Incorporation of the American Urologic Association symptom score may prevent unnecessary prostate needle biopsies
John M. Corman, M.D., Christopher R. Porter, M.D., Katherine Odem-Davis, Ph.D.*, and Erika M. Wolff, Ph.D.*
In order to minimize prostate needle biopsies (PNBs) performed in the absence of prostate cancer (PCa), models that estimate risk based on pre-biopsy variables must be developed. Absence of lower urinary tract symptoms (LUTS) has been correlated with presence of cancer on PNB.
Few studies have assessed whether the absence of LUTS improves the predictive power of commonly used pre-biopsy measures, such as age and PSA.
To evaluate the influence of LUTS on the ability to predict the risk of PCa.
We performed an analysis of a prospective database of 2,803 consecutive men who underwent transrectal ultrasound (TRUS)-guided PNB. The American Urologic Association Symptom Score (AUASS) was used to assess the severity of LUTS. An AUASS ≤7 was considered indicative of absence of LUTS (No LUTS).
Baseline characteristics were compared between patients with and without LUTS. Variables were compared by t-test (age), by Wilcoxon (PSA), and by Chi-square (all others).
A log-binomial regression model was fit to a random sample of 80% of eligible subjects. This model included family history, prior negative biopsy, age >65, PSA ≥3, no LUTS, and an interaction term of no LUTS & PSA ≥3.
Risk predictions from this model were compared to estimates from the Prostate Cancer Prevention Trial Risk Calculator (PCPTRC). A validation subset of 20% was used for model comparison, evaluated by decision curve analysis. 
Complete data were available for 1,482 subjects. Among these, 296 (20%) were reserved for validation.
Patients with no LUTS were slightly younger and had lower PSA (Table 1).
The association between cancer risk and PSA in our cohort depends on the absence of LUTS (Figure 1, Table 2).
No LUTS is most predictive of prostate cancer in men >65 years old with PSA ≥3 (Figure 1, Table 2).
Association between PSA and cancer risk depends on LUTS.
Absence of LUTS is most predictive of prostate cancer in older men with high PSA.
A clinical model using categorized pre-biopsy characteristics performed as well as the published PCPTRC and may support a simpler algorithm for in office PCa risk prediction.