| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Enzymes and Protein Markers |
1
Centro Nazionale Applicazione Biotecnologie in Oncologia and
2
Division of Urology, Regional Hospital, 30122 Venice, Italy.
3
Urology Clinic, University of Padua, 35128 Padua, Italy.
4
Central Laboratory, Regional Hospital, 35128 Padua,
Italy.
5
Central Laboratory, Le Molinette Hospital, 10100 Turin,
Italy.
6
Urology Clinic, University of Turin, 10100 Turin, Italy.
7
Urology Clinic and
8
Nuclear Medicine Unit,
University of Sassari, 07100 Sassari, Italy.
9
Central Laboratory and
10
Division of
Urology, Regional Hospital, 31100 Treviso, Italy.
a Address correspondence to this author at: Centro Regionale Indicatori Biochimici di Tumore, Ospedale Civile, 30122 Venezia, Italy. Fax 39-41-5294532; e-mail cnabo{at}provincia.venezia.it.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
4 µg/L with only 3.8% false-negative results. The post-test
probability of percent free PSA was, however, <50% in men 50–70
years of age, using cutoff points providing sensitivity from 99% to
80%. Percent free PSA is superior to total PSA in distinguishing
primary CaP from BPH in patients with total PSA between 2 and 30
µg/L. In men with low total PSA, the diagnostic performance of the
percent free PSA assay may be optimized by controlling methodological
variability. The percent free PSA assay is effective in reducing the
rate of unnecessary biopsies in men with total PSA >4 µg/L. However,
the post-test probability provided by percent free PSA is relatively
low in asymptomatic patients 50–70 years of age. |
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The serine protease prostate-specific antigen (PSA) was shown to be a useful prostate-specific tumor marker soon after it was first detected (5). However, several studies have shown that serum PSA concentrations between 4 and 10 µg/L were not accurate for CaP diagnosis (6)(7).
Several PSA derivatives have been evaluated with the goal of increasing the diagnostic accuracy of the antigen. PSA density, PSA velocity, and age-adjusted reference ranges have been studied extensively. However, none of the above PSA approaches could markedly improve the diagnostic accuracy of CaP in patients with PSA from 4 to 10 µg/L (6)(7)(8).
Lilja et al. (9) and Stenman et al. (10) showed
independently in 1991 that serum PSA exists in different molecular
forms, either free or complexed with serine protease inhibitors.
Immunodetectable PSA in serum is mainly bound with
1-antichymotrypsin, whereas a minor fraction is free. Several
studies showed that the percentage of free to total PSA was lower in
cancer than in non-cancer cases (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26). However, although
published data are in agreement in showing that percent free PSA
improves the diagnostic effectiveness of total immunoreactive PSA,
specificity and sensitivity figures are less established
(27)(28). This may be attributable to several causes
reviewed recently by Woodrum et al. (28), which include
preanalytical and analytical variables.
In 1996, we began a multicenter prospective study aimed at comparing the diagnostic effectiveness of percent free PSA and total PSA in patients referred to urological practices, using the IMMULITE total and free PSA assays from Diagnostics Products Corp. The aims of the study were as follows: (a) to establish the sensitivity and specificity of percent free PSA in nondisseminated CaP and in benign prostatic hypertrophy (BPH); (b) to evaluate the relationship between percent free PSA and clinical and pathological indicators in patients with CaP; (c) to assess the effect of the presence of concomitant prostatic diseases both in cancer and BPH; (d) to define proper reflex ranges; and (e) to evaluate how the likelihood of the diagnosis of CaP changes after percent free PSA was determined, calculating post-test probability in different age intervals and for differences in total PSA.
Notably, all the assays were carried out in the same laboratory by the same technician and using the same reagent batch to minimize methodological variability.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Blood sampling was performed before diagnostic procedures. Specimens were clotted at room temperature and centrifuged. Serum samples were stored in multiple fractions at -80 °C within 2 h after venipuncture. Serum samples were then shipped to the coordinating laboratory (Venice) in dry ice. The storage time from sampling to assay ranged from 1 to 8 months.
Overall, 85 patients were diagnosed with primary CaP (median age, 69 years; range, 53–93 years), 261 patients were diagnosed with BPH (median age, 67 years; range, 49–91 years), and 10 patients were diagnosed with prostatic intraepithelial neoplasia (median age, 60 years; range, 55–79 years). BPH was confirmed histologically in 160 cases, whereas in 101 cases diagnosis was based on clinical and/or transrectal ultrasonography findings. Only results obtained from biopsy-confirmed cases will be reported.
Sixty-four of 85 patients with CaP underwent radical prostatectomy; the remaining 21 cases were treated with radiotherapy. Fifty-one of 85 cancer patients also had different prostatic complications. BPH was complicated in 49 of 160 patients.
immunoassay
Both total immunoreactive (Third Generation) and free PSA were
assayed using the chemiluminescent immunoassay IMMULITE (Diagnostic
Products Corp.), according to the manufacturer's instructions. The two
methods are solid phase, two-site sequential chemiluminescent assays
that are fully processed on the IMMULITE Automated Analyzer. Briefly,
the patient sample is incubated in a test unit containing a polystyrene
bead coated with monoclonal anti-PSA antibody (or monoclonal antibody
specific for uncomplexed PSA in the case of free PSA assay). After a
30-min incubation at 37 °C, the unbound serum is then removed and an
alkaline phosphatase-labeled polyclonal goat anti-PSA antibody is
introduced. After a second incubation step (30 min), the unbound enzyme
conjugate is removed and the chemiluminescent substrate (a phosphate
ester of amantyl dioxetane) is added. The latter undergoes hydrolysis
in the presence of alkaline phosphatase and yields a light-emitting
intermediate product. Both total and free PSA were measured in all
samples by the same technician, using the same IMMULITE Analyzer
and the same reagent batch. The interassay and intrabatch precision was
evaluated in duplicate for 10 analytical runs under the procedural
conditions given above and measuring three pools obtained with clinical
samples. The coefficient of variation (CV) was <7.7% for total PSA
from 0.8 to 18.1 µg/L and <8.7% for free PSA from 0.5 to 17.6
µg/L.
All assays were performed without knowledge of the eventual diagnosis.
statistical analysis
ANOVA and Kruskal-Wallis one-way tests were used to assess the
differences in PSA among different groups. Total PSA and percent free
PSA were evaluated. The ROC curve was generated by plotting sensitivity
vs 1 - specificity, and the area under the curve (AUC) was
calculated using the Astute package (DDU Software).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Total PSA was significantly higher (F = 33.67; P
<0.0001; power = 0.999) and percent free PSA was significantly lower
(F = 101.15; P <0.0001; power = 1.000) in cancer
than in BPH. However, as shown in Fig. 1
, percent free PSA discriminated better between the two groups,
the overlap being narrower than total PSA. The ROC curve confirms that
percent free PSA had better diagnostic performance than total PSA (AUC,
0.915; 95% CI, 0.876–0.954 for percent free PSA; AUC, 0.712; 95% CI,
0.653–0.770 for total PSA; P <0.0001).
|
In 218 patients with total PSA between 2 and 30 µg/L, the AUCs were
0.690 (95% CI, 0.610–0.770) for total PSA and 0.905 (95% CI,
0.865–0.940) for percent free PSA (P <0.0001; Fig. 2
). Considering that approximately threefold more BPH than CaP
patients were available in the present patient series, each case of
cancer was matched with the closest case of BPH with reference to age
and total PSA. Sixty-nine matched pairs were thus selected. Free PSA
was still very effective, showing an AUC comparable to that found in
the overall patient series (AUC, 0.906; 95% CI, 0.857–0.954).
|
In 10 cases with prostatic intraepithelial neoplasia, the percent free PSA was significantly higher than in cancer (F = 20.64; P <0.0001; power = 0.994) and tended to be lower than in BPH (F = 2.02; P = 0.157; power = 0.292).
CaP was extraglandular in 29 cases and confined to the prostate in 34. Total PSA was significantly higher (P <0.001) in extraglandular malignancies (median, 16.3 µg/L) than in confined malignancies (median, 10.1 µg/L), whereas percent free PSA was not different in the two groups (median, 6.2% and 6.8%, respectively).
The diagnostic performance of percent free PSA was not affected by
prostate gland volume in 164 patients in which the latter information
was available. We grouped patients with prostate volume either 
40
cm or >40 cm, as indicated by Stephan
et al. (22). The gland volume was 40 cm in 20
patients with BPH and in 38 with CaP, whereas it was >40
cm in 84 BPH and 22 CaP patients. Total PSA was
significantly higher (P = 0.01) in BPH patients with
prostate volume 40 cm than in those with smaller glands.
On the other hand, percent free PSA was not significantly different in
the two groups (P = 0.085). Total PSA and percent free
PSA were not significantly affected by gland volume in CaP
(P = 0.421 and P = 0.167,
respectively). In our patient series, the percent free PSA was
significantly lower in CaP than in BPH in patients with prostate gland
volumes both equal to or larger than and equal to or smaller than 40
cm (P <0.0001 in both).
In patients with CaP, the presence of complications potentially related to variations of total PSA (BPH, urinary retention, prostatitis, and urinary tract infections) did not significantly affect either total PSA (P = 0.816), or percent free PSA (P = 0.174). However, only one case affected by prostatitis was present in the cancer group. Total PSA was significantly higher (P = 0.009), whereas percent free PSA tended to be lower (P = 0.091) in complicated (urinary retention, prostatitis, and urinary tract infections) than in noncomplicated BPH. The lowest percent free PSA concentrations were found in 19 patients affected by prostatitis.
Neither total PSA nor percent free PSA showed substantial variations related to tumor grade in the 62 cases in which the Gleason score was available (25 G1 tumors, 31 G2 tumors, and 6 G3 tumors).
patients with total psa between 4 and 10 µg/L
Total PSA was between 4 and 10 µg/L in 65 patients with BPH and
in 22 patients with CaP. In this subgroup of patients, total PSA did
not distinguish CaP from BPH (median, 6.7 and 6.5 µg/L,
respectively), whereas percent free PSA was still significantly
different between CaP and BPH (median, 7.2% and 15.8%, respectively).
The ROC analysis showed that the AUC was 0.899 (95% CI, 0.859–0.937)
for percent free PSA, whereas it was 0.510 (95% CI, 0.370–0.650) for
total PSA (P <0.0001).
patients with total psa between 2 and 4 µg/L
Total PSA was not able to distinguish CaP from BPH in this
subgroup of patients. However, percent free PSA was still significantly
lower in CaP than in BPH (AUC, 0.904; 95% CI, 0.757–1.000).
Considering that the number of BPH cases was redundant in comparison to
CaP, we matched each cancer patient with the closest BPH case according
to age and total PSA. In all five selected couples of matched cases,
percent free PSA was lower in CaP than in BPH.
decision levels in different clinical settings
Several possible cutoff points were calculated using the ROC
analysis to obtain 99%, 95%, 90%, and 80% specificity or
sensitivity, respectively. Specificity, sensitivity, and the likelihood
ratio positive were thus calculated for every fixed sensitivity and
specificity level. To find out a patient-tailored decision criteria
suitable for clinical use, we calculated the post-test probability
(29) as follows:
Likelihood ratio positive = sensitivity/(1 - specificity)
Pretest odds = prevalence/(1 - prevalence)
Post-test odds = pretest odds x likelihood ratio positive
Post-test probability = post-test odds/(post-test odds + 1)
The post-test probability of the disease is the proportion of patients with positive percent free PSA who have the disease. It was calculated for both patients referred to urology practice and men enrolled in screening programs. The detection rate was assumed to correspond to the prevalence of the disease in the evaluated clinical settings. However, we did not use the detection rate of CaP found in our patient series because the frequency of CaP was increased in the present study. The detection rate of CaP was taken from Cooner et al. (4) for urology practice and from Catalona et al. (3) for screening.
The post-test probability was calculated for three different age groups (50–59, 60–69, and 70–79 years). Moreover, the CaP detection rate in urology practice was also examined by PSA intervals, using detection rates reported for different groups of patients subdivided according to both age and PSA concentrations (4).
Results concerning patients referred to urological practice are
reported in Table 1
. The post-test probability of percent free PSA was low in
patients with total PSA 4 µg/L in each age category, using several
cutoff points for percent free PSA, because of the rate of
false-positive results and the relatively low cancer prevalence. An
acceptably high post-test probability was found in patients 60–69
years of age only when a 7.2% free percent PSA cutoff (99%
specificity and 59% sensitivity) was used and in those 70–79 years of
age only when an 8.4% cutoff (95% specificity and 69% sensitivity)
was used. In patients with total PSA between 4 and 10 µg/L, the
post-test probability was low in men 50–59 years of age, except when a
very low cutoff (7.2%) was used, whereas it became acceptably high in
patients 60–69 and 70–79 years of age when a cutoff of 11.8% (80%
specificity and 83% sensitivity) was used. In patients with PSA >10
µg/L, the percent free PSA showed a good post-test probability in all
of the age groups taken into consideration all the evaluated cutoff
points were used.
|
The results expected when percent free PSA is used for the general
population, i.e., for screening purposes, are shown in Table 2
. In this setting, prevalence was not subdivided according to
total PSA because this information in screening studies was only
available anecdotally. In this clinical scenario the post-test
probability was >50% only when 7.2% free percent PSA was used as the
cutoff (99% specificity and 59% sensitivity) in patients 50–69 years
of age and when 9.7% free percent PSA was used as the cutoff (90%
specificity and 76% sensitivity) for those 70–79 years of age.
|
These data suggest that percent free PSA should not be reported with reference to a defined cutoff point. Tentatively, it could be reported in association with a cancer probability scale calculated with reference to age, total PSA, and cancer prevalence in the clinical scenario in which the marker will be used, as has been suggested previously (12)(15)(17)(24).
percent free psa in the differential diagnosis between cap and bph
Patients were classified as positive or negative with reference to
percent free PSA, using two different thresholds set at 95%
sensitivity (16.1%) and at 95% specificity (8.4%), respectively.
Patients with percent free PSA between the latter decision levels were
defined as equivocal, as were patients with total PSA between 4 and 10
µg/L.
Total PSA was unable to distinguish CaP from BPH in 87 of 245 (35.5%) cases; percent free PSA also was unable to distinguish CaP from BPH in 87 of 245 (35.5%) cases. However, the rate of nonclassifiable patients dropped to the acceptable low rate of 32 of 245 cases (13.1%) when total PSA and percent free PSA were used in association.
On the basis of these findings, we calculated the probability of
reducing the rate of unnecessary biopsies in patients affected by BPH
(Table 3
).
|
The number of patients spared unnecessary biopsies could have been 55
of 117 (47.0%) in patients with total PSA 4 µg/L, with 3 of 78
(3.8%) cancers missed. In patients with PSA 10 µg/L, a unnecessary
biopsy could have been avoided in 25 of 55 cases (45.4%), whereas 3 of
56 (5.4%) cancers would had been missed. In patients with total PSA
<4 µg/L, percent free PSA was still very effective, showing
excellent true positive and true negative rates. However, the number of
patients in these subgroups (Table 3
) was too limited to allow for any
conclusion.
reflex range
The reflex range was recently defined as the optimal range of
total PSA in which the percent free PSA determination is indicated
(23). ROC curves were calculated for several total PSA
intervals. The results of these calculations are reported in Table 4
. Percent free PSA was superior to total PSA in all the dose
ranges evaluated, with AUCs from 0.894 (95% CI, 0.835–0.952) to 0.947
(95% CI, 0.871–1.000). Total PSA showed a different diagnostic
accuracy depending on its actual dose level; its AUC ranged from 0.510
(95% CI, 0.370–0.650) to 0.723 (95% CI, 0.506–0.936). The
differences between the AUCs of percent free PSA and those of total PSA
were greater in patients with low total PSA, in which the latter marker
has poor diagnostic value. In addition, the differences between the
AUCs of percent free PSA and total PSA tended to decrease when total
PSA was >10 µg/L because the diagnostic effectiveness of total PSA
tends to improve. From these data, either 2–10, 3–10, or 4–10 µg/L
total PSA reflex ranges could be accepted for the routine use of
percent free PSA.
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
1-antichymotrypsin
in patients with CaP than in those with BPH, provided a sound
biological rationale for a new diagnostic tool and prompted additional
investigations. The interest toward PSA fraction determination grew
dramatically in a short time, and a considerable number of papers and
meeting abstracts, as well as reviews and an editorial
(27)(28)(30), have been published.
However, decision criteria for the clinical application of percent free
PSA have yet to be established in spite of the bulk of published data.
The diagnostic performance, the proper cutoff points, and the
reflex range depend on several variables, including the method used and
the patient series examined (28). The majority of
investigators agree that percent free PSA improves the diagnostic
accuracy of CaP. However, the reported rates of sensitivity,
specificity, and predictive values are widely scattered
(27)(28). The analytical variability of assay
methods may represent a relevant cause of discrepancies among the
results of different studies. This issue deserves some comments. The
concentration of free PSA corresponds to 10–30% of that of total PSA
in the majority of subjects. For a total PSA concentration from 2 to 10
µg/L, the expected free PSA concentrations range from 0.2–0.6 to
1.0–3.0 µg/L. The precision of free PSA assays is therefore lower
than that of total PSA because immunoassays present a high imprecision
when the antigen concentration is low. A recently published study
reported a CV ~20% for a sample with 0.5 µg/L free PSA, with
results found by different laboratories in the range of 0.24–0.75
µg/L (31). When percent free PSA is calculated, the final
precision is further affected by the variability of the total PSA
assay. In addition, total variability may be increased when different
reagent batches are used, as may occur in practice (32).
Sample handling and storage are additional aspects that should be
considered when results from different studies are compared
(33)(34). The present study was carried out with the aim of minimizing preanalytical and analytical variability. Sample collection, transport, and storage were properly planned and carefully monitored. In addition, all of the assays were carried out in the same laboratory by the same technician, using the same IMMULITE instrument and the same reagent batch. Serum samples from CaP and BPH patients were matched in each analytical run.
We found that the percent free PSA was not associated with patient age, in agreement with Oesterling et al. (13), Filella et al. (14), Morgan et al. (18), and Jung et al. (19). A recent, prospective study by Catalona et al. (26) showed a statistically significant and clinically relevant direct association between percent free PSA and age. The results of Catalona et al. should be considered very reliable because of the excellent quality of their study design and the high number of evaluated cases. However, the cohort in the study by Catalona et al. and the present study are not easily comparable because that group studied a large patient series with total PSA between 4 and 10 µg/L, whereas we evaluated a smaller cohort of patients with total PSA from 2 to 30 µg/L.
We found no association between percent free PSA and total PSA, in agreement with Christensson et al. (11). Jung et al. (19) found a significant inverse correlation between percent free PSA and total PSA. However, from the data shown in their study, the correlation seems related to the occurrence of very high percent free PSA values in patients with total PSA <3 µg/L. This latter finding could have been affected by the high assay imprecision that occurs when very low concentrations of free PSA are measured. A significant, although weak, inverse association between percent free PSA and total PSA has been also reported by Catalona et al. (26). However, their data are hardly comparable with those found in the present study because of relevant differences in the patient series evaluated.
We found a significant direct association between total PSA and local tumor extension, whereas percent free PSA was not significantly different between confined CaP and extraglandular tumors. In addition, we did not find significant variations of percent free PSA between T1 and T2 tumors, in contrast with Filella et al. (14) and van Caugh et al. (17), but in agreement with Egawa et al. (21) and Stephan et al. (22).
In the present patient series, unlike the one studied by Stephan et al.
(22), the diagnostic effectiveness of percent free PSA was
not affected by prostate gland volume. In fact, we showed that percent
free PSA is significantly lower in CaP than in BPH in cases with a
prostate gland volume both larger and smaller than 40 cm.
Although Stephan et al. used the same method, they evaluated only 10
cases with a prostate gland volume 40 cm and 26 with
gland volume <40 cm. The corresponding number of patients
were 22 and 38 in the present study. In their patient series, the
percentage of cancer patients with total PSA between 4 and 10 µg/L
was 80% in the group with larger gland volumes and 54% in patients
with smaller prostate volumes. The corresponding percentages were 23%
and 24% in our study, making the comparison probably more homogeneous.
The above discrepancy could therefore be related to differences in the
case histories that, when dealing with small numbers of patients, may
seriously affect any conclusion.
In the present study, the percent free PSA was not associated with tumor grade, in agreement with Catalona et al. (12), Egawa et al. (21), and Stephan et al. (22), but in contrast with Elgamal et al. (25) and Catalona et al. (26).
In patients with prostatic intraepithelial neoplasia, the percent free PSA showed intermediate values between CaP and BPH, in agreement with Tarle and Kraljic (35).
Concomitant prostatic complications seem not to affect the percent free PSA in CaP, whereas it tends to be lowered by the presence of prostatitis in BPH, in agreement with Filella et al. (14). It must be emphasized that the effect of prostatitis was not evaluable in our cancer patients because of the limited number of patients affected by this complication.
We confirmed the diagnostic effectiveness of percent free PSA shown by several investigators. In addition, our AUCs are among the most favorable reported to date for percent free PSA. Catalona et al. (26), in a prospective study carried out in a wide cohort of patients with total PSA between 4 and 10 µg/L, showed results similar to those found in the present series but using a different assay method. It should be noted that the AUCs shown in the present study are slightly better than those found by Catalona et al. Differences could be ascribed to the differences in the characteristics of the enrolled patients. However, we believe that the very restrictive methodological approach adopted in the present study minimized the analytical variability of both total and free PSA assays. This probably improved the diagnostic accuracy of percent free PSA, especially in cases with low total PSA concentrations. The choice of working under optimal analytical conditions is far from perfect from the point of view of clinical application. It must be noted that the present optimal performance of percent free PSA might change under usual assay conditions because higher run-to-run and lot-to-lot assay variability is expected. However, the present strategy probably identified the optimum diagnostic performance of percent free PSA with the method evaluated.
We confirmed the role of percent free PSA in reducing the number of
unnecessary biopsies in BPH, reported in several studies [reviewed in
Refs. (27)(28)(30)]. In patients
with total PSA 4 µg/L, if a 16.1% cutoff point had been used,
47.0% of unnecessary biopsies could have been avoided, with a
false-negative rate of 3.8%. If the same cutoff had been used in
patients with total PSA 10 µg/L, 45.4% of biopsies could have been
avoided, with 5.3% missed cancers.
A second issue we dealt with was the assessment of the probability of any individual patient to have a cancer, given his percent free PSA. For this goal, we calculated the post-test probability of the disease, which indicates the proportion of patients with positive test results who have the disease. We used sensitivity and specificity values based on our patient series and the prevalence data available from both screening (3) and large-scale urology practice studies (4), according to the approach used previously to evaluate the actual tumor marker effectiveness in clinical practice (36). The post-test probability of percent free PSA was, however, <50% in men 50–70 years of age when cutoff points providing sensitivity values from 99% to 80% were used.
In conclusion, the percent free PSA is superior to total PSA in distinguishing primary CaP from BPH in patients with total PSA between 2 and 30 µg/L. In men with relatively low total PSA, assay method variability may affect the reliability of free PSA determination and may in part account for discrepancies reported thus far in the literature on the sensitivity and specificity of percent free PSA. Use of a percent free PSA assay might reduce the number of unnecessary biopsies indicated by total PSA >4 µg/L by ~50%, with a probably still acceptable 96% cancer detection rate. The post-test probability of CaP provided by percent free PSA is, however, relatively low in asymptomatic patients 50–70 years of age when the actual prevalence of CaP is properly taken into account. Therefore, although it is probably the most powerful biochemical tool presently available for CaP detection, free percent PSA must be interpreted cautiously in individual patients and used in association with clinical and instrumental evaluation in decision making.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Present address: Urology Unit, Regional Hospital, Udine, Italy. 
3 Present address: Urology Clinic, University of Bologna, Bologna,
Italy.
1 Nonstandard abbreviations: CaP, prostate cancer; PSA, prostate-specific antigen; BPH, benign prostatic hypertrophy; AUC, area under the ROC curve; and CI, confidence interval.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
-1-antichymotrypsin. Clin Chem 1991;37:1618-1625.
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  S. A.R. Kort, F. Martens, H. Vanpoucke, H. L. van Duijnhoven, and M. A. Blankenstein Comparison of 6 Automated Assays for Total and Free Prostate-Specific Antigen with Special Reference to Their Reactivity toward the WHO 96/670 Reference Preparation Clin. Chem., August 1, 2006; 52(8): 1568 - 1574. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Jung, U. Elgeti, M. Lein, B. Brux, P. Sinha, B. Rudolph, S. Hauptmann, D. Schnorr, and S. A. Loening Ratio of Free or Complexed Prostate-specific Antigen (PSA) to Total PSA: Which Ratio Improves Differentiation between Benign Prostatic Hyperplasia and Prostate Cancer? Clin. Chem., January 1, 2000; 46(1): 55 - 62. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Beckett, L. H. Cazares, A. Vlahou, P. F. Schellhammer, and G. L. Wright Jr. Prostate-specific Membrane Antigen Levels in Sera from Healthy Men and Patients with Benign Prostate Hyperplasia or Prostate Cancer Clin. Cancer Res., December 1, 1999; 5(12): 4034 - 4040. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
