U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Bradley LA, Palomaki G, Gutman S, et al. PCA3 Testing for the Diagnosis and Management of Prostate Cancer [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2013 Apr. (Comparative Effectiveness Reviews, No. 98.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of PCA3 Testing for the Diagnosis and Management of Prostate Cancer

PCA3 Testing for the Diagnosis and Management of Prostate Cancer [Internet].

Show details

Appendix KSummary of the Remaining Combined Analyses for KQ 1 and KQ 2

Comparator: Percent Free serum PSA (%fPSA)

Intermediate Outcome: Diagnostic Accuracy

Key Points -The extent of %fPSA elevations were compared with PCA3 scores to determine their diagnostic accuracy to predict prostate biopsy results (cancer / no cancer). Six of eight included studies (Table K1) enrolled only men with tPSA elevations in the grey zone (usually 2 to 10 ng/mL).

Table K1. Summary results for the eight available analytic comparisons of PCA3 versus %fPSA in matched populations of men having prostate biopsies.

Table K1

Summary results for the eight available analytic comparisons of PCA3 versus %fPSA in matched populations of men having prostate biopsies.

  • Area under the curve (AUC). Five studies1-5 reported AUC results for PCA3 and %fPSA (Table K2). Three studies1, 4, 5focused on the ‘grey zone’ of tPSA values. Two ‘grey zone’ studies1, 5found PCA3 to be better by 12 percent to 17 percent, while two studies that enrolled men with all levels of tPSA2, 3 found PCA3 and %fPSA to be similar (-3% and +6%). The third ‘grey zone’ study4 did not report the actual AUC for %fPSA, but did report the PCA3 AUC to be significantly higher.
  • Reported medians and standard deviations (SD). Three studies2, 3, 6 provided some data for analysis (Table K3). One study in the ‘grey zone’6 found PCA3 to be better (+0.86 z). Another2 found %fPSA to be better (-0.91 z) and the third3 found them to be nearly equivalent (+0.15 z). Both of these latter studies finding smaller effects were not restricted to the ‘grey zone’.
  • Performance at a PCA3 cut-off score of 35. Only one ‘grey zone’ study7 (Table K4) reported the sensitivity and specificity of PCA3 at this cut-off that could be compared with the %fPSA sensitivity. PCA3 sensitivity was higher by 17 percent.
  • ROC Curves - Sensitivity/Specificity. Five studies1, 3-5, 7 (Table K5) provided a ROC curve, or data representing an ROC curve, for both markers. Two datasets were available from one study3, providing results for both initial and repeat biopsy. At all specificities noted, the median (PCA3 - %fPSA sensitivities) showed PCA3 to be slightly better. However, four of the six datasets were from studies focusing on the ‘grey zone’.1, 4, 5, 7
  • Regression analysis. Two studies6, 7 (Table K6) provided sufficient data to apply the respective regression coefficients to create a relative odds ratio (OR) between the 25th and 75th centiles of the two distributions. Both studies had ORs showing PCA3 to perform better (2.88, 1.70), and both focused on the ‘grey zone’ of tPSA.
Table K2. Comparing performance of PCA3 levels and percent free PSA (%fPSA) measurements in matched studies via AUC analysis to correctly diagnose prostate cancer, as defined by a positive biopsy.

Table K2

Comparing performance of PCA3 levels and percent free PSA (%fPSA) measurements in matched studies via AUC analysis to correctly diagnose prostate cancer, as defined by a positive biopsy.

Table K3. Comparison of PCA3 and percent free PSA (%fPSA) differences in central estimates in men with positive and negative prostate biopsy results, after accounting for study-specific variability in measurements.

Table K3

Comparison of PCA3 and percent free PSA (%fPSA) differences in central estimates in men with positive and negative prostate biopsy results, after accounting for study-specific variability in measurements.

Table K4. A comparison of PCA3 and %fPSA in identifying a positive prostate biopsy among matched studies: difference in sensitivities at the fixed specific associated with the commonly used PCA3 score cut-off of 35.

Table K4

A comparison of PCA3 and %fPSA in identifying a positive prostate biopsy among matched studies: difference in sensitivities at the fixed specific associated with the commonly used PCA3 score cut-off of 35.

Table K5. Sensitivity Differences (PCA3 - %fPSA) at PCA3 False positive rates (1 – Specificity) from 20% to 80% in matched studies to identify positive biopsy men.

Table K5

Sensitivity Differences (PCA3 - %fPSA) at PCA3 False positive rates (1 – Specificity) from 20% to 80% in matched studies to identify positive biopsy men.

Table K6. Comparison of modeled univariate continuous odds ratios (OR) for PCA3 and %fPSA in matched studies.

Table K6

Comparison of modeled univariate continuous odds ratios (OR) for PCA3 and %fPSA in matched studies.

Study Characteristics

A total of seven studies1-7 reported PCA3 and %fPSA comparisons that could be used in one or more of the matched analyses (Table K1). The table is sorted by the number of enrolled men. Five of the seven studies1, 4-7 focused on the ‘grey zone’ of tPSA elevations. Given that there are likely to be correlations between tPSA and %fPSA, these studies might be expected to show a greater difference between PCA3 and %fPSA performance than studies in a more general population of men with tPSA elevations. Six1, 3-7 of the seven studies were consistent in finding the matched PCA3 estimate better at identifying positive prostate biopsies than the corresponding %fPSA estimate. The exception was a small study2 that enrolled men without regard to their tPSA measurements (i.e., not ‘grey zone’). Overall, the PCA3 and %fPSA comparative findings were likely generalizable to only men in the ‘grey zone’ of tPSA levels at about 2 to 10 ng/mL (Ferro4 2 to 20 ng/mL).

PCA3 and %fPSA: Area Under the Curve

Five studies1-5 reported data on AUCs, and are presented in Table K2 sorted by study size. Three1, 4, 5 of the five studies recruited only men in the ‘grey zone’ of tPSA (two of these found the AUC of PCA3 to be more than 10% higher, and the other did not report values, but did report that the AUC of PCA3 was significantly higher). In the two studies in the general population of PSA positive men2, 3, one found %fPSA to be slightly better (-0.027), and the other found PCA3 to be slightly better (0.058). Due to the variability in the study populations and the limited number of studies, no further analyses are presented.

PCA3 and %fPSA: Reported Medians and Standard Deviations

Four studies2, 3, 6, 7 provided at least some relevant data that is presented in Table K3, sorted by study size. Two of the studies6, 7 focus on the ‘grey zone’ of tPSA. One did not provide sufficient data to compute the z-scores7, but did provide the median levels of both markers in those with positive versus negative biopsies. Among the three studies with a z-score difference, one found %fPSA to be slightly better2, one found PCA3 to be slightly better6 and one found only a small difference3. Due to the limited number of studies, no further analyses are presented.

PCA3 and %fPSA: Performance at a PCA3 Cut-Off Score of 35

Only one study7 provided sufficient data to compare PCA3 sensitivity with the sensitivity of %fPSA at the specificity defined by the PCA3 cut-off score of 35 (Table K4). That study reported the PCA3 sensitivity to be 17 percent higher than the corresponding %fPSA sensitivity at the same (1-specificity) rate of 24 percent. With only a single study, no further analyses are presented.

PCA3 and %fPSA: ROC Curves - Sensitivity / Specificity

Five studies1, 3-5, 7 provided sufficient data to compare PCA3 sensitivity versus %fPSA sensitivity at (1-specificity) rates of 20 percent through 80 percent (Table K5). For one study3 two separate estimates were made; one for those men who were undergoing an initial biopsy, and another for those having a repeat biopsy. Although the majority of observations found PCA3 to have higher sensitivities at fixed specificities, half of six datasets found at least one point on the ROC curve at which %fPSA was better than PCA3. Due to the variability in the study populations and limited number of studies, no further analyses are presented.

PCA3 and %fPSA: Regression Analysis

To be included in this analysis, studies would have reported the odds ratio (OR) for each marker, when that marker was assumed to be a continuous variable. The coefficients were used to compute the ratio of the ORs at the 25th and 75th centiles for each marker. This ratio of ORs for PCA3 will then be divided by the corresponding ratio for %fPSA. Values greater than 1 indicate PCA3 provides more discrimination than %fPSA. This normalization also allows for comparisons between studies, where the coefficient is dependent on the range of %fPSA values studied.

Three studies1, 6, 7 reported results of regression analysis separately for PCA3 and for %fPSA elevations in the same population of men. Only two of the studies6, 7 provided sufficient data to complete the analysis (Table K6). Both found the ratio of ORs over the inter-quartile range to be greater for PCA3 than for %fPSA. All three studies included in this analysis restricted tPSA levels to < 10 ng/mL (‘grey zone’).

Two specific measures were sought regarding the independence of PCA3 and %fPSA as markers for prostate biopsy status. Bivariate correlations (parametric or non-parametric) and results of logistic regression. Other markers may also be included in the regression model (e.g., prostate volume). If both PCA3 and %fPSA coefficients remain essentially constant after adjusting for the other marker (and possibly, additional markers), this can be taken as evidence that the two markers together are more predictive that either alone.

Only two studies1, 6 provided some information regarding independence of PCA3 and %fPSA (Table K7). None reported information on correlation coefficients. Both reported the univariate and multivariate ORs for PCA3 which indicated little change when other makers were added. Similar results were found for %fPSA, but in one of the studies6, %fPSA was not included in the ‘best’ predictive model.

Table K7. Measures of independence of PCA3 and %fPSA in identifying men having a positive biopsy finding in matched studies.

Table K7

Measures of independence of PCA3 and %fPSA in identifying men having a positive biopsy finding in matched studies.

PCA3 and %fPSA Diagnostic Accuracy

PCA3 and %fPSA GRADE Strength of Evidence: Insufficient

  • Risk of Bias: HIGH
    %fPSA is not highly correlated with tPSA and, therefore, the partial verification bias and sampling bias identified for the tPSA analyses may not have a strong influence on the %fPSA analyses. However, studies are observational and rated poor.
  • Consistency: INCONSISTENT
    In general, ‘grey zone’ studies find PCA3 consistently higher while more generalizable studies find similar performance for the two markers (or find %fPSA slightly better). Given the small number of studies, it is not clear whether this is chance or that the %fPSA is subject to a higher bias than expected.
  • Directness: INDIRECT
    The ultimate outcome of interest is long-term morbidity / mortality, and diagnostic accuracy is an intermediate outcome that cannot be linked directly to health outcomes.
  • Precision: IMPRECISE
    A formal analysis of precision (e.g., confidence intervals) could not be perfored due to the matched nature of our analyses, the lack of original data and the limited number of included studies.

PCA3 and %fPSA Intermediate Outcome: Impact on Decisionmaking

No studies were identified that reported PCA3 and %fPSA levels along with outcomes related to the impact of decisionmaking on initial or repeat biopsy, or changing decisions leading to a reduction in unnecessary biopsies (false positives) with maintenance or improvement in cancer detection.

GRADE Strength of Evidence: Insufficient

PCA3 and %fPSA Intermediate Outcome: Biopsy-Related Harms

No studies were identified that reported PCA3 and %fPSA levels along with health outcomes related to harms of biopsy.

GRADE Strength of Evidence: Insufficient

PCA3 and %fPSA Long-Term Health Outcome: Morbidity/mortality

No studies were identified that reported PCA3 and %fPSA levels along with other long-term health outcomes, such as morbidity, mortality or quality of life.

GRADE Strength of Evidence: Insufficient

Comparator: Serum PSA Density (PSAD)

Intermediate Outcome: Diagnostic Accuracy

Key Points

PSAD measurements were compared with PCA3 scores to determine their diagnostic accuracy to predict prostate biopsy results (cancer/no cancer). Measures included in the analyses are the sensitivity, specificity (or the false positive rate equal to 1-specificity), and positive and negative predictive values.

  • Area under the curve (AUC). Three studies7, 9, 10 reported AUC estimates for PSAD and PCA3 in the same population. All found PCA3 scores to perform better than PSAD. However, two7, 8 of the three focused on the ‘grey zone’ of tPSA results. Given the potential for the PSAD performance to be reduced when enrollees with higher tPSA results are excluded, findings must, at best, be restricted to the ‘grey zone’ setting.
  • Reported medians and standard deviations (SD). Two studies7, 8 provided data, but only one8 provided sufficient data for analysis. The difference, reported as a z-score of 0.51, was in favor of PCA3.
  • Performance at a PCA3 cut-off score of 35. Two studies7, 10 reported the sensitivity and specificity of PCA3 at this cut-off and could be compared with PSAD sensitivity. One ‘grey zone’ study7 found PCA3 to be 17 percent higher, while one study in the general population9 found PCA3 to be 11 percent lower.
  • ROC Curves - Sensitivity / Specificity. Three studies7-9 provided a ROC curve, or data representing a ROC curve for both markers. Two7, 8 were ‘grey zone’ studies. At a specificity of 50 percent, the differences in corresponding (PCA3 – PSAD) sensitivities ranged from 3 to 12 percent.
  • Regression analysis. Only one study10 provided data regarding the correlation between PCA3 and PSAD among biopsy positive men, and the estimate was low (r = 0.13).

Study Characteristics

Four studies7-10 reported PCA3 and PSAD comparisons that could be used in one or more of the matched analyses (Table K8). Overall, the analyses were inconsistent in finding whether the PCA3 or PSAD measurements were better at identifying positive prostate biopsies, and any findings should be restricted to the ‘grey zone’ setting.

Table K8. Summary results for the five available analytic comparisons of PCA3 versus PSAD in matched populations of men having prostate biopsies.

Table K8

Summary results for the five available analytic comparisons of PCA3 versus PSAD in matched populations of men having prostate biopsies.

PCA3 and PSAD: Area Under the Curve

Three studies7-9 satisfied inclusion criteria and relevant data are presented in Table K9. Two found PCA3 to have a higher AUC, but both were ‘grey zone’ studies.7, 8 The third study9 allowed participants with high tPSA to be enrolled, and it found the two AUCs to be similar (-0.04). Due to the differences in the study populations and the limited number of studies, no further analyses are presented.

Table K9. Comparing PCA3 levels and PSA density (PSAD) measurements in matched studies via AUC analysis to correctly diagnose prostate cancer, as defined by a positive biopsy.

Table K9

Comparing PCA3 levels and PSA density (PSAD) measurements in matched studies via AUC analysis to correctly diagnose prostate cancer, as defined by a positive biopsy.

PCA3 and PSAD: Reported Medians and Standard Deviations

Two studies7, 8 satisfied the inclusion criteria, and relevant data are presented in Table K10. In one study8, the log standard deviation was estimated using the range of results divided by six. The difference in z-scores was 0.51, indicating a better separation between men with positive and negative biopsies for PCA3 compared to PSAD. In the other study7, it was not possible to estimate the log standard deviation. Due to the single study, no further analyses are presented.

Table K10. Comparison of PCA3 and PSA density (PSAD) in central estimates in men with positive and negative prostate biopsy results, after accounting for study-specific variability in measurements.

Table K10

Comparison of PCA3 and PSA density (PSAD) in central estimates in men with positive and negative prostate biopsy results, after accounting for study-specific variability in measurements.

PCA3 and PSAD: Performance at a PCA3 Cut-Off Score of 35

Two studies7, 9 provided sufficient data to compare PCA3 sensitivity with the sensitivity of PSAD at the specificity defined by the PCA3 cut-off score of 35 (Table K11). One study8 found PCA3 to be better, but that study focused on the ‘grey zone’. The second study9 was in the general population of tPSA/DRE positive patients, and found PSAD to be better. Due to the limited number of studies, no further analyses are presented.

Table K11. A comparison of PCA3 and PSAD in identifying a positive prostate biopsy among matched studies: difference in Sensitivities at the fixed specific associated with the commonly used PCA3 score cut-off of 35.

Table K11

A comparison of PCA3 and PSAD in identifying a positive prostate biopsy among matched studies: difference in Sensitivities at the fixed specific associated with the commonly used PCA3 score cut-off of 35.

PCA3 and PSAD: ROC Curves - Sensitivity / Specificity

Three studies7-9 provide sufficient data to compare PCA3 sensitivity versus PSAD sensitivity at (1-specificity) rates of 20 percent through 80 percent (Table K12). The results show the two tests to be similar, with point estimates sometimes higher for PCA3 and sometimes for PSAD. However, due to the differences in the populations studied and the limited number of studies, no further analyses are presented.

Table K12. Sensitivity Differences (PCA3 - PSAD) at PCA3 False positive rates (1 – Specificity) from 20% to 80% in matched studies to identify positive biopsy men.

Table K12

Sensitivity Differences (PCA3 - PSAD) at PCA3 False positive rates (1 – Specificity) from 20% to 80% in matched studies to identify positive biopsy men.

PCA3 and PSAD: Regression Analysis

Only one study10 reported results of the selected analyses regarding the independence of PCA3 and PSAD in predicting prostate biopsy results, and it only included cases (Table K14). The correlation was low (0.13).

Table K14. Comparing PCA3 and EVN scores in matched studies via AUC analysis to correctly diagnose prostate cancer, as defined by a positive needle biopsy.

Table K14

Comparing PCA3 and EVN scores in matched studies via AUC analysis to correctly diagnose prostate cancer, as defined by a positive needle biopsy.

PCA3 and PSAD Intermediate Outcome: Diagnostic Accuracy

  • Risk of Bias: HIGH.
    PSAD does not appear to be highly correlated with tPSA and, therefore, the partial verification bias, and sampling bias identified for the tPSA analyses are not expected to have a strong effect on the PSAD analyses. However, studies were observational and rated poor.
  • Consistency: UNKNOWN
    Two few data were available to assess consistency.
  • Directness: INDIRECT
    The ultimate outcome of interest is long-term morbidity / mortality, and diagnostic accuracy is an intermediate outcome that cannot be linked directly to health outcomes.
  • Precision: IMPRECISE
    A formal analysis of precision (e.g., confidence intervals) was not able to be computed due to the matched nature of our analyses, the lack of original data, and the limited number of included studies.

GRADE Strength of Evidence: Insufficient

PCA3 and PSAD - Intermediate Outcome: Impact on Decisionmaking

No studies were identified that reported PCA3 and PSAD levels along with outcomes related to the impact of decisionmaking on initial or repeat biopsy, or changing decisions leading to a reduction in unnecessary biopsies (false positives) with maintenance or improvement in cancer detection.

GRADE Strength of Evidence: Insufficient

PCA3 and PSAD - Intermediate Outcome: Biopsy-Related Harms

No studies were identified that reported PCA3 and PSAD levels along with health outcomes related to harms of biopsy.

GRADE Strength of Evidence: Insufficient

PCA3 and PSAD - Long-Term Health Outcome: Morbidity/Mortality

No studies were identified that reported PCA3 and PSAD levels along with long-term health outcomes, such as morbidity, mortality or quality of life.

GRADE Strength of Evidence: Insufficient

Comparator: Externally Validated Nomogram (EVN)

PCA3 and EVN Intermediate Outcome: Diagnostic Accuracy

Key points

Externally validated nomograms were compared with PCA3 scores to determine their diagnostic accuracy to predict prostate biopsy results (cancer / no cancer). Measures included in the analyses are the sensitivity, specificity (or the false positive rate equal to 1-specificity), and positive and negative predictive values.

  • Area under the curve (AUC). Three studies6, 11, 12reported AUC estimates for EVN and PCA3 in the same population (Table K14). Both found the better performance for PCA3 (AUC differences of 0.01 and 0.11).
  • Reported medians and standard deviations (SD). No studies provided sufficient data for analysis
  • Performance at a PCA3 cut-off score of 35. One study6 reported comparative results for PCA3 versus two separate EVN algorithms.
  • ROC Curves - Sensitivity / Specificity. Two studies6, 8 provided a ROC curve, or data representing a ROC curve for both markers. Only one of these6 reported data over the required range of specificities (20% to 80%). The second provided only limited data, with no information at the specificity of 50 percent.
  • Regression analysis. One study11 reported results and did not find a significant association.

Study Characteristics

Four studies6, 8, 11, 12 reported PCA3 and EVN comparisons that could be used in one or more of the matched analyses (Table K13). Comparisons included PCA3 to Chun's nomogram, the Prostate Cancer Prevention Trial (PCPT) model and other multivariate analyses that included variables such as tPSA level, age, DRE result, family history or race. Of the seven comparisons, all found PCA3 performed better than PSAD. The following sections provide detailed information about the five analyses performed and the specific findings.

Table K13. Summary results for the five analytic comparisons of PCA3 versus PSAD in matched populations of men having prostate biopsies.

Table K13

Summary results for the five analytic comparisons of PCA3 versus PSAD in matched populations of men having prostate biopsies.

PCA3 and EVN: Area Under the Curve

Three studies6, 12 11 satisfied inclusion criteria and relevant data are presented in Table K14. Both found that the AUC of PCA3 was greater than for EVN, but the smaller difference found for one study12 did not reach statistical significance. Due to the limited number of studies, no further analyses are presented.

PCA3 and EVN: Reported Medians and Standard Deviations

No studies were identified that provided relevant data.

PCA3 and EVN: Performance at a PCA3 Cut-Off Score of 35

One study6 satisfied the inclusion criteria and relevant data are presented in Table K15. That group used the Chun's and PCPT risk nomograms. Both found that PCA3 was better than either of the externally validated nomograms. Due to the limited number of studies, no further analyses are presented.

Table K15. Comparison of PCA3 and externally validated nomograms (EVN) in central estimates in men with positive and negative prostate biopsy results, after accounting for study-specific variability in measurements.

Table K15

Comparison of PCA3 and externally validated nomograms (EVN) in central estimates in men with positive and negative prostate biopsy results, after accounting for study-specific variability in measurements.

PCA3 and EVN: ROC Curves - Sensitivity / Specificity

Two studies6, 8 provide sufficient data to compare PCA3 sensitivity versus EVN sensitivity at (1-specificity) rates. Both used the risk algorithm by Chun. One6 provided data for specificities of 20 percent through 80 percent (Table K16) and found PCA3 to be better. The second study8 only reported results for 20 percent through 40 percent and found the EVN only slightly better. Due to the limited number of studies, no further analyses are presented.

Table K16. Sensitivity Differences (PCA3 - EVN) at PCA3 False positive rates (1 – Specificity) from 20% to 80% in matched studies to identify positive biopsy men.

Table K16

Sensitivity Differences (PCA3 - EVN) at PCA3 False positive rates (1 – Specificity) from 20% to 80% in matched studies to identify positive biopsy men.

PCA3 and EVN: Regression analysis

Two studies provided a comparison of an externally validated model with and without PCA3.11,12 In the first study12, the PCPT AUC increased from 0.653 to 0.696 with the inclusion of PCA3. However, this increase was not statistically significant. In the second study11, the inclusion of a dichotomized PCA3 score to tPSA and ‘standard of care covariates’ (age, DRE result, family history, race, number previous negative biopsies) resulted in an increased AUC from 0.707 to 0.733.

PCA3 and EVN - Intermediate Outcome: Diagnostic Accuracy

  • Risk of Bias: HIGH.
    EVNs often include tPSA and, therefore, the partial verification bias, and sampling bias identified for the tPSA analyses may have an impact. However, none of the models actually used tPSA (which are most subject to bias) and thus biasing the EVN analyses is not expected. However, studies were observational and rated poor.
  • Consistency: UNKNOWN
    Two few data were available to assess consistency.
  • Directness: INDIRECT
    The ultimate outcome of interest is long-term morbidity / mortality, and diagnostic accuracy is an intermediate outcome that cannot be linked directly to health outcomes.
  • Precision: IMPRECISE
    A formal analysis of precision (e.g., confidence intervals) was not able to be computed due to the matched nature of our analyses, the lack of original data, and the limited number of included studies.

GRADE Strength of Evidence: Insufficient

PCA3 and EVN - Intermediate Outcome: Impact on Decisionmaking

No studies were identified that reported PCA3 and EVN results along with outcomes related to the impact of decisionmaking on initial or repeat biopsy, or changing decisions leading to a reduction in unnecessary biopsies (false positives) with maintenance or improvement in cancer detection.

GRADE Strength of Evidence: Insufficient

PCA3 and EVN - Intermediate Outcome: Biopsy-Related Harms

No studies were identified that reported PCA3 and EVN results along with health outcomes related to harms of biopsy.

GRADE Strength of Evidence: Insufficient

PCA3 and EVN - Long-Term Health Outcome: Morbidity/Mortality

No studies were identified that reported PCA3 and EVN results along with long-term health outcomes, such as morbidity, mortality or quality of life.

GRADE Strength of Evidence: Insufficient

Comparator: Complexed PSA (cPSA)

PCA3 and cPSA - Intermediate Outcome: Diagnostic Accuracy

No studies were identified that provide matched data for PCA3 and cPSA levels in eligible study populations with biopsy results.

GRADE Strength of Evidence: Insufficient

PCA3 and cPSA - Intermediate Outcome: Impact on Decisionmaking

No studies were identified that reported PCA3 and cPSA levels along with outcomes related to the impact of decisionmaking on initial or repeat biopsy, or changing decisions leading to a reduction in unnecessary biopsies (false positives) with maintenance or improvement in cancer detection.

GRADE Strength of Evidence: Insufficient

PCA3 and cPSA - Intermediate Outcome: Biopsy-Related Harms

No studies were identified that reported PCA3 and cPSA levels along with health outcomes related to harms of biopsy.

GRADE Strength of Evidence: Insufficient

PCA3 and cPSA - Long-term Health Outcome: Morbidity/Mortality

No studies were identified that reported PCA3 and cPSA levels along with long-term health outcomes, such as morbidity, mortality or quality of life.

GRADE Strength of Evidence: Insufficient

Comparator: Total PSA Doubling Time (DT) and Total PSA Velocity (PSAV)

PCA3 and DT/PSAV - Intermediate Outcome: Diagnostic Accuracy

No studies were identified that provide matched data for PCA3 and DT/PSAV results levels in eligible study populations with biopsy results.

GRADE Strength of Evidence: Insufficient

PCA3 and DT/PSAV - Intermediate Outcome: Impact on Decisionmaking

No studies were identified that reported PCA3 and DT/PSAV results along with outcomes related to the impact of decisionmaking on initial or repeat biopsy, or changing decisions leading to a reduction in unnecessary biopsies (false positives) with maintenance or improvement in cancer detection.

GRADE Strength of Evidence: Insufficient

PCA3 and DT/PSAV - Intermediate Outcome: Biopsy-Related Harms

No studies were identified that reported PCA3 and DT/PSAV results along with health outcomes related to harms of biopsy.

GRADE Strength of Evidence: Insufficient

PCA3 and DT/PSAV - Long-term Health Outcome: Morbidity/Mortality

Key points

No studies were identified that reported PCA3 and DT/PSAV results along with long-term health outcomes, such as morbidity, mortality or quality of life.

GRADE Strength of Evidence: Insufficient

Comparator: Multivariate Models including PCA3 and tPSA

Study Characteristics

Three studies1, 6, 7 reported multivariate models (logistic regression) that included both tPSA elevations and PCA3 scores. Only one of these7 provided results for both a base model and the base model plus PCA3 scores. The base model included the man's age, DRE (categorical), tPSA (continuous), and prostate volume (continuous); all were statistically significant (p<0.01). PCA3 score as a continuous variable was then included. The coefficients in the base model were essentially unchanged with all point estimates being within the 95 percent CI of the base model estimates and all remained statistically significant. The p-value associated with PCA3 measurements was <0.001. This implied that the PCA3 scores added independent information.

However, there are several limitations to the study. Neither the tPSA nor PCA3 measurements were reported to have been transformed. An underlying assumption of logistic regression is that continuous variables are reasonably Gaussian and this was most likely violated by the modeling. Second, the authors provided no information regarding the change in sensitivity at fixed specificities. Rather, they provided a modest increase in predictive accuracy, a difficult estimate to interpret.

Multivariate Models Including PCA3 and tPSA - Outcome: Diagnostic Accuracy

GRADE Strength of Evidence: Insufficient

  • Risk of Bias: HIGH.
    Multivariate models often include tPSA and, therefore, the partial verification bias, and sampling bias identified for the tPSA analyses may have a modest impact. However, studies were observational and rated poor.
  • Consistency: UNKNOWN
    The limited number of studies (three) did not report results in a consistent manner (e.g., correlations or effect sizes).
  • Directness: INDIRECT
    The ultimate outcome of interest is long-term morbidity / mortality; the subject of Key Question 3.
  • Precision: IMPRECISE
    A formal analysis of precision (e.g., confidence intervals) was not able to be computed due to the matched nature of our analyses, the lack of original data, and the limited number of included studies.

Multivariate Models Including PCA3 and tPSA - Intermediate Outcome: Impact on Decisionmaking

No studies were identified that reported the multivariate model results along with outcomes related to the impact of decisionmaking on initial or repeat biopsy, or changing decisions leading to a reduction in unnecessary biopsies (false positives) with maintenance or improvement in cancer detection.

GRADE Strength of Evidence: Insufficient

Multivariate Models Including PCA3 and tPSA - Intermediate Outcome: Biopsy-Related Harms

No studies were identified that reported the multivariate model results along with health outcomes related to harms of biopsy.

GRADE Strength of Evidence: Insufficient

Multivariate Models Including PCA3 and tPSA - Long-term Health Outcome: Morbidity/Mortality

No studies were identified that reported the multivariate model results along with long-term health outcomes, such as morbidity, mortality or quality of life.

GRADE Strength of Evidence: Insufficient

References

1.
Aubin SM, Reid J, Sarno MJ, et al. PCA3 molecular urine test for predicting repeat prostate biopsy outcome in populations at risk: validation in the placebo arm of the dutasteride REDUCE trial. J Urol. 2010 Nov;184(5):1947–52. [PubMed: 20850153]
2.
Auprich M, Augustin H, Budaus L, et al. A comparative performance analysis of total prostate-specific antigen, percentage free prostate-specific antigen, prostate-specific antigen velocity and urinary prostate cancer gene 3 in the first, second and third repeat prostate biopsy. BJU Int. 2012 Jun;109(11):1627–35. [PubMed: 21939492]
3.
Bollito E, De Luca S, Cicilano M, et al. Prostate cancer gene 3 urine assay cutoff in diagnosis of prostate cancer: a validation study on an Italian patient population undergoing first and repeat biopsy. Analytical and quantitative cytology and histology / the International Academy of Cytology [and] American Society of Cytology. 2012 Apr;34(2):96–104. [PubMed: 22611765]
4.
Ferro M, Bruzzese D, Perdona S, et al. Predicting prostate biopsy outcome: prostate health index (phi) and prostate cancer antigen 3 (PCA3) are useful biomarkers. Clin Chim Acta. 2012 Aug 16;413(15-16):1274–8. [PubMed: 22542564]
5.
Ploussard G, Haese A, Van Poppel H, et al. The prostate cancer gene 3 (PCA3) urine test in men with previous negative biopsies: does free-to-total prostate-specific antigen ratio influence the performance of the PCA3 score in predicting positive biopsies? BJU Int. 2010 Oct;106(8):1143–7. [PubMed: 20230386]
6.
Perdona S, Cavadas V, Di Lorenzo G, et al. Prostate cancer detection in the “grey area” of prostate-specific antigen below 10 ng/ml: head-to-head comparison of the updated PCPT calculator and Chun's nomogram, two risk estimators incorporating prostate cancer antigen 3. European urology. 2011 Jan;59(1):81–7. [PubMed: 20947244]
7.
de la Taille A, Irani J, Graefen M, et al. Clinical evaluation of the PCA3 assay in guiding initial biopsy decisions. J Urol. 2011 Jun;185(6):2119–25. [PubMed: 21496856]
8.
Ochiai A, Okihara K, Kamoi K, et al. Prostate cancer gene 3 urine assay for prostate cancer in Japanese men undergoing prostate biopsy. Int J Urol. 2011 Mar;18(3):200–5. [PubMed: 21332814]
9.
Wu AK, Reese AC, Cooperberg MR, et al. Utility of PCA3 in patients undergoing repeat biopsy for prostate cancer. Prostate Cancer Prostatic Dis. 2012 Mar;15(1):100–5. [PubMed: 22042252]
10.
Durand X, Xylinas E, Radulescu C, et al. The value of urinary prostate cancer gene 3 (PCA3) scores in predicting pathological features at radical prostatectomy. BJU Int. 2012 Jul;110(1):43–49. [PubMed: 22221521]
11.
U.S. Food and Drug Administration. Summary of Safety and Effectiveness Data: PROGENSA PCA3 Assay. 2012. [July 16, 2012]. http://www​.accessdata​.fda.gov/cdrh_docs/pdf10/P100033b.pdf.
12.
Ankerst DP, Groskopf J, Day JR, et al. Predicting prostate cancer risk through incorporation of prostate cancer gene 3. J Urol. 2008 Oct;180(4):1303–8. discussion 08. [PubMed: 18707724]
Copyright Notice
Bookshelf ID: NBK132741
Contents

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.8M)

In this Page

Other titles in these collections

Recent Activity

ClearTurn OffTurn On

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...