Discussion
We investigated adherence to guideline-recommended annual surveillance PSA testing in a national cohort of veterans treated with definitive RT for prostate cancer. We found guideline concordance was initially high and decreased slightly over time. We also found guideline concordance with PSA surveillance varied based on a number of clinical and delivery system factors, including marital status, rurality, receipt of concurrent ADT, as well as whether the veteran was treated at his diagnosing facility. Taken together, these overall results are promising, however, also point to unique considerations for some patient groups and potentially those treated in the community.
Our finding of lower guideline concordance among nonmarried patients is consistent with prior research, including our study of patients undergoing surgery for prostate cancer.4 Addressing surveillance in this population is important, as they may have less social support than do their married counterparts. We also found surveillance was lower at the extremes of age, which may be appropriate in elderly patients with limited life expectancy but is concerning for younger men with low competing mortality risks.7 Future work should explore whether younger patients experience barriers to care, including employment challenges, as these men are at greatest risk of cancer progression if recurrence goes undetected.
Although rural patients are less likely to undergo definitive prostate cancer treatment, possibly reflecting barriers to care, in our study, surveillance was actually higher among this population than that for urban patients.9 This could reflect the VA’s success in connecting rural patients to appropriate services despite travel distances to maintain quality of cancer care.10 Given annual PSA surveillance is relatively infrequent and not particularly resource intensive, these high surveillance rates might not apply to patients with cancers who need more frequent survivorship care, such as those with head and neck cancer. Future work should examine why surveillance rates among urban patients might be slightly lower, as living in a metropolitan area does not equate to the absence of barriers to survivorship care, especially for veterans who may not be able to take time off from work or have transportation barriers.
We found guideline concordance was higher among patients with higher Gleason scores, which is important given their higher likelihood of failure. However, low- and intermediate-risk patients also are at risk for treatment failure, so annual PSA surveillance should be optimized in this population unless future studies support the safety and feasibility of less frequent surveillance.10-13 Our finding of increased surveillance in patients who receive concurrent ADT may relate to the increased frequency of survivorship care given the need for injections, often every 3 to 6 months. Future studies might examine whether surveillance decreases in this population once they complete their short or long-term ADT, typically given for a maximum of 3 years.
A particularly relevant finding given recent VA policy changes includes lower guideline concordance for patients receiving RT at a different facility than where they were diagnosed. One possible explanation is that a proportion of patients treated outside of their home facilities use Medicare or private insurance and may have surveillance performed outside of the VA, which would not have been captured in our study.14 However, it remains plausible that there are challenges related to coordination and fragmentation of survivorship care for veterans who receive care at separate VA facilities or receive their initial treatment in the community.15 Future studies can help quantify how much this difference is driven by diagnosis and treatment at separate VA sites vs treatment outside of the VA, as different strategies might be necessary to improve surveillance in these 2 populations. Moreover, electronic health record-based tracking has been proposed as a strategy to identify patients who have not received guideline concordant PSA surveillance.14 This strategy may help increase guideline concordance regardless of initial treatment location if VA survivorship care is intended.
Although our study examined receipt of PSA testing, it did not examine whether patients are physically seen back in radiation oncology clinics, or whether their PSAs have been reviewed by radiation oncology providers. Although many surgical patients return to primary care providers for PSA surveillance, surveillance after RT is more complex and likely best managed in the initial years by radiation oncologists. Unlike the postoperative setting in which the definition of PSA failure is straightforward at > 0.2 ng/mL, the definition of treatment failure after RT is more complicated as described below.
For patients who did not receive concurrent ADT, failure is defined as a PSA nadir + 2 ng/mL, which first requires establishing the nadir using the first few postradiation PSA values.15 It becomes even more complex in the setting of ADT as it causes PSA suppression even in the absence of RT due to testosterone suppression.2 At the conclusion of ADT (short term 4-6 months or long term 18-36 months), the PSA may rise as testosterone recovers.15,16 This is not necessarily indicative of treatment failure, as some normal PSA-producing prostatic tissue may remain after treatment. Given these complexities, ongoing survivorship care with radiation oncology is recommended at least in the short term.
Physical visits are a challenge for some patients undergoing prostate cancer surveillance after treatment. Therefore, exploring the safety and feasibility of automated PSA tracking15 and strategies for increasing utilization of telemedicine, including clinical video telehealth appointments that are already used for survivorship and other urologic care in a number of VA clinics, represents opportunities to systematically provide highest quality survivorship care in VA.17,18
