View
226
Download
6
Category
Preview:
Citation preview
Prostate specific antigen (PSA)Measurand characteristics, reference measurement
procedures and reference materials for PSA
JCTLM SymposiumSevres December 14-15, 2004
Ulf-Håkan StenmanDepartment of Clinical Chemistry
Helsinki University Central HospitalFinland
PSA• Why is standardization of PSA assays
important?• Characteristics of PSA• Reference materials• Reference measurement procedures• Conclusions
Prostate cancer, basic factsProstate cancer, basic facts
Most common (nonMost common (non--skin) cancer in menskin) cancer in men•• LifeLife--time risk of acquiring prostate cancer 7 time risk of acquiring prostate cancer 7 -- 17%17%
Second most common cause of cancer deathSecond most common cause of cancer death•• LifeLife--time risk of death from prostate cancer 2 time risk of death from prostate cancer 2 -- 4% 4% •• 5050--70% of cases advanced when giving rise to symptoms70% of cases advanced when giving rise to symptoms
Preclinical stage during which early diagnosis and Preclinical stage during which early diagnosis and curative therapy are possiblecurative therapy are possiblePotential target for screening Potential target for screening
UU--H Stenman 2004H Stenman 2004
Clinical use of PSA• Sensitive marker for prostate cancer
– Expressed by more than 99% of all prostate cancers
– Serum PSA increased by 1 gram tumor
– Prostate-specific rather than cancer specific
• Used for– Monitoring of response to therapy
– Detection of relapse
– Evaluation of prognosis
– Early diagnosis and screening (case finding)
Prostatasyöpäpotilaan seuranta
20001999199819971996.1
1
10
100
1000S-
PSA
(µg/
l)
Castration, androgen suppression
Radiation
PSA during monitoring of a prostate cancer patient
Early detection and screening for prostate cancer with PSA
Pathophysiological basis
Serum PSA at sampling and time to presentationSerum PSA at sampling and time to presentation
0-2-4-6-8-10-120,1
1
10
100
1000
Time before presentation (yr)
PS
A (µ
g/l)
2,5
Stenman et al. Lancet 1994
Prostate cancer incidence and mortality in Finland 1958-2001
20102000199019801970196019500
500
1000
1500
2000
2500
3000
3500
4000
IncidenceMortality
Year
Incidence and Mortality
Use of PSA and prostate cancer incidence in Norway
Prostate cancer incidence and mortality in Finland 1958-2001
20102000199019801970196019500
500
1000
1500
2000
2500
3000
3500
4000
IncidenceMortality
Year
Incidence and Mortality
Survival of cancer patients and serum PSA at samplingSurvival of cancer patients and serum PSA at sampling
2015105000,0
0,2
0,4
0,6
0,8
1,0
<4 µg/l>4 µg/l
Time (years)
PSA cut-off
Stenman et al. Lancet 1994Stenman et al. Lancet 1994
Development of prostate cancer, screening and age
PSA PSA (µg/l)(µg/l)
Lethal stageTumor vol 1 kg
Clinical diseaseTumor vol 32 g
Curable stageTumor vol 1 g
UU--H Stenman 1995H Stenman 1995
1009080706050401
10
100
1000
10000
Age (years)
Are we doing something wrong?• Are we over-diagnosing?• Yes, in 75% of those over 70 years of age• Are we finding the patients that need to be
cured?• Not all, 35% of patients with clinically localized
tumors experience a relapse• We need to improve our diagnostic methods
and procedures– Even earlier diagnosis– More cancer-specific methods– Avoid detection of clinically indolent tumors– Prognostic methods
Characteristics of measurand (PSA)• 30 kDa chymotrypsin-like serine proteinase
• Produced by epithelial cells of the prostate
• Prostate specific, not cancer specific
• Dissolves seminal clot by degrading semenogelin• PSA in plasma is heterogeneous
– Complexes with protease inhibitors– Partially degraded forms and proforms of free
PSA– Crossreacting substances
• hK2
Gel filtration of PSA in prostate cancer serum
7060504030200
1
2
3
4
5
6
7
8
PSA (µg/L)
Fraction no.
PSA
(µg/
L)
30 kDa
90 kDa
Stenman et al. J Nucl Med Allied Sci 1990
Assays for free PSA and PSA-ACT
Eu
PSA PSA
ACT
Sm
Forms of immunoreactive PSA in prostate cancer patients and men with BPH
PCa Controls• PSA-ACT 86% 79%• Free PSA 13% 19%• PSA- API (-AAT) 1% 2%• hK-2 1.5% 1%
• PSA-A2M 2% 4%
67 –43 –
30 –
20 –
14 –
67 –43 –30 –
20 –
14 –
M HIC GF A B C D EM HIC GF A B C D E
kDa kDa
SDS gel electrophoresis of PSA isoenzymes from seminal fluid separated by ion exchange chromatography
Nonreducing conditions Reducing conditions
10080604020200
100
200
300
400 PSA controlACT 24 hACT 240 h
Fraction no.
PSA (ng/mL)
A
B
C
D´ D E
F
PSA-ACT
Amino acid sequence of PSA
M WV PVVFLTL SVT WIGA APLILSR1 7 17 27 37 IVGG WE CEKHSQP W QV LVASRGRAVC GGVLVHPQ WV
LTAAHCIRNK 47 57 67 77 87 SVILLGRHSL FHPEDTGQVF QVSHSFPHPL YDMSLLKNRF LRPGDDSSHD 97 107 117 127 137 LMLLRLSEPA ELTDAVKVMD LPTQEPALGT TCYASGW GSI EPEEFLTPKK 147 157 167 177 187 LQCVDLHVIS NDVCAQVHPQ KVTKFMLCAG RWTGGKSTCS GDSGGPLVCN 197 207 217 227 237
Activation peptidein proPSA
-4 -2Signal peptide
SDS gel electrophoresis of PSA isoenzymes from seminal fluid separated by ion exchange chromatography
67 –43 –
30 –
20 –
14 –
M HIC GF A B C D E
kDaReducing conditions
10080604020200
100
200
300
400 PSA controlACT 24 hACT 240 h
Fraction no.
PSA (ng/mL)
A
B
C
D´ D E
F
PSA-ACT
A B C D E0
20
40
60
80
100Proportion of complex formation (%)
2-dimensional electrophoresis of free PSA in serum and isoelectric focusing of PSA isoenzymes isolated from seminal fluid
Isoelectric point
2-D electrophoresis of serum PSA
Charrier et al. Electrophoresis 20: 1075-81, 1999
Isoelectric focusing of seminal plasma PSA
- 8.15 -
- 7.35 -
- 6.85 -
- 6.55 -
A B C D E pI
Zhang et al. Clin Chem 41:1567-73 1995
Development of PSA standards• Stanford conferences organized by Thomas Stamey in
1992 and 1994
• Preparation of standards for
– Free PSA
– PSA-ACT and free PSA 90/10 mixture
• Establishment of WG on PSA standardization
• Adoption of standards as WHO reference materials
– WHO 96/668 (free PSA)
– WHO 96/670 (PSA-ACT/PSA 90/10)
Rafferty et al. Clinical Chemistry 46: 1310-1317, 2000
Development of PSA standards• Purification of PSA from seminal plasma
• Value assignment by amino acid analysis (mol)
• Determination of MW by mass spectrometry, 28430
• Calculation of mass concentration
• Formation of PSA-ACT in vitro
• A280 of PSA-ACT based on AA compostion = 1.0– Assignment of mass concentration for PSA-ACT
• Preparation of PSA-ACT-free PSA 90/10 mixture
Prestigiacomo AF, Chen Z, Stamey TA. A universal calibrator for prostate specific antigen (PSA). Scand J Clin Lab Invest Suppl 1995;221:57-9.
Figure 4. Potency of human serum sample (97/568) in individual assays calculated using IHR (A) and PSA 90:10 (96/670; B)
Figure 5. Potency of human serum sample (97/566) in individual assays calculated using IHR (A) and PSA 90:10 (96/670; B).
Epitope mapping of PSA
Summary Report of the TD-3 Workshop: Characterization of 83 Antibodies against Prostate-Specific Antigen.Stenman U-H, Paus E, Allard WJ, Andersson I, Andres C, Barnett TR, Becker C, Belenky A, Bellanger L, Pellegrino CM, Börmer OP, Davis G, Dowell B, Grauer LS, Jette DC, Karlsson B, Kreutz FT, van der Kwast T, Lauren L, Leinimaa M, Leinonen J, Lilja H, Linton HJ, Nap M, Nilsson O, Ng PC, Nustad K, Peter A, Pettersson K, Piironen T, Rapp J, Rittenhouse HG, Rye PD, Seguin P, Slota J, Sokoloff RL, Suresh MR, Very DL, Wang TJ, Wigheden I, Wolfert RL, Yeung KK, Zhang W, Zhou Z, Hilgers J. Tumour Biol 1999; 20 (Suppl 1): 1-12.
Epitope mapping of PSA
AA 158-163 Region 3
AA 86-91 Region 1
AA 80-85 Region 2
AA 58-64 Region 5
AA 3-11 Region 6
Catalytically active site
AA 58-64
AA 3-11
1
5a
5b
5c
6a
6b 4b
4a
3b
3a
2a
2c
2b
AA 158-163
AA 86-91
AA 80-85
Stenman et al. Tumour Biol 1999;20 (Suppl 1):1-12.
Correlation HybritechTandem E -Wallac AutoDelfia before 2003
100101.1.1
1
10
100
Access T-PSA
Delfi
a T-
PSA
201510500
5
10
15
20
Access T-PSA
Delfi
a T-
PSA
y = - 6.5112e-2 + 0.88453x R^2 = 0.994
Recalibrated AutoDelfia vs. Access T-PSA (2004)
Intercept : -0.084 [ -0.198 to 0.000 ]Slope : 0.994 [ 0.970 to 1.020 ]
Passing-Bablok agreement test N = 1470 4 8 12 16 20
AutoDelfia0
4
8
12
16
20 Access
Recalibrated AutoDelfia vs. DPC Immulite 2000 T-PSA (2004)
Intercept : -0.050 [ -0.228 to 0.071 ]Slope : 1.352 [ 1.315 to 1.392 ]
Passing-Bablok agreement test N = 1470 5 10 15 20 25 30
AutoDelf ia0
5
10
15
20
25
30 Immulite 2000
Recalibrated AutoDelfia vs. Access F-PSA (2004)
Intercept : 0.033 [ 0.010 to 0.055 ]Slope : 0.963 [ 0.933 to 1.000 ]
Passing-Bablok agreement test N = 1600 2 4 6 8 10 12
AutoDelfia0
2
4
6
8
10
12 Access
Recalibrated AutoDelfia vs. DPC Immulite 2000 F-PSA (2004)
Intercept : -0.022 [ -0.037 to 0.001 ]Slope : 0.787 [ 0.767 to 0.808 ]
Passing-Bablok agreement test N = 1600 2 4 6 8 10
AutoDelfia0
2
4
6
8
10 Immulite 2000
Correlation F/T
0,80,60,40,20,00,0
0,2
0,4
0,6
0,8
AutoDelfia
Beck
man
Acc
ess
y = 8,7969e-3 + 0,95312x R^2 = 1,000
0,80,60,40,20,00,0
0,2
0,4
0,6
0,8
AutoDelifa
DPC
Imm
ulite
y = - 5,8281e-3 + 0,57812x R^2 = 1,000
Probability of prostate cancer in relation to total and free PSAExcel formula avavailable at: www.finne.info
35302520151050
10
20
30
40
50
60
70
80
PSA 4 µg/lPSA 7 µg/lPSA 10 µg/lPSA 20 µg/l
F/T PSA (%)
Prob
abili
ty o
f fin
ding
pro
stat
e ca
ncer
on
biop
sy (%
)
Proportion of free PSA and tumor gradeProportion of free PSA and tumor grade
Bangma J Urol 1997; 157: 544
0.0
0.1
0.2
0.3
0.4
0.5
0.6
BPH Grade 1 Grade 2 Grade 3
Equimolarity of assays for PSA
• Ability to detect free and complexed PSA equally
• Assays based on polyclonal antibodies tend to overestimate free PSA
Effect of proportion of F-PSA on ratio between Access and Delfia total PSA
Tandem-E vs. Delfia Beckman Access vs. Delfia
60504030201000,75
1,00
1,25
1,50
1,75
2,00
F/T (%)
Acc
ess/
Delfi
a
1008060402000,75
1,00
1,25
1,50
1,75
2,00
F/T (%)
Ratio
Acc
ess/
Delfi
a
Recognition of free PSA by Delfia and Access assays
Control containing >90% F-PSA
DelfiaT-PSA
DelfiaF-PSA
Access T-PSA
Sample 1 2.12 2.00 4.1
Sample 2 19.6 19.1 28.0
Recognition of various forms of free PSA by AutoDelfia and Beckman Access assays
AutoDelfia Access RatioF-PSA T-PSA F/T T-PSA Access/AD
PSA-A 96,5 92,2 104,6 114,9 1,25PSA-B 59,5 57,4 103,6 72,7 1,27PSA-C 41,7 40,3 103,6 53,9 1,34PSA-D 55,0 61,6 89,2 77,9 1,26PSA-E 50,6 52,9 95,5 63,8 1,21ProPSA 22,7 23,9 94,9 30,9 1,29
Mean 1,269
Equimolarity of assays for PSA
• Beckman Access uses the same antibodies as Hybritech Tandem E
• What is the difference?
• Assay kinetics– Small molecules react faster
Frequency of aberrant results in 20000 samples analyzed with two methods
Median ratio Access:Delfia = 1.14
Access value >2-fold Delfia 13/2651 0.49%
Access value >1.5-fold Delfia 49/2651 1.8%
Access value <0.67-fold Delfia 1/2651 0.00%
Access value <0.50-fold Delfia 0/2651 -
Comparison of Delfia and Access on samples with aberrant results and benign biopsy
Sample Total PSA Total PSA Free PSA (%)Delfia Access Delfia
A 1,55 18,2 26%B 0,71 5,7 22%C 0,92 6,1 24%D 1,28 5,2 20%E 2,15 4,8 24%F 5,62 12,2 21%G 5,20 30,7 31%
Effect of mouse serum on interference
Beckman AccessT-PSA
Wallac Delfia T-PSA
Wallac Delfia F-PSA
Routine assay 6.33 0.73 0.26
Addition of mouse serum
2.74 1.08 0.26
Conclusions from assay comparisons
• Assay manufacturers compete with speed, capacity and price
• Quality suffers• Who is responsible?• Customer?
– Capable of judging quality?• Manufacturer?• Regulatory bodies?
Reference measurement proceduresAre they necessary?
PSA assays from major manufacturers are unusually well standardized
Room for improvement
Quality of assays from minor companies varies
Stenman U-H. Immunoassay standardization: is it possible,
who is responsible, who is capable? Clin Chem 2001;47(5):815-20.
Is it possible?• Identification of reference antibodies
– Known epitope specificity– Generally available
• Definition of assay format– Sandwich assay– Microtitration well format, sample volume 25 ul– Two-step incubation, 1 + 1 h incubation time– Applicable to any detection method
• ELISA, Delfia, IRMA, luminescence
• Reference laboratories• Calibrated samples• Mass spectrometry?
Who is capable?• I know some• Endagered species• Are they willing?• Who will pay for it?
– Not ”sexy science” that you get grants for
Who is responsible?• WHO?• IFCC?• JCTLM?• We are all responsible• We need to justify the expenses• Calculation of additional health care costs
caused by poor assay quality• Poor quality is expensive• Someone has to take the lead
Recommended