Long-term Outcomes from a Phase 1 Dose Escalation Study Using Stereotactic Body Radiotherapy for Patients with Low- or Intermediate-risk Prostate Cancer

Abstract

Background:

Ultrahypofractionated stereotactic body radiation therapy (SBRT) has become a standard treatment intervention for localized prostate cancer.

Objective:

To report final long-term tumor control outcomes and late gastrointestinal (GI) and genitourinary (GU) toxicities from a single-center phase 1 dose escalation study using SBRT for patients with low- or intermediate-risk prostate cancer.

Design, setting and participants:

Between 2009 and 2012, 136 patients were enrolled and treated. The initial dose level was 32.5 Gy in five fractions. Doses were then sequentially escalated to 35 Gy, 37.5 Gy, and 40 Gy in five fractions delivered every other day.

Outcome measurements and statistical analysis:

The primary endpoint was late treatment-related toxicity. Secondary endpoints included prostate-specific antigen (PSA) failure.

Results and limitations:

The median follow-up was 10.5 yr for the 32.5-Gy group, 9.9 yr for the 35-Gy group, 8.2 yr for the 37.5-Gy group, and 7.3 yr for the 40-Gy group. The 8-yr cumulative incidence of PSA failure was 26% for 32.5 Gy, 15% for 35 Gy, 3.4% for 37.5 Gy, and 6.6% for 40 Gy. Higher radiation dose (37.5–40 Gy) and favorable intermediate risk (vs unfavorable intermediate risk) were associated with better PSA recurrence rates (p = 0.011 and 0.002, respectively). The 8-yr actuarial probability rates for survival free from late grade ≥2 toxicity were 94% for GI toxicity and 86% for GU toxicity. No grade 4 events were recorded. Higher dose levels were not associated with higher rates of late grade ≥2 GI (p = 0.2) or GU (p > 0.9) toxicity.

Conclusions:

SBRT doses ranging from 32.5 to 40 Gy were associated with low incidence of moderate or severe toxicities. Higher doses resulted in superior disease control outcomes 8 yr after treatment.

Patient summary:

We investigated the association between the radiotherapy dose used and the rate of control of prostate cancer. We found that higher doses resulted in more favorable outcomes without excess toxicity.

This trial is registered on ClinicalTrials.gov as NCT00911118.

1. Introduction

Ultrahypofractionated stereotactic body radiation therapy (SBRT) has become a standard treatment intervention for localized prostate cancer. A growing body of evidence demonstrates excellent long-term prostate-specific antigen (PSA) recurrence–free survival (RFS) rates of 95–100% for low-risk, 85–100% for intermediate-risk, and 80% for high-risk disease with this treatment approach [1–6]. While cumulative rates of grade ≥2 genitourinary (GU) toxicities of 15–33% have been reported, most events are transient. Grade 3 complications have been infrequent, and grade 4 events are rare [1–4].

Two trials have reported early results for conventional or moderately hypofractionated regimens in comparison to ultrahypofractionation. In PACE-B, patients with low- or intermediate-risk prostate cancer were randomized to 78 Gy in 39 fractions or 62 Gy in 20 fractions versus SBRT at a dose of 36.25 Gy in five fractions. Both acute and 2-yr grade ≥2 toxicity rates (Radiation Therapy Oncology Group [RTOG] classification) were similar between the treatment groups. However, the rate of grade ≥2 GU toxicity events according to the Common Terminology Criteria for Adverse Events (CTCAE) criteria at 2 yr was higher with SBRT [7,8]. In the HYPO-RT-PC trial, patients diagnosed with intermediate- or high-risk disease were randomized to 78 Gy in 39 fractions or 42.7 Gy in seven fractions without the use of androgen deprivation therapy (ADT). The estimated 5-yr RFS rate was 84% in both groups. There was a trend for higher frequency of acute physician-reported RTOG grade ≥2 GU toxicity at treatment completion in the SBRT group. A higher rate of late GU toxicity among patients treated with SBRT was only detected at 1 yr after treatment (6% vs 2%). There were higher rates of patient-reported acute GU and GI symptoms in the SBRT group, but there were no significant differences in late symptoms, except for GU toxicity at 1 yr, similar to physician-evaluated toxicity [4].

We have previously published 5-yr results from a phase 1 trial of dose-escalated SBRT to the prostate, initiated at a prescription dose of 32.5 Gy in five fractions. Subsequent cohorts were treated using 2.5-Gy increments up to a dose of 40 Gy. The primary endpoint was toxicity, and secondary endpoints included PSA relapse and local tumor control according to prostate biopsies performed 2 yr after treatment. We observed higher incidence of late grade 2 GU toxicities in the 40-Gy cohort that gradually resolved. Only one late grade 3 GU event was recorded. Sequential dose tiers were significantly associated with better biopsy outcomes at 2 yr after treatment [9]. However, the low overall rate of biochemical failure precluded comparison between dose levels. The purpose of this extended follow-up is to report long-term outcomes and toxicities for this prostate SBRT dose escalation trial for the first time.

2. Patients and methods

This was a single-center prospective SBRT dose escalation trial for patients with low- or intermediate-risk prostate cancer. The original publication reported on the trial procedures and analyses in detail [9], so only a brief synopsis is included here.

Patients with a Karnofsky performance status score of ≥70 and prostate size ≤60 g who were diagnosed with low- or intermediate-risk prostate cancer were eligible for the study. Patients treated with neoadjuvant ADT or who had a baseline International Prostate Symptom Score of >15, a history of transurethral resection of the prostate, or a history of inflammatory bowel disease were excluded.

Between 2009 and 2012, 136 patients were enrolled and underwent computed tomography–based simulations. The clinical target volume (CTV) included the prostate and seminal vesicles. The planning target volume (PTV) was defined as a 5-mm margin around the CTV except for a 3-mm margin at the prostate-rectum interface. Doses were prescribed to the 100% isodose line to encompass the PTV with a maximum dose <110%. Dose constraints used in the study are listed in Table 1. Patients were immobilized using an Aquaplast device (Qfix, Avondale, PA, USA) and were treated with intensity-modulated external beam radiation therapy in the supine position using multiple coplanar fields with 6- or 15-MV photons. Image guidance was based on fiducial markers, Calypso beacons (Varian Medical Systems, Palo Alto, CA, USA), and intrafractional imaging. Post-treatment biopsies were recommended at 2 yr after completion of SBRT.

Table 1 –

Planning dose constraints for normal tissues

Dose volume for normal tissue a	Constraint
Rectal wall maximum	103% of Rx
Rectal wall dose (1 cm3)	38.5 Gy
Rectal wall 53% dose	24 Gy
Rectum volume (30.15 Gy)	8 cm3
Bladder wall maximum	105% of Rx
Bladder wall D (1 cm3)	42.00 Gy
Bladder wall 53% dose	24.00 Gy

Rx = prescription

^aThe rectal wall was defined as the outer circumference of the rectum minus the inner lumen extending 5 mm both superiorly and inferiorly from the planning target volume. The bladder wall was defined as the outer circumference of the visible bladder minus the inner lumen with a presumed wall thickness of 6 mm.

2.1. Statistical considerations

All endpoints were calculated from the time of SBRT completion, including overall survival (OS) until death from any cause and PSA failure. PSA failure was defined as the nadir + 2 ng/ml (Phoenix criterion) or as clinical scenarios with rising PSA and clinical evidence of disease progression on imaging or post-treatment biopsy without meeting this threshold. OS estimates and times to GI and GU toxicity were calculated using Kaplan-Meier curves, with subgroups compared using log-rank tests. PSA failure was estimated using cumulative incidence curves with death without PSA failure as a competing risk, and results were compared using Gray’s test. All patients who did not develop PSA failure or experience death were censored at the date of last follow-up. PSA trends were calculated for individual patients and correlated with dose tiers and analyzed both as relative changes in comparison to baseline and as the slope of the PSA change at 2 and 3 yr. The relative change in PSA and the slope of the PSA change were dichotomized according to their respective median values and used as stratifiers for Kaplan-Meier curves for OS and cumulative incidence curves for PSA failure, using landmark analyses. The Cochran-Armitage test was used to test for trends for categorical variables across the ordinal dose levels. The two lower-dose tiers (32.5 and 35 Gy) were merged as a “low dose” (LD) cohort and compared with the two higher-dose tiers (37.5 and 40 Gy), comprising the “high dose” (HD) cohort in several analyses. Toxicity was classified according to CTCAE v4. Considering the longer follow-up for the lower-dose tiers, all data past 10 yr of follow-up were censored. All statistical analyses were conducted using R v4.2.2 (R Foundation for Statistical Computing, Vienna, Austria), with a type I error rate (α) of 0.05.

3. Results

The median follow-up for survivors in the entire cohort and the 4 dose levels was 8.1 yr (interquartile range [IQR] 7.8–8.3) for the overall cohort, 10.5 yr (IQR 5.7–11.4) for the 32.5-Gy group, 9.9 yr (IQR 9.2–10.2) for the 35-Gy group, 8.2 yr (IQR 7.9–8.4) for the 37.5-Gy group, and 7.3 yr (IQR 7.2–7.8) for the 40-Gy group. Table 2 summarizes patient characteristics by dose cohort. The median age at enrollment was 67.5 yr and most patients had Gleason 7 disease (61.8%), T1c stage (76.5%), and National Comprehensive Cancer Network favorable intermediate risk (44.1%). There was a strong trend to a higher proportion of patients with Gleason 7 disease in the 40Gy cohort, and a significantly higher percentage of baseline positive biopsy cores.

Table 2 –

Patient characteristics overall and by radiation dose group

Parameter	All patients (n = 136)	32.5 Gy (n = 30)	35 Gy (n = 35)	37.5 Gy (n = 36)	40 Gy (n = 35)	p value
Age at SBRT (yr) a	67.5 (47.0–82.0)	68.0 (52.0–82.0)	67.0 (49.0–79.0)	66.0 (47.0–79.0)	68.0 (49.0–80.0)	0.44
Gleason score, n (%)						0.059
6	52 (38.2)	15 (50)	15 (42.9)	15 (41.7)	7 (20)
7	84 (61.8)	15 (50)	20 (57.1)	21 (58.3)	28 (80)
3 + 4	68 (50)	13 (43.3)	14 (40)	18 (50)	23 (65.7)
4 + 3	16 (11.8)	2 (6.7)	6 (17.1)	3 (8.3)	5 (14.3)
Positive cores (%) a	25 (4−100)	18 (7−63)	17 (6−83)	25 (4−75)	33 (8−100)	0.049
Baseline PSA (ng/ml) a	5.5 (0.3–33.2)	5.5 (0.9–12.8)	5.4 (0.9–33.2)	5.8 (0.3–13.9)	5.2 (1.4–13.0)	0.79
T stage, n (%)						0.58
T1c	104 (76.5)	21 (70)	25 (71.4)	31 (86.1)	27 (77.1)
T2a	20 (14.7)	5 (16.7)	6 (17.1)	4 (11.1)	5 (14.3)
T2b	10 (7.4)	2 (6.7)	4 (11.4)	1 (2.8)	3 (8.6)
T2c	2 (1.5)	2 (6.7)	0 (0)	0 (0)	0 (0)
NCCN risk group, n (%)						0.11
Low	45 (33.1)	15 (50)	11 (31.4)	13 (36.1)	6 (17.1)
Favorable intermediate	60 (44.1)	11 (36.7)	13 (37.1)	17 (47.2)	19 (54.3)
Unfavorable intermediate	31 (22.8)	4 (13.3)	11 (31.4)	6 (16.7)	10 (28.6)
Prostate volume (cm3) a	35.0 (11.6–70.0)	35.3 (19.3–59.0)	30.0 (12.6–58.0)	35.0 (11.6–70.0)	37.1 (16.0–56.0)	0.27

SBRT = stereotactic body radiation therapy; PSA = prostate-specific antigen; NCCN = National Comprehensive Cancer Network.

^aResults presented as median (range).

3.1. Late toxicity outcomes

Table 3 provides a summary of late toxicity events. Overall, seven patients (5.1%) experienced late grade 2 and one patient (0.7%) experienced late grade 3 GI toxicity (rectal bleeding requiring transfusion in a patient treated to 35 Gy). No late grade 4 events were recorded. The 8-yr probability of freedom from late grade ≥2 GI toxicity was 94%. Six patients had complete resolution of their toxicity (75%). As shown in Figure 1A, there was no significant difference in the probability of freedom from late GI toxicity between the LD and HD groups. The 5-yr probability of freedom from GI toxicity was 97% for the LD group and 91% for the HD group (p = 0.2). The incidence of late grade ≥2 GI toxicity was significantly correlated with the incidence of acute grade ≥2 GI toxicity (p = 0.041).

Table 3 –

Late GU and GI toxicity overall and by radiation dose group

Toxicity	Patients, n (%)
	All patients (n = 136)	32.5 Gy (n = 30)	35 Gy (n = 35)	37.5 Gy (n = 36)	40 Gy (n = 35)
GI, grade 2
Incidence	7 (5.1)	0 (0)	1 (2.9)	3 (8.5)	3 (8.6)
Resolved	6/7	NA	1/1	3/3	2/3
GI, grade 3
Incidence	1 (0.7)	0 (0)	1 (2.9)	0 (0)	0 (0)
Resolved	1/1	NA	1/1	NA	NA
GU, grade 2
Incidence	27 (19.9)	5 (16.7)	9 (25.7)	5 (13.9)	8 (22.9)
Resolved	18/27	1/5	7/9	4/5	6/8
GU, grade 3
Incidence	2 (1.5)	0 (0)	0 (0)	1 (2.8)	1 (2.9)
Resolved	2/2	NA	NA	1/1	1/1

GI = gastrointestinal; GU = genitourinary; NA = not applicable.

Fig. 1 – Toxicity-free probability by dose group for (A) gastrointestinal toxicity and (B) genitourinary toxicity.

Overall, 27 patients (19.9%) experienced late grade 2 GU toxicity, and two patients (1.5%) experienced late grade 3 GU toxicity (hematuria requiring cauterization in a patient treated to 37.5 Gy and urethral stricture requiring dilatation in a patient treated to 40 Gy). The 8-yr probability of freedom from late grade ≥2 GU toxicity was 86%. Most late GU toxicities had completely resolved at last follow-up (20/29, 70%). As shown in Figure 1B, there was no significant difference in the probability of freedom from late GU toxicity between the LD and HD groups. The 5-yr probability of freedom from GU toxicity was 93% for the LD group and 94% for the HD group (p > 0.9). The incidence of late grade ≥2 GU toxicity did not significantly correlate with the incidence of acute grade ≥2 GU toxicity (p = 0.085).

3.2. PSA response

The median relative change in PSA in comparison to baseline for the entire cohort was −89% (IQR −94% to −83%) at 2 yr and −91% (IQR −96% to −82%) at 3 yr after completion of SBRT. Increasing dose was associated with an increase in the proportion of patients who had PSA ≤0.5 ng/ml at 2 yr (33% vs 39% vs 54% vs 59%; p = 0.018) and lower relative PSA values in comparison to baseline at 3 yr after treatment (p < 0.001). HD treatment was associated with a more rapid drop in PSA in comparison to LD treatment at 2 yr (p = 0.027) and 3 yr (p < 0.001) and a more profound relative PSA decline at 3 yr versus baseline (p = 0.02; Fig. 2). Patients with a positive 2-yr prostate biopsy had a significantly smaller 2-yr PSA relative change in comparison to patients with a negative biopsy or a biopsy with profound treatment-related effects (p = 0.032).

Fig. 2 – Prostate-specific antigen (PSA) slope at 3 yr by dose group.

3.3. Tumor control outcomes

PSA failure occurred in 17 patients during follow-up, of whom 14 were in the LD group and three were in the HD group. Thirteen patients (76.5%) met the Phoenix criterion for biochemical failure. Four patients with unfavorable intermediate-risk disease and treated to 35 Gy had rising PSA not reaching the nadir + 2 ng/ml threshold, but had clinical evidence of recurrent disease.

The 8-yr estimated PSA failure rates were 26% (95% confidence interval [CI] 10–46%) for 32.5 Gy, 15% (95% CI 5.2–29%) for 35 Gy, 3.4% (95% CI 0.24–15%) for 37.5 Gy, and 6.6% (95% CI 1.1–19%) for 40 Gy. As shown in Figure 3A, the risk of PSA failure was significantly higher in the LD group than in the HD group (p = 0.011). The 8-yr incidence of PSA failure was 19% (95% CI 10–31%) for the LD group and 4.9% (95% CI 1.3–13%) for the HD group.

Fig. 3 – Cumulative incidence of prostate-specific antigen relapse by (A) dose group and (B) risk group across all doses.

As shown in Figure 3B, the risk of PSA failure was significantly higher for unfavorable intermediate-risk disease than for favorable intermediate-risk disease (p = 0.002). The estimated PSA failure rate at 8 yr was 35% (95% CI 18–53%) for unfavorable intermediate-risk disease and 7.8% (95% CI 2.4–17%) for favorable intermediate-risk disease.

An analysis of PSA failure rates by risk group and dose level was not feasible owing to the small number of events.

3.4. Salvage treatment

Twelve patients received salvage treatments for recurrent disease. Eleven patients had brachytherapy at a median time of 41.6 mo (range 29.7–111.2) after SBRT. One patient had SBRT to a prostate-specific membrane antigen–avid nodal recurrence. At last follow-up, 6/11 patients treated with salvage brachytherapy had no evidence of disease progression at a median of 52 mo (range 4.7–83.7) after their salvage treatment.

3.5. Distant metastasis, prostate cancer–specific survival, and OS

Three patients developed distant metastases. Overall, 15 patients died during follow-up. The estimated 10-yr OS rate was 87% (95% CI 80–94%). There were no significant differences in OS by dose level or risk group. No prostate cancer–related deaths were recorded.

4. Discussion

We report extended long-term outcomes from a dose escalation trial for patients with low- or intermediate-risk prostate cancer treated with SBRT doses ranging from 32.5 to 40 Gy in five fractions. Following results previously reported for this trial, 40 Gy delivered over five fractions became the standard dose fractionation scheme for treatment of localized prostate cancer at our institution. Over prolonged follow-up of up to 10 yr, we observed durable disease control rates and low incidence of persistent toxicities.

Despite concerns regarding late toxicity when treating to a higher dose per fraction, we did not observe significant associations when comparing LD and HD. This update represents the longest follow-up for patients treated with dose-escalated SBRT and demonstrates late toxicity rates that are comparable to previous results. In HYPO-RT-PC, late grade ≥2 and ≥3 GU and GI toxicities occurred in 18% and 4.2%, and in 10% and 1.5% of patients, respectively, randomized to SBRT [4]. In a pooled analysis, cumulative incidence rates were 11.2%, 12.3%, and 13.4% for late grade 2 GU toxicity, and 1.8%, 2.4%, and 3.2% for late grade 3 GU toxicity at 5, 7, and 10 yr, respectively. The cumulative rate of late grade 2 and 3 GI toxicity remained 4.5% and 0.4%, respectively, at 5 and 10 yr [2,3]. Fuller et al [1] reported cumulative incidence rates of 16.3% at 5 yr and 19.2% at 10 yr for late grade 2 GU toxicity. Grade ≥3 GU and GI toxicity remained stable at 2.6% and 0%, respectively, at 5 and 10 yr [1]. As previously reported, our relatively low rates of severe late toxicity with HD SBRT could be explained by tight margins around the CTV, intrafraction motion targeting, and a meticulous treatment planning process. Prescribing doses to the 100% isodose line resulted in homogeneous dose distributions while strictly adhering to dose-volume constraints.

In our experience, HD SBRT yielded better biochemical control outcomes than LD SBRT. The 8-yr cumulative incidence estimate for PSA failure was 4.9% in the HD group and 19% in the LD group. This difference seems to be mostly driven by the 32.5-Gy cohort. Of note, no cases of PSA failure occurred in patients with low-risk disease. The proportion of patients with favorable and unfavorable intermediate-risk disease who had PSA relapse was higher in the LD group than in the HD group, and although the small number of events does not support any regression models, the result is hypothesis-generating. Whether LD SBRT is noninferior to HD SBRT in low-risk disease and HD SBRT is superior to LD SBRT in intermediate-risk disease is beyond the scope of this trial, but warrants further investigation.

The estimated incidence of PSA failure was significantly higher for patients with unfavorable intermediate-risk disease than for patients with favorable intermediate-risk disease. This observation is in accordance with published SBRT data. Fuller et al [1] found that the 10-yr rate of freedom from biochemical failure was 84.3% for patients with favorable intermediate risk and 68.4% for patients with unfavorable intermediate risk. A pooled analysis revealed 7-yr biochemical recurrence rates of 8.6% for patients with favorable intermediate risk and 14.9% for patients with unfavorable intermediate risk [2]. In a multicenter trial of 40 Gy in five fractions, 5-yr disease-free survival was 100% for patients with favorable intermediate risk and 93.1% for patients with unfavorable intermediate risk [5]. It has been suggested that patients with unfavorable intermediate-risk disease may benefit from intensification of their definitive treatment, such as ADT or brachytherapy boost [10–12]. Whether dose-escalated SBRT could replace ADT in the management of unfavorable intermediate-risk prostate cancer requires further prospective investigation.

The safety of escalated whole-gland SBRT doses beyond 40 Gy in five fractions is questionable. In a phase 1/2 SBRT dose escalation trial, whole-gland doses of 45–50 Gy resulted in high disease control rates but unacceptable toxicities [13]. Focal escalation of doses to the dominant intraprostatic lesion (DIL) may prove to be a safer approach. In the FLAME trial, focal boosts to the DIL resulted in better biochemical disease–free survival without excess toxicity [14]. Our group previously showed that men with Prostate Imaging-Reporting and Data System (PI-RADS) 4–5 DILs have a higher risk of a positive post-SBRT prostate biopsy, most commonly within the DIL [15]. A DIL boost is also being evaluated in the SBRT setting. The hypo-FLAME trial is investigating a whole-gland dose of 35 Gy in five fractions with an integrated boost up to 50 Gy [16]. A recent report on toxicity and patient-reported quality-of-life outcomes trial for patients treated with 40G y in five fractions with a boost of 2.5–5 Gy in a phase 2 trial [17] revealed late grade ≥2 toxicity cumulative incidence rates at 24 mo of 13% for GU toxicity and 5% for GI toxicity. Our group is currently investigating the safety and efficacy of focal dose intensification using magnetic resonance imaging–guided linear accelerator–based adaptive planning in intermediate-risk prostate cancer without the use of ADT [18,19].

Our study has several limitations. While tumor control outcomes could be reliably reported, late toxicity may be under-represented. Although our long-term data continue to suggest that higher SBRT doses lead to more favorable tumor control rates, well-powered prospective trials would be useful in confirming these dose-dependent effects.

5. Conclusions

Dose-escalated SBRT seems to have resulted in superior biochemical control rates in comparison to lower doses, without excess late toxicity, and should become the standard of care for treatment of localized prostate cancer. Prospective randomized trials are necessary to identify the optimal SBRT dose and fractionation by risk group, and management for patients with unfavorable intermediate-risk disease to further define the standard of care.