Dose-escalated radiation therapy doesn’t impact survival within ten years

Zaorsky et al. conducted a meta-analysis of 12 randomized clinical trials covering data on 6884 patients treated with external beam radiation at various dose levels. Their goal was to determine whether increasing the delivered biologically effective dose made a difference in 5 or 10 year metastasis-free survival, prostate cancer specific survival, or overall survival.

They found that dose-escalated radiation had the following effects:

• ·      10-year freedom from biochemical failure improved by 9.6% in low-risk men.
• ·      10-year freedom from biochemical failure improved by 7.2% in intermediate-risk men.
• ·      There was no corresponding improvement in metastasis-free survival, prostate cancer specific survival, or overall survival out to 10 years.
• ·      Dose escalation was not correlated with increases in acute toxicities.
• ·      Late-term gastrointestinal toxicities increased in patients treated with 3D-CRT.
• ·      Late-term toxicities were lower among patients treated with IMRT despite higher dose levels.

The abstract makes no mention of dose-escalated radiation treatment of high-risk men. We discussed some conflicting  survival data on higher risk patients in a previous commentary (see this link). As we saw, even at the higher risk levels, ten years follow-up was not long enough to detect difference in survival due to dose escalation.

The authors conclude:

“Thus, freedom from biochemical failure is a poor surrogate of overall patient outcomes for trials of RT.”

This is an unwarranted conclusion from the data presented in the abstract. We discussed the issue of surrogate endpoints (like freedom from biochemical failure) and length of follow-up in a previous commentary (see this link). For a newly diagnosed intermediate-risk man, the time frame for development of distant metastases could easily be upwards of 10 years, and a lot longer for low-risk men. The only valid conclusion one can draw from their analysis is that 10 years is too short a time frame to detect any effect of dose escalation on metastasis-free survival, prostate cancer survival or overall survival in these risk groups. Their analysis makes the argument for using surrogate endpoints, rather than against them. Given the long natural history of prostate cancer progression in these risk groups, how else can we gauge the impact of dose escalation within a practical followup timeframe?

The other interesting conclusion is that dose escalation, when delivered with IMRT technology, had no impact on acute or late-term toxicities. This argues for IMRT-delivered dose escalation: since it did not increase toxicity risk, and may increase long-term cancer control, there is no reason not to use it. This holds true even among low risk men who, for whatever reason, have elected not to engage in active surveillance. It also holds true for men with fewer than ten years of life expectancy who, for whatever reason, have elected not to engage in watchful waiting.

Survival benefit of dose escalation for higher risk patients

As discussed in a previous article, there is a seeming discrepancy between the findings of Kalbasi et al. and RTOG 0126, at least for intermediate-risk patients. Kalbasi et al. found that higher dose radiation is associated with higher overall survival rates for intermediate and high-risk patients (but not low-risk patients), while RTOG 0126 found no such association. Kalbasi et al. was a retrospective database analysis, while RTOG 0126 was a randomized clinical trial. Which is right?

I’ve had a chance to review the full text of the Kalbasi et al. study, but the RTOG 0126 trial has not yet been published. This article provides some additional observations based on what is currently available.

One shortcoming common to both is that they were attempting to find differences in overall survival after just 10 years. Because prostate cancer is often detected very early, has a long natural history, and there are many medicines now that extend survival, 10 years may not be enough to observe a difference. They also both used overall survival rather than cause-specific survival as endpoints, so we don’t know how many of the deaths were incidental to rather than because of prostate cancer. The database analysis did not include data on prostate cancer–related deaths, while in RTOG 0126, the prostate cancer-specific mortality was only 3%. RTOG 0126 did find differences in biochemical failure, distant metastases, and local progression related to higher doses of radiation. It is likely that differences in survival may emerge with longer follow up.

RTOG 0126 excluded the least favorable intermediate risk patients from their study. Those with Gleason score of 7 and stage T2c, and those with a PSA of 15-20 were excluded, and may be the group most likely to benefit from dose escalation.

Kalbasi et al. point out that RTOG 0126 may have been underpowered to detect the dose response. They also conjecture that the highly selected clinical trial patient sample of RTOG 0126 may not reflect the wider patient population – over 300,000 patients - in their data analysis. On the other hand, the retrospective nature of their study allows for confounding. Those patients who received lower doses of radiation may be the ones who had serious co-morbidities that precluded a higher dose. We can never establish a cause/effect relationship based on a retrospective study.

As often happens with long-term radiation studies, the findings become irrelevant by the time we get them. The finding that low-risk patients don’t need the higher dose is becoming increasingly irrelevant as those patients are more safely managed with active surveillance. The finding that intermediate risk (at least unfavorable intermediate risk) and high-risk patients do better with higher doses is irrelevant because IMRT doses of at least 80 Gy have become the standard of care already.

Kalbasi et al. suggest that even higher doses of radiation may improve survival. The problem with their suggestion is a limitation of IMRT – doses beyond 80 Gy become increasingly toxic. The solution is not to push the envelope on IMRT, but to use other methods. In a recent article, we saw that the combination of low dose rate brachytherapy and external beam radiation was able to push the effective radiation dose higher than IMRT alone, and resulted in increased oncological control. Toxicity, however, was higher than with monotherapy. As discussed there, adding a high dose rate brachytherapy boost has proven to be an equally effective technique for escalating dose. While SBRT has been used to increase the effective dose in intermediate-risk patients, and does so with extremely low toxicity, it has not yet been widely used in high-risk patients. In a recent article, we discussed the clinical trials that may prove to be a game changer in the management of such patients.

Androgen deprivation therapy (ADT) and escalated dose in radiation therapy (RT)

Several recent studies shed light on the optimal use of androgen deprivation therapy (ADT) used in conjunction with radiation therapy (RT), including new learning about timing of ADT, RT dose, and their use in various risk categories.

When external beam radiation doses of around 70 Gy were used in the 1990s, it was shown that androgen deprivation therapy (ADT) used with it could improve oncological outcomes. However, it was not at all clear that ADT provided any additional benefit when higher doses radiation of about 80 Gy were used. DART 01/05 (Zapatero et al.) was a randomized clinical trial to determine the optimal duration of ADT supplementation.

DART 01/05 was a multi-institutional Spanish trial among intermediate and high risk men receiving primary treatment with 3D conformal radiation therapy (3DCRT) between 2005 and 2010. The patients were randomized to receive either 4 months (short term) or 28 months (long term) of ADT.

• Everyone received 2 months of ADT before and 2 months during their 3DCRT
•  Everyone received goserelin, an LHRH agonist, throughout, and also received 2 months of anti-androgen therapy (bicalutamide or flutamide) at the beginning.
•  173 patients received short-term ADT, 171 patients received long-term ADT

o   90 were high risk on long-term ADT

o   91 were high risk on short-term ADT

o   83 were intermediate risk on long-term ADT

o   78 were intermediate risk on short-term ADT

• Everyone received a median radiation dose of 78 Gy.

The 5-year outcomes were as follows:

• Biochemical disease-free survival was significantly better with long-term compared to short-term ADT: 90% vs. 81%

o   The difference was only significant among high risk patients: 88% vs. 76%.

•  Metastasis-free survival was significantly better with long-term compared to short-term ADT: 94% vs. 83%.

o   The difference was only significant among high risk patients: 94% vs. 79%.

• Overall survival was significantly better with long-term compared to short-term ADT: 95% vs. 86%.

o   The difference was only significant among high risk patients: 96% vs. 82%.

o   There were 5 deaths due to prostate cancer, all among men on short-term ADT.

• There were no significant differences in acute or late term rectal or urinary toxicities.

The authors conclude:

“Compared with short-term androgen deprivation, 2 years of adjuvant androgen deprivation combined with high-dose radiotherapy improved biochemical control and overall survival in patients with prostate cancer, particularly those with high-risk disease, with no increase in late radiation toxicity. Longer follow-up is needed to determine whether men with intermediate-risk disease benefit from more than 4 months of androgen deprivation.”

For high risk patients, at least, this establishes that dose-escalated RT with long-term ADT is preferable to short term. It leaves several open questions about optimum radiation treatment for this group:

• What is the optimal duration of ADT? We know from an earlier randomized clinical trial (Nabid et al.) that 18 months of adjuvant ADT is as good as 36 months, even with lower dose RT. So the optimal duration is somewhere between 6 months and 18 months.
•  Is IMRT with a brachytherapy boost preferable, and is that enhanced by ADT? (See this link)
•  Is SBRT monotherapy preferable, with or without adjuvant ADT? (This was discussed in a recent article.)
• What is the effect on erectile function?
• Should the pelvic lymph nodes be treated as well? This is the subject of an ongoing clinical trial.

Another randomized clinical trial presented at the Genitourinary (GU) Conference found more support for the addition of ADT to RT for intermediate risk patients. While DART 01/05 looked at long-term vs, short-term ADT with RT and found no difference for the intermediate risk subset, Nabid et al. looked at short-term vs. no additional ADT with RT for intermediate risk patients. They also examined the effect of radiation dose.

Their study consisted of 600 intermediate risk men treated with external beam radiation at several hospitals in Quebec between 2000 and 2010. The 3 arms of their study were treated under the following protocols:

1.    Arm 1: 6 months of ADT + 70 Gy of RT
2. Arm 2: 6 months of ADT + 76 Gy of RT
3. Arm 3: 76 Gy of RT

Those who received ADT were treated with 6 months of both goserelin and Casodex (bicalutamide) beginning 4 months before their RT began. After a median follow up of 76 months, the researchers found that:

• Biochemical failure was significantly higher in Arm 3, but not statistically different between arms 1 and 2.

o   Arm 1: 12.5%

o   Arm 2:   8.0%

o   Arm 3: 21.5%

• 10-year disease-free survival was significantly lower in arm3, but not statistically different between arms 1 and 2.

     

o   Arm 1: 77%

o   Arm 2: 90%

o   Arm 3: 64.5%

• 10-year overall survival was not statistically different between any of the arms.

o   Arm 1: 64%

o   Arm 2: 70%

o   Arm 3: 78% 

• There were only 6 deaths (1%) attributable to prostate cancer, not enough to discern a difference among treatment arms.

The authors conclude:

“In patients with intermediate risk prostate cancer, the use of short term ADT in association with RT, even at lower doses, leads to a superior biochemical control and DFS as compared to dose-escalated RT alone. These outcomes did not translate into an improved overall survival.”

I hope the authors will attempt a sub-group analysis to determine if there were significant differences when favorable vs. unfavorable intermediate risk (see below) is taken into account. It will also be interesting to look at the side effect profile in the 3 arms.

A randomized clinical trial (RTOG 0126) of low dose (70 Gy) vs. high dose (79 Gy) radiation in intermediate risk patients, but without ADT, found improvements in the risk of biochemical failure, distant metastases, and time to local progression in those treated with the higher dose. However, they found no improvement in overall survival with 10 years of observation. Those treated with the higher dose did experience higher rates of urinary and rectal toxicity, however.

One must consider whether the higher rates of urinary and rectal toxicities are still applicable with modern IGRT/IMRT techniques. The men in the above studies were treated with 3DCRT – an older, less precise radiotherapy. As often occurs with long-term clinical trials of radiation therapies, the results may become irrelevant by the time they are reported because of technological advances.

I think 10 years is too short a follow-up period to detect significant differences in survival among intermediate risk men, and especially among favorable intermediate risk men. It also begs the question of whether those men require immediate treatment at all. Some of the sub-groups, including some who are older with co-morbidities, some with favorable PSA kinetics, low volume of cancer, and some with Gleason score≤ 3+4, may be better off with expectant management.

In contrast to the lack of survival benefit to the escalated dose found in RTOG 0126, a retrospective analysis reported at the GU Conference by Kalbasi et al. looked at 12,848 low risk patients, 14,966 intermediate risk patients, and 14,587 high risk patients After a median 73 months of follow up, they found a significant dose response for both the intermediate risk and the high risk patients, but not the low risk patients. For every 2 Gy increase in dose, there was a reduction in the hazard of death of 9% and 7% among the intermediate and high risk patients, respectively.

Perhaps sub-group analysis will explain the difference in the dose response between the two studies. I will report on both further when more detailed findings become available.

I don’t think it will come as any surprise that radiation added to androgen deprivation has better oncological outcomes than androgen deprivation alone. In a randomized clinical trial among 1,205 locally advanced prostate cancer patients treated between 1995 and 2005 with ADT and with or without low dose (64-69 Gy) RT, Mason et al., with median 8 years of follow-up, found that the addition of RT reduced prostate cancer mortality by about half.

Favorable vs. Unfavorable Intermediate Risk

In an earlier article, we noted that Dr. D’Amico raised a caution that the results may look very different if the intermediate risk men were divided into favorable and unfavorable groups. It may be that with further follow-up time, significant differences will appear among the intermediate risk men, and particularly among those with unfavorable features. In a retrospective study by Castle et al. where intermediate risk men were divided into favorable or unfavorable intermediate risk, favorable risk patients had no discernable benefit from the addition of ADT. Unfavorable intermediate risk patients had significantly higher 5-yr freedom from failure if they also received ADT, 74% vs. 94%, respectively. Similarly, Edelman et al. found that ADT combined beneficially with RT only in intermediate risk patients with GS 4+3, more than 50% positive cores, or multiple intermediate risk factors.

Another retrospective study by Keane et al. confirming that finding was presented at the recent Genitourinary Conference. They analyzed the oncological outcomes of 2,668 intermediate risk men (71% favorable, 29% unfavorable) treated between 1997 and 2013 with dose-escalated RT and with and without adjuvant ADT (median 4 months). After a median follow-up of 7.8 years, they found that there was a significant amelioration of the risk of prostate cancer-specific mortality among the unfavorable risk patients who also received ADT, but adding ADT did not make a difference to prostate cancer-specific mortality in those men categorized as favorable intermediate risk.

ADT Sequencing

The conventional wisdom is that neo-adjuvant ADT (ADT started at least two months before the start of radiation) and ADT given concurrently with RT have a different functional benefit from adjuvant ADT (ADT given after the completion of RT). Neoadjuvant and concurrent ADT is thought to radiosensitize the cancer to the radiation treatment, while the adjuvant ADT is thought to function as “clean-up,” killing off small amounts of hormone-sensitive stray cancer cells that may already be systemic. A new study by Weller et al. is calling that model into question.

They analyzed the records of intermediate and high risk patients treated from 1995 to 2002 who had either neoadjuvant and concurrent ADT with their dose-escalated RT (311 patients) or only adjuvant ADT immediately after their dose-escalated RT (204 patients). Ten-year biochemical recurrence-free survival was 61%, distant metastasis-free survival was 80%, and overall survival was 66%. There were no significant differences in any of those measures based on the sequencing of ADT.

The authors conclude:

“the synergy between RT and androgen deprivation is independent of the sequencing of both modalities and the initiation of RT does not need to be delayed for a course of neoadjuvant ADT.”

I think these findings have to be confirmed by a randomized clinical trial. It raises interesting questions about the way ADT and radiation interact to kill cancer cells, perhaps supporting the hypothesis that ADT sustains the immune response to the radiation-induced increase in cancer antigens. If the abscopal effect turns out to be of major importance in the ADT/radiation killing of cancer cells, various immunotherapies, like Provenge, Prostvac, Yervoy, and Keytruda, may improve the oncological benefits still further.

For very high-risk patients, EBRT + BT is superior to surgery or EBRT only

A retrospective analysis of oncological outcomes among modern-era patients with a Gleason score of 9 or 10 demonstrates a clear advantage to a combination of external beam radiation therapy (EBRT) with a brachytherapy (BT) boost to the prostate and short–term androgen deprivation therapy (ADT).

Kishan et al. reported on 487 patients with biopsy Gleason scores of 9 or 10 who were consecutively treated between 2000 and 2013 at the University of California Los Angeles and Fox Chase Cancer Center. The patient characteristics were as follows:

• 170 were treated with radical prostatectomy (RP).
• 230 were treated with EBRT only.
• 87 were treated with EBRT + BT, and most of the BT was high dose rate.
• All patients were Gleason 9 or 10 on biopsy.
• RP patients were younger (median 62 years of age) compared to all radiation patients (median 70 years of age).
• RP patients had more favorable disease characteristics: lower initial PSA, and lower clinical stage.

The patient characteristics by treatment category are listed below.

For the RP patients:

• 11% had pre-surgery ADT or chemotherapy.
• 55% had adjuvant or salvage radiation therapy (68 Gy).
  • 39% of them had adjuvant ADT with radiation for a median of 22 months if adjuvant radiation, 12 months if salvage radiation.
• 85% with biochemical recurrence and no detected distant metastases had salvage radiation.
• 21% had a lower Gleason score (7 or 8) on final pathology, but 91 percent had any Gleason pattern 5 on final pathology.
• 78% were stage T3 or T4 on final pathology (vs. 12 percent clinically).
• 41% had positive surgical margins.
• 16% had positive lymph nodes
  • Among those, 64% received no immediate treatment because of patient preference.

For the EBRT patients:

• Median dose of radiation was 76.4 Gy.
• 94% had ADT starting before EBRT.
  • The median duration of ADT was 24 months.
• 76% had pelvic lymph nodes treated.
• 2 patients received salvage cryotherapy.

For the EBRT + BT patients:

• The median equivalent dose of radiation was 88.7 Gy
• 86% had ADT starting before radiation.
  • The median duration of ADT was 8 months.
• 78% had pelvic lymph nodes treated.
• 1 patient received salvage cryotherapy.

After a median follow-up of 4.6 years, the oncological outcomes were as follows:

• The 10-year biochemical recurrence rates (BCRs) were
  • 84% for RP
  • 40% for EBRT
  • 30% for EBRT + BT
  • Differences between RP and EBRT and between RP and EBRT + BT were statistically significant.
• Percentages of patients who began lifelong ADT after therapy failure were
  • 31% for RP
  • 21% for EBRT
  • 16% for EBRT + BT
  • Differences between RP and EBRT and between RP and EBRT + BT were statistically significant.
• The 10-year rates of distant metastases were
  • 39% for RP
  • 33% for EBRT
  • 10% for EBRT + BT
  • Differences between EBRT + BT and the two others were statistically significant, while the differences between RP and EBRT were not.
• The 10-year rates of prostate cancer-specific mortality were
  • 22% for RP
  • 20% for EBRT
  • 12% for EBRT + BT
  • None of the differences were statistically significant.
• The 10-year rates of overall survival were
  • 75% for RP (they were younger and healthier)
  • 65% for EBRT
  • 59% for EBRT + BT
  • None of the differences were statistically significant.

The authors conclude:

These data suggest that extremely dose-escalated radiotherapy with ADT might be the optimal upfront treatment for patients with biopsy GS 9–10 prostate cancer.

It will come as no surprise to readers that EBRT + BT boost has better outcomes than EBRT alone (see this link and this one). Dose escalation has been found to improve outcomes, and the use of ADT to radiosensitize the cancer, and to systemically clear up micrometastasis, seems to improve outcomes still further. However, ADT for as long as 2 years could not compensate for the lower radiation dose of EBRT used by itself. Longer duration of ADT was not associated with improved outcomes after accounting for the dose effect.

Those who were treated with EBRT + BT were at a considerable disadvantage in this study: they were older, had worse disease characteristics, and were given less local salvage, yet they performed much better. When controlling for those disparities, the total radiation dose emerged as the single most important variable, affecting biochemical recurrence, metastases-free survival, and prostate cancer-specific survival. No other variable – neither duration of ADT nor adjuvant/salvage radiation – was statistically significant.

Prostate cancer-specific mortality rates were cut in half by combining EBRT with a BT boost. While the combination therapy did not make a statistically significant difference in prostate cancer-specific survival, the study was probably under-powered to detect that with statistical significance. The survival curves between EBRT + BT and the other two therapies did consistently diverge throughout the follow-up period, so the difference might well be statistically significant on a larger sample size or longer follow-up.

Not everyone in this study received optimal therapy. The EBRT-only dose was sometimes low by today’s standards, salvage radiation was under-utilized, use of concurrent ADT with adjuvant/salvage radiation was low (see this link) and of too-short duration.  However, most were treated according to the standards of care. The authors looked at the subset of patients who were treated optimally and found no difference in conclusions. The conclusions were robust even excluding those who were lymph-node positive.

What is new here is the comparison of the three potentially curative treatments for very high-risk prostate cancer in the 21st Century. There have been several long-term database analyses that compared surgery to radiation therapy as offered in the 1990s, when radiation doses were often inadequate to achieve cures. We recently saw a comparative benefit to radiation over surgery in the modern era among high-risk patients at the University of Alabama Birmingham (see this link). Ideally, we would like to see a randomized comparative trial between surgery and radiation, but that is unlikely to occur. Meanwhile, this kind of analysis is about the best we have to inform our treatment decisions.

We understand that a future, expanded analysis will include data from other institutions, including Harvard, the Cleveland Clinic, and Memorial Sloan-Kettering Cancer Center. That analysis will also include toxicity data. We will certainly report on that when it is published.

note: Thanks to Drs. King and Kishan for allowing me to see the full text of this analysis, and responding to questions.