the perturbation effect ofirradiated area: the output values of the nine detectors (Field C, Fig. 5) arepresented. The Y1 jaw position was changed in steps of 1 mm with the Y1jaw moving from −1 to 0.5 cm. For each detector, the change in the signalwith the respective change in the position of the Y1 is plotted (upper). Forthree detectors (−10, 0, 10, Field C, Fig. 5), the change in the signal with therespective change in the position of the Y1 jaw based on the no buildup and1 mm thickness of the buildup (middle) and 1 mm change in buildup for themeasurements at 5 cm depth (lower) are plotted.

movements of the Y1 jaw. The chamber at the position0.75 cm was never covered by the projection of the collima-tors but still its signal was influenced by the Y1 jaw, whichmodifies the linear accelerator’s head scattered radiation com-ponent. The results show similar patterns of response of thedetectors in Fields A and B since the chambers inside thefields are able to detect millimetric (or even submillimetric for1% array readout relative uncertainty) field changes as shownin Fig. 5. The measurement showed that the geometrical

Medical Physics, Vol. 41, No. 5, May 2014

051704-11 Markovic et al.: 2D ion chamber detector array for plan verification 051704-11

separation of the detectors does not contribute to the loss ofthe sensitivity. The investigation showed that there is no no-ticeable perturbation effect of the irradiated area of the Oc-tavius Detector 1000 SRS.

Figure 12 (middle and lower) shows the sensitivity of theOctavius Detector 1000 SRS regarding the changes in thethickness of the built up used during the measurement. Forthis experiment, the measurements were acquired by chang-ing the thickness of the buildup in 1 mm increments andkeeping the effective point of measurement of the detector at100 cm from the source (Fig. 2, upper). The same dose wasdelivered in every measurement and the data were recorded.In Fig. 12 (middle and lower), the response of the three de-tectors of the array at different depths is shown when 1 mmchange in buildup was applied. As noted before the relationdose to detector signal varies with depth, because at largerdepths the array signal exhibits higher variations to somesmall change of geometry or depth. At a shallow depth, 1 mmchange in buildup thickness results in notably a higher changein sensitivity compared with the same change in buildupthickness at a larger depth. While the change in sensitivityat shallow depths can be attributed predominantly to the sec-ondary charged particles, the change in sensitivity at largerdepths is predominantly caused by the change in energy spec-trum (beam hardening). The measurements that were per-formed indicate that for depths of the effective point of mea-surement up to 50 mm, the change in sensitivity due to thechange of 1 mm of buildup is noticeable.

3.H. Directional (angular) dependence

For the purpose of this investigation, we focused on a sin-gle liquid-filled ion chamber in the middle of the detectorarray. Figure 3 shows the detector setup for directional (an-gular) dependence measurements. Figure 13 shows the mea-sured and calculated values for the field size 5 × 5 cm2 forthe number of gantry angles (270◦, 255◦, 240◦, 225◦, 210◦,195◦, 180◦, 165◦, 150◦, 135◦, 120◦, 105◦, 90◦) together withthe percentage dose difference on the beam CAX detector asa function of gantry angle.

It is noticeable that when the beam incidence moves tothe rear of the array, the considerable absolute deviation be-comes apparent. Although the TPS (Phillips Pinnacle 9.2) ispredicting only slight differences between absorbed doses formirrored beam angles (e.g., 225◦, 135◦), the measurementsshow a considerably smaller signal. There are few possiblecauses for the asymmetry of the data. The different thicknessof the buildup on the top and bottom side of the array in re-spect to the inverse direction radiation. The array constructionand the nonwater equivalent materials on beam path can causedeviation in the measurements. It is also worth to mention thetolerance35 of the radiation isocenter that can contribute to theobserved deviation. All the measurements were accompaniedwith the air ion chamber measurements. There is noticeabledifference between the Octavius Detector 1000 SRS phantomand air ion chamber phantom setup because of their differentstructure and average density. Separate dose calculations forboth setups are therefore required. In Fig. 13, it is observed

that the CAX detector measurements underestimated the doseof the beam for the open field half-arc treatment deliveriesby about −4%. To account for deviations, due to the use ofthe detector in such configurations, correction factors shouldbe applied. Measurements to determine the angular depen-dence correction factors for the use of the detector in clinicalsetup should be performed and the relevant correction factorsshould be determined. However, the use of rotating phantomssuch as PTW Octavius 4D where the surface of the array isalways perpendicular to the beam will eliminate the need forangular dependence corrections.

3.I. Clinical application

Figures 14 and 15 show the results of the comparisons be-tween Octavius Detector 1000 SRS, pinpoint ion chamber,and film for open and wedged profiles of different field sizesusing the setup shown in the upper scheme of Fig. 2.

The data were normalized at the central axis. The Oc-tavius Detector 1000 SRS data fit well the pinpoint ionchamber curve and it appears to accurately measure thepenumbra region. The Octavius Detector 1000 SRS data andfilm-measured data show discrepancies in the penumbra re-gion, which can be attributed to the limited number of pointsof the calibration curve in the low dose region, to the fact thatEDR2 film overresponds in the low dose region and the lat-eral response function of the detector. The profiles measuredfor the smallest field size (2 × 2 cm2) are of interest becausethey show the advantage of the sampling capabilities of theOctavius Detector 1000 SRS. Figure 16 shows the dosimetriccomparison between the Octavius Detector 1000 SRS and theSRS diode for the open off-axis field profiles for different fieldsizes (2 × 2 cm2 and smaller). The gamma factors that werecalculated with tolerances of 2% and 2 mm yield 100% agree-ment between the Octavius detector and SRS diode. Conse-quently, corrections for detector’s lateral response were notapplied.

4. DISCUSSION

In this work, the characterization of the two-dimensionalliquid-filled ion chamber detector array Octavius Detector1000 SRS is investigated. The Octavius Detector 1000 SRSis found to be easy to use for treatment verification in radio-therapy. Its weight of 5.4 kg makes it practical for daily use.The setup procedures and experimental positioning are simpleand straightforward and the whole system can be managed ina short period of time.

The measurements performed in this study showed that theOctavius Detector 1000 SRS response regarding short-termreproducibility standard deviation of the mean does not ex-ceed 0.2%, while standard deviation of the medium and long-term reproducibility does not exceed 0.5%. The system lin-earity with dose rate throughout a wide range of dose de-livered to the detector, showed maximum deviation of 3%which makes this detector very reliable when it comes to thehigh oscillation in the dose delivery for the fields deliveringvery high doses per segment. The charge collection efficiencymeasurements showed that the maximum dose per pulse of

Medical Physics, Vol. 41, No. 5, May 2014

051704-12 Markovic et al.: 2D ion chamber detector array for plan verification 051704-12

FIG. 13. Isotropical behavior of the liquid filled ionization chamber of the Octavius Detector 1000 SRS. Measurements at different incidence angles in respectto the device were acquired and compared to an air ionization chamber. The percentage differences are also plotted.

FIG. 14. Comparison between Octavius Detector 1000 SRS, pinpoint ion chamber, and film measured profiles for open fields of different field sizes.

FIG. 15. Comparison between Octavius Detector 1000 SRS, pinpoint ion chamber, and film measured profiles for wedged fields of different sizes.

FIG. 16. Dosimetric comparison between the Octavius Detector 1000 SRS and the SRS diode for the off-axis field sizes.

Medical Physics, Vol. 41, No. 5, May 2014

051704-13 Markovic et al.: 2D ion chamber detector array for plan verification 051704-13

2 × 10−4 Gy/pulse yields charge collection efficiency higherthan 0.98. The measurements and comparison of the outputfactors with the pinpoint chamber for the field sizes from 3× 3 to 10 × 10 cm2 showed an agreement within ±0.3%.The output factor agreement between the Octavius detec-tor and the SRS diode for field sizes from 3 × 3 down to1 × 1 cm2, are within 2%. The output factor agreement be-tween the Octavius detector and the MicroLion chamber forfield sizes from 10 × 10 down to 1.5 × 1.5 cm2 is ±0.8% andfor the field size 1 × 1 cm2 is less than 2%. There are a fewmajor factors that influence the measurements such as detec-tor’s material, the volume effect of detectors, and positioningaccuracy, which has greater influence on smaller field sizes.Increasing positioning accuracy is possible by using someadd on devices although these methods are time consumingand not practical. Overall, the results obtained from the mea-surement of the output factors showed field sizes down to1 × 1 cm2 are measurable to a good fidelity with the Oc-tavius detector. This could be considered of great importancedue to the fact that it is hard to attain good output factors forthe small field sizes when using matrices of detectors. Theminimal energy dependant response of the Octavius Detector1000 SRS also presents an attractive advantage. For the mea-surement of the small segments delivered during the IMRTmeasurements, this is an important characteristic. The mea-surements that tested the detector sensitivity to small col-limator position changes as well as the field perturbationeffect showed that the Octavius Detector 1000 SRS could suc-cessfully detect millimetric positional movements of the sec-ondary collimators of the linear accelerator. The experimentaldata showed the relation dose to detector signal varies withdepth, at more depth the array signal exhibits higher varia-tion to some small change of geometry or depth. It was foundthat even 1 mm change in buildup up to 50 mm depth couldbe detected. The dose profiles measured by the Octavius De-tector 1000 SRS fitted very well the measurements obtainedwith the pinpoint ion chamber and SRS diode for differentfield sizes. The spacing of 2.5 mm between ion chambers inthe central area up to the field size of 5.5 × 5.5 cm2 is a greatadvantage in cases involving the measurement of the high-resolution profiles especially in high dose gradient regions.With the Octavius Detector 1000 SRS the workload can bereduced compared with some other verification techniques.Dose distribution are acquired, shown, and can be processedon the fly, which can be a great advantage since time is alwaysan important factor. The graphical environment for the com-parison and evaluation of the dose maps can be accomplishedby the acquisition software. Also, the measured data can besaved in different formats for the analysis.

5. CONCLUSIONS

The measurements and data analysis of this investigationprovide sufficient proof that Octavius Detector 1000 SRS is adosimetrically accurate and sensitive array detector and it canbe used for quality assurance and verification of treatments inradiotherapy. The high detector resolution and sensitivity aswell as dose for its linearity and minimum energy dependence

allows this detector to be considered as a useful tool not justfor the check of the treatment delivery in radiotherapy but alsofor the quality assurance of the output and the accuracy of thelinear accelerator.

a)Electronic mail: markovic@livemail.uthscsa.edu1J. U. Wuerfel, “Dose measurements in small fields,” Med. Phys. Int. 1, 81–90 (2013).

2J. F. Fowler, “Intercomparisons of new and old schedules in fractionatedradiotherapy,” Semin. Radiat. Oncol. 2, 67–72 (1992).

3S. Webb, Intensity Modulated Radiation Therapy (Institute of Physics,Bristol, 2000).

4M. J. Zelefsky, Z. Fuks, M. Hunt, Y. Yamada, C. Marion, C. C. Ling,H. Amols, E. S. Venkatraman, and S. A. Leibel, “High-dose intensity mod-ulated radiation therapy for prostate cancer: Early toxicity and biochemicaloutcome in 772 patients,” Int. J. Radiat. Oncol., Biol., Phys. 53, 1111–1116(2002).

5M. K. M. Kam, R. M. C. Chau, J. Suen, P. H. K. Choi, and P. M. L. Teo,“Intensity-modulated radiotherapy in nasopharyngeal carcinoma: Dosimet-ric advantage over conventional plans and feasibility of dose escalation,”Int. J. Radiat. Oncol., Biol., Phys. 56, 145–157 (2003).

6E. A. Krueger, B. A. Fraass, D. L. McShan, R. Marsh, and L. J. Pierce, “Po-tential gains for irradiation of chest wall and regional nodes with intensitymodulated radiotherapy,” Int. J. Radiat. Oncol., Biol., Phys. 56, 1023–1037(2003).

7K. A. Paskalev, J. P. Seuntjens, H. J. Patrocinio, and E. B. Podgorsak,“Physical aspects of dynamic stereotactic radiosurgery with very smallphoton beamlets (1.5 and 3 mm in diameter),” Med. Phys. 30, 111–118(2003).

8J. M. Galvin and G. Bednarz, “Quality assurance procedures for stereotac-tic body radiation therapy,” Int. J. Radiat. Oncol., Biol., Phys. 71, S122–S125 (2008).

9B. Poppe, A. Blechschmidt, A. Djouguela, R. Jollhoff, A. Rubach,K. C. Willborn, and D. Harder, “Two-dimensional chamber arrays forIMRT plan verification,” Med. Phys. 33, 1005–1015 (2006).

10S. Li, A. Rashid, S. He, and D. Djajaputra, “A new approach in dose mea-surement and error analysis for narrow photon beams (beamlets) shapedby different multileaf collimators using a small detector,” Med. Phys. 31,2020–2032 (2004).

11G. J. Kutcher, L. Coia, M. Gillin, W. F. Hanson, S. Leibel, R. J. Mor-ton, J. R. Palta, J. A. Purdy, L. E. Reinstein, G. K. Svenson, M. Weller,and L. Wingfield, “Comprehensive QA for radiation oncology: Report ofAAPM Radiation Therapy Committee Task Group 40,” Med. Phys. 21,581–618 (1994).

12B. Poppe, A. Blechschmidt, A. Djouguela, K. C. Willborn, A. Ruhmann,and D. Harder, “Spatial resolution of 2D ion chamber arrays for IMRT doseverification: Single detector size and sampling step width,” Phys. Med.Biol. 52, 2921–2935 (2007).

13B. Fraass, K. Doppke, M. Hunt, G. Kutcher, G. Starkschall, R. Stern, andJ. V. Dyke, “American Association of Physicists in Medicine RadiationTherapy Committee Task Group 53: Quality assurance for clinical radio-therapy treatment planning,” Med. Phys. 25, 1773–1829 (1998).

14D. A. Low, W. B. Harms, S. Mutic, and J. A. Purdy, “A technique forthe quantitative evaluation of dose distributions,” Med. Phys. 25, 656–661(1998).

15J. S. Tsai, D. E. Wazer, M. N. Ling, J. K. Wu, M. Fagundes, T. DiPetrillo,B. Kramer, M. Koistinen, and M. J. Engler, “Dosimetric verification ofthe dynamic intensity modulated radiation therapy of 92 patients,” Int. J.Radiat. Oncol., Biol., Phys. 40, 1213–1230 (1998).

16D. Boehmer, J. Bohsung, I. Eichwurzel, A. Moys, and V. Budach, “Clini-cal and physical quality assurance for intensity modulated radiotherapy ofprostate cancer,” Radiat. Oncol. 71, 319–325 (2004).

17E. Spezi, A. L. Angelini, F. Romani, and A. Ferri, “Characterization of a2D ion chamber array for the verification of radiotherapy treatments,” Phys.Med. Biol. 50, 3361–3373 (2005).

18D. M. Gonzalez-Castano, F. Gomez, L. Brualla, J. V. Rosello, D. Planes,M. Sanchez, and M. Pombar, “A liquid-filled ionization chamber for highprecision relative dosimetry,” Phys. Med. 27, 89–96 (2011).

19J. Pardo, J. V. Rosello, F. Sanchez-Doblado, and F. Gomez, “Verificationof intensity modulated profiles using a pixel segmented liquid-filled lineararray,” Phys. Med. Biol. 51, N211–N219 (2006).

Medical Physics, Vol. 41, No. 5, May 2014

051704-14 Markovic et al.: 2D ion chamber detector array for plan verification 051704-14

20J. Pena, L. Franco, F. Gomez, A. Iglesias, J. Pardo, and A. Pazos, “Liquid-filled ionization chamber temperature dependence,” Nucl. Instrum. Meth-ods Phys. Res. A 8, 560–584 (2006).

21J. Pardo, L. Franco, F. Gómez, A. Iglesias, R. Lobato, J. Mosquera, A. Pa-zos, J. Pena, M. Pombar, A. Rodríguez, and J. Sendón, “Free ion yieldobserved in liquid isooctane irradiated by gamma rays: Comparison withthe Onsager theory,” Phys. Med. Biol. 49, 1905–1914 (2004).

22A. V. Esch, C. Clermont, M. Devillers, M. Iori, and D. P. Huyskens,“On-line quality assurance of rotational radiotherapy treatment delivery bymeans of a 2D ion chamber array and the Octavius phantom,” Med. Phys.34, 3825–3837 (2007).

23J. W. Boag and J. Currant, “Current collection and ionic recombination insmall cylindrical ionization chambers exposed to pulsed radiation,” Br. J.Appl. Phys. 3, 471–478 (1980).

24P. R. Almond, “Use of Voctoreen 500 electrometer to determine ionizationchamber collection efficiencies,” Med. Phys. 8, 901–904 (1981).

25E. Chung, S. Davis, and J. Seuntjens, “Experimental analysis of general ionrecombination in a liquid-filled ionization chamber in high-energy photonbeams,” Med. Phys. 40, 062104 (7pp.) (2013).

26J. Pardo-Montero and F. Gomez, “Determining charge collection efficiencyin parallel-plate liquid ionization chambers,” Phys. Med. Biol. 54, 3677–3689 (2009).

27H. Tölli, R. Sjögren, and M. Wendelsten, “A two-dose-rate method forgeneral recombination correction for liquid ionization chambers in pulsedbeams,” Phys. Med. Biol. 55, 4247–4260 (2010).

28J. W. Boag, “Ionization measurements at very high intensities,” Br. J. Ra-diol. 23, 601–611 (1950).

29C. Martens, C. De Wagter, and W. De Neve, “The value of the LA48 lin-ear ion chamber array for characterization of intensity-modulated beams,”Phys. Med. Biol. 46, 1131–1148 (2001).

30B. Poppe, T. S. Stelljes, H. K. Looe, N. Chofor, D. Harder, and K. Willborn,“Performance parameters of a liquid filled ionization chamber array,” Med.Phys. 40(8), 082106 (14pp.) (2013).

31J. M. Lydon, “Photon dose calculations in homogeneous media for a treat-ment planning system using a collapsed cone superposition convolutionalgorithm,” Phys. Med. Biol. 43, 1813–1822 (1998).

32H. K. Looe, D. Harder, and B. Poppe, “Experimental determina-tion of the effective point of measurement for various detectors usedin photon and electron dosimetry,” Phys. Med. Biol. 56, 4267–4290(2011).

33F. Haryanto, M. Fippel, W. Laub, O. Dohm, and F. Nüsslin, “Inves-tigation of photon beam output factors for conformal therapy: MonteCarlo simulations and measurements,” Phys. Med. Biol. 47, N133–N143(2002).

34W. U. Laub and T. Wong, “The volume effect of detectors in the dosimetryof small fields used in IMRT,” Med. Phys. 30, 341–347 (2003).

35E. E. Klein, J. Hanley, J. Bayouth, F. F. Yin, W. Simon, S. Dresser, C. Ser-ago, F. Aguirre, L. Ma, B. Arjomandy, C. Liu, C. Sandin, and T. Holmes,“Task Group 142 Report: Quality assurance of medical accelerators,” Med.Phys. 36(9), 4197–4212 (2009).

Medical Physics, Vol. 41, No. 5, May 2014