Download A genetic risk score to guide age-specific, personalized prostate cancer screening PDF

TitleA genetic risk score to guide age-specific, personalized prostate cancer screening
Author
LanguageEnglish
File Size1.2 MB
Total Pages25
Document Text Contents
Page 1

1

A genetic risk score to guide age-specific, personalized prostate cancer
screening
Tyler M. Seibert, MD1,2
Chun Chieh Fan, MD1,3
Yunpeng Wang, PhD4
Verena Zuber, PhD4
Roshan Karunamuni, PhD1,2
J. Kellogg Parsons, MD5
Rosalind A. Eeles, PhD6,7
Douglas F. Easton, PhD8
ZSofia Kote-Jarai, PhD6
Ali Amin Al Olama, PhD8
Sara Benlloch Garcia, PhD8
Kenneth Muir, PhD9,10
Henrik Gronberg, Prof11
Fredrik Wiklund, PhD11
Markus Aly, PhD11,12
Johanna Schleutker, PhD13,14
Csilla Sipeky, PhD15
Teuvo LJ Tammela, Prof16
Børge G. Nordestgaard, Prof17,18
Sune F. Nielsen, PhD17,18
Maren Weischer, MD18
Rasmus Bisbjerg, MD19
M. Andreas Røder, MD20
Peter Iversen, Prof17,20
Tim J. Key, DPhil21
Ruth C. Travis, DPhil21
David E. Neal, FMedSci22,23
Jenny L. Donovan, PhD24
Freddie C. Hamdy, FMedSci25
Paul Pharoah, Prof26
Nora Pashayan, MD27,26
Kay-Tee Khaw, FRCP28
Christiane Maier, PhD29
Walther Vogel, Prof29
Manuel Luedeke, PhD29
Kathleen Herkommer, MD30
Adam S. Kibel, MD31
Cezary Cybulski, MD32
Dominika Wokolorczyk, PhD32
Wojciech Kluzniak, MS32
Lisa Cannon-Albright, PhD33,34
Hermann Brenner, Prof35,36,37

All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint. http://dx.doi.org/10.1101/089383doi: bioRxiv preprint first posted online Nov. 25, 2016;

http://dx.doi.org/10.1101/089383

Page 2

2

Katarina Cuk, PhD35
Kai-Uwe Saum, PhD35
Jong Y. Park, PhD38
Thomas A. Sellers, PhD39
Chavdar Slavov, DMSc40
Radka Kaneva, PhD41
Vanio Mitev, DSc41
Jyotsna Batra, PhD42
Judith A. Clements, PhD42
Amanda Spurdle, PhD43
Australian Prostate Cancer BioResource 42,44
Manuel R. Teixeira, PhD45,46
Paula Paulo, PhD45
Sofia Maia, PhD45
Hardev Pandha, FRCP47
Agnieszka Michael, PhD47
Andrzej Kierzek, PhD47
David S. Karow, MD1,48
Ian G. Mills, PhD4,49
Ole A. Andreassen, MD4
Anders M. Dale, PhD1,48,50
The PRACTICAL consortium*

Affiliations

1Center for Multimodal Imaging & Genetics, University of California, San Diego, La Jolla,
CA, USA
2Department of Radiation Medicine & Applied Sciences, University of California, San
Diego, La Jolla, CA, USA
3Department of Cognitive Science, University of California, San Diego, La Jolla, CA,
USA
4University of Oslo, Oslo, Norway
5Department of Surgery, University of California, San Diego, La Jolla, CA, USA
6The Institute of Cancer Research, London, SM2 5NG, UK
7Royal Marsden NHS Foundation Trust, London, SW3 6JJ, UK
8Centre for Cancer Genetic Epidemiology, Department of Public Health and Primary
Care, University of Cambridge, Strangeways Research Laboratory, Worts Causeway,
Cambridge CB1 8RN, UK
9Institute of Population Health, University of Manchester, Manchester, UK
10Warwick Medical School, University of Warwick, Coventry, UK
11Department of Medical Epidemiology and Biostatistics, Karolinska Institutet,
Stockholm, Sweden
12Department of Molecular Medicine and Surgery, Karolinska Institutet, and Department
of Urology, Karolinska University Hospital, Solna, 171 76 Stockholm

All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint. http://dx.doi.org/10.1101/089383doi: bioRxiv preprint first posted online Nov. 25, 2016;

http://dx.doi.org/10.1101/089383

Page 12

12

PSA screening is controversial, but most guidelines recommend individualized
discussion between physicians and patients12,14–16. PHS affords a quantitative and
understandable way to evaluate individual risk that could prove pivotal in these
discussions. The results here suggest PHS can be used to identify a large percentage
of men for whom forgoing or delaying PSA screening makes sense. At the same time,
PHS can also identify men with a high risk of developing PCa at a young age and who
therefore may benefit from early detection through PSA screening.
Another concern with PSA screening is overtreatment of indolent disease. Genetic
prediction of aggressive PCa alone has proven elusive19, and the problem is
compounded by the propensity of initially low-risk tumors to progress20,21. Datasets
describing the initial tumor characteristics are not enough—tumors that develop
aggressive features over time must also be identified if aggressive-PCa-only predictors
are to be effective. Active surveillance is one answer to overtreatment that avoids up-
front aggressive treatment but still allows intervention if the tumor progresses. Indeed,
most tumors will eventually require treatment21,22, and earlier treatment prevents
development of metastatic disease22. Hence, avoiding screening altogether in patients
who may develop PCa at a young age does carry risk of considerable morbidity.
While PHS was applied here to PSA screening, PHS itself is not specific to PSA testing.
Rather, the PHS is predictive of patients’ underlying risk of PCa at a given age and
therefore relates to pre-test probability—and, by extension, positive predictive value—
within any screening strategy that might be adopted.
Cost effectiveness is a prominent concern in all discussions of healthcare policy. PHS
has the potential to improve screening efficiency while also reducing overall costs. PHS
need only be calculated once and is valid for a lifetime. The genotyping chip assay can
be run for costs in the range of that for single-gene testing (e.g., BRCA mutation),
informs multiple diseases23, and a saliva sample suffices. PSA screening and
subsequent biopsies could thus be limited to those men at higher risk, while delaying or
forgoing screening in those whose genetic makeup confers a low risk.
Prior studies have used GWAS-associated polymorphisms to predict risk of PCa using a
case/control design24–26. However, epidemiologic data show that PCa risk is not a
simple dichotomy of cases and controls, but rather is highly dependent on increasing
age. Therefore, we opted for a survival analysis approach optimized for genetic
prediction of age of PCa onset. The PHS can then be used in clinical decisions, where
age plays a critical role. If a man has a high risk of developing prostate cancer at age
95, this is a very different clinical situation from a man at high risk at age 55.
There are several limitations to this study. It is inherently a retrospective analysis, but
the Discovery Set data come from large studies carried out in multiple institutions and
nations; the Validation Set, too, comes from an independent, large, prospective trial.
The absolute risk models shown in Figure 5 are only as accurate as the population data
upon which they are based, which reflect diagnosed prostate cancer in the U.S. It is
important to note, though, that PHS is a measure of hazard and therefore could be
readily applied to other population incidence curves to estimate absolute risk in those

All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint. http://dx.doi.org/10.1101/089383doi: bioRxiv preprint first posted online Nov. 25, 2016;

http://dx.doi.org/10.1101/089383

Page 13

13

populations. The age range of the Validation Set is limited to only 50-70 years;
fortunately, this includes the age where screening is believed to have the most
benefit12,14–16. Finally, race in this PHS model is limited to European ancestry. Validation
of PHS in other racial groups—and, if necessary, custom models for each—is needed;
our group plans to investigate this important question.
In conclusion, we describe here the development of a new polygenic hazard score for
personalized genetic assessment of individual, age-associated prostate cancer risk.
This score has been validated in an independent data set, demonstrating accurate
prediction of prostate cancer onset. Moreover, PHS is shown to predict the utility of PSA
testing for an individual patient and can guide the decision of whether and when to
screen for prostate cancer.



All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint. http://dx.doi.org/10.1101/089383doi: bioRxiv preprint first posted online Nov. 25, 2016;

http://dx.doi.org/10.1101/089383

Page 24

24

Table 2: Prostate cancer-risk (PCaR) age
PHS percentilea DAgeb

[95% CI]
PCaR age when

true age is 50 years
Age when risk is that of

typical 50-year-oldc
0.1 -35

[-35,-19]
15

[15,31]
85

[69,101]
1 -11

[-15,-6]
39

[35,44]
61

[56,65]
5 -7

[-11,-2]
43

[39,48]
57

[52,60]
20 -3

[-7,1]
47

[43,51]
53

[49,57]
50 0

[-4,4]
50

[46,54]
50

[46,54]
80 3

[0,7]
53

[50,57]
47

[43,50]
95 7

[3,10]
57

[53,60]
43

[40,47]
99 9

[5,13]
59

[55,63]
41

[37,45]
99.9 13

[9,16]
63

[59,66]
37

[34,41]
aPHS percentile among young (<70 years old) controls within Discovery Set.
bDAge = PCaR age – true age.
cRisk of typical 50-year-old defined as overall population incidence at age 50.






All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint. http://dx.doi.org/10.1101/089383doi: bioRxiv preprint first posted online Nov. 25, 2016;

http://dx.doi.org/10.1101/089383

Page 25

25

Research in Context


Evidence before this study
Prostate cancer (PCa) screening with prostate-specific-antigen (PSA) testing can lead
to early detection of PCa and allow for curative treatment, but universal screening also
has considerable disadvantages for men who may never develop life-threatening
disease. Whom to screen and at what age to do so remain unclear.

Genetic studies have shown that single-nucleotide polymorphisms (SNPs) have modest
predictive value for PCa risk, but that a combination of genotype information from
multiple SNPs can give a more robust PCa risk prediction. The practical, clinical utility of
SNP-based prediction of expected age of PCa onset is not well understood.

Added value of this study
This study presents and validates a novel polygenic hazard score that is an indicator of
PCa-free survival. This polygenic hazard score (PHS) offers a relatively inexpensive
assessment of an individual man’s age-specific PCa risk.

Implications of all the available evidence
SNP-based polygenic hazard scores can provide objective, readily interpretable
information to guide the decision of whether a given patient might benefit from PCa
screening at each age in his life



All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint. http://dx.doi.org/10.1101/089383doi: bioRxiv preprint first posted online Nov. 25, 2016;

http://dx.doi.org/10.1101/089383

Similer Documents