Cell free DNA: Disclosures The Popular Press vs The Evidence Mary - - PowerPoint PPT Presentation

6/5/2014 1

Cell free DNA: The Popular Press vs The Evidence

Mary E Norton, MD Professor of Obstetrics, Gynecology & Reproductive Sciences; UCSF Antepartum and Intrapartum Management

Disclosures

• Principal Investigator of ongoing clinical trial
• n cfDNA supported by Ariosa Diagnostics
• Unpaid clinical consultant for CellScape

and Natera

• Research support from Natera
• No personal financial involvement in any of

the cfDNA companies

6/5/2014 2

20 40 60 80 100 120

Detection rate of prenatal screening for Down syndrome has improved over time

Detection Rate (%)

Cell free fetal DNA

• Cell free fetal DNA (cffDNA) is made up of short

segments of fetal DNA (<200 base pairs) that circulate in maternal plasma

• Origin of these fragments is thought to be primarily

placenta

Maternal DNA Fetal DNA

Cell free DNA results from apoptosis

6/5/2014 3

Companies currently offering cfDNA screening (in order of appearance):

Noninvasive Prenatal Testing (NIPT)

• Detection requires accurate quantification of

DNA from a specific chromosome

• Somewhat different methods are utilized by

different laboratories

Analysis of fetal DNA

Zhong, X, Holzgreve, W, Glob. libr. women's med 2009

Wts (%) 100

99.1 (98.3 - 99.6)

2.5 94.4 (72.7 - 100) 1.2 100 (63.1 - 100) 3.5 100 (86.3 - 100) 4.1 100 (88.4 - 100) 1.6 100 (71.5 - 100) 4.9 100 (90.3 - 100) 11.0 100 (95.6 - 100) 1.2 100 (63.1 - 100) 1.6 100 (71.5 - 100) 2.3 100 (79.4 - 100) 12.0 100 (95.9 - 100) 6.8 100 (92.9 - 100) 1.9 100 (75.3 - 100) 28.4 98.6 (95.9 - 99.7) 5.3 100 (91.0 - 100) 50 60 70 80 90100 DR % (95% CI) (%) 11.6 100 (95.8 - 100)

Pooled analysis

Verweij et al., 2013 [18] Song et al., 2013 [8] Nicolaides et al., 2013 [25] Guex et al., 2013 [30] Zimmerman et al., 2012 [11] Sparks et al., 2012 [36] Norton et al., 2012 [81] Nicolaides et al., 2012 [8] Lau et al., 2012 [11] Jiang et al., 2012 [16] Bianchi et al., 2012 [89] Ashoor et al., 2012 [50] Sehnert et al., 2011 [13] Palomaki et al., 2011 [212] Ehrich et al., 2011 [39] Chiu et al., 2011 [86] Author DR (95% CI) 3 6 9 12 FPR % (95% CI) (%)

0.08 (0.03 - 0.17)

FPR (95% CI) Wts (%) 100 0.00 (0.0 - 0.7) 5.8 0.00 (0.0 - 0.2) 13.9 0.00 (0.0 - 1.8) 2.7 0.00 (0.0 - 2.5) 2.0 0.00 (0.0 - 2.7) 1.8 0.00 (0.0 - 2.8) 1.8 0.04 (0 - 0.2) 18.0 0.00 (0.0 - 0.2) 14.8 0.00 (0.0 - 3.7) 1.3 0.00 (0.0 - 0.4) 9.0 0.00 (0.0 - 0.9) 4.9 0.00 (0.0 - 1.1) 4.3 0.00 (0.0 - 10.3) 0.5 0.20 (0.0 - 0.6) 12.6 0.24 (0.0 - 1.4) 4.9 2.06 (0.4 - 5.9) 2.0

T21: n=733 11,475 non-T21

Trisomy 21 performance cfDNA testing: meta-analysis (Gil et al, Fetal Diagn Ther, 2014)

6/5/2014 4

Wts (%) 100 99.1 (98.3 - 99.6) 2.5 94.4 (72.7 - 100) 1.2 100 (63.1 - 100) 3.5 100 (86.3 - 100) 4.1 100 (88.4 - 100) 1.6 100 (71.5 - 100) 4.9 100 (90.3 - 100) 11.0 100 (95.6 - 100) 1.2 100 (63.1 - 100) 1.6 100 (71.5 - 100) 2.3 100 (79.4 - 100) 12.0 100 (95.9 - 100) 6.8 100 (92.9 - 100) 1.9 100 (75.3 - 100) 28.4 98.6 (95.9 - 99.7) 5.3 100 (91.0 - 100) 50 60 70 80 90100 DR % (95% CI) (%) 11.6 100 (95.8 - 100)

Pooled analysis

Verweij et al., 2013 [18] Song et al., 2013 [8] Nicolaides et al., 2013 [25] Guex et al., 2013 [30] Zimmerman et al., 2012 [11] Sparks et al., 2012 [36] Norton et al., 2012 [81] Nicolaides et al., 2012 [8] Lau et al., 2012 [11] Jiang et al., 2012 [16] Bianchi et al., 2012 [89] Ashoor et al., 2012 [50] Sehnert et al., 2011 [13] Palomaki et al., 2011 [212] Ehrich et al., 2011 [39] Chiu et al., 2011 [86] Author DR (95% CI) 3 6 9 12 FPR % (95% CI) (%)

0.08 (0.03 - 0.17)

FPR (95% CI) Wts (%) 100 0.00 (0.0 - 0.7) 5.8 0.00 (0.0 - 0.2) 13.9 0.00 (0.0 - 1.8) 2.7 0.00 (0.0 - 2.5) 2.0 0.00 (0.0 - 2.7) 1.8 0.00 (0.0 - 2.8) 1.8 0.04 (0 - 0.2) 18.0 0.00 (0.0 - 0.2) 14.8 0.00 (0.0 - 3.7) 1.3 0.00 (0.0 - 0.4) 9.0 0.00 (0.0 - 0.9) 4.9 0.00 (0.0 - 1.1) 4.3 0.00 (0.0 - 10.3) 0.5 0.20 (0.0 - 0.6) 12.6 0.24 (0.0 - 1.4) 4.9 2.06 (0.4 - 5.9) 2.0

T21: n=733 11,475 non-T21

Trisomy 21 performance cfDNA testing: meta-analysis (Gil et al, Fetal Diagn Ther, 2014)

DR: 99.1% (98.3 - 99.6)

Wts (%) 100 99.1 (98.3 - 99.6) 2.5 94.4 (72.7 - 100) 1.2 100 (63.1 - 100) 3.5 100 (86.3 - 100) 4.1 100 (88.4 - 100) 1.6 100 (71.5 - 100) 4.9 100 (90.3 - 100) 11.0 100 (95.6 - 100) 1.2 100 (63.1 - 100) 1.6 100 (71.5 - 100) 2.3 100 (79.4 - 100) 12.0 100 (95.9 - 100) 6.8 100 (92.9 - 100) 1.9 100 (75.3 - 100) 28.4 98.6 (95.9 - 99.7) 5.3 100 (91.0 - 100) 50 60 70 80 90100 DR % (95% CI) (%) 11.6 100 (95.8 - 100)

Pooled analysis

Verweij et al., 2013 [18] Song et al., 2013 [8] Nicolaides et al., 2013 [25] Guex et al., 2013 [30] Zimmerman et al., 2012 [11] Sparks et al., 2012 [36] Norton et al., 2012 [81] Nicolaides et al., 2012 [8] Lau et al., 2012 [11] Jiang et al., 2012 [16] Bianchi et al., 2012 [89] Ashoor et al., 2012 [50] Sehnert et al., 2011 [13] Palomaki et al., 2011 [212] Ehrich et al., 2011 [39] Chiu et al., 2011 [86] Author DR (95% CI) 3 6 9 12 FPR % (95% CI) (%)

0.08 (0.03 - 0.17)

FPR (95% CI) Wts (%) 100 0.00 (0.0 - 0.7) 5.8 0.00 (0.0 - 0.2) 13.9 0.00 (0.0 - 1.8) 2.7 0.00 (0.0 - 2.5) 2.0 0.00 (0.0 - 2.7) 1.8 0.00 (0.0 - 2.8) 1.8 0.04 (0 - 0.2) 18.0 0.00 (0.0 - 0.2) 14.8 0.00 (0.0 - 3.7) 1.3 0.00 (0.0 - 0.4) 9.0 0.00 (0.0 - 0.9) 4.9 0.00 (0.0 - 1.1) 4.3 0.00 (0.0 - 10.3) 0.5 0.20 (0.0 - 0.6) 12.6 0.24 (0.0 - 1.4) 4.9 2.06 (0.4 - 5.9) 2.0

T21: n=733 11,475 non-T21

Trisomy 21 performance cfDNA testing: meta-analysis (Gil et al, Fetal Diagn Ther, 2014)

DR: 99.1% (98.3 - 99.6)FPR: 0.08% (0.03 - 0.17)

6/5/2014 5

NIPT: Clinical Challenges

False positives:

• Unrecognized or vanishing twin
• Placental mosaicism
• Low level maternal mosaicism, esp sex chromosomal
• Maternal malignancy

False negatives:

• Genetic variants
• Placental mosaicism

Failed results:

• Increased BMI
• Failure to extract adequate material
• Individual variation in amount of cell free fetal DNA
• Fetal aneuploidy

Published Trials of NIPT: failure rates

Trial Failure rate Detection False positive rate Chiu et al (2011) 11/764 (1.4%) 86/86 3/146 Ehrich et al. (2011) 18/467 (3.8%) 39/39 1/410 Palomaki et al. (2011) 13/1696 (0.8%) 209/212 3/1471 Bianchi et al. (2012) 30/532 (3.0%) 89/89 0/404 Norton et al (2012) 148/3228 (4.6%) 81/81 1/2888 Zimmermann et al (2012) 21/166 (12.6%) 11/11 0/145 All 241/6853 (3.5%) 424/427 (99.3%) 8/5319 (0.15%)

Fetal fraction of DNA and test failure

Up to 5% of samples do not provide a result

• Low fraction fetal DNA, failed sequencing, high

variability in counts

• Some association with gestational age (<10 wks)
• Low fetal fraction associated with maternal BMI
• 20% at >250 lbs
• 50% at >350 lbs

Low fetal fraction appears to be associated with aneuploidy

Repeating test will provide a result in some cases

6/5/2014 6

Why is NIPT not diagnostic?

• Confined Placental Mosaicism
• THIS IS A PLACENTAL, AND NOT A FETAL TEST

False positive False negative

NIPT: Trisomy 13 and Trisomy 18

Author T13 T18 DR FPR DR FPR Palomaki ’11 11/12 (92%) 16/1688 (0.97%) 59/59 (100%) 5/1688 (0.28%) Bianchi ’12 11/14 (79%) 0/488 35/36 (99%) 0/460 (0%) Norton ’12

• 37/38 (97.4%) 2/2888(0.07%)

Zimmermann 2/2 (100%) 0/145 3/3 (100%) 0/145 ‘12 Porreco ’14 14/16 (92%) 0/3322 36/39 (92%) 0/3322 Bianchi ’14 1/1 (100%) 1/899 (0.1%) 2/2 (100%) 3/1905 (0.2%) TOTAL 39/45 (87%) 17/6542 (.03%) 172/177 (97%) 10/10,408 (0.1%)

Sex Chromosomal Aneuploidy

Author Cases: Controls DR FPR No result Samango- Sprouse 16:185 92% 7% Bianchi 20:532 75% 0.2% 20% Nicolaides 59:118 88% 0.8% 2.7% Total 95:835 86% 0.6% 10%

6/5/2014 7 Professional Society Opinions: ACOG; ACMG; International Society of Prenatal Diagnosis; National Society of Genetic Counselors

Common themes: There are recognized benefits, but…

• Not diagnostic
• Needs confirmation
• “Advanced screening test”
• Only detects common trisomies (vs invasive testing)
• Requires comprehensive genetic counseling
• Should only be used in validated groups (eg high risk)
• Need a low risk study before introducing into general

population screening

6/5/2014 8

How Does Test Performance Differ with Risk?

Low Risk (age 25; 1/1000) High Risk (age 38; 1/100)

N=1000 1 T21 999 not T21 1 TP, 0 FN 2 FP, 998 TN OAPR = 1/3 N=1000 10 T21 990 not T21 10 TP, 0 FN 2 FP, 988 TN OAPR = 10/12 or 5/6

(Assume >99% sensitivity and 99.8% specificity)

How Does Test Performance Differ with Risk?

Low Risk (age 25; 1/8,000) High Risk (age 38; 1/1000) N=10,000 1 T13 9,999 not T13 1 TP, 0 FN 100 FP, 9900 TN

OAPR = 1/100

N=10,000 10 T13 9,990 not T13 9 TP, 1 FN 100 FP, 9890 TN

OAPR = 9/100

(T13: Assume 99% sensitivity and 99% specificity)

• N=1914 women undergoing standard screening
• Mean maternal age = 29.6 yrs
• Primary outcome = false positive rates for T18 and T21

6/5/2014 9

cfDNA vs Standard Screening

Bianchi et al, NEJM, 2014 FPR PPV cfDNA 0.3% 45.5% p<.001 Standard 3.6% 4.2%

• Only 8 aneuploidy cases in the cohort (5: T21, 2: T18,

and 1: T13)

• All were detected

Where does cfDNA fit?

Is this an outstanding screening test or an imperfect diagnostic test? Is this best used as a secondary screening test, or as a first tier screening test? Are we ready to abandon current screening in favor of cfDNA?

Secretary’s Advisory Committee

• n Genetics, Health and Society
• Analytic validity: ability of test to measure

particular genetic characteristics (eg DNA sequence) accurately and reliably in a given specimen

• Clinical validity: test’s accuracy in detecting the

presence of, or predicting risk for, a health condition or phenotype Clinical utility: balance between health related benefits and harms that can ensue from a genetic test

Personalized Medicine 2008

6/5/2014 10

NIPT is more precise for T18, 21

cfDNA

Current NT + serum screen

cfDNA

Current NT + serum screen Other abnormalities

NIPT is more precise for T18, 21 NIPT is more precise for T18, 21

cfDNA

FTS Other abnormalities 8/8 T21 2/3 T18; 1/3 no result 8/8 T21 3/3 T18 7/7 others (45X; triploidy; deletions and duplications)

Nicolaides et al, 2012

6/5/2014 11

NIPT is more precise for T18, 21

cfDNA

FTS Other abnormalities 8/8 T21 2/3 T18; 1/3 no result 8/8 T21 3/3 T18 7/7 others (45X; triploidy; deletions and duplications) 55% (10/18) 100% (18/18)

Nicolaides et al, 2012

DS and T18 are 2/3 of aneuploidies detectable by karyotype

Aneuploidies Detected by Prenatal Diagnostic Testing

Tri 21: 53.2% Sex chrom: 8.2% Tri 13: 4.6% Tri 18: 17.0%

Other 16.9%

Norton et al, SMFM, 2014

Chromosome Abnormalities by Maternal Age

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% < 25 25 to 29 30 to 34 35 to 39 40 to 44 ≥ 45 Percent Detected Age Group (Years)

NIPT Detectable

Yes No P<0.01

Norton et al, SMFM, 2014

6/5/2014 12

NIPT Detection Rate

• ~83% of chromosomal abnormalities

detected by current screening can potentially be identified by NIPT

• This varies by maternal age
• Lower detection in younger women (75-80%)
• Greater detection in older women, but still only

90%

Disorders potentially detectable by serum screening and NIPT

NIPT Current Screening

• Trisomy 21
• Trisomy 18
• Trisomy 13
• Some sex

chromosomes

• Trisomy 21
• Trisomy 18
• Trisomy 13
• Some sex chromosomes
• Triploidy
• Other rare aneuploidies
• Congenital heart defects
• Noonan syndrome
• Neural tube defects
• Ventral wall defects
• Congenital adrenal hypoplasia
• Smith Lemli Opitz syndrome
• Steroid sulfatase deficiency
• Adverse OB outcomes (IUGR, PreE, PTB)

Disorder Prevalence Common trisomies (13,18,21) 0.2% Other chromosome abnormalities 0.4% Microdeletions and duplications 1.5% Mendelian Genetic Disorders 0.4% Congenital heart defects 0.3% Other structural defects 3% Adverse OB outcomes 15-20% Total ~25%

Causes of Birth Defects and Other Adverse Perinatal Outcomes: It’s Not All Down Syndrome

6/5/2014 13

NIPT: Expanded panels

Laboratories have added other trisomies and microdeletions

• Trisomies 16 and 22
• Microdeletion syndromes
• 22q (diGeorge)
• 5p (cri-du-chat)
• 1p36
• 15q (Prader Willi)
• 4p (Wolf-Hirshhorn)

What is a microdeletion?

• 1MB (megabase) = 1 million base pairs
• Microdeletions are 100kb to several MB
• Karyotype can usually only visually detect >7-10 MB

Outcome will depend on the size & the genes involved

Microdeletion syndromes

Syndrome Frequency Features

22q11.2 (DiGeorge) 1/4K Varies: cardiac, palatal, immune, intellectual disability 1q36 1/5-10K Severe intellectual disability (ID), +/- obvious structural anomalies Angelman 1/12-20K Severe ID, seizures, speech delay Prader-Willi 1/10-30K Obesity, ID, behavioral problems Cri-du-chat 1/20-50K Microcephaly, ID, +/- CHD Wolf- Hirshhorn 1/50K ID, seizures, +/- CL/CP

6/5/2014 14

Prevalence in 100,000 Live Births

20 40 60 80 100 120 140

53

NIPT for Rare Disorders

N=100,000 2 Wolf-Hirschhorn 99,998 not WHS 2 TP; 0 FN 800 FP; 99,198 TN

OAPR = 1/400

Population Risk = 1/50,000

(Wolf-Hirschhorn, 4p-: Assume 99% sensitivity and 99.2% specificity)

Chromosomal Microarray (CMA) for Prenatal Diagnosis

6/5/2014 15 Karyotype

Resolution: >7-10 Million Base Pairs (7-10 Mb) Resolution: < 0.5 Million Base Pairs (< 500 kb)

Chromosomal Microarray

Genomic imbalance detected by CMA but not karyotype

Miller et al, 2010, AJHG

Diagnostic Yield in Cases with Normal Karyotype

Indication for Testing Clinically Relevant

(N=96)

U/S Anomaly

N=755

6.0% AMA

N=1,966

1.7% Positive Screen

N=729

1.7% Other

N=372

1.3%

In patients with fetal structural abnormalities undergoing prenatal diagnosis, microarray is recommended.

6/5/2014 16

New “menu” in prenatal testing

Screening test for common aneuploidies (1/500) VS Invasive diagnostic testing with CMA (1/60)

NIPT and chromosomal microarray

IF: CMA detects an abnormality in 1.7% of cases (about 1/60) AND: NIPT detects T13,18, 21 – about 1/500 pregnancies THEN: If NIPT is the routine screening test, it will detect only about 12% of diagnosable chromosomal abnormalities

6/5/2014 17

Cost sensitivity analysis of NIPT

• Calculated marginal costs of Down syndrome detection

compared to current screening

• If NIPT costs $1000, increased cost per case detected is $3.6

million more than first trimester combined screening Universal NIPT screening will only become cost effective if costs drop substantially Contingent screening of highest risk 10-20% is recommended Cuckle et al, Prenat Diag 2013

6/5/2014 18

Summary

• cfDNA is a better test for Down syndrome than

current screening

• For patients that obtain a result, a positive or negative test is

near diagnostic

• Test failure indicates an increased risk for aneuploidy
• These patients require counseling and follow up
• The detected disorders are fewer than with

traditional screening or diagnostic testing

• Patients need to be carefully counseled about

the trade-offs of lower false positives but fewer disorders tested If only it were this simple…

Thank You!