To the Editor:

King and Kapp¹ recently hypothesized that the apparent benefits of whole pelvic radiotherapy (WPRT) noted on RTOG 9413 in terms of improved prostate-specific antigen (PSA) control might well be explained by a higher dose of radiation scattered to the testes. Their rather lengthy commentary is backed by examples from a series of studies demonstrating the well-known fact that the testes are relatively radiosensitive organs and scattered radiation can have a modest impact on serum testosterone levels. The most relevant data available for addressing their hypothesis is not the “updated” analysis they allude to but rather the data on the basis of the initial report, given that their hypothesis relates to a delayed recovery of testosterone after androgen deprivation therapy (ADT) of 4 months duration is discontinued, not a biologic response 10 years later.^2,3 Thus, we will address their hypothesis, focusing primarily on these data and, using these data, show why we know their hypothesis cannot possibly be true.

Arms 3 and 4 of RTOG 9413 included patients treated with large and small fields and adjuvant ADT respectively and with a median follow-up of nearly 5 years showed no difference in the PSA control rates. Thus, field size per se using the techniques required in patients treated on RTOG 9413 did not affect outcome, which disproves their hypothesis. This should not be surprising because per protocol all patients were required to have a urethrogram performed, and their inferior border was defined such that patients treated with WPRT had the same inferior border as those treated with prostate-only radiation therapy (PORT). Contrary to the data that King and Kapp allude to, roughly 90% of patients treated on RTOG 9413 were treated with nonconformal radiotherapy (unpublished data) because few institutions had three-dimension planning equipment in 1994. These patients were treated with relatively large nonconformal fields (median field size was 11 × 10 cm and maximum 11 × 11 cm) with little blocking.⁴ Patients treated with WPRT generally had fields that were significantly taller, reaching up to L5-S1, but did not extend lower into the pelvis. Thus, it is exceedingly unlikely that there were large differences in testicular doses. To make this point more clearly, we compared the doses received by the testes for a single treatment in a patient (not treated on RTOG 9413 but planned per the protocol) using either four-field PORT (9 × 9 cm with corner blocks) to patients treated to a WPRT (Fig 1A and 1B). As shown in the dose volume histogram (Fig 1C ), the mean dose for WPRT compared with PORT was only slightly greater, being .0356 Gy versus .0260 Gy for this randomly selected patient. Although the authors concede that a single randomly selected patient cannot prove that there were no differences between the dose to the testes for patients treated with WPRT compared with PORT, if the patients were treated per protocol, the differences in the dose to the testes would likely to be small, particularly considering that the median field size among patients treated on the prostate-only arm was 11 × 10 (larger than the 9 × 9 field used in the example case above).

In addition, when comparing patients treated on RTOG 9413 with neoadjuvant hormonal therapy (NHT) versus adjuvant ADT both receiving PORT (arms 2 and 4 respectively), no differences were noted in outcomes. Thus, their hypothesis that “there is a synergistic effect leading to longer time to testosterone recovery” is also not supported by the data.

King and Kapp quote a secondary analysis of RTOG 9413 comparing the timing of ADT (NHT v adjuvant hormonal therapy [AHT]).⁵ The two arms differed solely in the timing of the total of 4 months of total androgen deprivation: arm 1 (320 patients administered NHT) and arm 3 (319 patients, administered as AHT). We assessed the influence of testosterone levels (measured at baseline and yearly thereafter). Of all analyzed patients, 75 (38.1%) and 73 (38.4%) in arms 1 and 3, respectively, did not have a return to normal testosterone levels. The difference in the time to testosterone recovery was significant (P = .01) between arm 1 (mean, 11.3 months) and arm 3 (mean, 14.5 months). Thus the time to testosterone recovery was significantly slower in the adjuvant arm (mean difference, 3.2 months; P = .0103), the exact opposite of what King and Kapp suggest in their hypothesis. This data clearly disproves their hypothesis.

Furthermore, not considered in our secondary analysis (or by King and Kapp), is the fact that the time to PSA failure is biased against the NHT arms of the study because the date of PSA failure is measured from the date of random assignment. This bias results because PSA failure is more likely to occur after discontinuation of ADT. Patients treated on arm 1 completed ADT at the end of radiotherapy occurring roughly 4 months after random assignment. In contrast, patients treated on arm 3 completed ADT roughly 6 months after random assignment. Thus, if there were an artificial suppression of testosterone as a result of the use of NHT ADT (3 months) and a 2-month timing bias in favor of the AHT WPRT arm, there could be a total of a 5-month bias against the benefits of NHT and WPRT. Despite this timing of NHT and time to testosterone recovery biases, the progression-free survival (PFS) benefits of NHT and WPRT were nearly 2 years as shown in Figure 2.⁶

Even if WPRT were associated with higher doses of radiation to the testes and a delayed recovery of testosterone, is there any evidence that this would affect PSA failure? We recently performed an analysis on men treated on RTOG 9413 and 9202 assessing the impact of baseline serum testosterone (BST) on outcome.⁷ All patients with a BST measurement and complete data (n = 2,478) were included in this analysis and divided into four categories: very low BST, ≤ 16.5th percentile of BST (≤ 248 ng/dL; n = 408); low BST, > 16.5th percentile and ≤ 33rd percentile (> 248 ng/dL but ≤ 314 ng/dL; n = 415); average BST, > 33rd percentile and ≤ 67th percentile (314 to 437 ng/dL; n = 845); and high BST, > 67th percentile (> 437 ng/dL; n = 810). On multivariable analysis, BST had no effect on PSA failure or any other outcome measure. If BST did not affect PSA failure, it seems implausible that a dose of scattered radiation that has only a modest effect on serum testosterone would affect PSA failure rates. This observation is supported by at least one other study assessing the impact of pretreatment testosterone levels on outcome.⁸

Finally, what about the assertion that there is a “lack of biologic understanding of why it is the results would change with longer follow-up?” Again, King and Kapp are confused about the findings from RTOG 9413. PSA failure continues to favor the NHT and WPRT (arm 1) of RTOG 9413 even with long follow-up. However, the definition of PFS used in the initial report included death as a result of any cause, and at 5 years, there were few deaths from any cause. At 10 years, many men (median age ∼70) have died of other causes, and consequently PFS is no longer the most robust way to assess the impact of field size on outcome.³ Despite this fact, using the most current standard definition for biochemical failure, an advantage is still seen at 10 years for NHT and WPRT. As shown in the update, PFS and biochemical failure by the Phoenix definition continue to favor WPRT and NHT (P = .034 and P = .0098, respectively).⁹

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: Colleen A. Lawton, AstraZeneca Research Funding: None Expert Testimony: None Other Remuneration: None