Why did biochemical control not translate into a survival increase after brachy boost therapy?

The first randomized clinical trial to prove that brachy boost (BB) therapy had better oncological outcomes among high risk patients was Sathya et al. (2005). After 5 years, 36% of those high-risk patients who received the brachy boost had a PSA recurrence vs. 66% of those who received external beam radiation (EBRT) only. In an update, the authors report that overall survival was not significantly different in the two groups. This seems to call into question whether PSA recurrence is a useful surrogate endpoint for survival, or if it is, under what circumstances?

Dayes et al. provided a 14-year median update on the original study and added further comments in this "Beyond the Abstract" essay. The 104 patients in the original study who were treated between 1992 and 1997 had the following characteristics and treatments:

• Median age was 66
• 60% were high risk, 40% intermediate risk
• All had a negative pelvic lymph node dissection, negative bone scan and CT
• Brachy boost (BB) comprised 35 Gy of Ir 192 over 48 hours plus 40 Gy of EBRT in 20 fractions for a total of 75 Gy [sic].
• EBRT-only compromised 66 Gy delivered in 33 fractions using 2DRT (an outmoded external beam technology).
• None received androgen deprivation as part of their radiation therapy, nor afterwards unless PSA reached 20 ng/ml.

As of the update on the 104 patients (with only 5 lost to follow-up):

• Mortality from any cause was 67% among the BB patients, 77% among the EBRT-only patients -- not significantly different
• Prostate cancer-specific mortality was 18% among the BB patients, 23% among the EBRT-only patients - not significantly different
• Incidence of metastases was 20% among the BB patients, 28% among the EBRT-only patients - not significantly different
• Improvement in PSA control was maintained: 47% higher rate of biochemical recurrence-free survival among the BB group

There was a biopsy given 2 years after treatment to 87 of the 104 men in the original study

• In the BB group, 24% had a positive biopsy and 6% were metastatic
• In the EBRT-only group, 51% had a positive biopsy and 6% were metastatic

The authors conclude:

Despite ongoing benefit with respect to biochemical disease control, long term follow up out to 2 decades failed to demonstrate improvements in other important outcomes such as development of metastatic disease, deaths from prostate cancer and deaths from any cause. 

Increased biochemical (PSA) control usually translates into increased survival later on. That correlation is well-characterized. So why did it not in this case?

This study, with a sample size of only 104 (51 BB, 53 EBRT-only), was not large enough to detect statistically significant survival differences. We note that directionally there was an improvement in survival even though the difference wasn't big enough for 95% confidence. Also, 40% were intermediate risk patients who are slower to have detectable metastases and are more likely to die of other causes. By contrast, the ASCENDE-RT trial of LDR brachy boost therapy recruited 398 men, 30% were intermediate risk, and may eventually be able to demonstrate overall survival differences with longer follow-up.

We have to acknowledge that the doses delivered in this study were below what is now considered curative, and the findings here are to a large extent irrelevant. I am at a loss to explain how a hot iridium implant could be left in a patient for 48 hrs without doing serious damage or cooking the prostate to a crisp.  Perhaps they used cooler implants back then.  I can only trust that Dr. Sathya is correct in not making a correction for the lack of fractionation, which would be typical. It seems the BB dose was sub-optimal as demonstrated by the fact that in a quarter of men, the cancer was left alive in the prostate. EBRT-only was worse - leaving cancer alive in the prostates of twice as many men. Although they dissected some pelvic lymph nodes that they could find, we now know that even with improved modern lymph node detection methods, we miss 44% of positive lymph nodes (see this link). The 6% who were metastatic might have been caught with some of our new PET scans. So, in both groups, there was a lot of cancer left behind. Many high-risk radiation patients today would have had whole-pelvic radiation and would have had hormone therapy for up to two years. This highlights the importance of expanding the treated area, using escalated doses, and adding systemic therapy when the probability is high that the cancer might have escaped the prostate.

Even though BB wasn't curative for many high risk patients, it is disappointing that death was not delayed by reducing the tumor burden. There are several clinical trials of treating the prostate (with surgery or radiation) even after metastases have been detected, thereby hoping to prolong survival by reducing the load of cancer cells. Metastasis-directed radiation is sometimes given in this hope as well. Both of those therapies decrease PSA, at least temporarily. But only treating PSA serves no purpose if that is the only outcome. If this study is any indication, the cancer will catch up and replace the killed cells with no net survival benefit. I hope that is not the case.