1. The Next Generation of Prostate Cancer Detection: How to Make PSA a Better Marker Linda C. Rogers, PhD, DABCC, FACB Scientific Affairs Beckman Coulter, Inc.

2. Introduction PSA Limitations & Controversy
Elevated PSA may not indicate clinically significant cancer
Low specificity results in high costs
Result: Prostate cancer is over diagnosed and over treated.
How can we make PSA a better marker? 2

3. Prostate Cancer Most common type of cancer found in men
Second most common cause of cancer mortality in men
Now estimated that one in six men will have prostate cancer
High cost to manage patients
Approximately 1.6 million men in the US undergo prostate biopsies
yields only a 20-40% positive cancer rate
Estimates in the US alone over a million negative biopsies annually
Prostate cancer testing: (where did we start, where are we going?)
1986: First FDA approved PSA assay with Hybritech Tandem-R
mid-1990’s: Moved from a monitoring claim to a detection claim
– Hybritech established the 4 ng/mL cutoff
Late-1990’s: Free PSA helped to improve specificity for cancer
– Hybritech established the 4 to 10 ng/mL PSA Range
Other derivatives of PSA utilized (e.g. velocity, density)
New biomarkers (p2PSA)
Currently in the USA about 1.6 million men per year undergo prostate biopsies, and roughly 80 percent of these men have negative results Conversely, roughly 15 percent of men with prostate cancer go undetected because their PSA levels are below the cutoff level 3 3

4. We want to “see” what’s going on to make accurate clinical decisions regarding prostate disease
Can PSA help us do that?

5. Cancer Easily Recognized
Benign Prostate
Prostate Cancer
Only way to confirm cancer is to look at prostate cells:
Normal cells are differentiated; Cancer cells are undifferentiated

6. The non-specificity of PSA
Prostate Cancer
Even if prostate cancer is there, PSA cannot distinguish indolent from aggressive cancers
Benign Prostatic Hyperplasia (BPH)
Present in 4% at 40 yr, 30% at 50 yr, 75% at 80 yr
Other Conditions Affecting PSA
Acute and chronic prostatitis
Physical trauma, inflammation

7. Prostate Cancer Markers: Optimal Characteristics
Specificity
Must be able to differentiate cancerous from normal prostate
Separate prostate cancer from other cancers and disease types (e.g. BPH, prostatitis)
“Tumor Specific”
Sensitivity
Identify all individuals with prostate cancer
Early marker of disease
Easily Detectable with Non-Invasive Methods

8. Sensitivity/Specificity
Ideal Situation
Specificity = 100%
True Negative = 100%
False Positive = 0%
Sensitivity = 100%
True Positive = 100%
False Negative = 0%
Healthy
Individuals
Cancer
Patients
Cut Off

9. Sensitivity/Specificity
Real World
Healthy
Individuals
Cancer
Patients
False Negatives
False Positives
Cut Off
We must choose a cut off that maximizes both
sensitivity and specificity

10. Normal Physiological Pathway
pPSA
Prostate
luminal
epithelial
cell
PSA
Seminal
Plasma
Blood
PSA promotes
sperm motility
(keeps semen in its liquid form)
Seminal Vesicle
Fluid
pro
Proenzyme secreted
Active enzyme
Like most secreted peptide enzymes, PSA is produced initially as an inactive proPSA molecule that includes a 7-amino acid leader peptide sequence. Human kallikrein 2 (hK2) activates this proPSA molecule by clipping off the 7-amino acid leader peptide sequence.
Cleaved by enzymes
Why PSA is a good tumor marker

11. PSA Improves Detection of Early Cancers Expected
At this point, PSA screening had been done for about 4-5 years
Clinically unimportant (80%)
Clinically unimportant (80%)
PSA screening
results: detect
more & earlier
Cancers detected by
physical examination
Early stage
(10%)
Advanced stage
(10%)
Cancers to worry about
Adapted from Scientific American,1996: 215(3):115. Illustration by Christiansen J in
“Does Screening for Prostate Cancer Make Sense?” by Hanks GE and Scardino PT.

12. PSA Improves Detection of Early Cancers Real World
Clinically unimportant (80%)
PSA
screening
results
Cancers detected by
physical examination
Early stage
(10%)
Advanced stage
(10%)
(not to scale)
BPH
We are detecting a lot more cancers that don’t need to be treated
Adapted from Scientific American,1996: 215(3):115. Illustration by Christiansen J in
“Does Screening for Prostate Cancer Make Sense?” by Hanks GE and Scardino PT.

13. Applications of PSA Testing
Screening for cancer
Monitoring for
recurrence
Follow-up for
suspected cancer
Monitoring
progression
Managing
BPH therapy
Am J Clin Pathol 2000;113:

14. Prostate Cancer Detection: Clinical Dilemma
Total PSA (ng/mL)
0-2
2-4
4-10
>10
DRE -
No Biopsy
(1% PCa)
Biopsy?
(15% PCa)
(25% PCa)
Biopsy
(>50% PCa)
DRE +
(5% PCa)
(20% PCa)
(45% PCa)
(>75% PCa)
Comparison: Cancer rate for general population of men >50 yr= 4%
Grey Zone: The area that creates concern over “false positive” results
Adapted from Catalona, JAMA, 1998 11

15. The Chance of Finding Prostate Cancer Increases with PSA Levels (ng/mL)
(Adapted from Dr. Alan Partin)

16. Circulating Forms of PSA
From: Rev Urol Spring; 6(2): 58–72.

17. Major PSA Forms in Serum
ACT
MAC
f-PSA
PSA-ACT
1-Antichymotrypsin 2-Macroglobulin
5-50 % %
PSA-MAC
Not measured
No binding sites
Next: Measuring free PSA.
Complexed PSA
Total PSA

18. Probability of Cancer Based on PSA and % fPSA Results (Men with Non-Suspicious DRE Results, Any Age)
Total
PSA
Probability of Cancer %
free PSA
Probability
of Cancer
0-2 ng/mL 1%
0-10%
56%
2-4 ng/mL
15%
10-15%
28%
4-10 ng/mL
25%
15-20%
20%
>10 ng/mL
>50%
20-25%
16%
>25% 8%
Free PSA helps us gain some specificity, but not enough to significantly affect the clinical decision-making process.

19. “PSA has reached the end!”
“We urgently need a serum marker that reflects prostate cancer in the current PSA range of 2 to 10 ng/mL.”
- Dr. Thomas Stamey, Stanford University
“Using the 2 to 10 ng/mL PSA range is the contemporary era of PSA screening”
- Dr. William Catalona, Northwestern University

20. Getting More Out of PSA: How can we increase its specificity?
Free PSA
Catalona WJ (1998) JAMA
PSA Velocity (Δ over time)
Smith DS (1994) J Urology
PSA Density (ng/mL/cm3)
Zlotta (1997) J Urology
Beduschi (1997) Urology Clinical NA
Age-Specific Reference Ranges
Battikhi (2006) Intl Urology Nephrology
Novel Biomarkers
PSA Isoforms
EPCA-2
PCA-3
TMPRSS2-ERG
EN-2 (Engrailed-2)
ACCEPTED
UNDER
INVESTIGATION
(or currently in practice) 20

21. Novel Biomarkers [-2]proPSA (p2PSA) EPCA-2 PCA3 TMPRSS2-ERG
EN-2 (Engrailed-2)

22. Disease-Associated Forms of PSA
pro
pro
This inactive precursor of PSA is associated with PCa
Mikolajczyk, Urology 2002 22

23. Major PSA Forms in Serum
ACT
MAC
f-PSA
PSA-ACT
1-Antichymotrypsin 2-Macroglobulin
5-50 % %
PSA-MAC
Not measured
No binding sites
Next: Measuring free PSA.
Complexed PSA
Total PSA 23

24. Molecular Forms of PSA in Serum
intact
non-native
PSA
39%
[-2]pro PSA 6%
Complexed
PSA
[-4]pro PSA
10%
Free PSA
15%
proPSA
32%
[-5/7]pro PSA
17%
BPH-A
28%
Mikolajczyk, Urology 2002 24

25. Receiver Operating Characteristic (ROC) Curve
True Positive rate vs. False Positive rate for different cut-off points
The closer the ROC is the upper left corner, the higher the overall accuracy of the test.
Relate two or more biomarkers to each other.
Area Under the Curve
> Significant

26. Subjects Inclusion criteria Results
2005, NCI formed the EDRN (independent investigators to identify and validate new cancer biomarkers)
Subjects
566 men, in a prospective PCa detection study at four National Cancer Institute Early Detection Research Network clinical validation centers
Inclusion criteria
40 years and older, no prior prostate surgery, biopsy or history of PCa, no use of 5-α reductase inhibitors, and at least a 10 core template biopsy (post-enrollment)
Results
Calclulated %[-2]proPSA ([-2]proPSA/free PSA) – found this had value
In the 2 to 10 ng/mL PSA range, %[−2]proPSA outperformed %fPSA (AUC 0.76 vs. 0.66)
EDRN: Early detection research network
Research group, 5 year grant
ID and validate potential markers for cancers: prostate and urol., lung, breast and gyn, GI
Prostate arm
-first published in 2008, final in 2010
Has another 5 year funding
Sokoll, Cancer Epidemiol Biomarkers Prev, 2010

27. ROC analysis (n = 195) comparing PSA, %fPSA, %[−2]proPSA, and a logistic regression model
(0.58)
Most labs can’t calculate
A regression model that is difficult to understand
(0.66)
(0.70)
% -2proPSA
(0.76)
Sokoll, 2010

28. Prostate Health Index (phi)
Access will give this derived result 28

29. Subjects Inclusion criteria Results
892 men with no history of prostate cancer, in a prospective multi-institutional trial
Inclusion criteria
Normal DRE, pre-study PSA 2.0–10 ng/mL and 6-core or greater prostate biopsy
Results
At 80% to 95% sensitivity, the AUC of phi exceeded PSA & fPSA
phi = 0.724; fPSA = 0.670
4.7-fold increased risk of PCa and a 1.61-fold increased risk of a Gleason score greater than or equal to 4+3=7 disease on biopsy
Suggests a detection of indolent vs. aggressive cancers….higher phi, higher Gleason score

30. PSA, fPSA, [-2]proPSA, free-to-total PSA and phi ROC curves in 2 to 10 ng/mL PSA range
Specificities & Sensitivities
phi = 0.724
For cancer detection
Greatest at high sensitivity

31. Novel Biomarkers [-2]proPSA (p2PSA) EPCA-2 PCA3 TMPRSS2-ERG
EN-2 (Engrailed-2)

32. EPCA-2 Serum Marker Early Prostate Cancer Antigen (EPCA)
Proteomic approach focused on nuclear structure
Changes in the cell nucleus are hallmarks of cancer
Highly specific and sensitive for prostate cancer
Leman, Adult Urology 2007 32

33. EPCA-2 Serum Marker
Leman, Adult Urology 2007 33

34. EPCA-2 Serum Marker EPCA-2 early-ELISA assay by Dr Getzenberg
Used a cut-off of 30 ng/mL
385 subjects
Results were reported as:
--92% specificity (CI 0.85 – 0.96)
--94% sensitivity (CI 0.93 – 0.99)
Research invalid
Showed best data, not all, found by auditing agency
Leman, Adult Urology 2007 34

35. Novel Biomarkers [-2]proPSA (p2PSA) EPCA-2 PCA3 TMPRSS2-ERG
EN-2 (Engrailed-2)

36. PCA3
Discovered by researchers in The Netherlands and Johns Hopkins University
Prostate-specific non-coding RNA which is overexpressed in 95% of prostate cancers
The PCA3 assay is a gene-based (mRNA) diagnostic test
Urine test
CE-marked (2006) and available as the PROGENSA PCA3 Assay in Europe
Not available in the US
Genprobe nearly done with clinical study
May submit this year 36

37. Attentive DRE to release sufficient # of prostate cells into the urine
Gen-Probe PCA3
Attentive DRE is critical otherwise invalid test
About $400
Attentive DRE to release sufficient # of prostate cells into the urine
Transport to PCA3 specialty lab where mRNA molecules are amplified (RT-PCR) and PCA3 score is calculated
Both PCA3 & PSA mRNA separately quantified:
the ratio of PCA3 to PSA mRNA is called the “PCA3 Score”

38. Gen-Probe PCA3 Study Characteristic Result Patient Sample Population
529 men
Average Serum PSA Level
7.9 +/ ng/mL**
May help with repeat neg biopsies
** PCA3’s clinical utility in the diagnostic “gray zone” of 2-10 ng/mL is not apparent due to their cohort’s high levels of PSA (PSA values ranged from 0.3 to 484 ng/mL PSA)
Gen-Probe EN Rev. A.1 38

39. PCA3: AUA 2011 Presentation
PCA3 AS A PREDICTOR OF UNFAVORABLE CONFIRMATORY BIOPSY PATHOLOGY IN CANDIDATES FOR ACTIVE SURVEILLANCE
Objective: Assessed ability of PCA3 to predict unfavorable confirmatory biopsy pathology in comparison to both PSA and PSA density (PSAD) in men with prostate cancer being considered for active surveillance. Subjects: 71 candidates with clinical stage T1c, Gleason score 6
Method: Predictive ability of these parameters was determined by ROC analysis. Results: ROC AUC: 0.65 for PCA3 overall, 0.65 for PCA3>35, 0.63 for PSA and 0.72 for PSA density. Conclusions: These preliminary results suggest that PCA3 is no better than either PSA or PSAD in predicting the results of confirmatory biopsies in prostate cancer patients being considered for active surveillance.
Not positive 39

40. Novel Biomarkers [-2]proPSA (p2PSA) EPCA-2 PCA3 TMPRSS2-ERG
EN-2 (Engrailed-2)

41. Gene Fusion Product: TMPRSS2-ERG
Gene fusions are genetic alterations – described in other types of tumors (e.g. BCR-ABL is present in 100% of chronic myelogenous leukemia)
TMPRSS2: Androgen-related transmembrane serine protease preferentially expressed in normal prostate tissue.
In PCa, TMPRSS2 may fuse with a transcription factor, ERG, which modulates the transcription of genes involved in cell growth, transformation and apoptosis.
The result of gene fusion with ERG is a mechanism for neoplastic transformation (i.e. cancer).
Very specific not sensitive
TMPRSS2 protein's function in prostate carcinogenesis relies on overexpression of ETS transcription factors, such as ERG and ETV1 through gene fusion. TMPRSS2-ERG fusion gene is the most frequent, present in 40% - 80% of prostate cancers in humans
Rubio-Briones. Urology 2010 41

42. Tomlins, Neoplasia 2008, Salagierski, Cancers, 2010
TMPRSS2-ERG
Tomlins, Neoplasia 2008, Salagierski, Cancers, 2010 42

43. Gene Fusion Product: TMPRSS2-ERG
Fusion genes may be detected in tissue or urine.
>20 TMPRSS2-ERG fusion transcripts have been described
Observed in 33% - 80% of PCa cases
RT-PCR assay
Current Status: Accurate gene fusion detection is complex, assays have not been standardized and once they are, larger studies will need to determine their clinical utility.
Rubio-Briones. Urology 2010 43

44. TMPRSS2-ERG Study – Survival Rates
PSA can be used as a prognostic indicator in positive pts
Negative: PSA not as useful
Gen Probe has a research assay
Kaplan-Meier plots byTMPRSS2-ERG gene fusion status
Results/Conclusions:
TMPRSS2-ERG detected in ~50% of subjects
Suggest possibility of using TMPRSS2-ERG status to classify different PCa prognostic groups.
May be relevant to refine therapeutic strategies
Rubio-Briones. Urology 2010 44

45. Novel Biomarkers [-2]proPSA (p2PSA) EPCA-2 PCA3 TMPRSS2-ERG
EN-2 (Engrailed-2)

46. Morgan et al, Clinical Cancer Research, 2011
Engrailed-2 (EN2)
A protein (transcription factor) expressed in prostate cancer (PC) cell lines and secreted into the urine
Protein measurement via ELISA
Results: EN2 was expressed and secreted by PC cell lines and PC tissue but not by normal prostate tissue.
The presence of EN2 in urine was highly predictive of PC, with a sensitivity of 66% and a specificity of 88.2%.
No correlation with PSA levels.
Conclusions: Urinary EN2 is a highly specific and sensitive candidate biomarker of prostate cancer.
Independent of PSA
Morgan et al, Clinical Cancer Research, 2011 46

47. RT-QPCR analysis of EN2 in biopsies
Engrailed-2 (EN2)
RT-QPCR analysis of EN2 in biopsies
that were found to be histologically positive ("PC") or negative ("Non PC") for prostate cancer.
Small N
Needs further study
Morgan et al, Clinical Cancer Research, 2011 47

48. Morgan et al, Clinical Cancer Research, 2011
Engrailed-2 (EN2)
ROC analysis of urine EN2 concentrations in men with biopsy proven prostate cancer versus the men that were a biopsy did not find PC.
Morgan et al, Clinical Cancer Research, 2011 48

49. Morgan et al, Clinical Cancer Research, 2011
Engrailed-2 (EN2)
EN2 can be detected in 100 mL of unprocessed urine, collected without DRE and uses a simple enzymatic detection method.
A larger, multicenter study is planned to determine whether EN2 could be used as a monitoring tool and whether levels of urinary EN2 correlate with tumor stage and Gleason grade.
Morgan et al, Clinical Cancer Research, 2011 49

50. Summary of Biomarkers Engrailed-2 (EN2) TMPRSS2-ERG PCA3 EPCA-2
Research stage, simple detection
Larger studies needed to determine utility
TMPRSS2-ERG
Research stage, complex detection
PCA3
CE Marked - Available in Europe
Promising biomarker, mixed results, complex detection
EPCA-2
Falsified data tainted this biomarker
Continued research in process
[-2]proPSA & phi
Promising biomarker & index for PCa detection

51. Overall Summary & Conclusions
Prostate cancer detection is a clinical dilemma
PSA is limited by poor clinical specificity
Separating indolent from aggressive cancers is a key unmet clinical need
There is a clear need for novel biomarkers that not only detect prostate cancer, but identify those men with aggressive cancer that would benefit from treatment.