Abstract
OBJECTIVE: To investigate whether prostate-specific antigen (PSA) fluctuation correlates with a prostate cancer and to assess whether PSA fluctuation could be used for diagnosis of prostate cancer. MATERIALS AND METHODS: Our study included 229 patients who were performed a prostate biopsy (non-cancer group, 177; prostate cancer group, 52). Enrolled patients were provided twice PSA tests within 6 months. PSA fluctuation (%/month) was defined as a change rate of PSA per a month. Independent t test was used to compare between two groups. Receiver operator characteristic curve was used to assess the availability as a differential diagnostic tool and the correlation. Simple linear regression was performed to analyze a correlation between PSA fluctuation and other factors such as age, PSA, PSA density, and prostate volume. RESULTS: There were significant differences in PSA, PSA density, percentage of free PSA, and PSA fluctuation between two groups. PSA fluctuation was significantly greater in non-cancer group than prostate cancer group (19.95 ± 23.34%/month vs 9.63 ± 8.57%/ month, P=0.004). The most optimal cut-off value of PSA fluctuation was defined as 8.48%/month (sensitivity, 61.6%; specificity, 59.6%; AUC, 0.633; P=0.004). In a simple linear regression model, only PSA level was significantly correlated with PSA fluctuation. CONCLUSION: Patients with wide PSA fluctuations, although baseline PSA levels are high, might have a low risk of diagnosis with prostate cancer. Thus, serial PSA measurements could be an option in patients with an elevated PSA level.