Short-Term PSA Testing to Calculate PSADT is Error-Prone
VANCOUVER—Relying only on PSA doubling time (PSADT) from a year of active surveillance can lead to as many as one in five patients being advised to have surgery that is actually unnecessary, new research has revealed.
Frederik B. Thomsen, MD, at University of Copenhagen in Denmark, and his colleagues found that using one year of PSA testing results can lead to 10%-20% of patients being misclassified as having a high risk of prostate cancer progression.
“Our study shows that the statistical uncertainty of the PSA doubling time leads to a substantial risk of patients being misclassified as having an aggressive tumor using the currently applied active surveillance criteria,” Dr. Thomsen said after presenting results at the 33rd Congress of the Societé Internationale d'Urologie. “Therefore, PSA doubling time should not be used as a trigger for recommending curative therapy in these patients.” He presented other results demonstrating the unreliability of PSADT earlier in the year at the 28th annual congress of the European Association of Urology.
His team performed the analysis because a PSADT of less than three years is widely accepted as the cut-off for recommending definitive treatment in both clinical practice and research. They sought to demonstrate that the accuracy of the use of PSADT depends on a number of factors, including the frequency and length of PSA testing.
“We chose to analyze the probability of being falsely classified as belonging to the high-risk of progression group because this has obvious clinical consequences: such patients are normally recommended curative therapy,” Dr. Thomsen said.
Dr. Thomsen and his colleagues modeled data from 221 patients who had at least four PSA measurements during active surveillance. The team focused on three scenarios: patients receiving PSA testing every three months for one year, PSA testing every three months for one year and subsequently every six months for two years, and testing every three months for one year and subsequently eery six months for two years.
Dr. Thomsen and his co-investigators calculated the PSADTs using the Memorial Sloan-Kettering Cancer Center guidelines. Furthermore, they determined patients' probability of misclassification using the median of the root mean square errors from an earlier regression analysis.
They found that if clinicians only use one year of PSA testing, 20% of patients with a true PSADT of five years and 10% of those with a true PSADT of nine years will be misclassified and recommended curative therapy. However, if physicians wait two years to make this determination, 5% of patients with a true PSADT of five years and 1% of those with a true PSADT of nine years will be misclassified. The probability of misclassification drops still further if clinicians usd three years of PSA testing, to 2% among those with a true PSADT of five years and 0.6% of those with a true PSADT of nine years.