Total reference air kerma: To what extent can it predict intracavitary volume enclosed by isodose surfaces during multiple high-dose rate brachytherapy?

Brachytherapy 2 (2003) 91–97

Total reference air kerma: To what extent can it predict intracavitary volume enclosed by isodose surfaces during multiple high-dose rate brachytherapy? Niloy R. Datta1,*, Koilpillai J. Maria Das1, Rimpa Basu1, Uttam Singh2 1

Department of Radiotherapy, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India Department of Biostatistics, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India

2

Abstract

Background: The International Commission on Radiation Units and Measurements (ICRU) report 38 recommends reporting of total reference air kerma (TRAK) and reference ICRU isodose volumes during intracavitary brachytherapy (ICBT) in cancer of the cervix. The present study attempts to estimate the volumes enclosed by isodose surfaces from TRAK and evaluate its utility to represent doses to organs of interest. Material and Methods: Volumes encompassed by isodose surfaces of 3 Gy, 6 Gy, 9 Gy, and 12 Gy were obtained for 90 high-dose rate (HDR) ICBT procedures. These were used to derive a relation between isodose volumes, TRAK/dose (K/D), and rectal and bladder doses. Results: Actual volumes (V) encompassed by isodose surfaces were reflected as a quadratic function of K/D (r2 ⫽ 0.998) and the expression, V ⫽ ⫺23.09 ⫹ 1295.99(K/D) ⫹ 5661.65(K/D)2 gave the best estimates for various volumes. No correlation was observed between TRAK and bladder (r2 ⫽ 0.086) or rectal doses (r2 ⫽ 0.082). Conclusions: Estimates of volumes encompassed by different dose levels from TRAK could be derived with reasonable certainty. However, TRAK fails to correlate with rectal and bladder doses. 쑖 2003 American Brachytherapy Society. All rights reserved.

Keywords:

Cancer cervix; Intracavitary brachytherapy; ICRU report 38; TRAK; High-dose rate; Rectal dose; Bladder dose

Introduction The International Commission on Radiation Units and Measurements (ICRU) in its report, ICRU report 38, recommended that total reference air kerma (TRAK) as a part of the dose and volume specification for reporting intracavitary brachytherapy (ICBT) in carcinoma of the cervix (1). This was in addition to doses at specified reference points related to the organs at risk and also the reference volume that would be encompassed by a specified isodose surface. This volume was considered to be a combined teletherapy and brachytherapy absorbed dose level of 60 Gy, pertaining to classical low dose rate (LDR) brachytherapy. It was, however, left to the radiation therapist for medium and high-dose rate (HDR) to “indicate the dose level which he believes to Received 23 January 2003; received in revised form 7 May 2003; accepted 13 May 2003. * Corresponding author. Department of Radiotherapy, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Rae Barelli Road, Lucknow 226014, India. Tel.: ⫹91-522-2668004 to 8, 2668700, Ext. 2449; fax: ⫹91522-2668017. E-mail address: [email protected] (N.R. Datta).

be equivalent to 60 Gy delivered at classical low dose rate.” Thus, the equivalence of doses for various dose rates has been a subject of much discussion and a major factor for ambiguity among radiation oncologists, which has lead to its infrequent reporting. TRAK, even though is a well-quantified parameter without much ambiguity, has also found relatively less penetration in the literature pertaining especially to HDR ICBT. Po¨tter et al. (2) in a recent survey observed that TRAK was recorded only in 14% of HDR in clinical practice and 0% in literature pertaining to HDR ICBT. The corresponding figures for LDR were 43% and 10%, respectively. Similarly, the 60 Gy ICRU reference volume was recorded in 18% of HDR users compared with 51% of LDR applications in clinical practice, whereas in the literature the respective figures were 0% and 10%. Nevertheless, attempts have been made by several authors to derive a mathematical expression for various volumes at different dose levels with TRAK as one of the variables (3–5). The reference volume as proposed by ICRU Report 38 is a 60 Gy combined isodose surface for both teletherapy and ICBT. However, external radiotherapy contributions

1538-4721/03/$ – see front matter 쑖 2003 American Brachytherapy Society. All rights reserved. doi:10.1016/S 1 53 8 - 47 2 1 ( 03 ) 0 00 9 6 - 5

92

N.R. Datta et al. / Brachytherapy 2 (2003) 91–97

may vary depending on the treatment protocols used at different centers for various stages. Moreover, the use of different dose rate systems and the problem of their radiobiological equivalence might further complicate the choice of the specific isodose level that would be equivalent to a LDR system. Further, volumes derived merely as products of the height, width and thickness of ICBT reference isodose have found to differ significantly from those computed from the treatment planning systems (TPS) (5). Therefore, it would be of interest to evaluate the physical volumes as computed from TPS for different isodoses solely for ICBT. Because TRAK is a quantity that can be derived during ICBT treatment planning, it might be worthwhile to examine the appropriateness of using the various expressions reported by others to estimate the physical volumes for various dose levels. This study attempts to quantify the volume estimated from the expressions proposed by other authors (3–5) and compare with actual and estimated volumes obtained from the data set of ICBT applications carried out during this study. The correlation of TRAK with bladder and rectal doses is also examined.

Methods Patient population A total of 90 ICBT applications were evaluated for this study. These were carried out on 30 patients who had carcinoma of the cervix who were considered for ICBT after completion of teletherapy. All patients were treated with a four-field box technique to a dose of 50 Gy delivered in 25 fractions over a period of 5 weeks by telecobalt or 6 MV/10 MV photons. Following 2 weeks of teletherapy, all cases received three ICBT applications of 6 Gy each at weekly intervals either using a Standard or Rotterdam intracavitary applicator which was used with a remote HDR brachytherapy unit (HDR Microselectron, Nucletron B.V., Vandeenlaar, The Netherlands). The unit houses a 192Ir source of a maximum 10 Ci activity. The choice of the applicator was random and based on availability of the applicators, as individual patients had to be treated with the same applicator for all the 3 consecutive sessions of ICBT. Thus, 90 applications were performed in 30 patients using either a Standard (n ⫽ 15 patients, 45 applications) or Rotterdam (n ⫽ 15 patients, 45 applications) applicator. The ICBT dose was prescribed at point A, defined as 2-cm above the cervical os marker and 2-cm perpendicular to the uterine axis along the plane of the uterus (6). Although a dose of 6 Gy was prescribed for each session, treatment was individualized to treat a better geometry, to reduce doses to bladder and rectum, and take care of the variations in the applicator geometry that could occur during each of the three applications (7). Thus, the source loading and dwell times were different in each application, leading to varying TRAK values in three applications even in a given patient.

Calculations of the volumes encompassed by different isodose levels The volumes estimated in this study pertained to the volumes for respective isodose during each of the three ICBT insertions. Although a dose of 6 Gy at point A was used for each application, the volumes for various dose levels ⫺3 Gy, 6 Gy, 9 Gy, and 12 Gy were calculated directly from the treatment planning system (Plato treatment planning system with brachytherapy software version 13.7.2 and evaluation version 2.6; Nucletron B.V.). A grid of 15 cm ⫻15 cm ⫻ 15 cm in X, Y, and Z axes was used and dose volume histograms were generated using random sampling method and optimized for 10,000 points. TRAK values for each application were recorded for each of the 90 ICBT applications. These were used to estimate the volume, V, for each of the corresponding dose levels using the various expressions given by the following authors: Wilkinson and Ramachandran (3), V ⫽ 160 (K/D)3/2 ml; Eisbruch et al. (4), V ⫽ [104.8 ⫺ 8.103 (M/D) ⫹ 0.437 (M/ D)2] (M/D)1.635 ml; and Deshpande et al. (5), V ⫽ 156 (K/D)1.55 ml where, K ⫽ TRAK in mGy, D ⫽ dose in Gy, and M/D ⫽ mgRaEq-hour/cGy. The conversion factor from M/D to K/ D was based on the relation, 7140 mgRaEq-hour ⫽ 51.62 mGy·m2·h⫺1 (8). Statistical analysis Scatter plots obtained between the ICRU volume and ratio of TRAK and dose levels (K/D) were best fitted using a quadratic fit. Consequently, a stepwise forward linear regression was performed to determine the influence of the determinants, K/D and its square, (K/D)2, on the actual volume, V, obtained from the treatment planning system. The step-wise model retained only those parameters having significance at a probability of ⱕ0.05. The regression coefficient (β) for the various determinants along with standard error and 95% confidence intervals were obtained. The predicted values of isodose volumes using the regression equation were compared with those obtained from the other models (3–5). A paired Student’s t test was carried out to test the significance of differences between the actual and the estimated volumes for each dose levels. Analysis was carried out using the Statistical Package for Social Sciences (SPSS) for windows (Version 10.0; SPSS, Chicago, IL). Results Relation between isodose volume and TRAK at different dose levels Of the 90 ICBT applications in 30 patients, an equal number of procedures were carried out with either the Rotterdam or Standard applicator. In any given patient, all the 3

N.R. Datta et al. / Brachytherapy 2 (2003) 91–97

93

ICBT were carried out with the same applicator. The mean TRAK values for Standard and Rotterdam applicators, respectively, were 0.46 cGy at 1 m (SD, ⫾0.02) and 0.43 cGy at 1 m (SD, ⫾0.03), p ⬍ 0.001, for a dose prescription of 6 Gy at point A. The volumes for various dose levels at 3 Gy, 6 Gy, 9 Gy, and 12 Gy were significantly different for the two applicators (Table 1). However, the relation between isodose volume and K/D were almost identical for both these applicators as evident from Fig. 1. Both these were fitted well using quadratic fit (r2 for both Standard and Rotterdam applicators being 0.998). Thus, it was apparent from the plots that although there could be differences in the TRAK used in these two different applicators, their relations with the respective isodose volume were nearly identical and followed a quadratic function of K/D. Relation between TRAK and doses to bladder and rectum Consequent to the inherent differences in the angulations of the intrauterine tandem, there were considerable differences in the corresponding bladder and rectal doses delivered by these applicators. Average bladder doses were noticeably higher for the Rotterdam applicator (mean ⫾ SD being 654.22 ⫾ 316.72 cGy vs. 489.02 ⫾ 158.83 cGy, p ⫽ 0.002). On the other hand, patients treated with Standard applicator received a greater rectal dose (mean ⫾ SD being 404.12 ⫾ 71.65 cGy vs. 338.09 ⫾ 75.72 cGy, p ⬍ 0.001). However, the bladder and rectal doses failed to show any significant correlation with TRAK as expected from ICRU Report 38 (Figs. 2 and 3). Estimation of the isodose volumes using regression equation Regression analysis performed using actual isodose volumes as a dependent parameter further substantiated the earlier observation that the volumes were related to a quadratic function of K/D (Table 2). The adjusted R2 of the expression was 0.999 and this was used to predict the volume for different dose levels using the TRAK values for each individual application. The set of 360 such predictive values obtained from the regression equation, V ⫽ ⫺23.09 ⫹ 1295.99 (K/D) ⫹ 5661.65 (K/D)2 (90 for each of the dose

Fig. 1. Scatter plot between the isodose volume and TRAK/dose for all patients and those with brachytherapy performed with either Rotterdam or Standard Microselectron applicators. The fitted line represents a quadratic fit. TRAK ⫽ total reference air kerma.

levels of 3 Gy, 6 Gy, 9 Gy, and 12 Gy) was used for comparison with the actual volumes at each dose level. Comparison of estimates of isodose volumes obtained from different methods The estimated isodose volumes were obtained from each of four methods – namely from the regression equation from 90 insertions of this study, Wilkinson and Ramachandran (3), Eisbruch et al. (4), and Deshpande et al. (5). All these values showed a high degree of correlation between those estimated and the actual values. However, the values estimated from the models proposed by Wilkinson and Ramachandran

Table 1 Isodose volumes for different dose levels for Standard and Rotterdam applicators Isodose volume (mean ⫾ SD) (mL) Dose level

Standard applicator

3 6 9 12

309.64 109.53 56.15 33.49

Gy Gy Gy Gy

⫾ ⫾ ⫾ ⫾

22.01 7.53 3.89 2.66

SD ⫽ standard deviation. a Student’s t test.

Rotterdam applicator 276.68 98.91 51.84 31.82

⫾ ⫾ ⫾ ⫾

28.72 9.10 4.44 2.74

p valuea ⬍0.001 ⬍0.001 ⬍0.001 0.004 Fig. 2. Scatter plot between rectal dose and TRAK during the 90 intracavitary brachytherapy applications. TRAK ⫽ total reference air kerma.

94

N.R. Datta et al. / Brachytherapy 2 (2003) 91–97

with the model proposed by Eisbruch et al. (4), and 2.2% to 8.2% (skewness, ⫺0.536; kurtosis, 0.769) with the model derived by Deshpande et al. (5). Using the quadratic model with K/D, obtained from the present data set of 90 applications, differences in 95% of the estimates between the actual and predicted volumes were within ±5.32% (skewness, 1.092; kurtosis, 3.102).

Discussion

Fig. 3. Scatter plot between bladder dose and TRAK during the 90 intracavitary brachytherapy applications. TRAK ⫽ total reference air kerma.

(3), Eisbruch et al. (4), and Deshpande et al. (5) were significantly different from the true volumes for each dose level obtained from the treatment planning system. A comparative estimate of these volumes along with the observed volumes is presented in Table 2. Paired sample t test to test the significance of difference between the observed and the estimated values by each of these methods indicate a highly significant difference between the observed and the estimated values with the above three methods proposed by others for each of the four dose levels (p ⬍ 0.001) (3–5). However, isodose volumes estimated from the quadratic relation of K/D were not significantly different from the actual values. The normal distribution curves based on the histograms of the percentage differences between actual and the estimated isodose volume further depicted the differences between actual and estimated volumes (Fig. 4) with different models studied. Thus, differences in 95% of the estimates between the actual and predicted isodose volumes by different models ranged between ⫺7.90% and 3.90% (skewness, ⫺0.891; kurtosis, 0.546) for Wilkinson and Ramachandran’s (3), ⫺8.59% and 0.59% (skewness, ⫺0.374; kurtosis, 0.910), Table 2 Coefficients of the regression analysis of the actual isodose volume 95% C.I. of β Determinants

Coefficient (β)

S.E.

Lower

Upper

p Value

Constant K/D (K/D)2

⫺23.09 1295.99 5661.65

1.26 32.87 168.87

⫺25.57 1231.34 5329.55

⫺20.61 1360.64 5993.75

⬍0.001 ⬍0.001 ⬍0.001

Adjusted R2 ⫽ 0.999. C.I. ⫽ confidence interval; K/D ⫽ TRAK / dose; S.E. ⫽ standard error; TRAK ⫽ total reference air kerma.

Because intracavitary brachytherapy is an important part of radiation therapy in carcinoma of the cervix, the ICRU in 1985 provided recommendations for a uniform system of dose and volume specification for reporting. However, as evident from the recent survey from Po¨tter et al. (2), radiation oncologists have been rather reluctant to follow the guidelines, both in their clinical practice and in publications. Some of the concepts, like TRAK, could have physical implications from radiation protection, while clinical implications of specification of ICRU reference volumes and dose specifications at organs at risk continue to remain unresolved (2). The relations between TRAK and the respective reference ICRU volumes at different dose levels have been evaluated by some authors and each has tried to come up with a different expression based on the intracavitary system practiced at a given center (3–5, 9). Wilkinson and Ramachandran (3) have used the Manchester method of intracavitary brachytherapy and TRAK was calculated and tabulated as a function of the prescription of the standard Manchester source configurations. They predicted that for traditional Manchester geometries, depending on the source distributions, there could be a random error of ⫾3% in estimates of reference isodose volumes. Their expression was quite simple, with V, the volume enclosed by the isodose surface, being estimated as 160 (K/D)3.2 ml, where K is the TRAK in mGy, and D is the isodose surface in Gy. Eisbruch et al. (4) analyzed over 200 implants from more than 100 patients and found the same relationship characterized ICBT dose–volume histograms for both Fletcher-Suit geometries, tandem, and cylindrical implants. They used curve-fitting techniques and proposed that the volume encompassed by each isodose level could be predicted by a modified power-law function of the mgRaEq-hour/dose ratio. Their relation of volume estimates was therefore, V ⫽ [104.8⫺8.103 (M/D) ⫹ 0.437 (M/D)2](M/D)1.635 ml, M/D being mgRaEq-hour/cGy. They observed that volume estimates with their model were accurate within ⫾10% in 95% of the cases when M/D was 0.8. Deshpande et al. (5) used the Fletcher-Suit applicator with a Selectron-LDR remote afterloading device to derive the relation between ICRU reference volume, TRAK and dose levels. Of all patients treated, they selected 12 cases on the basis of satisfactory geometry representing various

N.R. Datta et al. / Brachytherapy 2 (2003) 91–97

95

Fig. 4. Histogram with the normal distribution curve for the percent difference between the estimated isodose volume and the actual isodose volumes with each of the four different methods based on (a) regression analysis from the present study, (b) Wilkinson and Ramachandran (3), (c) Eisbruch et al. (4), and (d) Deshpande et al. (5).

combinations of tandem and colpostat to derive the above relation. Their expression was very similar to that of Wilkinson and Ramachandran (3), with V ⫽ 156 (K/D)1.55 ml. They observed that the percentage variation between actual and the estimated 25 Gy ICRU reference volume ranged from ⫺2% to ⫺9%. Ragnhult et al. (9) used data from 200 cervical cancer patients being treated on a HDR unit using 60Co sources. They discussed the correlation between kerma and volume for the single-source, double-source, and triple-source configurations. They estimated an uncertainty of ⫾5% in the determination of treatment volumes based on the treatment methods used in their clinics. The isodose volumes as estimated in this study pertain only to the ICBT volume. This is in contrast to the ICRU report 38 recommendations that the absorbed dose level of 60 Gy (including dose delivered by teletherapy) should be considered, since this level is considered as appropriate reference level for classical low dose-rate therapy. Thus, in situations where teletherapy is used, the isodose level should be the difference between 60 Gy and the dose delivered by teletherapy. Because in most cases a dose of approximately 50 Gy is delivered to the pelvis for stages IIB and beyond, the ICBT portion would get limited to a mere

10 Gy or its equivalence. Moreover, in ICRU report 38, the equivalence of 10 Gy LDR for a multifractionated HDR, which is usually delivered at doses of 5 Gy to 8 Gy per session, is left to the choice of the radiotherapist. Thus, in this study, instead of the reference isodose volumes of 60 Gy, an isodose volume encompassed by 3 Gy to 12 Gy has been evaluated, as 6 Gy was the prescribed dose at point A for each ICBT session. The present observations indicated that although TRAK was statistically different for two different applicators, the relation between isodose volumes and K/D was very well represented as a quadratic function (Fig. 1). Thus, isodose volumes at different dose levels were estimated using the expression, V ⫽ ⫺23.09 ⫹ 1295.99 (K/D) ⫹ 5661.65 (K/D)2 ml (Table 2). There were significant differences in the estimated and actual ICRU reference volumes derived from expressions proposed by the three studies (Table 3; (3–5)) comparing the predicted and actual values of respective isodose volumes. Although there were nonsignificant differences in the predicted and actual volume estimates for 3 Gy, 6 Gy, and 12 Gy using the quadratic model, there were still significant differences between estimated and actual 9 Gy volumes (p ⫽ 0.021). This could be because 9 Gy volumes fell in

96

N.R. Datta et al. / Brachytherapy 2 (2003) 91–97

Table 3 Comparative isodose volumes at different dose levels as actually observed, and those estimated, by regression analysis from present study, and other methods proposed by Wilkinson and Ramachandran (3), Eisbruch et al. (4), and Deshpande et al. (5) Observed volume (ml)

Dose level

Number of Patients Mean ⫾ SD

3 6 9 12

90 90 90 90

Gy Gy Gy Gy

293.16 104.22 54.00 32.65

⫾ ⫾ ⫾ ⫾

Estimated volume (ml)*

30.36 292.98 ⫾ 30.50 9.88 103.84 ⫾ 10.93 4.68 54.92 ⫾ 6.33 2.81 32.59 ⫾ 4.39

Estimated p Value* volume (ml)†

0.794 0.144 0.021 0.845

288.26 101.91 55.47 36.03

⫾ ⫾ ⫾ ⫾

29.93 10.58 5.76 3.74

Estimated p Value† volume (ml)‡

Estimated p Value‡ volume (ml)§

⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001

⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001

305.70 105.62 55.77 35.28

⫾ ⫾ ⫾ ⫾

31.71 11.42 6.12 3.90

286.68 97.90 52.22 33.43

⫾ ⫾ ⫾ ⫾

p Value§

30.77 ⬍0.001 10.51 ⬍0.001 5.60 ⬍0.001 3.59 0.001

The p values indicate the paired sample Student’s t test between the each pairs of observed isodose volume and the volumes estimated by regression analysis from the *present study, methods proposed by †Wilkinson et al. (3), ‡Eisbruch et al. (4) and §Deshpande et al. (5).

the region where the relation between isodose volume and K/D changed from linear to quadratic (Fig. 1). Because estimates of ICBT volume are dependent on dose level, the extent and pattern of variation in the ratio of estimated and actual specified isodose volume would be of interest (Fig. 5). In case of absolute agreement between the observed and estimated doses, the ratio between them is 1.0. As evident from Fig. 5, the curve fitted on the individual data points (not shown in this graph for the sake of clarity), for these ratios for dose levels 3–12 Gy from the present study, closely followed 1.0 and was nearly horizontal. On the contrary, a curve fitted from estimates of the ratio of estimated and actual isodose volume from all the other three models showed a “U” shaped pattern which was either below or above the ratio of 1.0 at low dose levels but curved upwards and crossed the unity level near 9 Gy. This highlights that the accuracy of estimates of these volumes depends on the specific isodose volume. All these factors could

Fig. 5. Ratios of the estimated and actual isodose volumes for dose levels 3 Gy, 6 Gy, 9 Gy, and 12 Gy. The estimated volumes were obtained either from the regression analysis from the present study or expressions proposed by Wilkinson and Ramachandran (3), Eisbruch et al. (4), and Deshpande et al. (5). The individual data points have not been depicted for the sake of clarity of the figure and the fitting is based on quadratic fit.

lead to a high level of uncertainty in deriving these volumes using TRAK. The polynomial equation obtained by fitting the various volumes to K/D in this study is relatively easily understood than those from the different power-law relations proposed by the other authors. The estimates obtained from this polynomial relation although provide a better prediction of the actual volumes at different isodose levels, it should not be construed as a preferred equation over others that could be universally applied in any given situation. The estimates obtained from the other relations indicate subtle differences (Table 3) from the actual volumes. Even though this may be a physical quantity of statistical significance, but its ultimate impact on the various treatment related outcome parameters is questionable. Moreover, it has been well documented that the volumes obtained from TRAK values would depend on a number of factors, such as time distribution of dose, relative source distribution, type of applicator, ovoid shielding, and applicator geometry. Thus, it would be inappropriate to favor any particular expression of volumes of different isodose levels obtained on TRAK that could be universally applied to any situation. Moreover, with the present day treatment planning systems, volumes enclosed by any isodose levels can be computed using algorithms for dose–volume histogram computations. Thus, attempts to estimate isodose volumes using TRAK values alone should be discouraged. The other interest in a detailed evaluation of the TRAK is to look for its correlation with the clinical outcomes, both in terms of disease control and doses to adjoining critical organs. In this study, TRAK has failed to show any correlation with the rectal and bladder doses (Figs. 2 and 3). This could be because the bladder and rectal doses estimated at the respective ICRU report 38 recommended points depend critically on their proximity to the active source position, which certainly would have a much greater effect than TRAK. This is in line with the ICRU 38 caution that TRAK cannot be used to estimate doses at specific points close to the sources. Similarly, Grigsby et al. (10) in a large retrospective analysis of 1253 patients have failed to show any positive correlation of TRAK and 60 to 160 Gy reference volumes

N.R. Datta et al. / Brachytherapy 2 (2003) 91–97

to primary control. The treatment consisted of teletherapy and brachytherapy in most cases. On the other hand, in a recent abstract, Hunter et al. (11) observed a significant correlation of TRAK and morbidity in 249 patients with small volume early cancer of the cervix that were treated with ICBT alone. Thus, the utility of TRAK as an important indicator of outcomes of clinical relevance remains contested. It is highly difficult to comment on the utility of reference isodose volume in view of its limited reporting in the literature, especially with HDR ICBT (2). However, an area of concern for reference ICRU isodose level as highlighted previously consists of the equivalence of 60 Gy especially when using dose rates other than LDR. Moreover, it has also been demonstrated that the ICRU reference volume pertaining to ICBT is dependent on components of the applicator and their spatial position in the pelvis (7, 12, 13). This would be significant with multiple HDR applications as these would introduce inherent variation in the geometry and position of the applicators. It would introduce areas within the cervix and the adjoining structures, where some regions would get equal contributions from all the multiple ICBT applications, while there would be other regions which would fall in the varied region of intersections of different isodoses during the multiple applications (7). This factor is totally dependent on the spatial position of the applicator and would lead to considerable inhomogeneity in the cumulative doses to these structures with multiple HDR insertion. Thus, even though the TRAK might be identical for each of these multiple ICBT applications, the summated reference isodose delivered to tumor and organs of interest would vary. This adds to the complexities in relating the clinical outcomes to TRAK and ICRU reference volume.

Conclusions Thus, to conclude, TRAK and volumetric characteristics of isodose surfaces appear to have limited utility in dose specification for ICBT application in carcinoma of the cervix. Even though TRAK and its obsolete equivalent mg-hour had first been demonstrated by Wilkinson and Ramachandran [3] and Eisbruch et al. (4) to accurately predict the volume of tissue irradiated by a specified brachytherapy dose for LDR brachytherapy, to date the problem of registering cumulative ICBT and teletherapy distributions to underlying anatomy has remained unresolved. No doubt, TRAK and mg-hour have served as key prescription parameters in some of the most clinically efficacious treatment systems documented in the literature (14–16). However, even though a mathematical expression might be derived between TRAK and different isodose volumes, it may be difficult to generalize these for routine use. Further, with widespread availability of brachytherapy treatment planning

97

software, actual ICBT isodose volumes and TRAK can be computed for each application, thus limiting the use of any specific mathematical expression to compute the volumes of specific interest using TRAK. Moreover, the final utility of TRAK as a variable for outcome parameters is debatable, although TRAK could still continue to serve as a useful parameter for radiation protection. References [1] International Commission on Radiation Units and Measurements (ICRU). Dose and volume specification for reporting intracavitary therapy in gynecology (ICRU Report 38). Bethesda, Maryland: ICRU; 1985. [2] Po¨tter R, Limbergen EV, Gerstner N, et al. Survey of the use of the ICRU 38 in recording and reporting cervical cancer brachytherapy. Radiother Oncol 2001;58:11–18. [3] Wilkinson JM, Ramachandran TP. The ICRU recommendations for reporting intracavitary therapy in gynecology and the Manchester method of treating cancer of the cervix uteri. Br J Radiol 1989;62: 362–365. [4] Eisbruch A, Williamson JF, Dickson DR, et al. Estimation of tissue volume irradiated by intracavitary implants. Int J Radiat Oncol Biol Phys 1993;25:733–744. [5] Deshpande DD, Shrivastava SK, Nehru RM, et al. Treatment volume from total reference air kerma (TRAK) in intracavitary applications and its composition with ICRU reference volume. Int J Radiat Oncol Biol Phys 1994;28:499–504. [6] Tod MC, Meredith WJ. Treatment of cancer of the cervix uteri: A revised “Manchester method”. Br J Radiol 1953;26:252–257. [7] Datta NR, Kumar S, Maria Das KJ, et al. Variations of intracavitary applicator geometry during multiple HDR brachytherapy insertions in carcinoma cervix and its influence on reporting as per ICRU report 38. Radiother Oncol 2001;60:15–24. [8] Williamson JF, Grigsby PW, Perez C, et al. Reply to Dr. Wilkinson. Int J Radiat Oncol Biol Phys 1994;29:217–218. [9] Ragnhult I, Holmberg E, Mattsson S. Relationship between kerma and treatment volume in intracavitary radiation therapy. Acta Oncol 1990;29:307–312. [10] Grigsby PW, Williamson JF, Clifford Chao KS, et al. Cervical tumor control evaluated with ICRU 38 reference volumes and integrated reference air kerma. Radiother Oncol 2001;58:19–23. [11] Hunter R, Wilkinson J, Swindell R. Volume in intracavitary therapy: The first proof that it is clinically significant. Radiother Oncol 2002; 64(Suppl 1):S47–S48. [12] Elhanafy OA, Das RK, Paliwal BR, et al. Anatomic variation of prescription points and treatment volume with fractionated high-doserate gynecological brachytherapy. J Appl Clin Med Phys 2002;2: 114–120. [13] Hoskin PJ, Cook M, Bouscale D, et al. Changes in applicator position with fractionated high dose rate gynaecological brachytherapy. Radiother Oncol 1996;40:59–62. [14] Fletcher GH. Textbook of radiotherapy 3rd ed. Philadelphia: Lea & Febiger, 1980;p. 720–789. [15] Perez C, Kuske RR, Camel HM, et al. Analysis of pelvic tumor control and impact on survival in carcinoma of the uterine cervix treated with bradiation therapy alone. Int J Radiat Oncol Biol Phys 1988;14: 613–621. [16] Barillot I, Horiot JC, Pigneux J, et al. Carcinoma of the intact uterine cervix treated with radiotherapy alone: A French cooperative study update and multivariate analysis of prognostic factors. Int J Radiat Oncol Biol Phys 1997;38:969–978.