Health Services Research
PV QA 3 - Poster Viewing Q&A 3
TU_38_2969 - Clinical Predictors and Disparities in Surveillance PSA Utilization After Prostate Cancer Treatment
Tuesday, October 23
1:00 PM - 2:30 PM
Location: Innovation Hub, Exhibit Hall 3
Christina Chapman, MD, MS
University of Michigan: Assistant Professor: Employee; Veterans Affairs Ann Arbor Healthcare System: Staff Physician: Employee
National Cancer Institute: Research Grants
National Medical Association: Vice Chair for Membership, Post Graduate Section; RSNA: Education Study Section
Clinical Predictors and Disparities in Surveillance PSA Utilization After Prostate Cancer Treatment
C. H. Chapman1,2, J. Burns1, A. Tsodikov3, M. G. Chang4, C. Deville Jr5, B. Hollenbeck6, and T. Skolarus1,3; 1Veterans Administration Ann Arbor Healthcare System, Ann Arbor, MI, 2Department of Radiation Oncology, University of Michigan, Ann Arbor, MI, 3University of Michigan, Ann Arbor, MI, 4Hunter Holmes McGuire VA Medical Center, Richmond, VA, 5Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, 6Department of Urology, University of Michigan, Ann Arbor, MI
Purpose/Objective(s): Guidelines recommend prostate cancer patients undergo PSA surveillance at least annually after treatment to identify disease recurrence. Few studies have assessed adherence to these recommendations, and the extent to which early post-treatment PSA values impact guideline concordance is unknown.
Materials/Methods: We used the Veterans Affairs Central Cancer Registry to identify 11660 patients treated definitively for prostate cancer (surgery, radiotherapy (RT) +/- androgen deprivation therapy (ADT)) with at least 5 years of follow up through 2011. We defined guideline-concordant PSA surveillance (GCPS) as receipt of at least one PSA annually for 5 years post-treatment. We used logistic regression models adjusted for demographic and clinical variables to examine associations with GCPS. We also determined whether post-treatment PSA values were associated with GCPS.
Results: Surveillance PSA utilization was >90% in years 1-2 across all modalities, and >83% in years 3-5. Gleason score 8-10 was associated with GCPS for men treated with surgery (adjusted Odds Ratio (aOR) 1.70, p<0.001) and RT+ADT (aOR 1.49, p=0.002). Higher comorbidity was associated with GCPS for men treated with surgery (aOR 1.29, p=0.005), with a trend for RT (aOR 1.18, p=0.063). Younger age predicted for GCPS for RT+ADT (aOR 0.98, p=0.005). Black (v. white) men treated with RT were more likely to have GCPS (aOR 1.33, p=0.001), with a trend for lower likelihood of GCPS in the surgical (aOR 0.86 p=0.077), and RT+ADT groups (aOR 0.82, p=0.080). Other factors associated with GCPS after surgery included: marital status and positive surgical margins, with a trend for initial PSA. Higher year 1 PSA levels predicted for GCPS for surgery and RT (OR 1.07 & 1.02, p’s < 0.03), and remained significant for surgical patients with PSAs of < 0.5 (OR 4.07, p<0.001) and 1.0 (OR 2.20, p=0.002).
Conclusion: Guideline-concordant PSA surveillance is high in the first 5 years after prostate cancer treatment. We found that surgical patients with higher comorbidity had higher GCPS raising concerns about missed opportunities to salvage low comorbidity patients with longer life expectancy. Higher year 1 PSA levels were associated with GCPS for surgical and RT patients, potentially mediated through monitoring and treating recurrent disease. On the other hand, early low PSA levels may provide false reassurance in patients at risk of failure. Better understanding the impact of both broad interventions (e.g., cancer survivorship plans, electronic reminders) and risk-adapted initiatives to increase guideline concordance while decreasing racial disparities appear warranted.
Author Disclosure: C.H. Chapman: Research Grant; NCI. J. Burns: None. A. Tsodikov: Research Grant; NCI. M.G. Chang: None. B. Hollenbeck: Research Grant; NIH, AHRQ, CMS. T. Skolarus: Research Grant; NIH, Veterans Affairs Healthcare System.