An interim analysis of a validation study suggests that urine prostate cancer antigen 3 (PCA3) levels and urine TMPRSS2:ERG (T2:ERG) have additive utility in predicting clinically significant prostate tumor volume, and their use together may enhance the selection and monitoring of candidates with low-volume/low-grade disease for active surveillance.
Atlanta-An interim analysis of a validation study suggests that urine prostate cancer antigen 3 (PCA3) levels and urine TMPRSS2:ERG (T2:ERG) have additive utility in predicting clinically significant prostate tumor volume, and their use together may enhance the selection and monitoring of candidates with low-volume/low-grade disease for active surveillance.
New markers are needed to select patients for active surveillance given the lack of specificity of PSA testing. Previous studies have shown that urine PCA3 levels correlated with positive versus negative biopsies and total tumor volume in radical prostatectomy specimens, said first author John Davis, MD, assistant professor of urology at the University of Texas MD Anderson Cancer Center, Houston, one of three sites participating in the study.
Of the planned 300 participants, 257 had completed data for interim analysis. Eighty-three percent were treated after a first biopsy; 86% had at least 10 biopsy cores. Biopsy Gleason 6 cancer was detected in 41% and pathologic Gleason 6 in 17%. Sixty-two percent were upgraded at biopsy from Gleason 6 to Gleason ≥7. Seventy percent had organ-confined disease (pT2) and 8% had lymph node disease (N1). Sixty-three percent had tumor volume >0.5 cc. Some 87% had significant cancer, defined as tumor volume >0.5 cc, Gleason score >6, or non-organ-confined cancer.
Serum PSA level, urine PCA3, and urine T2 were significantly higher in specimens with significant cancer. Median values for insignificant versus significant cancer were 3.0 ng/mL versus 5.0 ng/mL (p<.0001) for PSA, 18 versus 39 (p<.0001) for urine PCA3 score, and 6 versus 23 (p=.0262) for urine T2:ERG score.
Similarly, for cancers that were pathologic Gleason score <7 versus ≥7, urine PCA3 score (19 vs. 39; p<.0001) and urine T2:ERG score (7 vs. 24; p=.011) were significantly lower. Both scores were also significantly lower when comparing organ-confined versus non-organ-confined cancers and biopsy Gleason 6 versus upgraded disease.
Of the 105 patients with a biopsy Gleason 6 score who are candidates for active surveillance or treatment, the median PCA3 score was 31 and the median T2:ERG score was 18.
Of the patients with biopsy Gleason 6 disease, 40 (38%) had pathologic Gleason 6 disease and 65 (62%) were upgraded. The median PCA3 score in those not upgraded was 19 versus 37 in those who were upgraded (p=.0003), and the median T2:ERG score was 5 in those not upgraded versus 24 in those upgraded (p=.0862). The receiver operating characteristic area under the curve for all three markers combined for upgrading was 0.7184, which was superior to the markers individually (PCA3=0.7164, T2:ERG=0.6066, PSA=0.6521).
The PCA3 score increased significantly when tumor volume increased from <0.5 cc to 0.5 to 2.0 cc, "but the increase in PCA3 score for tumors >2.0 cc was not significant," said Dr. Davis.
Using area under the curve relative to significant cancer, the data suggest a meaningful improvement when PCA3 and T2:ERG scores are added to serum PSA, with a difference in sensitivity of 12.7% at 90% specificity, but this improvement was not statistically significant.
High predictive values for PCA3, T2 scores
Predictive values were high for very high PCA3 and T2 scores: For a PCA3 cutoff of 70, the positive predictive value (PPV) was 98.3%, and for a T2:ERG cutoff of 160, the PPV was 100%.
"The clinical utility might be that if the PSA biopsy scores are telling you the patient is a good surveillance candidate but you're off the charts on the other two [PCA3 and T2:ERG], it may be a red flag that it's not the proper candidate and needs to be looked at further," he said.
"Overall, the vision in using these [urinary markers] is in what I call 'sifting populations.' When the biopsy looks good for surveillance, it may be an indication to re-sift that population with additional urinary markers, and people might incorporate imaging and other strategies," he added. "The idea is that most Gleason 6 biopsy patients are good surveillance candidates but a few of them can come up and bite you, and to the extent that we can find those and treat them properly, it's an important advance."