Serum prostate-specific antigen profile following radiotherapy for prostate cancer: implications for patterns of failure and definition of cure.

OBJECTIVES

A reference range of prostate-specific antigen (PSA) values compatible with cure following radiotherapy (RT) for prostate cancer (PCa) has yet to be established. Various thresholds, as low as 0.5 ng/mL, have been used to define biochemical disease-free status. We report PSA profiles in 118 patients who were systematically biopsied following standard RT, with a minimum 4-year follow-up.

METHODS

One hundred eighteen patients were treated with standard external beam RT from May 1987 to October 1991, and were followed prospectively with transrectal ultrasound (TRUS)-guided biopsies and measurement of serum PSA levels. Stage distribution was as follows: T1b: 25 patients, T2a: 27 patients, T2b/c: 42 patients, T3: 23 patients, T4: 1 patient. Median follow-up for patients without clinical failure is 68 months (range 48 to 108). Treatment failures were categorized as biochemical (biochemical failure [chemF]: PSA level of 2.0 ng/mL or more and greater than 1 ng/mL over nadir), local (local failure [LF]: positive biopsy and PSA level greater than 2.0), and distant failure (DF).

RESULTS

PCa recurred in 55% of patients: 38% LF (n = 45; 30 isolated and 15 with DF), 25% DF (n = 30; 15 isolated and 15 with LF), and 4% chemF (n = 5). Mean PSA nadir was 0.4 for patients with no evidence of disease (NED) and occurred at 33 months, 3.2 for LF at 17 months, 7.7 for DF at 12 months, and 1.4 for chemF at 24 months. After reaching the nadir, PSA in patients with recurrence followed first-order kinetics, rising exponentially over time. The mean PSA doubling time was 12.6 months for LF, 5.2 months for DF, and 21.8 months for chemF (P = 0.004). At last follow-up, the median PSA for patients without evidence of disease is 0.5 ng/mL. Four such patients had PSA values that rose to between 1 and 2 ng/mL for 5 to 38 months, but these eventually fell again to less than 1 ng/mL. Three patients had PSA values between 2 and 3 ng/mL, but 2 now have decreasing levels and the third has a rising level. All patients whose PSA levels rose to greater than 3 ng/mL exhibited a persistently rising pattern and ultimate tumor recurrence.

CONCLUSIONS

There is a range of PSA values following RT for PCa that is compatible with cure. A definition of biochemical disease-free status at any absolute threshold of PSA level less than 3 ng/mL will overdiagnose failure in a significant proportion of patients. Patients with a PSA level between 1.5 and 3 ng/mL should be observed until there is unequivocal evidence of disease recurrence. In the absence of known biopsy status, PSA doubling time can be a useful indicator of whether failure is local or distant.