Dose escalation in permanent brachytherapy for prostate cancer: dosimetric and biological considerations*

Transcription

1 INSTITUTE OF PHYSICS PUBLISHING Phys. Med. Biol. 48 (2003) PHYSICS IN MEDICINE AND BIOLOGY PII: S (03) Dose escalation in permanent brachytherapy for prostate cancer: dosimetric and biological considerations* XAllen Li 1,JianZWang 1,Robert D Stewart 2 and Steven J DiBiase 1 1 Department of Radiation Oncology, University of Maryland, School of Medicine, 22 South Greene Street, Baltimore, MD , USA 2 School of Health Sciences, Purdue University, West Lafayette, IN , USA ali001@umaryland.edu Received 16 April 2003 Published 12 August 2003 Online at stacks.iop.org/pmb/48/2753 Abstract No prospective dose escalation study for prostate brachytherapy (PB) with permanent implants has been reported. In this work, we have performed a dosimetric and biological analysis to explore the implications of dose escalation in PB using 125 I and 103 Pd implants. The concept of equivalent uniform dose (EUD), proposed originally for external-beam radiotherapy (EBRT), is applied to low dose rate brachytherapy. For a given 125 Ior 103 Pd PB, the EUD for tumour that corresponds to a dose distribution delivered by EBRT is calculated based on the linear quadratic model. The EUD calculation is based on the dose volume histogram (DVH) obtained retrospectively from representative actual patient data. Tumour control probabilities (TCPs) are also determined in order to compare the relative effectiveness of different dose levels. The EUD for normal tissue is computed using the Lyman model. A commercial inverse treatment planning algorithm is used to investigate the feasibility of escalating the dose to prostate with acceptable dose increases in the rectum and urethra. The dosimetric calculation is performed for five representative patients with different prostate sizes. A series of PB dose levels are considered for each patient using 125 I and 103 Pd seeds. It is found that the PB prescribed doses (minimum peripheral dose) that give an equivalent EBRT dose of 64.8, 70.2, 75.6 and 81 Gy with a fraction size of 1.8 Gy are 129, 139, 150 and 161 Gy for 125 I and 103, 112, 122 and 132 Gy for 103 Pd implants, respectively. Estimates of the EUD and TCP for a series of possible prescribed dose levels (e.g., 145, 160, 170 and 180 Gy for 125 I and 125, 135, 145 and 155 for 103 Pd implants) are tabulated. The EUD calculation was found to depend strongly on DVHs and radiobiological parameters. The dosimetric calculations suggest that the dose to prostate can be escalated without a substantial increase in both rectal and urethral dose. For example, increasing the PB prescribed dose from 145 * This work was presented at the 2002 ASTRO annual meeting /03/ $ IOP Publishing Ltd Printed in the UK 2753

2 2754 XALiet al to 180 Gy increases EUD for the rectum by only 3%. Our studies indicate that the dose to urethra can be kept within % of the prescription dose for all the dose levels studied. In conclusion, dose escalation in permanent implant for localized prostate cancer may be advantageous. It is dosimetrically possible to increase dose to prostate without a substantial increase in the dose to the rectum and urethra. Based on the results of our studies, a prospective dose escalation trial for prostate permanent implants has been initiated at our institution. 1. Introduction Transperineal ultrasound guided prostate brachytherapy (PB), using 125 I and 103 Pd permanent implants, has become increasingly popular for the treatment of early stage prostate cancer. The convenience of PB and its limited hospitalization has promoted the growth of this treatment. It has produced encouraging long-term outcome with relatively low morbidity. Kwok et al (2002) have reported a five-year freedom from biochemical failure (FFBF) rate of 85% at a medium follow-up of seven years. Such results are comparable to both surgery and externalbeam radiotherapy (EBRT), as also reported by many other investigators (Martinez et al 2000, D Amico et al 1998, Stokes 2000, Zagars et al 1991, Zagars and von Eschenbach 1993, Partin et al 1993, Fuks et al 1991, Kuban et al 1989, Koprowski et al 1991, Blasko et al 1991, 1993, 1995, Litwin et al 1995, Shrader-Bogen et al 1997). In the last several years, the dose escalation in the radiotherapeutic management of prostate cancer with EBRT has become increasingly important. Accumulated dose escalation studies with EBRT have shown that, for localized prostate cancer, the treatment outcome can be improved with higher doses. Hanks et al (1996) reported on a series of 375 patients treated with conformal EBRT and noted a dose response at 71 Gy with a FFBF rate at two years of 72% at doses less than 71 Gy versus 85% for patients receiving doses 71 Gy. Recent studies by these investigators confirmed that the favourable results with dose escalations remain in alonger term (Hanks et al 1997, 1998). The benefit of higher doses may be particularly evident in patients with poor risk disease. In a randomized trial that compared 70 versus 78 Gy (Pollack et al 2000), the higher dose was associated with a significant improvement in five-year FFBF in men with a pretreatment PSA (prostate specific antigen) >10 ng ml 1 (75 versus 48%) but not in men with a pretreatment PSA <10 ng ml 1. Zelefsky et al (1996) demonstrated that patients receiving doses 75.6 Gy had a 93% chance of achieving nadir PSA 1 bytwo years compared to 80% for those receiving lower doses. In a more recent study, Zelefsky et al (1998) confirmed that 90% of patients receiving 75.6 or 81.0 Gy achieved a PSA nadir of 1.0 ng compared with 76% and 56% for those treated with 70.2 Gy and 64.8 Gy, respectively. They reported that the five-year actuarial FFBF was significantly improved in patients with intermediate and unfavourable prognosis receiving dose 75.6 Gy (21). Furthermore, a recent review by Valicenti et al (2000)ofprostate cancer patients treated on radiation therapy oncology group (RTOG) trials showed an overall survival advantage for patients receiving higher doses of EBRT. As such, the RTOG has recently conducted a multi-institutional dose escalation trial evaluating higher doses of EBRT in a phase I/II study. Prospective data regarding the appropriate dose in PB are not available at present. Based on early work conducted at Memorial Sloan Kettering Cancer Center (MSKCC), an arbitrary dose of 145 Gy (160 Gy pre-tg43 standardization) was established as the appropriate dose for 125 I permanent implants (Hilaris et al 1974). This dose is equivalent to an EBRT dose of

3 Dose escalation for prostate implant Gy (2.0 Gy per fraction). However, we now know that 70 Gy of EBRT is inadequate to effectively cure localized prostate cancer for intermediate- and high-risk patient groups (see below). Consequently, the equivalent PB dose of 145 Gy may also be inadequate for these patient groups. Recent retrospective studies suggest that doses higher than 145 Gy may be beneficial in PB. In a retrospective analysis of 134 patients treated with 125 Iimplant, Stock et al (1998) showed that the four-year FFBF rates for patients with D90 (dose covers 90% of the target volume) in the range of <100 Gy, Gy, Gy, Gy and 160 Gy were 53, 82, 80, 95 and 89%, respectively. Such improvement on outcome may be simply because higher doses overcome higher tumour burden, although it is possible that higher doses lead to better coverages of the target or higher doses reduce dosimetric cold spots in the target. Recently, Stock et al (2000) updated their PB series and revealed that improved biochemical control rates were shown with dose up to 180 Gy with acceptable urinary toxicity. Higher doses above 180 Gy in this retrospective study did not produce higher control rates, but noticeably increased toxicity. Similar to 125 I, PB with 103 Pd implant achieved a high rate of biochemical and clinical control in patients with clinically organ confined disease as reported by Blasko et al (2000). A study performed by Ling et al (1995) showed that the relative biological effectiveness (RBE) of 125 I and 103 Pd is 1.4 and 1.9, respectively, compared to 60 Co irradiation. Based on the RBE of 103 Pd compared to 125 I, current treatment protocols use a prescribed dose of 125 Gy for 103 Pd implants (115 Gy pre TG-43 era). However, recent analyses on clinical high dose rate (HDR) brachytherapy data indicated that the RBE effect for the PB may not be clinically significant (Brenner et al 2002, Wang et al 2003b). Nonetheless, dose escalation studies have not been performed to determine the maximum tolerated dose for 103 Pd using modern implant and treatment-outcome evaluation techniques. Maximum tolerated doses for PB are largely unknown at present. Unfortunately, prospective clinical trials evaluating dose escalation in PB have not been performed. To provide further biological and dosimetric justification for prospective clinical trials to determine the maximum tolerated dose, we have studied the implications of dose escalation in PB using 125 I and 103 Pd implants. A biological model was used to determine the PB dose levels that correspond to those reported for EBRT. To ensure that the escalated doses can be safely delivered, dosimetric techniques to achieve a higher target dose without significant increases of the dose to rectum and urethra were studied. 2. Methods and materials 2.1. Linear quadratic survival formula The linear quadratic (LQ) formalism is the most prevalent model used to predict the radiation killing of cells in clinical applications. The mechanistic basis for the standard LQ model has been discussed extensively in the literature (reviewed by Brenner et al (1998) and Sachs et al (1997)). Extensions to the LQ formalism have also been proposed to account for cell-cycle, repopulation and reoxygenation effects (Brenner et al 1995, Zaider et al 1996). The LQ model with corrections for tumour repopulation effects can be expressed as and S = e E (1) E = αd + βgd 2 γt (2)

4 2756 XALiet al where S is the surviving fraction, D is the total dose delivered within the effective treatment time T, α and β characterize intrinsic radiosensitivity, G is the protraction factor which corrects for sub-lethal damage repair, γ is the repopulation rate (γ = ln(2)/t d,where T d is the doubling time). For EBRT, G = 1/n and D = nd, where n is the number of dose fractions, and d is the dose per fraction. The treatment time/duration (T) forebrt can be simply calculated as the number of treatment fractions multiplied by 1.4 (seven days per week divided by five fractions per week). For PB, the dose delivered within the treatment time T is given by D = R 0 λ (1 e λt ) (3) and the dose protraction factor is 2R 2 [ 0 1 G = D 2 (µ λ) 2λ (1 e 2λT ) 1 ] µ + λ (1 e (µ+λ)t ) (4) T = 1 ( ) γ λ ln. (5) αr 0 Here, R 0 is the initial dose rate, D 0 is the prescribed dose (D 0 = R 0 /λ), µ is the repair rate of tumour cells (µ = ln(2)/t r,where T r is the characteristic repair half-time of cells with sub-lethal damage) and λ is the decay constant for the implanted isotopes (λ = ln(2)/t s, where T s is the half-time of the isotope). T s is equal to 60.2 days for 125 I and 17 days for 103 Pd implants. Except where explicitly noted otherwise, the following LQ parameters were used to estimate prostate tumour cell survival as a function of dose and dose rate: α = 0.15 Gy 1, α/β = 3.1 Gy, T r = 16 min and T d = 42 days. These parameters are derived from an LQ analysis of compiled multi-institutional clinical datasets obtained with multi-radiotherapy modality (Wang et al 2003a). Very recently, these parameters are validated based on an analysis of an independent single-institutional outcome data of single modality (high dose rate brachytherapy) (Wang et al 2003b). To assess the dependence of the results on the parameters, calculations were also performed using several other sets of LQ parameters, including α/β = 1.5 Gy as reported by Fowler et al (2001), Brenner et al (2002) and Brenner and Hall (1999), and α/β = 10 Gy as normally assumed for tumour Calculation of the equivalent uniform dose The concept of equivalent uniform dose (EUD) was proposed originally for EBRT by Niemierko (1997). The EUD is defined as the biologically equivalent dose that, if given uniformly, will lead to the same biological effect as a given non-uniform dose distribution. EUD can be applied to both tumour and normal tissue (Niemierko 1999, Wu et al 2002). The EUD has been recently applied for low dose rate (LDR) brachytherapy (Wang and Li 2003), and it is a convenient quantity to compare EBRT and PB dosing schemes. By definition, the EUD for EBRT or PB that is referenced to EBRT can be formulated by log(s) EUD = α + βd 1.4γ/d. (6) To account for dose heterogeneity in both EBRT and PB, the survival fraction was calculated based on dose volume histogram (DVH) by S = i V i V 0 S(D i ) (7)

5 Dose escalation for prostate implant 2757 where V 0 is the tumour volume, V i is the sub volume corresponding to dose bin D i in DVH. Because the quantitative distribution of tumour clonogens is not available, the density of clonogens throughout the tumour is presumed constant. Except where explicitly noted otherwise, the EUD calculation for prostate cancer was performed using the radiosensitivity parameters derived recently from clinical data (Wang et al 2003a). All the reported EUD values in this paper correspond to a daily dose fraction of 1.8 Gy delivered in a few minutes to mimic a conventional EBRT Calculation of tumour control probability Tumour control probability (TCP) was calculated from the clonogenic surviving fraction S using the Poisson distribution: TCP = e KS (8) where K is the cell number of tumour clonogens. In a previous study, Wang et al (2003a)have derived the clonogen numbers for different patient risk groups based on published clinical data ( , and for low-, intermediate- and high-risk patient groups, respectively). These clonogen numbers were used in this study EUD for normal tissue For an organ at risk (OAR), one is mainly interested in effects of hot spots. The above EUD calculation for tumour becomes invalid. In this work, we define the EUD for normal tissue as the biologically equivalent dose that, if given uniformly, will lead to the same normal tissue complication probability (NTCP) as the actual non-uniform dose distribution. The Lyman model (Lyman 1985)wasused to calculate NTCP: NTCP = 1 t e x2 2 dx (9) 2π t = D TD 50(ν) m TD 50 (ν) (10) TD(ν) = TD(1) ν n (11) where ν is the fraction of OAR volume irradiated, TD is the tolerance dose, TD 50 is the tolerance dose with 50% complication, m and n are the model parameters. For a non-uniform dose distribution, the method of effective volume (V eff )(Kutcher and Burman 1989)wasused. That is, ν = V eff = ( ) V 1/n i Di = ( ) 1/n Di ν i (12) V 0 V i 0 D max D i max where V 0 is the total volume irradiated, D max is the maximum dose in the OAR. To yield the same NTCP value, the EUD should produce the same t value, i.e., t(eud) = EUD TD 50(1). (13) m TD 50 (1) That is EUD TD 50 (1) = D max TD 50 (ν). (14) m TD 50 (1) m TD 50 (ν)

6 2758 XALiet al Then we have EUD = ( i ν i D 1/n i ) n. (15) The calculation of EUD for normal tissue involves only one adjustable parameter, denoted n, for each OAR. It is independent of TD 50 and m. Theformula derived here is based on Lyman NTCP model but has the same form as that in the generalized EUD proposed by Niemierko (1999) based on the power law. The parameter a used by Niemierko (1999) is replaced here by the inverse of n. It has been reported that the n value for rectum is 0.12 (Burman et al 1991), which corresponds to an a value of 8.3. This value is slightly different from that of a = 6 used by Wuet al (2002). To determine the model sensitivity on parameter n, calculation was performed using both n values corresponding to a = 8.3 and 6 for both rectum and urethra. It was found that the EUD with n = 0.17 was 7% lower than that with n = 0.12 for rectum, while only 1% difference between the use of the two n values was observed for urethral EUD. For the purpose of this work, we are only interested in the relative changes in rectal and urethral EUD with prostate dose escalation. The calculation showed that there was almost no difference in the relative change for rectal and urethral EUDs when the two n values were used. Therefore, the n value of 0.12 was selected for both rectum and urethra in the calculations presented in this paper Dose escalation studies Two doses are assumed to be biologically equivalent if they lead to the same EUD. The EUDs corresponding to the doses of 64.8, 70.2, 75.6 and 81 Gy delivered in the MSKCC EBRT dose-escalation studies (Zelefsky et al 1998) were calculated. The equivalent PB doses that would generate the same EUD if delivered by 125 Ior 103 Pd implant were obtained. A series of PB dose levels suitable for the PB dose escalations as well as their corresponding EUDs were provided. The EUDs for rectum and urethra for these PB dose levels were also determined. In order to compare the relative effectiveness of different dose levels, the TCP values were determined for low-, intermediate- and high-risk patient groups. The definitions for these risk groups are the same as those defined by Levegrun et al (2001), i.e., low-risk: Gleason score 6 and Stage < T2c; intermediate-risk: Gleason score > 6orStage T2c; high-risk: Gleason score > 6 and Stage T2c. Note that clinical data for the same risk groups were used to derive the LQ parameters by Wang et al (2003a). Representative DVHs for EBRT and BT were used for all the dose-escalation calculations (see figure 1), except for the prospective dosimetric studies. These DVHs were acquired by averaging a series of DVHs obtained retrospectively based on patient CT images. DVHs from both 125 I and 103 Pd implants were used to calculate the average. The representative DVHs were scaled according to the prescribed dose considered. In the calculation of EUD for PB, a dose threshold of 110 Gy for 125 Ior95Gyfor 103 Pd implants was used. The tumour volume with dose delivered under the threshold is ignored in the calculation. This approximation was required because the cell killing produced by a dose under the threshold is so low such that it dramatically changes the mean survival fraction. The same approach was also used by other investigators (King and Fowler 2001). This approximation can be justified by the fact that the volume with dose under the threshold is very small, usually less than 0.1% of total volume, and this volume is normally near the edge of the target volume where there may not be any clonogenic cells.

7 Dose escalation for prostate implant 2759 Target Volume (%) EBRT BT Dose Normalized to Prescribed Dose(%) Figure 1. Representative dose volume histograms obtained retrospectively based on patient CT images for external beam radiotherapy (EBRT) and prostate permanent brachytherapy (PB) Dosimetric considerations A commercial treatment planning system (Strata, Rosses Medical) was used to perform the dosimetry calculations based on patient ultrasound images. In order to see whether it is possible to escalate dose to the prostate without a significant increase in dose to the rectum and to maintain the urethra dose at an acceptable level, the inverse planning algorithm provided in the Strata system was used. For a given seed strength, the seed locations and the number of seeds required to achieve a desired target dose were determined using the inverse treatment planning algorithm. During the planning process, the urethral dose was constrained to not exceed 120% of the prescribed dose and the rectal dose was minimized. The 120% constraint for urethra was selected based on a reported retrospective analysis (Wallner et al 1995)aswell as on practical consideration. Seed strength was selected by trial and error for each prostate dose level in order to ensure that a proper number of seeds were used. Treatment plans with a small number of seeds will lead to non-conformal dose distribution, while those with a large number of seeds are often impractical. Particularly, effort was spent on selecting the source strength to ensure a tight dose coverage on the posterior margin of the prostate. The seed locations determined by the inverse planning algorithm near the posterior prostate region were carefully adjusted to avoid significant increases in rectal dose in the case of a higher prescribed dose. The calculation was performed on five representative patients of different prostate sizes (ranging from 20 to 38 cc) for a series of PB dose levels using either 125 Ior 103 Pd seeds. 3. Results 3.1. Dose to prostate The corresponding values of EUD as well as the PB prescribed doses of 125 Ior 103 Pd implant were calculated, based on the EBRT dose escalation studies of MSKCC. The results are tabulated in table 1. The representative DVHs presented in figure 1 were used in the calculation. Note that the EUD is slightly higher than the corresponding EBRT prescribed dose due to

8 2760 XALiet al EUD (Gy) Pd 125 I Prescribed Dose for Implants (Gy) Figure 2. EUDs as a function of the prescribed dose for both 125 Iand 103 Pd implants. The EUD values are corresponding to EBRT doses with 1.8 Gy pre-fraction. Table 1. The values of EUD and the prescribed doses of prostate permanent brachytherapy corresponding to the prescribed EBRT doses used in the MSKCC dose escalation studies. All values are in Gy. The LQ parameters reported by Wang et al (2003a) are used to compute the EUD. Prescribed dose for EBRT (1.8 Gy/fr) EUD (1.8 Gy/fr) Prescribed dose for 125 Iimplant Prescribed dose for 103 Pd implant the slight dose inhomogeneity revealed in the EBRT DVH. It is also clear that the currently prescribed PB doses are estimated here to be of lower effectiveness than the escalated EBRT doses, suggesting the need for PB dose escalation. Figure 2 presents EUD values as a function of the prescribed dose for both 125 I and 103 Pd implants. The representative DVHs for both EBRT and PB presented in figure 1 were used. It is seen that the prescribed dose for 125 Ior 103 Pd implants should be 177 or 146 Gy, respectively, for an EUD of 90 Gy. The TCP values calculated presently for the three patient risk groups are plotted against the dose escalations in figure 3 for (a) 125 I and (b) 103 Pd implants. These data clearly suggest, again, that the current prescribed doses (145 Gy for 125 I and 125 Gy for 103 Pd) may be inadequate, at least for the intermediate- and high-risk patient groups. Based on these results, the dose steps of 145, 160, 170 and 180 Gy for 125 I and 125, 135, 145 and 155 Gy for 103 Pd implant would be reasonable in a prospective PB dose escalation study. With these dose escalation steps, the TCP would be elevated gradually from the steep-variation region into the plateau region for the intermediate- and high-risk groups (figure 3). Table 2 lists the corresponding values of EUD and TCP for these dose levels. Again, the representative DVHs presented in figure 1 were used to generate these EUD and TCP data. Also, it should be pointed out that, although predicted by the model, the TCP may never be equal to 100% from aclinical point of view. This is particularly true for the intermediate- and high-risk groups, in whom there is a risk of pre-existing occult metastatic disease.

9 Dose escalation for prostate implant (a) (b) TCP (%) low-risk intermediate-risk high-risk low-risk intermediate-risk high-risk Prescribed Dose of 125 I Implants (Gy) Prescribed Dose of 103 Pd Implants (Gy) Figure 3. TCP values calculated presently for the three patient risk groups as a function of proposed prescribed dose for (a) 125 Iand(b) 103 Pd implants. Table 2. A prospective dose escalation for permanent brachytherapy of prostate cancer. The LQ parameters reported by Wang et al (2003a) are used to compute the EUD. I-125 implant Pd-103 implant Prescribed dose (Gy) EUD (Gy) (1.8 Gy/fr) TCP (Low-risk) TCP (Intermediate-risk) TCP (High-risk) The EUD calculation for prostate cancer was found to strongly depend on the LQ parameters. For example, the EUDs calculated with two parameter sets ((a): α/β = 1.5 Gy, α = 0.04 Gy 1, T r = 114 min and T d =,assuggested by Brenner et al (2002) and Fowler et al (2001); and (b): α/β = 10 Gy, α = 0.3 Gy 1, T r = 16 min and T d = 42 days, as normally suggested for tumour) were found to be up to 20% higher than those obtained with the parameter set of Wang et al (2003a). However, the calculations with these two parameter sets may overestimate the EUD values, given the fact that the 125 Iimplant with 145 Gy is equivalent to 70 Gy of EBRT based on accumulated clinical data. Although the EUD values for a given prescribed dose are very different with different sets of parameters, the relative increases remain approximately the same with the proposed dose escalations for all parameter sets used. To assess the dependence on DVH, calculations were performed by using several clinically possible DVHs. It was found that, for a given prescribed dose, the EUD varied as much as ±5% with different DVHs used. However, the relative changes in EUD with the proposed dose escalations remain approximately same for these possible DVHs. Because of the dependence of model parameters and DVH, the uncertainty on the absolute EUD values presented can be as high as 25% Dose to rectum and urethra Prospective dosimetry calculations were performed on patient ultrasound images to investigate whether it was possible to escalate dose to prostate without significantly increasing the dose to the rectum and urethra. As an example of such calculation, DVHs calculated for four 125 I implants with the prescribed dose of 145, 160, 170 and 180 Gy are shown in figure 4. The

10 2762 XALiet al Cumulative Volume (%) Rectum Urethra 145 Gy 160 Gy 170 Gy 180 Gy Prostate Dose (Gy) Figure 4. DVHs calculated for four 125 Iimplants with the prescribed dose of 145, 160, 170 and 180 Gy based on ultrasound images. The data included are for prostate, rectum and urethra, with ultrasound volumes of 23, 9.3 and 0.2 cc, respectively. data included are for prostate, rectum and urethra, with ultrasound volumes of 23, 9.3 and 0.2 cc, respectively. The rectal volume was defined as the volume of rectal wall drawn from ultrasound images excluding the rectal probe. Only the portion of rectum directly underneath the prostate was considered. The total activities (the number of seeds times the activity per seed) used were , , and mci for implants with prescriptions of 145, 160, 170 and 180 Gy, respectively. It is seen from figure 4 that, for the four dose levels, rectal doses are slightly changed. For example, the rectum volumes covered by 50 and 100 Gy changed from 30% to 31% and from 2.0% to 1.5% when the prescribed dose increased from 145 to 180 Gy. The urethral dose is kept within % of the prescription. That is, urethral dose can be maintained below 220 Gy for the dose escalation. It has been reported that complications on urethra are not significant if the urethral dose is kept below 400 Gy for an 125 IPB(Wallner et al 1995). Based on the DVHs in figure 4, theeud values for the prostate, the rectum and the urethra were calculated. Figure 5 compares the changes of these EUD values when the prescribed dose is escalated from 145 to 180 Gy. Since we are only interested in the relative change, all EUD values are normalized to its value corresponding to the prescription of 145 Gy. When the prescribed dose is increased from 145 to 180 Gy, the EUD for the prostate increased 10%, while the EUD for the rectum changed only 3%. Thus, the proposed dose escalation does not necessarily worsen the rectal complications. For all other cases studied, similar results were generally found, although slightly higher rectal and urethral doses for cases with larger prostates were observed. It should be noted that the above dosimetric data were based on pre-planning and may not represent the actual results on patients. During an implant procedure, there are many practical/technical factors that can influence the quality of the implant. 4. Discussion and conclusion Accumulated EBRT dose escalation studies have shown that, for localized prostate cancer of intermediate- and high-risk patient groups, treatment outcome can be improved with higher

11 Dose escalation for prostate implant Prostate Rectum Urethra Relative EUD Prescribed Dose (Gy) Figure 5. Relative changes of the EUD values when the prescribed dose for an 125 Iimplant escalated from 145 to 180 Gy. All EUD values are normalized to its value corresponding to the prescription of 145 Gy. doses. Analysis on the EBRT dose escalations using the present biological models reveals that the doses currently used in prostate permanent brachytherapy may be too low. Our studies suggest that dose levels up to 180 Gy for 125 I and 155 Gy for 103 Pd may be advantageous. Dosimetric calculations on selected patient cases show that the escalated doses can in principle be safely delivered. It is dosimetrically possible to increase the dose to prostate without a substantial change in rectal dose. The urethral dose can be kept just slightly above the escalated dose. Recently, Stock et al (2000) performed a retrospective analysis on their patients of the low-risk group who underwent monotherapy of 125 Iimplants. They reported that improved biochemical control rates were associated with a higher dose up to 180 Gy with acceptable urinary toxicity. Higher doses above 180 Gy in this retrospective study did not produce higher control rates, but noticeably increased toxicity. The results obtained presently from the biological modelling, that employs the model parameters derived from clinical data, are consistent with this retrospective analysis. Based on the results of our studies, a prospective dose escalation trial for prostate permanent implants has been initiated at our institution. The predictions presented in this work require validation from such a clinical trial. The results generated in this study, although based on the latest clinical data, may be quantitatively limited by the approximations used in the models. In particular, the reliability of the radiobiological parameters involved is under debate at present (Fowler et al 2001, Brenner et al 2002, Brenner and Hall 1999). As recognized by several groups (Tucker and Taylor 1996, Kendal 1998, Zaider and Minerbo 2000), the Poisson model used to calculate TCP is an approximation, especially when considering repopulation effect. The assumption of uniform tumour-clonogen distribution may be too simple. In addition, the effect of RBE for 125 I and 103 Pd implants is ignored, although recent comparisons with the HDR brachytherapy data indicate that this RBE effect may not be clinically significant. Uncertainties on the absolute EUD values presented can be as high as 25%. Therefore, caution needs to be exercised in using the data presented for clinical decision-making purposes.

12 2764 XALiet al Acknowledgment The authors would like to thank A Niemierko, PhD for the helpful discussions on the subjects. References Blasko J C, Grimm P D and Ragde H 1993 Brachytherapy and organ preservation in the management of carcinoma of the prostate Semin. Radiat. Oncol Blasko J C, Ragde H and Grimm P D 1991 Transperineal vs. guided implantation of the prostate: morbidity and complications Scand. J. Urol. Nephrol. 137 (Suppl) Blasko J C et al 1995 Prostate specific antigen based disease control following ultrasound guided 125 Iimplantation for stage T1/T2 prostatic carcinoma J. Urol Blasko J C et al 2000 Palladium-103 brachytherapy for prostate carcinoma Int. J. Radiat. Oncol. Biol. Phys Brenner D J and Hall E J 1999 Fractionation and protraction for radiotherapy of prostate carcinoma Int. J. Radiat. Oncol. Biol. Phys Brenner D J et al 1995 A convenient extension of the linear-quadratic model to include redistribution and reoxygenation Int. J. Radiat. Oncol. Biol. Phys Brenner D J et al 1998 The linear-quadratic model and most other common radiobiological models result in similar predictions of time-dose relationships Radiat. Res Brenner D J et al 2002 Direct evidence that prostate tumors show high sensitivity to fractionation (low α/β ratio), similar to late-responding normal tissue Int. J. Radiat. Oncol. Biol. Phys Burman C et al 1991 Fitting of normal tissue tolerance data to an analytic function Int. J. Radiat. Oncol. Biol. Phys D Amico A V et al 1998 Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer J. Am. Med.Assoc Fowler J, Chappell R and Ritter M 2001 Is α/β for prostate cancer really low? Int. J. Radiat. Oncol. Biol. Phys Fuks Z et al 1991 The effect of local control on metastatic carcinoma of the prostate: long term results in patients treated with I-125 Int. J. Radiat. Oncol. Biol. Phys Hanks G E et al 1996 Conformal technique dose escalation for prostate cancer: improved cancer control with higher doses in patients with pretreatment PSA 10 ngm/ml Int. J. Radiat. Oncol. Biol. Phys Hanks G E et al 1997 Optimization of conformal radiation treatment of prostate cancer: report of a dose escalation study Int. J. Radiat. Oncol. Biol. Phys Hanks G E et al 1998 Dose escalation with 3D conformal treatment: five year outcomes, treatment optimization, and future directions Int. J. Radiat. Oncol. Biol. Phys Hilaris B S et al 1974 Radiation therapy and pelvic node dissection in the management of cancer of the prostate Am. J. Roentgenol. Radium Ther. Nucl. Med Kendal W S 1998 A closed-form description of tumour control with fractionated radiotherapy and repopulation Int. J. Radiat. Biol King C R and Fowler J F 2001 A simple analytic derivation suggests that prostate cancer α/β ratio is low Int. J. Radiat. Oncol. Biol. Phys Koprowski C D et al 1991 External beam irradiation versus 125 iodine implant in the definitive treatment of prostate carcinoma Int. J. Radiat. Oncol. Biol. Phys Kuban D A, El-Mahdi A M and Schellhammer P F 1989 I-125 interstitial implantation for prostate cancer. What have we learned 10 years later? Cancer Kutcher G J and Burman C 1989 Calculation probability factors for non-uniform normal tissue irradiation: the effective volume method Int. J. Radiat. Oncol. Biol. Phys Kwok Y et al 2002 Risk group stratification in patients undergoing permanent 125 Iprostate brachytherapy as monotherapy Int. J. Radiat. Oncol. Biol. Phys Levegrun S et al 2001 Fitting tumor control probability models to biopsy outcome after three-dimensional conformal radiation therapy of prostate cancer: pitfalls in deducing radiobiologic parameters for tumors from clinical data Int. J. Radiat. Oncol. Biol. Phys Ling C C, Li W X and Anderson L L 1995 The relative biological effectiveness of I-125 and Pd-103 Int. J. Radiat. Oncol. Biol. Phys Litwin M S et al 1995 Quality of life outcomes in men treated for localized prostate cancer J. Am. Med. Assoc

13 Dose escalation for prostate implant 2765 Lyman J T 1985 Complication probability as assessed from dose volume histograms Radiat. Res. 104 S13 9 Martinez A A et al 2000 A comparison of external beam radiation therapy versus radical prostatectomy for patients with low risk prostate carcinoma diagnosed, staged, and treated at a single institution Cancer Niemierko A 1997 Reporting and analyzing dose distributions: a concept of equivalent uniform dose Med. Phys Niemierko A 1999 A generalized concept of equivalent uniform dose (EUD) (Abs) Med. Phys Partin A W et al 1993 Serum PSA with anatomic radical prostatectomy. The Johns Hopkins experience after 10 years Urol. Clin. North Am Pollack A et al 2000 Preliminary results of a randomized radiotherapy dose-escalation study comparing 70 Gy with 78 Gy for prostate cancer J. Clin. Oncol Sachs R K, Hahnfeld P and Brenner D J 1997 Review: the link between low-let dose-response relations and the underlying kinetics of damage production/repair/misrepair Int. J. Radiat. Biol Shrader-Bogen C L et al 1997 Quality of life and treatment outcomes: prostate cancer patients perspectives after prostatectomy or radiation therapy Cancer Stock R G et al 1998 A dose response study for I-125 prostate implants Int. J. Radiat. Oncol. Biol. Phys Stock R G et al 2000 What is the optimal dose for I-125 prostate implant? A dose response analysis of long-term urinary symptoms, biochemical control and post-treatment biopsy Int. J. Radiat. Oncol. Biol. Phys. 48 (Suppl) Stokes S H 2000 Comparison of biochemical disease-free survival of patients with localized carcinoma of the prostate undergoing radical prostatectomy, transperineal ultrasound-guided radioactive seed implantation, or definitive external beam irradiation Int. J. Radiat. Oncol. Biol. Phys Tucker S L and Taylor J M 1996 Improved models of tumour cure Int. J. Radiat. Biol Valicenti R et al 2000 Survival advantage from higher-dose radiation therapy for clinically localized prostate cancer treated on the radiation therapy oncology group trials J. Clin. Oncol Wallner K, Roy J and Harrison L 1995 Dosimetry guidelines to minimize urethral and rectal morbidity following transperineal I-125 prostate brachytherapy Int. J. Radiat. Oncol. Biol. Phys Wang J Z, Guerrero M and Li X A 2003a How low is the α/β ratio for prostate cancer? Int. J. Radiat. Oncol. Biol. Phys Wang J Z and Li X A 2003 Evaluation of external beam radiotherapy and brachytherapy for localized prostate cancer using equivalent uniform dose Med. Phys Wang J Z, Li X A, Yu C X and DiBiase S J 2003b The low α/β ratio for prostate cancer: what does the clinical outcome of HDR brachytherapy tell us? Int. J. Radiat. Oncol. Biol. Phys. 57 at press Wu Q et al 2002 Optimization of intensity-modulated radiotherapy plans based on the equivalent uniform dose Int. J. Radiat. Oncol. Biol. Phys Zagars G K, Sherman N E and Babaian R J 1991 Prostate-specific antigen and external beam radiation therapy in prostate cancer Cancer Zagars G K and von Eschenbach A C 1993 Prostate-specific antigen: an important marker for prostate cancer treated by external beam radiation Cancer Zaider M and Minerbo G N 2000 Tumour control probability: a formulation applicable to any temporal protocol of dose delivery Phys. Med. Biol Zaider M, Wuu C S and Minerbo G N 1996 The combined effects of sublethal damage repair, cellular repopulation and redistribution in the mitotic cycle: I. Survival probabilities after exposure to radiation Radiat. Res Zelefsky M J et al 1996 Impact of radiation dose on achieving a nadir PSA level after 3-dimensional conformal radiotherapy for patients with localized prostate cancer Int. J. Radiat. Oncol. Biol. Phys Zelefsky M J et al 1998 Dose escalation with three-dimensional conformal radiation therapy affects the outcome in prostate cancer Int. J. Radiat. Oncol. Biol. Phys