- Email: [email protected]
Is complexed PSA a valuable marker for routine diagnostic?
725 IS
COMPLEXED DIAGNOSTIC?
PSA
A VALUABLE
MARKER
FOR
ROUTINE
Oremek G., Saooutzis N., Pelekanos C., Eden F., Jonas D.
Haese A.‘. Dworschack R.T.‘, Partin A.W.’
University, Urology, Frankfurt i Main, Germany INTRODUCTION complex form of for early detection the ratio of fPSA value of cPSA in
726 UTILISATION OF PERCENT FREE PSA IN THE TOTAL PSA (TPSA) RANGE 2-4 NC/ML DOES NOT SUBSTANTIALLY INCREASE THE NUMBER OF BIOPSIES NEEDED TO FIND SIGNIFICANT PROSTATE CANCER
& OBJECTIVES: Prostatic tumour tissue produce more PSA (cPSA) than free PSA (fPSA). investigating possibilities of prostate cancer cPSA is supposed to be more sensitive than and tPSA. The aim of the study was to evaluate the diagnostic early detection of malignant prostatic tumour.
MATERIAL & METHODS: We evaluated the new cPSA test comparing it with the already routinely used fPSA and tPSA test. The study comprised a total of 250 patients with different urological affections attending the Division of Urology, University of Frankfurt, Germany. Biopsy and histologic examination were done in all cases. The histologic report was compared with the laboratory results done by immunologic and electrochemiluminescence methods. cPSA was measured on the ACS CENTAUR ANALYSER of BAYER CORP. The results for fPSA and tPSA on the Bayer analyser were compared to the results obtained from the Roche’s Elecsys System.
‘Johns Hopkins Medical Institution, The Brady Urological Institute, Baltimore, United States of America. ‘Roche Inc. Clinical Research, Indianapolis, United States of America
INTRODUCTION & OBJECTIVES: Percent free PSA (%fPSA) is a useful adjuct for selection of patients for prostate biopsy in the PSA-range 4.lOng/ml. More recently it has been shown that up to 25% of men with PSA-levels between 2.6-4ng/ml harbour significant prostate cancers. %fF’SAhas been suggested as an aid in the decision to biopsy in this range as well, however, concerns exist that the number of biopsies needed to detect one cancer in
this range may be inappropriately high. In a prospective referral population, we evaluated sensitivity and specificity of various %fF’SAcutoffs and determined the number of biopsies needed to detect one cancer in the PSA-range 2.0.4ng/ml in men with a benign DRE. The results were compared to those obtained from the PSA-range 4.long/ml with benign DRE from the same referral cohort MATERIAL & METHODS: Total and free PSA were measured using the Elecsys 2010 PSA-assay. %fF’SAwas calculated. From the initial Database of 1602 men, 756 had a benign DRE and a PSA 4.IOngiml, and 219 had a bemgn DRE and a PSA>Z-3.99nylml. For the PSA-range 2.0-3.99, sensitivity, specificity, numbers of true positive (cancer detection) and false positive (no ewdence for cancer) biopsies were determined. The ratio of true positive
to false positive determined the number of biopsies needed to detect one cancer The procedure was repeated for the PSA-range 4-lOng/ml and results were compared.
RESULTS: 102 had a benign tumour of the prostatic gland and 148 had a malignant process diagnosed by histological examination. The cut-off level for cPSA was at 2.3 ng/ml. Within a range of 2.0-4.0 ng/ml tPSA and a cut-otT of 2.3 ng/ml for cPSA we found a tumour sensitivity of the test in 92%, proved by the histologic report. cPSA allows us a detection of malignant process at an earlier stage than tPSA and even within a “normal-range” of tPSA between 2.04.0 ngiml. At a concentration of tPSA between 4.1 and 10 ng/ml we talk about a grey zone. With measurement of cPSA we achieved in 71% of our cases a more specific result than in obtaining the PSA ratio. We did not observe any adverse result in the concentration of cPSA concerning manipulation of the prostatic gland as seen before in the measurement of fPSA. CPSA is much more stable regarding to transportation and storage. There is no loss of concentration level observed up to -2O”C, whereas t?SA shows a loss of 10%.
CONCLUSIONS:
CONCLUSIONS: We conclude that the new tumour marker cPSA has a higher value for specificity in the concentration range of tPSA from 2.0-4.0 ng/ml and as well in the grey zone between 4.0-10 ng/ml.
4ng/ml, a cutoff for %fPSA between I8-22% detects every second cancer and 3-5 biopsies are needed to detect one cancer. We conclude, that %PSA can be applied to the PSA-range between 2.4ng/ml to detect prostate cancer and only moderately increases the biopsies needed to detect one aigmficant cancer.
727
728
COMPLEXED PROSTATE CLINICAL PARAMETERS CANCER
SPECIFIC ANTIGEN COMBINED WITH ENHANCES PREDICTION OF PROSTATE
RESULTS:
7!4l cancers detected in the PSA-range 2.4ng/ml had a Gleason Score ~5, 28/41 had a Gleason Score of 6,514l had a Gleason Score of 7 and I patient had a Gleason Score of 8. In the PSA-range 4-long/ml, high sensitivity for prostate cancer detection is critical and 3 biopsies are needed to detect one cancer. In the PSA-range below
AGE RELATED REFERENCE RANGES FOR TOTAL SERUM PSA IN AN ASYMPTOMATIC EARLY PROSTATE CANCER DETECTION POPULATION
Bartsch G.‘, Cheli C.‘. Thiel R.‘, Klocker H.‘. Hormnger W.’
Srberi G.,. Hominger W.‘. Rcrger A ‘_Robertson C.‘. Bartsch G.‘, Rovlc P.’
‘University of Innsbruck, Urology, Innsbruck, Austria. ‘Bayer Corporation, Bayer, Tarrytown, United States of America. ‘Thiel Statistic Consultants, Thiel, Oxford, CT, United States of America
‘European Institute of Oncology. Urology, Innsbruck, Austria
INTRODUCTION & OBJECTIVES: Biopsy decisions are often based on absolute cutoff values for total PSA. Complexed PSA (cPSA) has been recently shown to enhance specificity over traditional total PSA testing. We sought to develop a model that would enhance prostate cancer detection using patient clinical parameters and their values of total PSA and cPSA. MATERIAL & METHODS: Data from 2,312 men who underwent prostate biopsies including DRE results, age, prostate volume, cPSA and total PSA levels were used to develop a model for prediction of biopsy outcome using logistic regression analysis with a split-sample approach for cross-validation. 60% of the patients were randomly drawn and a binary logistic regression analysis was performed using step-wise logistic selection of variables based on the Wald statistic. The remaining 40% of patients were used to validate the model. RESULTS: The significant variables associated with biopsy outcome by order of significance were cPSA followed by prostate volume, DRE, and age. Total PSA was not found to be a significant predictor in the model. Using this model, individual probabilities for prostate cancerqetection were computed. Receiver operating characteristic curve (ROC) analysis showed that the area under the curve (AUC) was 0.77. At an individual probability using a cutoff of 0.117 resulted in a sensitivity of 90% and a specificity of 37.6%. The model parameters determined in the training set were used to calculate individual probabilities in the validation set. As in the training set, using the cutoff of 0.117 resulted in a sensitivity of 90% and a specificity of 38.6%. In comparison, the AUC for cPSA as a single parameter provided an AUC of 0.69 with a sensitivity of 90% and specificity of 24.0%. The data further indicated that a positive DRE, increasing cPSA level, and increasing age increased the risk of predicting prostate cancer. However, a decreasing prostate volume measurement indicated a higher risk for the presence of cancer. CONCLUSIONS: These results indicate that cPSA is more associated with the risk of finding prostate cancer on biopsy than total PSA. The combination of clinical indicators signiticantly increases the precision of cPSA in the detection of prostate cancer. European
Urology Supplements
2 (2003)
No. 1, pp.
184
Milan, Milan, Italy, ‘University
of Innsbruck.
INTRODUCTION & OBJECTIVES: The aim of the following study was to evaluate “normal total semm PSA ranges” for various age groups using the database of the Tyrol PSA Screening project. MATERIAL & METHODS: A total of 228792 PSA tests were evaluated between 1993 and September 2000. The I3254 tests carried out because the man reported symptoms were excluded. We fitted a linear model to describe the relationship between the logarithm of PSA and age using Fractional Polynomials to derive the percentiles of the PSA distribution at each given age. RESULTS: We found a quadratic relationship between the logarithm of total semm PSA and age. The percentiles of the PSA distribution at a variety of ages are shown in the table below:
Median total serum PSA increases with age from 0.7 ngiml among men aged 45 to I .6 among men aged 75. Ninety-five per cent of men aged 55 have a total serum PSA in the range 0 - 4.0 ngiml, while the corresponding range for men aged 65 is 0 7.1 ng/mI.
.
CONCLUSIONS: The ranges reported here are based on an extremely large number of PSA tests carried out on a population of asymptomatic men. These data show that both median PSA and the range of PSA increases with age among an asymptomatic screened population. These results can be used to assess whether a total serum PSA value at a certain age can be considered as normal. By themselves these data do not provide cut-off points for deciding if patients should have a biopsy, as the data have not yet been linked with the results of a biopsy. In conjunction with rules for biopsy on the basis of PSA these data can be used to predict the expected number ofbiopsies to be carried out in future PSR screening projects.
COMPLEXED DIAGNOSTIC?
PSA
A VALUABLE
MARKER
FOR
ROUTINE
Oremek G., Saooutzis N., Pelekanos C., Eden F., Jonas D.
Haese A.‘. Dworschack R.T.‘, Partin A.W.’
University, Urology, Frankfurt i Main, Germany INTRODUCTION complex form of for early detection the ratio of fPSA value of cPSA in
726 UTILISATION OF PERCENT FREE PSA IN THE TOTAL PSA (TPSA) RANGE 2-4 NC/ML DOES NOT SUBSTANTIALLY INCREASE THE NUMBER OF BIOPSIES NEEDED TO FIND SIGNIFICANT PROSTATE CANCER
& OBJECTIVES: Prostatic tumour tissue produce more PSA (cPSA) than free PSA (fPSA). investigating possibilities of prostate cancer cPSA is supposed to be more sensitive than and tPSA. The aim of the study was to evaluate the diagnostic early detection of malignant prostatic tumour.
MATERIAL & METHODS: We evaluated the new cPSA test comparing it with the already routinely used fPSA and tPSA test. The study comprised a total of 250 patients with different urological affections attending the Division of Urology, University of Frankfurt, Germany. Biopsy and histologic examination were done in all cases. The histologic report was compared with the laboratory results done by immunologic and electrochemiluminescence methods. cPSA was measured on the ACS CENTAUR ANALYSER of BAYER CORP. The results for fPSA and tPSA on the Bayer analyser were compared to the results obtained from the Roche’s Elecsys System.
‘Johns Hopkins Medical Institution, The Brady Urological Institute, Baltimore, United States of America. ‘Roche Inc. Clinical Research, Indianapolis, United States of America
INTRODUCTION & OBJECTIVES: Percent free PSA (%fPSA) is a useful adjuct for selection of patients for prostate biopsy in the PSA-range 4.lOng/ml. More recently it has been shown that up to 25% of men with PSA-levels between 2.6-4ng/ml harbour significant prostate cancers. %fF’SAhas been suggested as an aid in the decision to biopsy in this range as well, however, concerns exist that the number of biopsies needed to detect one cancer in
this range may be inappropriately high. In a prospective referral population, we evaluated sensitivity and specificity of various %fF’SAcutoffs and determined the number of biopsies needed to detect one cancer in the PSA-range 2.0.4ng/ml in men with a benign DRE. The results were compared to those obtained from the PSA-range 4.long/ml with benign DRE from the same referral cohort MATERIAL & METHODS: Total and free PSA were measured using the Elecsys 2010 PSA-assay. %fF’SAwas calculated. From the initial Database of 1602 men, 756 had a benign DRE and a PSA 4.IOngiml, and 219 had a bemgn DRE and a PSA>Z-3.99nylml. For the PSA-range 2.0-3.99, sensitivity, specificity, numbers of true positive (cancer detection) and false positive (no ewdence for cancer) biopsies were determined. The ratio of true positive
to false positive determined the number of biopsies needed to detect one cancer The procedure was repeated for the PSA-range 4-lOng/ml and results were compared.
RESULTS: 102 had a benign tumour of the prostatic gland and 148 had a malignant process diagnosed by histological examination. The cut-off level for cPSA was at 2.3 ng/ml. Within a range of 2.0-4.0 ng/ml tPSA and a cut-otT of 2.3 ng/ml for cPSA we found a tumour sensitivity of the test in 92%, proved by the histologic report. cPSA allows us a detection of malignant process at an earlier stage than tPSA and even within a “normal-range” of tPSA between 2.04.0 ngiml. At a concentration of tPSA between 4.1 and 10 ng/ml we talk about a grey zone. With measurement of cPSA we achieved in 71% of our cases a more specific result than in obtaining the PSA ratio. We did not observe any adverse result in the concentration of cPSA concerning manipulation of the prostatic gland as seen before in the measurement of fPSA. CPSA is much more stable regarding to transportation and storage. There is no loss of concentration level observed up to -2O”C, whereas t?SA shows a loss of 10%.
CONCLUSIONS:
CONCLUSIONS: We conclude that the new tumour marker cPSA has a higher value for specificity in the concentration range of tPSA from 2.0-4.0 ng/ml and as well in the grey zone between 4.0-10 ng/ml.
4ng/ml, a cutoff for %fPSA between I8-22% detects every second cancer and 3-5 biopsies are needed to detect one cancer. We conclude, that %PSA can be applied to the PSA-range between 2.4ng/ml to detect prostate cancer and only moderately increases the biopsies needed to detect one aigmficant cancer.
727
728
COMPLEXED PROSTATE CLINICAL PARAMETERS CANCER
SPECIFIC ANTIGEN COMBINED WITH ENHANCES PREDICTION OF PROSTATE
RESULTS:
7!4l cancers detected in the PSA-range 2.4ng/ml had a Gleason Score ~5, 28/41 had a Gleason Score of 6,514l had a Gleason Score of 7 and I patient had a Gleason Score of 8. In the PSA-range 4-long/ml, high sensitivity for prostate cancer detection is critical and 3 biopsies are needed to detect one cancer. In the PSA-range below
AGE RELATED REFERENCE RANGES FOR TOTAL SERUM PSA IN AN ASYMPTOMATIC EARLY PROSTATE CANCER DETECTION POPULATION
Bartsch G.‘, Cheli C.‘. Thiel R.‘, Klocker H.‘. Hormnger W.’
Srberi G.,. Hominger W.‘. Rcrger A ‘_Robertson C.‘. Bartsch G.‘, Rovlc P.’
‘University of Innsbruck, Urology, Innsbruck, Austria. ‘Bayer Corporation, Bayer, Tarrytown, United States of America. ‘Thiel Statistic Consultants, Thiel, Oxford, CT, United States of America
‘European Institute of Oncology. Urology, Innsbruck, Austria
INTRODUCTION & OBJECTIVES: Biopsy decisions are often based on absolute cutoff values for total PSA. Complexed PSA (cPSA) has been recently shown to enhance specificity over traditional total PSA testing. We sought to develop a model that would enhance prostate cancer detection using patient clinical parameters and their values of total PSA and cPSA. MATERIAL & METHODS: Data from 2,312 men who underwent prostate biopsies including DRE results, age, prostate volume, cPSA and total PSA levels were used to develop a model for prediction of biopsy outcome using logistic regression analysis with a split-sample approach for cross-validation. 60% of the patients were randomly drawn and a binary logistic regression analysis was performed using step-wise logistic selection of variables based on the Wald statistic. The remaining 40% of patients were used to validate the model. RESULTS: The significant variables associated with biopsy outcome by order of significance were cPSA followed by prostate volume, DRE, and age. Total PSA was not found to be a significant predictor in the model. Using this model, individual probabilities for prostate cancerqetection were computed. Receiver operating characteristic curve (ROC) analysis showed that the area under the curve (AUC) was 0.77. At an individual probability using a cutoff of 0.117 resulted in a sensitivity of 90% and a specificity of 37.6%. The model parameters determined in the training set were used to calculate individual probabilities in the validation set. As in the training set, using the cutoff of 0.117 resulted in a sensitivity of 90% and a specificity of 38.6%. In comparison, the AUC for cPSA as a single parameter provided an AUC of 0.69 with a sensitivity of 90% and specificity of 24.0%. The data further indicated that a positive DRE, increasing cPSA level, and increasing age increased the risk of predicting prostate cancer. However, a decreasing prostate volume measurement indicated a higher risk for the presence of cancer. CONCLUSIONS: These results indicate that cPSA is more associated with the risk of finding prostate cancer on biopsy than total PSA. The combination of clinical indicators signiticantly increases the precision of cPSA in the detection of prostate cancer. European
Urology Supplements
2 (2003)
No. 1, pp.
184
Milan, Milan, Italy, ‘University
of Innsbruck.
INTRODUCTION & OBJECTIVES: The aim of the following study was to evaluate “normal total semm PSA ranges” for various age groups using the database of the Tyrol PSA Screening project. MATERIAL & METHODS: A total of 228792 PSA tests were evaluated between 1993 and September 2000. The I3254 tests carried out because the man reported symptoms were excluded. We fitted a linear model to describe the relationship between the logarithm of PSA and age using Fractional Polynomials to derive the percentiles of the PSA distribution at each given age. RESULTS: We found a quadratic relationship between the logarithm of total semm PSA and age. The percentiles of the PSA distribution at a variety of ages are shown in the table below:
Median total serum PSA increases with age from 0.7 ngiml among men aged 45 to I .6 among men aged 75. Ninety-five per cent of men aged 55 have a total serum PSA in the range 0 - 4.0 ngiml, while the corresponding range for men aged 65 is 0 7.1 ng/mI.
.
CONCLUSIONS: The ranges reported here are based on an extremely large number of PSA tests carried out on a population of asymptomatic men. These data show that both median PSA and the range of PSA increases with age among an asymptomatic screened population. These results can be used to assess whether a total serum PSA value at a certain age can be considered as normal. By themselves these data do not provide cut-off points for deciding if patients should have a biopsy, as the data have not yet been linked with the results of a biopsy. In conjunction with rules for biopsy on the basis of PSA these data can be used to predict the expected number ofbiopsies to be carried out in future PSR screening projects.