Recent Developments in Prenatal Screening

1
Recent Developments in Prenatal Screening

• Dave Wright
• March 2005

2
Diagnostic Tests for Chromasomal Abnormalities

• Chorionic Villus Sampling (CVS)
• Amniocentesis
• These invasive tests increase the risk of
  miscarriage.
• If they were applied on a population basis to
  500,000 pregnancies, they would lead to about
  5,000 miscarriages in unaffected pregnancies.
• More than 5 miscarriages for every Down syndrome
  pregnancy.

3
Screening Tests
Ultrasound image data maternal serum
measurements
Bayes Theorem
Risk Pr Affected Maternal Age Markers
High Risk Offer CVS or Amniocentesis (About 5
are classified as high risk 25,000 invasive
tests and 250 miscarriages)
Low Risk
Reassurance
4
Biases

• Results based on secondary data provided in the
  SURUSS study of Wald et al (2003 and 2004)
• Without the raw data, we cant explore the
  effects of data cleaning that will cause biases
  in assessment of screening performance
• These biases are important
• There is a need for further prospective studies

5
Some Common Protocols(detection rate fixed at
85)
Assuming a population of 500,000 pregnancies
screened.
Proportion of unaffected pregnancies classified
as high risk
Note parameters from Wald et al (2003) with
refinements of Wald et al (2004) in brackets.
6
Wald et al (2003)

• The integrated test offers the most effective and
  safe method of screening.

7
Repeated MeasuresWright and Bradbury (2005)
PAPP-A2 denotes first and second trimester
measurements of PAPP-A.
8
Why were repeated measures protocols overlooked?

• Widely held misconceptions
• Best combination combination of best
• A measurement, such as PAPP-A in the second
  trimester, that has no discriminatory value on
  its own is of no value in combination with other
  markers
• High within class correlations are a sign of
  redundancy

9

• Best combination combination of best
• A measurement, such as PAPP-A in the second
  trimester, that has no discriminatory value on
  its own is of no value in combination with other
  markers
• High within class correlations are a sign of
  redundancy

10
Problems with integrated tests

• No assessment of risk in the first trimester
• Inefficient

11
Sequential ScreeningWright et al. (2004)
12
Sequential Screening
Detection rate fixed at 85
13
A Repeated Measures Sequential Strategy
Disposition of unaffected pregnancies
10 Weeks NT PAPP-A uE3
14 to 20 weeks PAPP-A uE3
14
Disposition of affected pregnancies
10 Weeks NT PAPP-A uE3
14 to 20 weeks PAPP-A uE3
15
Problems with NT

• Insufficient resources to measure NT in all
  pregnancies

16
Stage 1 Biochemical Tests
A three stage sequential screening policy (Wright
et al, in press)
Risk 1 in b
Screen negative
Stage 2 Nuchal Translucency
Risk 1 in c1
Risk 1 in c2
Screen negative
Most of the benefits of the full integrated
version of the test can be achieved with 70 of
pregnancies screened negative in the first stage.
Screen positive
Stage 3 Second Trimester
Risk 1 in d
Risk lt 1 in d
Screen negative
Screen positive
17
Illustration
Disposition of unaffected pregnancies
Screen negative
Screen positive
18
Illustration
Disposition of affected pregnancies
19
Concluding remarks

• Contingent sequential screening is being used in
  centres in Canada, the US and in the UK
• Trials with repeated measures are being planned