Critical analysis of prostate-specific antigen doubling time calculation methodology

Area of Science:

• Oncology
• Biostatistics
• Prostate Cancer Research

Background:

• Prostate-specific antigen doubling time (PSADT) is a key surrogate marker for disease progression and survival in prostate cancer patients.
• Existing literature presents various methods for calculating PSADT, leading to potential inconsistencies in clinical application.
• Accurate PSADT calculation is crucial for predicting outcomes in men with androgen-independent prostate carcinoma.

Purpose of the Study:

• To identify the most accurate method for describing prostate-specific antigen (PSA) growth over time.
• To determine which PSADT calculation method best predicts disease-specific survival in patients with androgen-independent prostate carcinoma.

Main Methods:

• Calculated PSADT for 122 patients with androgen-independent prostate carcinoma using best-line fit (BLF) and first and last observations (FLO) methods.
• Employed random coefficient linear (RCL) and random coefficient quadratic (RCQ) models for PSADT calculation.
• Utilized statistical analysis to compare the fitting accuracy of PSA profiles and predictive ability for disease-specific survival.

Main Results:

• The RCQ model demonstrated the best fit for patient PSA profiles (P ≤ 0.002).
• PSADT estimates from FLO, RCL, and RCQ models significantly predicted disease-specific survival (P < 0.001), unlike the BLF method (P = 0.66).
• RCQ and RCL models showed improved correlation with disease-specific survival (R² = 0.55) compared to FLO (R² = 0.11) and BLF (R² = 0.003).

Conclusions:

• Random coefficient methods (RCL and RCQ) provide a more reliable fit of PSA profiles.
• These models are superior to existing methods for predicting disease-specific survival in androgen-independent prostate cancer.
• RCL or RCQ models should be considered for future PSADT assessments as predictive parameters.