Size and positioning reproducibility of an 192 Ir brachytherapy stepping sourceA. Berndt, D. W. Rickey, S. Rathee, and J. Bews Citation: Medical Physics 27, 129 (2000); doi: 10.1118/1.598874 View online: http://dx.doi.org/10.1118/1.598874 View Table of Contents: http://scitation.aip.org/content/aapm/journal/medphys/27/1?ver=pdfcov Published by the American Association of Physicists in Medicine Articles you may be interested in Beta versus gamma dosimetry close to Ir-192 brachytherapy sources Med. Phys. 28, 1875 (2001); 10.1118/1.1395038 Monte Carlo dosimetry of a new 192 Ir high dose rate brachytherapy source Med. Phys. 27, 2521 (2000); 10.1118/1.1315316 A Monte Carlo investigation of the dosimetric characteristics of the VariSource 192 Ir high dose ratebrachytherapy source Med. Phys. 26, 1498 (1999); 10.1118/1.598645 Experimental determination of dosimetry functions of Ir-192 sources Med. Phys. 25, 2279 (1998); 10.1118/1.598457 Monte Carlo and TLD dosimetry of an 192 Ir high dose-rate brachytherapy source Med. Phys. 25, 1975 (1998); 10.1118/1.598371

Size and positioning reproducibility of an 192Ir brachytherapystepping source

A. Berndt, D. W. Rickey, S. Rathee, and J. BewsCancercare Manitoba, 100 Olivia Street, Winnipeg, Manitoba, R3E 0V9, Canada andDepartment of Physics, University of Manitoba

~Received 7 May 1999; accepted for publication 12 October 1999!

In this paper we describe techniques for measuring the dimensions and position reproducibility ofan192Ir brachytherapy stepping source. Measurements were carried out using a 0.253103152 mm3

collimator placed in front of a detector of our own design. The brachytherapy source was translatedpast the collimator in 0.025 mm increments using a stepper motor. The source was found to be 3.58mm long and 0.69 mm wide, which is in good agreement with the manufacturer’s values of 3.530.6mm2. The source position was reproducible to within 0.12 mm. ©2000 American Association ofPhysicists in Medicine.@S0094-2405~00!01101-9#

Key words: brachytherapy, source dimensions

In stepping source brachytherapy, it is essential that thesource be positioned reproducibly. In addition, the sourcelength must be accurately known for Sievert integral dosedistribution calculations. In this paper, we measure thesource position reproducibility of the192Ir source of a highdose rate~HDR! brachytherapy unit~Microselectron, Nucle-tron Oldeft Corporation!. We also measure the192Ir sourcedimensions.

For all measurements, the source was collimated usingtwo 243723152 mm3 lead blocks, one of which had a 0.253103152 mm3 slit milled into it, as shown in Fig. 1. Theg-rays were detected using a 0.27530.831.0 cm3 CdWO4

scintillating crystal mounted on a photodiode~S5668-02Hamamatsu!. The current produced by the photodiode wasintegrated, amplified, and digitized to a precision of 16 bits,using electronics of our own design. An integration time of0.098 s was used for each measurement of the detector volt-age. The detector voltage, due to dark current~approximately20.24 V!, was subtracted. Data were acquired by sendingthe192Ir source to a predetermined position in a flexible cath-eter, which was attached to a linear stage~Series A2500Unislide, Velmex, Inc.!. The source was then translated pastthe stationary collimator in increments of 0.025 mm using acomputer-controlled stepping motor. Noise was reduced byaveraging between 40 and 100 measurements. For plots re-quiring a larger number of data points, a smaller number ofmeasurements was averaged, as the source could only remainat one position for a maximum of 999.9 s.

The reproducibility of the stage position was first charac-terized by translating the leading edge of the source past thecollimator slit five times without retracting the source intothe treatment unit. As the leading edge of the source is welldefined, any positioning inaccuracies will be readily appar-ent. Detector output voltage measurements were obtained at27 different stage positions per trial. This series of measure-ments was repeated five times with the translation stage re-turned to its home position between trials. To remove back-lash in the translation stage when homed, it was translatedbeyond the home position, and then returned. The reproduc-

ibility measurements, consisting of five trials each, were re-peated three times, with the source being retracted into thebrachytherapy unit after each of the three measurement se-ries. The source retraction after five trials was necessitatedby the restriction on the length of time the source couldremain at one position, namely 999.9 s. The data from thethird measurement series are plotted in Fig. 2.

These data were analyzed by fitting a straight line to therising portion of the curve of each of the 15 trials. Thisserved to average the noise in the detector output. The posi-tion at which the fitted line crossed 0.05 V~approximatelyhalf-way up the curves in Fig. 2! was found for each trial.The standard deviations of this position for each measure-ment series were calculated and are summarized in Table I.As can be seen, the translation stage position is very repro-ducible with a maximum imprecision of 0.003 mm.

Source position reproducibility was measured using atechnique similar to that described above, except that thesource was retracted between trials. Three series of reproduc-ibility measurements were performed, as shown in Table II.The first two series of reproducibility measurements used the‘‘interrupt’’ and ‘‘resume’’ features to retract the source andreturn it to the predetermined location. As this did not con-stitute reprogramming a treatment, the source could only ir-radiate one position for a maximum of 999.9 s, thereby lim-iting the number of trials~see Table II! that could constitutea measurement series. For the third series of reproducibilitymeasurements, each trial was initiated with a new treatmentprogram that returned the source to the predetermined posi-tion. The number of trials restriction therefore did not existfor this measurement series. The detector voltages for thethird source position reproducibility measurement series areplotted in Fig. 3 as a function of stage position. The result ofall three reproducibility measurement series are summarizedin Table II. It can be seen that the error introduced by thetranslation stage is negligible compared to the source posi-tion reproducibility error. Thus the standard deviations listedin Table II represent the source positioning error.

The length of the source was measured by translating the

129 129Med. Phys. 27 „1…, January 2000 0094-2405/2000/27 „1…/129/3/$17.00 © 2000 Am. Assoc. Phys. Med.

source lengthwise past the collimator slit. The source widthwas measured by translating the source widthwise past thecollimator slit.

Figure 4~a! shows the normalized detector output voltageplotted as a function of stage position for the source length

FIG. 1. Schematic showing the detector, the source, and the lead blocks usedto collimate the source. The dimensions of the slit are 0.253103152 mm3.The detector and the collimator are held stationary, while the source istranslated past the slit.

FIG. 2. Detector output voltage plotted as a function of translation stageposition verifying that the translation stage returns reproducibly to a prede-termined home position. The source was not retracted into the treatment unitbetween trials. The vertical offset is due to scattered radiation reaching thedetector. The error in the measurements is smaller than the plotted points.

FIG. 3. Detector output voltage plotted as a function of translation stageposition. The source was retracted and extended between trials. The verticaloffset is due to scattered radiation reaching the detector.

FIG. 4. Plot of the normalized detector output voltage as a function oftranslation stage position for the~a! source length and~b! source widthmeasurements. The vertical offset due to scattered radiation reaching thedetector has been subtracted to allow comparison of the measured data~points! with the calculated convolutions~lines!. The error in the measure-ments is smaller than the plotted points.

TABLE I. Translation stage reproducibility measurements.

Measurementseries

No. oftrials

Standard deviation instage position~mm!

1 5 0.00232 5 0.000 913 5 0.0031

TABLE II. Source position reproducibility measurements.

Measurementseries

No. oftrials

Standard deviation insource position~mm!

Source retractionmethod

1 6 0.12 interrupt and resume2 7 0.030 interrupt and resume3 15 0.097 new treatment

130 Berndt et al. : 192Ir source size 130

Medical Physics, Vol. 27, No. 1, January 2000

measurement. Taking the full-width at half-maximum~FWHM! to be the source dimension, the length of thesource is 3.58 mm. This agrees well with the length of 3.5mm specified by the manufacturer. The effect of the collima-tor slit width is inherently cancelled out in the FWHM mea-surement, assuming that both the slit and the source can bemodeled as rect functions. The solid line in Fig. 4~a! is theconvolution of two rect functions of width 0.25 and 3.58mm, representing the slit and the source, respectively. Theexcellent agreement between the measured data and the con-volved rect functions indicates that the rect function model-ing is valid.

Similarly, the detector output for the source width mea-surements is plotted in Fig. 4~b!. In this case, the effect of thecollimator slit width is not inherently cancelled out in theFWHM measurement and was empirically determined byconvolving a disc and a rect function, which represent thesource and the slit, respectively. This was accomplished byconvolving disk functions of diameters 0.6 and 0.7 mm witha rect function 0.25 mm in width. The relationship betweenthe FWHM of the convolution (FWHM0.6 mm50.485 mmand FWHM0.7 mm50.566 mm! and the disc diameterd wasfound to be d0.6 mm51.2373FWHM0.6 mm and d0.7 mm

51.2363FWHM0.7 mm. The FWHM of Fig. 4~b! is 0.562mm, which corresponds to a disc diameter, or source width,of 0.69 mm. This value agrees reasonably well with thewidth of 0.6 mm specified by the manufacturer. Determining

the source width by deconvolving the measured data with arect function representing the slit is inherently difficult, andwas therefore not used.

The convolution arguments used in this paper to extractsource dimensions from measured data assume that the de-tector does not record photons prior to the leading edge ofthe 192Ir source reaching the collimator slit. Transmissionthrough the edges of the collimator, in addition to the angularresponse of the collimator, are two effects that weaken thisassumption. Although the length of the collimator~152 mm!minimized these effects, neither could be completely elimi-nated, and as such constitute the main sources of error inthese measurements.

The techniques described here can be used to accuratelymeasure the position reproducibility and the dimensions ofan 192Ir stepping source. The imprecision in the source posi-tion was found to be less than 0.12 mm. The source lengthand width were found to be 2% and 15% larger than thosespecified by the manufacturer. None of these errors are ofsignificance in terms of the calculation of isodose distribu-tions for clinical treatments.

The authors are grateful for the financial support of theManitoba Health Research Council. We would also like tothank the Medical Devices and Nuclear Electronics Sectionfor their expert assistance.

131 Berndt et al. : 192Ir source size 131

Medical Physics, Vol. 27, No. 1, January 2000