- Email: [email protected]
Updated Results and Patterns of Failure in a Randomized Hypofractionation Trial for High-risk Prostate Cancer
International Journal of
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Genitourinary Cancer
Updated Results and Patterns of Failure in a Randomized Hypofractionation Trial for High-risk Prostate Cancer Stefano Arcangeli, MD,* Lidia Strigari, PhD,y Sara Gomellini, MD,* Biancamaria Saracino, MD,* Maria Grazia Petrongari, MD,* Paola Pinnaro`, MD,* Valentina Pinzi, MD,* and Giorgio Arcangeli, MD* *Department of Radiation Oncology and yLaboratory of Medical Physics and Expert Systems, Regina Elena National Cancer Institute, Rome, Italy Received Dec 30, 2011, and in revised form Feb 21, 2012. Accepted for publication Feb 21, 2012
Summary This article reports the updated results of a randomized study comparing an escalated conventional fractionation with a hypofractionation schedule in localized, high-risk prostate cancer. The 2 arms were hypothesized to be isoeffective with regard to tumor control, assuming a fairly low a/b ratio of 1.5-1.8 Gy. After median follow-up of 70 months, there were no statistically significant differences between hypofractionation and conventional fractionation in the actuarial freedom from all types of failure (biochemical, local, and distant).
Purpose: To report long-term results and patterns of failure after conventional and hypofractionated radiation therapy in high-risk prostate cancer. Methods and Materials: This randomized phase III trial compared conventional fractionation (80 Gy at 2 Gy per fraction in 8 weeks) vs hypofractionation (62 Gy at 3.1 Gy per fraction in 5 weeks) in combination with 9-month androgen deprivation therapy in 168 patients with high-risk prostate cancer. Freedom from biochemical failure (FFBF), freedom from local failure (FFLF), and freedom from distant failure (FFDF) were analyzed. Results: In a median follow-up of 70 months, biochemical failure (BF) occurred in 35 of the 168 patients (21%) in the study. Among these 35 patients, local failure (LF) only was detected in 11 (31%), distant failure (DF) only in 16 (46%), and both LF and DF in 6 (17%). In 2 patients (6%) BF has not yet been clinically detected. The risk reduction by hypofractionation was significant in BF (10.3%) but not in LF and DF. We found that hypofractionation, with respect to conventional fractionation, determined only an insignificant increase in the actuarial FFBF but no difference in FFLF and FFDF, when considering the entire group of patients. However, an increase in the 5-year rates in all 3 endpointsdFFBF, FFLF, and FFDFdwas observed in the subgroup of patients with a pretreatment prostate-specific antigen (iPSA) level of 20 ng/mL or less. On multivariate analysis, the type of fractionation, iPSA level, Gleason score of 4þ3 or higher, and T stage of 2c or higher have been confirmed as independent prognostic factors for BF. High iPSA levels and Gleason score of 4þ3 or higher were also significantly associated with an increased risk of DF, whereas T stage of 2c or higher was the only independent variable for LF. Conclusion: Our results confirm the isoeffectiveness of the 2 fractionation schedules used in this study, although a benefit in favor of hypofractionation cannot be excluded in the subgroup of patients with an iPSA level of 20 ng/mL or less. The a/b ratio might be more appropriately evaluated by FFLF than FFBF results, at least in high-risk disease. Ó 2012 Elsevier Inc.
Reprint requests to: Giorgio Arcangeli, MD, Department of Radiation Oncology, Regina Elena Cancer Institute, Via Elio Chianesi 53, 00144
Int J Radiation Oncol Biol Phys, Vol. 84, No. 5, pp. 1172e1178, 2012 0360-3016/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ijrobp.2012.02.049
Rome, Italy. Tel: þ39 0652665602; Fax: þ39 0652665667; E-mail: [email protected] Conflict of interest: none.
Volume 84 Number 5 2012
Hypofractionation vs conventional fractionation 1173
Introduction New advanced radiation therapy techniques have led to an improvement in conformality of high radiation doses to the target tissues and to novel approaches for prostate cancer treatment. These strategies include dose escalation, as well as hypofractionation. Dose escalation, which increases the total dose, has improved the freedom from failure without further complications (1-3). Hypofractionation allows one to obtain an equivalent biological effective dose using a dose per fraction higher than the conventional dose (4, 5); the clinical experience on hypofractionation is still limited, but some studies have reported promising results on biochemical control with acceptable late toxicity (6-11). We have previously reported the results of a randomized trial comparing a total escalated dose (80 Gy) by conventional fractionation with a hypofractionation schedule using 3-dimensional conformal radiation therapy (3D-CRT) in localized high-risk prostate cancer. The early results of this trial have shown a significant biochemical benefit of hypofractionation over conventional fractionation schedules without increasing the side effects (12, 13), thus supporting the radiobiological assumptions of high sensitivity of prostate tumors to large dose fractions. However, biochemical improvement does not necessarily mean better local tumor control because prostate-specific antigen (PSA) failure does not indicate whether failure was local or distant, the latter being a frequent event in the high-risk disease group, where early micrometastatic spread may have already occurred by the time of local treatment. Nevertheless, for this biochemical improvement with hypofractionation to translate into better clinical results, a longer follow-up would be required. In fact, although the PSA endpoint is presently considered an early and useful surrogate for treatment efficacy, the relationship and timeline between PSA failure, clinical failure, and survival need to be accurately documented for an optimal estimation of the treatment benefit. Patients in our study are presently approaching a median follow-up of 70 months, and the data on clinical failure have now matured, enabling the examination and comparison of the failure pattern in the 2 fractionation groups, as well as in subgroups of patients with different prognostic factors.
Methods and Materials
Simulation and treatment The simulation and treatment procedures used have been extensively reported in a previous article (13). In brief, all patients received 9-month androgen deprivation therapy consisting of a daily oral administration of 50 mg of bicalutamide for 270 days, as well as a quarterly subcutaneous injection of a luteinizing hormone-releasing hormone analog depot. The 3D-CRT started at day 67 from the first bicalutamide intake. The clinical target volume, which included the prostate and seminal vesicles, was expanded in 3 dimensions with a 1-cm margin to obtain the planning target volume, except at the prostate-rectal interface, where a 0.6-cm margin was used to decrease rectal involvement. All patients were treated with 15-MV photons by a Varian 2100 linear accelerator (Varian Medical Systems, Palo Alto, CA) equipped with a 120-leaf collimator. The prescription dose to 95% of the clinical target volume was 80 Gy in 2-Gy fractions and 62 Gy in 3.1-Gy fractions for the conventional and hypofractionation arms, respectively.
Study endpoints and statistical analysis
Study design and protocol eligibility This phase III, single-institution, randomized trial was designed to compare 2 dose fractionation schedules using 3D-CRT in patients Table 1
with localized, high-risk prostate cancer. Between 2002 and 2005, 168 patients were randomized to receive 80 Gy in 40 fractions within 8 weeks at 2.0 Gy per fraction (arm A, conventional fractionation) vs 62 Gy in 20 fractions within 5 weeks (4 fractions per week) at 3.1 Gy per fraction (arm B, hypofractionation) to the prostate and seminal vesicles. We tested the hypothesis that hypofractionation would result in a lower rate of late complications and would be just as effective as conventional fractionation. In brief, the 2 arms were hypothesized to be isoeffective with regard to tumor control, assuming a fairly low a/b ratio of 1.5-1.8 Gy. However, as regards the late complications, assuming an a/ b value of 3 Gy for normal tissue, the total dose by hypofractionation should be equivalent to approximately 75 Gy given at 2 Gy per fraction, which is considerably lower than 80 Gy by conventional fractionation. To achieve 80% power, it was estimated that a total of 168 patients were required to detect a significant difference (P<.05, 2- sided). The study design and inclusion criteria have been described previously (13). Patient characteristics are reported in Table 1.
The freedom from biochemical failure (FFBF) was defined as the time interval from the first day of radiation therapy to the biochemical relapse (biochemical failure [BF]) according to the Phoenix definition of nadir PSA plus 2 ng/mL (14). The secondary
Patient characteristics after conventional treatment and hypofractionated treatment Conventional fractionation
No. of patients Age [median (range)] (yr) Biopsy-derived Gleason score (3þ4) (4þ3) cT stage
Hypofractionation
85 75 (54-83)
83 75 (61-82)
41 (48%) 44 (52%)
42 (51%) 41 (49%)
P value (t test) .8647 .8788
.3222 48 (56%) 37 (44%) 19/13 (2-128)
Abbreviation: iPSA Z pretreatment prostate-specific antigen.
54 (65%) 29 (35%) 26/16 (4-157)
.0627
1174 Arcangeli et al.
International Journal of Radiation Oncology Biology Physics
Fig. 1. Forest plot of hazard ratios (HRs) and confidence intervals (CIs) for biochemical (A), local (B), and distant (C) failures stratified for prognostic factors. CI Z confidence interval; Conv Z conventional arm; Evts Z events; GS Z Gleason score; HR Z hazard ratio; Hypo Z hypofractionation arm; iPSA Z pretreatment prostate-specific antigen; pts Z patients.
objectives included local failure (LF), distant failure (DF), overall survival (OS), and cancer-specific survival (CSS). LF was defined as any evidence of disease documented by positron emission tomography (PET) with 18F-choline, by dynamic contrastenhanced magnetic resonance imaging, or by biopsy in patients with rising serum PSA level and/or palpable abnormality. DF was defined as any evidence of disease in extrapelvic nodes or other organs (bone, lung, liver, and so on) documented by bone scan, PET with 18F-choline, magnetic resonance imaging, or computed tomography (CT) imaging. OS was defined as death from any cause
and CSS as death from prostate cancer or death at the time of progressive metastatic disease. Kaplan-Meier actuarial analysis was used to determine the freedom from failure and survival. For all measured endpoints, patients were censored at the time of the specific event (BF, LF, DF, or survival). Comparison between the actuarial curves was made by use of the log-rank test. Statistical significance between the groups was tested by the c2 test and t test for categorical and continuous variables, respectively. A Cox proportional hazards model was applied for multivariate analysis and for hazard ratio
Volume 84 Number 5 2012
Hypofractionation vs conventional fractionation 1175 (HR) estimation. The HR and 95% confidence interval (CI) were also shown by forest plot. The procedures followed were in accordance with the ethical standards of the institutional responsible committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2000.
Results No difference was found in the incidence and severity of late gastrointestinal or genitourinary toxicity between the 2 fractionation schedules. The results of acute and late toxicity have been reported in detail in a previous article (12).
Patterns of failure The patterns of failure for the 2 fractionation arms are shown in Fig. 1. At a median follow-up of 70 months, BF occurred in 35 of the 168 patients (21%) in the study, 22 of 85 (26%) in the conventional group and 13 of 83 (15%) in the hypofractionation group (Fig. 1A). LF was detected in 10 patients (12%) and 7 patients (8%), respectively (Fig. 1B). DF was detected in 13 of 85 patients (15%) and 9 of 83 patients (5%), respectively (Fig. 1C). In 2 patients in the conventional group and 4 patients in the hypofractionation group, LF and DF were detected simultaneously. LF was detected by PET-choline images as an intense, persistent uptake in the prostate. In 5 patients LF was confirmed by a positive biopsy specimen. Biopsy findings were negative in 8 patients, and 4 patients refused to undergo biopsy. In 5 patients with negative biopsy findings, LF was confirmed by magnetic resonance imaging. The significance of the differences between the 2 fractionation schedules for all patients and subgroups of patients stratified for different levels of prognostic factors can be appreciated by the forest plots of the HRs and 95% CIs displayed in Fig. 1. An HR of less than 1 indicates a decreased risk of failure in the hypofractionation schedule. The risk reduction of 10.2% in BF was statistically significant for the whole group (HR, 0.34; 95% CI, 0.21-0.56); the reduction was significant in the subgroups of patients with either PSA level of 20 ng/mL or less (HR, 0.15; 95%
Fig. 2. Freedom from biochemical (a), local (b), and distant (c) failure in the 2 treatment groups: conventional arm (A, dashed line) and hypofractionation arm (B, solid line). Pts Z patients.
Table 2
Univariate analysis of failures 5-y freedom-from-failure rate Biochemical
Prognostic factor iPSA 20 ng/mL >20 ng/mL Gleason score (3þ4) (4þ3) T stage <2c 2c All patients
Local
Distant
Conv
Hypo
P value
Conv
Hypo
P value
Conv
Hypo
P value
83 72
95 73
.02 .39
92 91
100 85
.01 .51
87 84
98 81
.04 .96
94 66
88 83
.50 .01
97 87
95 91
.41 .13
98 76
91 89
.24 .08
85 74 79
89 83 85
.30 .08 .065
93 90 91
100 89 93
.05 .92 .33
91 82 86
89 91 90
.86 .19 .29
Abbreviations: Conv Z conventional fractionation; Hypo Z hypofractionation; iPSA Z pretreatment prostate-specific antigen.
International Journal of Radiation Oncology Biology Physics
1176 Arcangeli et al. Table 3
Multivariate Cox proportional hazards analysis of biochemical failure Variable
b
SE
P value
Exp(b)
Lower 95% confidence limit
Upper 95% confidence limit
Hypofractionation (yes/no) iPSA (continuous) (ng/mL) Gleason score (4þ3) (yes/no) T stage 2c (yes/no)
0.9145 0.0248 0.9117 0.9486
0.3607 0.0053 0.3699 0.3909
.0112 .0000 .0137 .0152
0.4007 1.0251 2.4885 2.5822
0.1983 1.0146 1.2097 1.2049
0.8097 1.0358 5.1190 5.5338
Abbreviation: iPSA Z pretreatment prostate-specific antigen. The P value of the model is <.0001.
CI, 0.03-0.71) or Gleason score of 4þ3 or higher (HR, 0.09; 95% CI, 0.03-0.31) (Fig. 1A). However, no significant difference was detected in LF between hypofractionation and conventional fractionation, either in all patients or the subgroups of patients stratified for prognostic factors (Fig. 1B). The risk reduction in DF in favor of hypofractionation was marginally significant (HR, 0.57; 95% CI, 0.33-0.98) for the whole group and significant in the subgroup of patients with a Gleason score of 4þ3 or higher (Fig. 1C). The actuarial analysis of failures is shown in Fig. 2. Contrary to our previous report (13), the actuarial analysis of FFBF, after a 70-month median follow-up, showed no statistically significant difference between the hypofractionation and conventional fractionation arms. At 5 years, 85% of patients in the former group were free from BF as compared with 79% of patients in the conventional fractionation group (PZ.065) (Fig. 2a). Furthermore, no differences between the 2 fractionation arms were detected in the actuarial freedom from local failure (FFLF) (5-year rate of 93% vs 91%) and freedom from distant failure (FFDF) (5-year rate of 90% vs 86%) (Figs. 2b and c). The actuarial 5-year rates of freedom from failure for all patients and each prognostic subgroup of patients are shown in Table 2. No significant difference in any type of failure between the 2 treatment arms was detected in any of the prognostic subgroups, with the exception of the subgroup of patients with a pretreatment prostatespecific antigen (iPSA) level of 20 ng/mL or less, in whom hypofractionation was significantly better than conventional fractionation in terms of the 3 failure outcomes (FFBF, 95% vs 83 [PZ.02]; FFLF, 100% vs 92% [PZ.01]; and FFDF, 98% vs 87% [PZ.04]), as well as the subgroup with a Gleason score of 4þ3 or higher, in which the difference in favor of hypofractionation was detected only for FFBF (83% vs 66%, PZ.01). On multivariate Cox proportional hazards analysis, the type of fractionation, iPSA level (taken as a continuous variable), Gleason score of 4þ3 or higher, and T stage of 2c or higher have been confirmed as independent prognostic factors for BF (Table 3). T stage of 2c or higher was the only independent variable for LF (Table 4), whereas higher iPSA levels and a Gleason score of 4þ3 or higher were also significantly associated with an increased risk of DF (Table 5).
Table 4
Cancer death and survival Prostate cancer death among the 35 patients with BF is shown in Fig. 3. The timeline from treatment to biochemical and clinical failure and then to prostate cancer death is reported for each patient. Death from prostate cancer occurred in 8 patients, and DF developed in 22 patients. In 16 of these 22 patients (73%), metastases occurred without a documented LF, with a median detection time of 29 months. In patients who had distant metastases as the first clinical failure, death occurred within 4.3 years from the initiation of treatment, with a median time to death of 38 months. The actuarial analysis of survival is shown in Fig. 4. After a 70-month median follow-up, no significant difference was detected between the 2 fractionation arms, either in OS (Fig. 4a) or in CSS (Fig. 4b), with 5-year rates of hypofractionation vs standard fractionation of 92% vs 82% and of 98% vs 92% for OS and CSS, respectively.
Discussion At present, only 2 randomized studies comparing radiation schedules with different fraction sizes in patients with prostate cancer have been published (4, 5). Both of these trials, designed and performed with no attempt to make the arms isoeffective, provided similar outcomes and toxicity rates between the conventional and hypofractionation schedules used. However, the total dose delivered in these trials is now considered insufficient in the treatment of prostate cancer, and therefore no definitive conclusion can be derived with regard to either outcome or toxicity. To our knowledge, this is the first randomized trial reporting the results of dose-escalated conventional fractionation in comparison to an equivalent hypofractionation schedule in patients with high-risk prostate cancer. The main hypothesis of the trial was that the FFBF rate of the hypofractionation schedule was not inferior to that of the conventional fractionation schedule. The early results of our study showed a better, though only marginally significant, incidence of FFBF in the hypofractionation
Multivariate Cox proportional hazards analysis of local relapse
Variable
b
SE
P value
Exp (b)
Lower 95% confidence limit
Upper 95% confidence limit
T stage 2c (yes/no)
1.1400
0.5837
.0508
3.1267
1.0018
9.7583
The P value of the model is .0349. Variables not included in the model were as follows: hypofractionation (yes/no), pretreatment prostate-specific antigen (continuous, in nanograms per milliliter), and Gleason score of 4þ3 or higher (yes/no).
Volume 84 Number 5 2012 Table 5
Hypofractionation vs conventional fractionation 1177
Multivariate Cox proportional hazards analysis of distant failure Variable
b
SE
P value
Exp (b)
Lower 95% confidence limit
Upper 95% confidence limit
iPSA (continuous) (ng/mL) Gleason score (4þ3) (yes/no)
0.0182 1.3200
0.0059 0.5107
.0020 .0097
1.0184 3.7436
1.0068 1.3829
1.0301 10.1340
Abbreviation: iPSA Z pretreatment prostate-specific antigen. The P value of the model is .0010. Variables not included in the model were as follows: hypofractionation (yes/no) and T stage of 2c or higher (yes/no).
arm compared with the conventional fractionation arm (13). The 3-year FFBF rate in the conventional fractionation arm of our study compared favorably with rates reported in the high-dose arm of dose-escalation randomized trials of radiation therapy alone (1-3) or in combination with androgen suppression (15, 16) for high-risk patients. However, even though PSA failure is a strong indicator of disease status (17), when one is assessing local treatment efficacy, LF should also be monitored because it is often only detected clinically several months after BF has occurred. Because of the short follow-up, the pattern of clinical failure could not be previously shown in the early results from this trial. The updated results reported in this study at a median followup of 70 months show that, although there is still a trend in favor of hypofractionation, the significance of the difference in FFBF between the 2 fractionation schedules was lost. No difference in FFLF was observed between the 2 groups, confirming the equivalence of the 2 fractionation schedules.
Fig. 3.
Although DF developed in 16 of 22 patients (73%) in a median time of 29 months without a documented LF, despite the concomitant/adjuvant association with androgen deprivation therapy given to all patients to minimize the risk of a bias due to early micrometastases in high-risk disease, no difference between the 2 groups could be detected in FFDF or in OS and CSS. However, the univariate subgroup analysis showed that in the subgroup of patients with an iPSA level of 20 ng/mL or less, all 3 failure endpointsdFFBF, FFLF, and FFDFdwere significantly improved by hypofractionation with respect to conventional fractionation. This might be because of the better local effect of hypofractionation compared with standard fractionation on a smaller tumor burden in such patients, as well as the achievement of local tumor control before the occurrence of metastatic spread. Our study shows that, in patients with high-risk prostate cancer, similar clinical local control can be achieved by high-dose
Clinical failures and prostate cancer death in patients with biochemical failures.
1178 Arcangeli et al.
Fig. 4. Overall survival (OS) (a) and cancer-specific survival (CSS) (b) in conventional arm (A, dashed line) and hypofractionation arm (B, solid line). Pts Z patients.
conventional fractionation and hypofractionation schedules, calculated to be equivalent by use of an a/b ratio of 1.5-1.8 Gy. Although in high-risk disease the events are more likely to occur earlier with respect to low- and intermediate-risk disease, we cannot exclude the possibility that some additional LFs may still be detected later. However, we consider that any other events that may be detected in a longer follow-up will likely be equally distributed between the 2 treatment groups and will not alter the isoeffectiveness of the 2 fractionation schedules. Whether this equivalence is because of a low a/b ratio (18, 19) or, as recently suggested, the reduced overall treatment time by a hypofractionation schedule (20), or both, will be a matter of debate, and the findings of this study may contribute to shedding some light on this debate. A suggestion that can be derived from the results of our study is that the a/b ratio should be more appropriately evaluated by long-term results of FFLF rather than those of FFBF, which could be affected, at least in high-risk disease, by the presence of occult metastases before treatment.
References 1. Pollack A, Zagars GK, Starkschall G, et al. Prostate cancer radiation dose response: results of the M.D. Anderson Phase III randomized trial. Int J Radiat Oncol Biol Phys 2002;53:1097-1105. 2. Peeters STH, Heemsbergen WD, Koper PCM, et al. Dose-response in radiotherapy for localized prostate cancer: results of the Dutch multicenter randomized Phase III trial comparing 68 Gy of radiotherapy with 78 Gy. J Clin Oncol 2006;24:1990-1996.
International Journal of Radiation Oncology Biology Physics 3. Zietman AL, Bae K, Slater JD, et al. Randomized trial comparing conventional-dose with high-dose conformal radiation therapy in early-stage adenocarcinoma of the prostate: long-term results from Proton Radiation Oncology Group/American College of Radiology 95-09. J Clin Oncol 2010;28:1106-1111. 4. Lukka H, Hayter C, Julian JA, et al. A randomized trial comparing two fractionation schedules for patients with localized prostate cancer. J Clin Oncol 2005;23:6132-6138. 5. Yeoh EE, Botten RJ, Butters J, et al. Hypofractionated versus conventionally fractionated radiotherapy for prostate carcinoma: final results of phase III randomized trial. Int J Radiat Oncol Biol Phys 2011;81:1271-1278. 6. Kupelian PA, Willoughby TR, Reddy CA. Hypofractionated intensitymodulated radiotherapy (70 Gy at 2.5 Gy per fraction) for localized prostate cancer. Cleveland Clinic experience. Int J Radiat Oncol Biol Phys 2007;63:1424-1430. 7. Soete G, Arcangeli S, De Meerleer G, et al. Phase II study of a four week hypofractionated external beam radiotherapy regimen for prostate cancer: report on acute toxicity. Radiother Oncol 2006;80:78-81. 8. Pollack A, Hanlon AL, Horwitz ER, et al. Dosimetry and preliminary acute toxicity in the first 100 men treated for prostate cancer on a randomized hypofractionation dose escalation trial. Int J Radiat Oncol Biol Phys 2006;64:518-526. 9. Leborgne F, Fowler J. Late outcomes following hypofractionated conformal radiotherapy vs. standard fractionation for localized prostate cancer: a non randomized contemporary comparison. Int J Radiat Oncol Biol Phys 2009;74:1441-1446. 10. Strigari L, Arcangeli G, Arcangeli S, et al. Mathematical model for evaluating incidence of acute rectal toxicity during conventional or hypofractionated radiotherapy courses for prostate cancer. Int J Radiat Oncol Biol Phys 2009;73:1454-1460. 11. Arcangeli S, Strigari L, Soete G, et al. Clinical and dosimetric predictors of acute toxicity after 3-week hypofractionated external beam radiotherapy regimen for prostate cancer: results from a multicentric prospective trial. Int J Radiat Oncol Biol Phys 2009;73:39-45. 12. Arcangeli G, Fowler J, Gomellini S, et al. Acute and late toxicity in a randomized trial of conventional versus hypofractionated threedimensional conformal radiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 2011;79:2013-2021. 13. Arcangeli G, Saracino B, Gomellini S, et al. A prospective phase III randomized trial of hypofractionation versus conventional fractionation in patients with high-risk prostate cancer. Int J Radiat Oncol Biol Phys 2010;78:11-18. 14. Roach M III, Hanks G, Thames H Jr., et al. Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: recommendations of the RTOG-ASTRO Phoenix Consensus Conference. Int J Radiat Oncol Biol Phys 2006;65:965-974. 15. Zapatero A, Valca´rcel F, Calvo FA, et al. Risk-adapted androgen deprivation and escalated three-dimensional conformal radiotherapy for prostate cancer: does radiation dose influence outcome of patients treated with adjuvant androgen deprivation? A GICOR study. J Clin Oncol 2005;23:6561-6568. 16. Dearnaley DP, Sydes MR, Graham JD, et al. Escalated-dose versus standard-dose conformal radiotherapy in prostate cancer: first results from the MRC RT01 randomized controlled trial. Lancet Oncol 2007; 8:475-487. 17. Crook JM, Choan E, Perry GA, et al. Serum prostate-specific antigen profile following radiotherapy for prostate cancer: implications for patterns of failure and definition of cure. Urology 1998;51:566-572. 18. Brenner DJ, Hall EJ. Fractionation and protraction for radiotherapy of prostate carcinoma. Int J Radiat Oncol Biol Phys 1999;43:1095-1101. 19. Fowler JF, Chapel RJ, Ritter MA. Is the a/b ratio for prostate tumors really low? Int J Radiat Oncol Biol Phys 2001;50:1021-1031. 20. Thames HD, Kuban D, Levy LB, et al. The role of overall treatment time in the outcome of radiotherapy of prostate cancer: an analysis of biochemical failure in 4839 men treated between 1987 and 1995. Radiother Oncol 2010;96:6-12.
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Genitourinary Cancer
Updated Results and Patterns of Failure in a Randomized Hypofractionation Trial for High-risk Prostate Cancer Stefano Arcangeli, MD,* Lidia Strigari, PhD,y Sara Gomellini, MD,* Biancamaria Saracino, MD,* Maria Grazia Petrongari, MD,* Paola Pinnaro`, MD,* Valentina Pinzi, MD,* and Giorgio Arcangeli, MD* *Department of Radiation Oncology and yLaboratory of Medical Physics and Expert Systems, Regina Elena National Cancer Institute, Rome, Italy Received Dec 30, 2011, and in revised form Feb 21, 2012. Accepted for publication Feb 21, 2012
Summary This article reports the updated results of a randomized study comparing an escalated conventional fractionation with a hypofractionation schedule in localized, high-risk prostate cancer. The 2 arms were hypothesized to be isoeffective with regard to tumor control, assuming a fairly low a/b ratio of 1.5-1.8 Gy. After median follow-up of 70 months, there were no statistically significant differences between hypofractionation and conventional fractionation in the actuarial freedom from all types of failure (biochemical, local, and distant).
Purpose: To report long-term results and patterns of failure after conventional and hypofractionated radiation therapy in high-risk prostate cancer. Methods and Materials: This randomized phase III trial compared conventional fractionation (80 Gy at 2 Gy per fraction in 8 weeks) vs hypofractionation (62 Gy at 3.1 Gy per fraction in 5 weeks) in combination with 9-month androgen deprivation therapy in 168 patients with high-risk prostate cancer. Freedom from biochemical failure (FFBF), freedom from local failure (FFLF), and freedom from distant failure (FFDF) were analyzed. Results: In a median follow-up of 70 months, biochemical failure (BF) occurred in 35 of the 168 patients (21%) in the study. Among these 35 patients, local failure (LF) only was detected in 11 (31%), distant failure (DF) only in 16 (46%), and both LF and DF in 6 (17%). In 2 patients (6%) BF has not yet been clinically detected. The risk reduction by hypofractionation was significant in BF (10.3%) but not in LF and DF. We found that hypofractionation, with respect to conventional fractionation, determined only an insignificant increase in the actuarial FFBF but no difference in FFLF and FFDF, when considering the entire group of patients. However, an increase in the 5-year rates in all 3 endpointsdFFBF, FFLF, and FFDFdwas observed in the subgroup of patients with a pretreatment prostate-specific antigen (iPSA) level of 20 ng/mL or less. On multivariate analysis, the type of fractionation, iPSA level, Gleason score of 4þ3 or higher, and T stage of 2c or higher have been confirmed as independent prognostic factors for BF. High iPSA levels and Gleason score of 4þ3 or higher were also significantly associated with an increased risk of DF, whereas T stage of 2c or higher was the only independent variable for LF. Conclusion: Our results confirm the isoeffectiveness of the 2 fractionation schedules used in this study, although a benefit in favor of hypofractionation cannot be excluded in the subgroup of patients with an iPSA level of 20 ng/mL or less. The a/b ratio might be more appropriately evaluated by FFLF than FFBF results, at least in high-risk disease. Ó 2012 Elsevier Inc.
Reprint requests to: Giorgio Arcangeli, MD, Department of Radiation Oncology, Regina Elena Cancer Institute, Via Elio Chianesi 53, 00144
Int J Radiation Oncol Biol Phys, Vol. 84, No. 5, pp. 1172e1178, 2012 0360-3016/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ijrobp.2012.02.049
Rome, Italy. Tel: þ39 0652665602; Fax: þ39 0652665667; E-mail: [email protected] Conflict of interest: none.
Volume 84 Number 5 2012
Hypofractionation vs conventional fractionation 1173
Introduction New advanced radiation therapy techniques have led to an improvement in conformality of high radiation doses to the target tissues and to novel approaches for prostate cancer treatment. These strategies include dose escalation, as well as hypofractionation. Dose escalation, which increases the total dose, has improved the freedom from failure without further complications (1-3). Hypofractionation allows one to obtain an equivalent biological effective dose using a dose per fraction higher than the conventional dose (4, 5); the clinical experience on hypofractionation is still limited, but some studies have reported promising results on biochemical control with acceptable late toxicity (6-11). We have previously reported the results of a randomized trial comparing a total escalated dose (80 Gy) by conventional fractionation with a hypofractionation schedule using 3-dimensional conformal radiation therapy (3D-CRT) in localized high-risk prostate cancer. The early results of this trial have shown a significant biochemical benefit of hypofractionation over conventional fractionation schedules without increasing the side effects (12, 13), thus supporting the radiobiological assumptions of high sensitivity of prostate tumors to large dose fractions. However, biochemical improvement does not necessarily mean better local tumor control because prostate-specific antigen (PSA) failure does not indicate whether failure was local or distant, the latter being a frequent event in the high-risk disease group, where early micrometastatic spread may have already occurred by the time of local treatment. Nevertheless, for this biochemical improvement with hypofractionation to translate into better clinical results, a longer follow-up would be required. In fact, although the PSA endpoint is presently considered an early and useful surrogate for treatment efficacy, the relationship and timeline between PSA failure, clinical failure, and survival need to be accurately documented for an optimal estimation of the treatment benefit. Patients in our study are presently approaching a median follow-up of 70 months, and the data on clinical failure have now matured, enabling the examination and comparison of the failure pattern in the 2 fractionation groups, as well as in subgroups of patients with different prognostic factors.
Methods and Materials
Simulation and treatment The simulation and treatment procedures used have been extensively reported in a previous article (13). In brief, all patients received 9-month androgen deprivation therapy consisting of a daily oral administration of 50 mg of bicalutamide for 270 days, as well as a quarterly subcutaneous injection of a luteinizing hormone-releasing hormone analog depot. The 3D-CRT started at day 67 from the first bicalutamide intake. The clinical target volume, which included the prostate and seminal vesicles, was expanded in 3 dimensions with a 1-cm margin to obtain the planning target volume, except at the prostate-rectal interface, where a 0.6-cm margin was used to decrease rectal involvement. All patients were treated with 15-MV photons by a Varian 2100 linear accelerator (Varian Medical Systems, Palo Alto, CA) equipped with a 120-leaf collimator. The prescription dose to 95% of the clinical target volume was 80 Gy in 2-Gy fractions and 62 Gy in 3.1-Gy fractions for the conventional and hypofractionation arms, respectively.
Study endpoints and statistical analysis
Study design and protocol eligibility This phase III, single-institution, randomized trial was designed to compare 2 dose fractionation schedules using 3D-CRT in patients Table 1
with localized, high-risk prostate cancer. Between 2002 and 2005, 168 patients were randomized to receive 80 Gy in 40 fractions within 8 weeks at 2.0 Gy per fraction (arm A, conventional fractionation) vs 62 Gy in 20 fractions within 5 weeks (4 fractions per week) at 3.1 Gy per fraction (arm B, hypofractionation) to the prostate and seminal vesicles. We tested the hypothesis that hypofractionation would result in a lower rate of late complications and would be just as effective as conventional fractionation. In brief, the 2 arms were hypothesized to be isoeffective with regard to tumor control, assuming a fairly low a/b ratio of 1.5-1.8 Gy. However, as regards the late complications, assuming an a/ b value of 3 Gy for normal tissue, the total dose by hypofractionation should be equivalent to approximately 75 Gy given at 2 Gy per fraction, which is considerably lower than 80 Gy by conventional fractionation. To achieve 80% power, it was estimated that a total of 168 patients were required to detect a significant difference (P<.05, 2- sided). The study design and inclusion criteria have been described previously (13). Patient characteristics are reported in Table 1.
The freedom from biochemical failure (FFBF) was defined as the time interval from the first day of radiation therapy to the biochemical relapse (biochemical failure [BF]) according to the Phoenix definition of nadir PSA plus 2 ng/mL (14). The secondary
Patient characteristics after conventional treatment and hypofractionated treatment Conventional fractionation
No. of patients Age [median (range)] (yr) Biopsy-derived Gleason score (3þ4) (4þ3) cT stage
Hypofractionation
85 75 (54-83)
83 75 (61-82)
41 (48%) 44 (52%)
42 (51%) 41 (49%)
P value (t test) .8647 .8788
.3222 48 (56%) 37 (44%) 19/13 (2-128)
Abbreviation: iPSA Z pretreatment prostate-specific antigen.
54 (65%) 29 (35%) 26/16 (4-157)
.0627
1174 Arcangeli et al.
International Journal of Radiation Oncology Biology Physics
Fig. 1. Forest plot of hazard ratios (HRs) and confidence intervals (CIs) for biochemical (A), local (B), and distant (C) failures stratified for prognostic factors. CI Z confidence interval; Conv Z conventional arm; Evts Z events; GS Z Gleason score; HR Z hazard ratio; Hypo Z hypofractionation arm; iPSA Z pretreatment prostate-specific antigen; pts Z patients.
objectives included local failure (LF), distant failure (DF), overall survival (OS), and cancer-specific survival (CSS). LF was defined as any evidence of disease documented by positron emission tomography (PET) with 18F-choline, by dynamic contrastenhanced magnetic resonance imaging, or by biopsy in patients with rising serum PSA level and/or palpable abnormality. DF was defined as any evidence of disease in extrapelvic nodes or other organs (bone, lung, liver, and so on) documented by bone scan, PET with 18F-choline, magnetic resonance imaging, or computed tomography (CT) imaging. OS was defined as death from any cause
and CSS as death from prostate cancer or death at the time of progressive metastatic disease. Kaplan-Meier actuarial analysis was used to determine the freedom from failure and survival. For all measured endpoints, patients were censored at the time of the specific event (BF, LF, DF, or survival). Comparison between the actuarial curves was made by use of the log-rank test. Statistical significance between the groups was tested by the c2 test and t test for categorical and continuous variables, respectively. A Cox proportional hazards model was applied for multivariate analysis and for hazard ratio
Volume 84 Number 5 2012
Hypofractionation vs conventional fractionation 1175 (HR) estimation. The HR and 95% confidence interval (CI) were also shown by forest plot. The procedures followed were in accordance with the ethical standards of the institutional responsible committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2000.
Results No difference was found in the incidence and severity of late gastrointestinal or genitourinary toxicity between the 2 fractionation schedules. The results of acute and late toxicity have been reported in detail in a previous article (12).
Patterns of failure The patterns of failure for the 2 fractionation arms are shown in Fig. 1. At a median follow-up of 70 months, BF occurred in 35 of the 168 patients (21%) in the study, 22 of 85 (26%) in the conventional group and 13 of 83 (15%) in the hypofractionation group (Fig. 1A). LF was detected in 10 patients (12%) and 7 patients (8%), respectively (Fig. 1B). DF was detected in 13 of 85 patients (15%) and 9 of 83 patients (5%), respectively (Fig. 1C). In 2 patients in the conventional group and 4 patients in the hypofractionation group, LF and DF were detected simultaneously. LF was detected by PET-choline images as an intense, persistent uptake in the prostate. In 5 patients LF was confirmed by a positive biopsy specimen. Biopsy findings were negative in 8 patients, and 4 patients refused to undergo biopsy. In 5 patients with negative biopsy findings, LF was confirmed by magnetic resonance imaging. The significance of the differences between the 2 fractionation schedules for all patients and subgroups of patients stratified for different levels of prognostic factors can be appreciated by the forest plots of the HRs and 95% CIs displayed in Fig. 1. An HR of less than 1 indicates a decreased risk of failure in the hypofractionation schedule. The risk reduction of 10.2% in BF was statistically significant for the whole group (HR, 0.34; 95% CI, 0.21-0.56); the reduction was significant in the subgroups of patients with either PSA level of 20 ng/mL or less (HR, 0.15; 95%
Fig. 2. Freedom from biochemical (a), local (b), and distant (c) failure in the 2 treatment groups: conventional arm (A, dashed line) and hypofractionation arm (B, solid line). Pts Z patients.
Table 2
Univariate analysis of failures 5-y freedom-from-failure rate Biochemical
Prognostic factor iPSA 20 ng/mL >20 ng/mL Gleason score (3þ4) (4þ3) T stage <2c 2c All patients
Local
Distant
Conv
Hypo
P value
Conv
Hypo
P value
Conv
Hypo
P value
83 72
95 73
.02 .39
92 91
100 85
.01 .51
87 84
98 81
.04 .96
94 66
88 83
.50 .01
97 87
95 91
.41 .13
98 76
91 89
.24 .08
85 74 79
89 83 85
.30 .08 .065
93 90 91
100 89 93
.05 .92 .33
91 82 86
89 91 90
.86 .19 .29
Abbreviations: Conv Z conventional fractionation; Hypo Z hypofractionation; iPSA Z pretreatment prostate-specific antigen.
International Journal of Radiation Oncology Biology Physics
1176 Arcangeli et al. Table 3
Multivariate Cox proportional hazards analysis of biochemical failure Variable
b
SE
P value
Exp(b)
Lower 95% confidence limit
Upper 95% confidence limit
Hypofractionation (yes/no) iPSA (continuous) (ng/mL) Gleason score (4þ3) (yes/no) T stage 2c (yes/no)
0.9145 0.0248 0.9117 0.9486
0.3607 0.0053 0.3699 0.3909
.0112 .0000 .0137 .0152
0.4007 1.0251 2.4885 2.5822
0.1983 1.0146 1.2097 1.2049
0.8097 1.0358 5.1190 5.5338
Abbreviation: iPSA Z pretreatment prostate-specific antigen. The P value of the model is <.0001.
CI, 0.03-0.71) or Gleason score of 4þ3 or higher (HR, 0.09; 95% CI, 0.03-0.31) (Fig. 1A). However, no significant difference was detected in LF between hypofractionation and conventional fractionation, either in all patients or the subgroups of patients stratified for prognostic factors (Fig. 1B). The risk reduction in DF in favor of hypofractionation was marginally significant (HR, 0.57; 95% CI, 0.33-0.98) for the whole group and significant in the subgroup of patients with a Gleason score of 4þ3 or higher (Fig. 1C). The actuarial analysis of failures is shown in Fig. 2. Contrary to our previous report (13), the actuarial analysis of FFBF, after a 70-month median follow-up, showed no statistically significant difference between the hypofractionation and conventional fractionation arms. At 5 years, 85% of patients in the former group were free from BF as compared with 79% of patients in the conventional fractionation group (PZ.065) (Fig. 2a). Furthermore, no differences between the 2 fractionation arms were detected in the actuarial freedom from local failure (FFLF) (5-year rate of 93% vs 91%) and freedom from distant failure (FFDF) (5-year rate of 90% vs 86%) (Figs. 2b and c). The actuarial 5-year rates of freedom from failure for all patients and each prognostic subgroup of patients are shown in Table 2. No significant difference in any type of failure between the 2 treatment arms was detected in any of the prognostic subgroups, with the exception of the subgroup of patients with a pretreatment prostatespecific antigen (iPSA) level of 20 ng/mL or less, in whom hypofractionation was significantly better than conventional fractionation in terms of the 3 failure outcomes (FFBF, 95% vs 83 [PZ.02]; FFLF, 100% vs 92% [PZ.01]; and FFDF, 98% vs 87% [PZ.04]), as well as the subgroup with a Gleason score of 4þ3 or higher, in which the difference in favor of hypofractionation was detected only for FFBF (83% vs 66%, PZ.01). On multivariate Cox proportional hazards analysis, the type of fractionation, iPSA level (taken as a continuous variable), Gleason score of 4þ3 or higher, and T stage of 2c or higher have been confirmed as independent prognostic factors for BF (Table 3). T stage of 2c or higher was the only independent variable for LF (Table 4), whereas higher iPSA levels and a Gleason score of 4þ3 or higher were also significantly associated with an increased risk of DF (Table 5).
Table 4
Cancer death and survival Prostate cancer death among the 35 patients with BF is shown in Fig. 3. The timeline from treatment to biochemical and clinical failure and then to prostate cancer death is reported for each patient. Death from prostate cancer occurred in 8 patients, and DF developed in 22 patients. In 16 of these 22 patients (73%), metastases occurred without a documented LF, with a median detection time of 29 months. In patients who had distant metastases as the first clinical failure, death occurred within 4.3 years from the initiation of treatment, with a median time to death of 38 months. The actuarial analysis of survival is shown in Fig. 4. After a 70-month median follow-up, no significant difference was detected between the 2 fractionation arms, either in OS (Fig. 4a) or in CSS (Fig. 4b), with 5-year rates of hypofractionation vs standard fractionation of 92% vs 82% and of 98% vs 92% for OS and CSS, respectively.
Discussion At present, only 2 randomized studies comparing radiation schedules with different fraction sizes in patients with prostate cancer have been published (4, 5). Both of these trials, designed and performed with no attempt to make the arms isoeffective, provided similar outcomes and toxicity rates between the conventional and hypofractionation schedules used. However, the total dose delivered in these trials is now considered insufficient in the treatment of prostate cancer, and therefore no definitive conclusion can be derived with regard to either outcome or toxicity. To our knowledge, this is the first randomized trial reporting the results of dose-escalated conventional fractionation in comparison to an equivalent hypofractionation schedule in patients with high-risk prostate cancer. The main hypothesis of the trial was that the FFBF rate of the hypofractionation schedule was not inferior to that of the conventional fractionation schedule. The early results of our study showed a better, though only marginally significant, incidence of FFBF in the hypofractionation
Multivariate Cox proportional hazards analysis of local relapse
Variable
b
SE
P value
Exp (b)
Lower 95% confidence limit
Upper 95% confidence limit
T stage 2c (yes/no)
1.1400
0.5837
.0508
3.1267
1.0018
9.7583
The P value of the model is .0349. Variables not included in the model were as follows: hypofractionation (yes/no), pretreatment prostate-specific antigen (continuous, in nanograms per milliliter), and Gleason score of 4þ3 or higher (yes/no).
Volume 84 Number 5 2012 Table 5
Hypofractionation vs conventional fractionation 1177
Multivariate Cox proportional hazards analysis of distant failure Variable
b
SE
P value
Exp (b)
Lower 95% confidence limit
Upper 95% confidence limit
iPSA (continuous) (ng/mL) Gleason score (4þ3) (yes/no)
0.0182 1.3200
0.0059 0.5107
.0020 .0097
1.0184 3.7436
1.0068 1.3829
1.0301 10.1340
Abbreviation: iPSA Z pretreatment prostate-specific antigen. The P value of the model is .0010. Variables not included in the model were as follows: hypofractionation (yes/no) and T stage of 2c or higher (yes/no).
arm compared with the conventional fractionation arm (13). The 3-year FFBF rate in the conventional fractionation arm of our study compared favorably with rates reported in the high-dose arm of dose-escalation randomized trials of radiation therapy alone (1-3) or in combination with androgen suppression (15, 16) for high-risk patients. However, even though PSA failure is a strong indicator of disease status (17), when one is assessing local treatment efficacy, LF should also be monitored because it is often only detected clinically several months after BF has occurred. Because of the short follow-up, the pattern of clinical failure could not be previously shown in the early results from this trial. The updated results reported in this study at a median followup of 70 months show that, although there is still a trend in favor of hypofractionation, the significance of the difference in FFBF between the 2 fractionation schedules was lost. No difference in FFLF was observed between the 2 groups, confirming the equivalence of the 2 fractionation schedules.
Fig. 3.
Although DF developed in 16 of 22 patients (73%) in a median time of 29 months without a documented LF, despite the concomitant/adjuvant association with androgen deprivation therapy given to all patients to minimize the risk of a bias due to early micrometastases in high-risk disease, no difference between the 2 groups could be detected in FFDF or in OS and CSS. However, the univariate subgroup analysis showed that in the subgroup of patients with an iPSA level of 20 ng/mL or less, all 3 failure endpointsdFFBF, FFLF, and FFDFdwere significantly improved by hypofractionation with respect to conventional fractionation. This might be because of the better local effect of hypofractionation compared with standard fractionation on a smaller tumor burden in such patients, as well as the achievement of local tumor control before the occurrence of metastatic spread. Our study shows that, in patients with high-risk prostate cancer, similar clinical local control can be achieved by high-dose
Clinical failures and prostate cancer death in patients with biochemical failures.
1178 Arcangeli et al.
Fig. 4. Overall survival (OS) (a) and cancer-specific survival (CSS) (b) in conventional arm (A, dashed line) and hypofractionation arm (B, solid line). Pts Z patients.
conventional fractionation and hypofractionation schedules, calculated to be equivalent by use of an a/b ratio of 1.5-1.8 Gy. Although in high-risk disease the events are more likely to occur earlier with respect to low- and intermediate-risk disease, we cannot exclude the possibility that some additional LFs may still be detected later. However, we consider that any other events that may be detected in a longer follow-up will likely be equally distributed between the 2 treatment groups and will not alter the isoeffectiveness of the 2 fractionation schedules. Whether this equivalence is because of a low a/b ratio (18, 19) or, as recently suggested, the reduced overall treatment time by a hypofractionation schedule (20), or both, will be a matter of debate, and the findings of this study may contribute to shedding some light on this debate. A suggestion that can be derived from the results of our study is that the a/b ratio should be more appropriately evaluated by long-term results of FFLF rather than those of FFBF, which could be affected, at least in high-risk disease, by the presence of occult metastases before treatment.
References 1. Pollack A, Zagars GK, Starkschall G, et al. Prostate cancer radiation dose response: results of the M.D. Anderson Phase III randomized trial. Int J Radiat Oncol Biol Phys 2002;53:1097-1105. 2. Peeters STH, Heemsbergen WD, Koper PCM, et al. Dose-response in radiotherapy for localized prostate cancer: results of the Dutch multicenter randomized Phase III trial comparing 68 Gy of radiotherapy with 78 Gy. J Clin Oncol 2006;24:1990-1996.
International Journal of Radiation Oncology Biology Physics 3. Zietman AL, Bae K, Slater JD, et al. Randomized trial comparing conventional-dose with high-dose conformal radiation therapy in early-stage adenocarcinoma of the prostate: long-term results from Proton Radiation Oncology Group/American College of Radiology 95-09. J Clin Oncol 2010;28:1106-1111. 4. Lukka H, Hayter C, Julian JA, et al. A randomized trial comparing two fractionation schedules for patients with localized prostate cancer. J Clin Oncol 2005;23:6132-6138. 5. Yeoh EE, Botten RJ, Butters J, et al. Hypofractionated versus conventionally fractionated radiotherapy for prostate carcinoma: final results of phase III randomized trial. Int J Radiat Oncol Biol Phys 2011;81:1271-1278. 6. Kupelian PA, Willoughby TR, Reddy CA. Hypofractionated intensitymodulated radiotherapy (70 Gy at 2.5 Gy per fraction) for localized prostate cancer. Cleveland Clinic experience. Int J Radiat Oncol Biol Phys 2007;63:1424-1430. 7. Soete G, Arcangeli S, De Meerleer G, et al. Phase II study of a four week hypofractionated external beam radiotherapy regimen for prostate cancer: report on acute toxicity. Radiother Oncol 2006;80:78-81. 8. Pollack A, Hanlon AL, Horwitz ER, et al. Dosimetry and preliminary acute toxicity in the first 100 men treated for prostate cancer on a randomized hypofractionation dose escalation trial. Int J Radiat Oncol Biol Phys 2006;64:518-526. 9. Leborgne F, Fowler J. Late outcomes following hypofractionated conformal radiotherapy vs. standard fractionation for localized prostate cancer: a non randomized contemporary comparison. Int J Radiat Oncol Biol Phys 2009;74:1441-1446. 10. Strigari L, Arcangeli G, Arcangeli S, et al. Mathematical model for evaluating incidence of acute rectal toxicity during conventional or hypofractionated radiotherapy courses for prostate cancer. Int J Radiat Oncol Biol Phys 2009;73:1454-1460. 11. Arcangeli S, Strigari L, Soete G, et al. Clinical and dosimetric predictors of acute toxicity after 3-week hypofractionated external beam radiotherapy regimen for prostate cancer: results from a multicentric prospective trial. Int J Radiat Oncol Biol Phys 2009;73:39-45. 12. Arcangeli G, Fowler J, Gomellini S, et al. Acute and late toxicity in a randomized trial of conventional versus hypofractionated threedimensional conformal radiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 2011;79:2013-2021. 13. Arcangeli G, Saracino B, Gomellini S, et al. A prospective phase III randomized trial of hypofractionation versus conventional fractionation in patients with high-risk prostate cancer. Int J Radiat Oncol Biol Phys 2010;78:11-18. 14. Roach M III, Hanks G, Thames H Jr., et al. Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: recommendations of the RTOG-ASTRO Phoenix Consensus Conference. Int J Radiat Oncol Biol Phys 2006;65:965-974. 15. Zapatero A, Valca´rcel F, Calvo FA, et al. Risk-adapted androgen deprivation and escalated three-dimensional conformal radiotherapy for prostate cancer: does radiation dose influence outcome of patients treated with adjuvant androgen deprivation? A GICOR study. J Clin Oncol 2005;23:6561-6568. 16. Dearnaley DP, Sydes MR, Graham JD, et al. Escalated-dose versus standard-dose conformal radiotherapy in prostate cancer: first results from the MRC RT01 randomized controlled trial. Lancet Oncol 2007; 8:475-487. 17. Crook JM, Choan E, Perry GA, et al. Serum prostate-specific antigen profile following radiotherapy for prostate cancer: implications for patterns of failure and definition of cure. Urology 1998;51:566-572. 18. Brenner DJ, Hall EJ. Fractionation and protraction for radiotherapy of prostate carcinoma. Int J Radiat Oncol Biol Phys 1999;43:1095-1101. 19. Fowler JF, Chapel RJ, Ritter MA. Is the a/b ratio for prostate tumors really low? Int J Radiat Oncol Biol Phys 2001;50:1021-1031. 20. Thames HD, Kuban D, Levy LB, et al. The role of overall treatment time in the outcome of radiotherapy of prostate cancer: an analysis of biochemical failure in 4839 men treated between 1987 and 1995. Radiother Oncol 2010;96:6-12.