Original Paper Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
Received: July 31, 2013 Accepted after revision: December 27, 2013 Published online: June 13, 2014
BAC Chromosomal Microarray for Prenatal Detection of Chromosome Anomalies in Fetal Ultrasound Anomalies: An Economic Evaluation Sarah C. Hillman a, e Pelham M. Barton b Tracy E. Roberts b Eamonn R. Maher a Dominic M. McMullan c Mark D. Kilby a, d, e a
School of Clinical and Experimental Medicine, College of Medicine and Dentistry, b Health Economics Unit, School of Health and Populations Science, c West Midlands Genetics Laboratory and d Fetal Medicine Centre, Birmingham Women’s Foundation Trust, and e Centre for Women’s and Children’s Health, University of Birmingham, Birmingham, UK
Abstract Introduction: To determine the cost-effectiveness of prenatal chromosomal microarray (CMA) when performed for structural anomalies on fetal ultrasound scan over conventional techniques. Method: A decision tree was populated using data from a prospective cohort of women undergoing testing when a fetal ultrasound scan showed a structural abnormality. Nine strategies of testing were modeled including combinations of the tests: QFPCR, G-band karyotyping, CMA and FISH for DiGeorge (22q) microdeletion. Results: When CMA costs GBP 405 and using a 1-Mb BAC array it would cost GBP 24,600 for every additional case detected by CMA over a combination of QFPCR, followed by G-band karyotype, followed lastly by FISH (for DiGeorge syndrome). If CMA is performed instead of conventional karyotyping alone it costs GBP 33,000 for every additional case detected. However, if the cost of CMA is reduced to GBP 360 than when CMA is performed instead of conventional karyotyping alone it would cost GBP 9,768 for every additional case detected. Discussion: The use of a prenatal BAC CMA is not currently cost-effective when compared to other testing strategies.
© 2014 S. Karger AG, Basel 1015–3837/14/0361–0049$39.50/0 E-Mail [email protected] www.karger.com/fdt
However, as CMA costs decrease and resolution (and detection rates) increase it is likely to become the cost-effective option of the future. © 2014 S. Karger AG, Basel
Introduction
Prenatal analysis of fetal chromosomes has been available since the 1960s. Full G-band karyotyping has classically been used to detect chromosomal anomalies at a resolution of between 5 and 10 Mb. This technology is now being supplemented, and in some instances replaced, by chromosomal microarray (CMA) which is capable of examining chromosomes to a resolution of 1 kb, smaller than the average gene. CMA would be appropriate when there is a ‘high-risk’ result found on aneuploidy screening, there is a previous history of chromosomal anomalies, advanced maternal age or for parental anxiety. But its highest detection rate and arguably its most effective use in the prenatal setting is for detecting submicroscopic chromosome deletions and duplications when a congenital anomaly is detected on fetal ultrasound scan (USS) [1, 2]. The advantages of CMA lie in its ability to detect smaller potentially pathological chromosomal variants that are Prof. Mark D. Kilby School of Clinical and Experimental Medicine College of Medical and Dental Sciences University of Birmingham, Birmingham B15 2TT (UK) E-Mail m.d.kilby @ bham.ac.uk
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Key Words Chromosomal microarray · Fetal anomaly · Ultrasound · Economic evaluation
Table 1. Different strategies of chromosomal testing that were analysed within the model
Option
Test 1
Followed by test 2
Followed by test 3
1 2 3 4 5 6 7 8 9
QFPCR G-band karyotyping CMA QFPCR QFPCR G-band karyotyping QFPCR QFPCR QFPCR
– – – Then G-band karyotyping (if QFPCR-negative) Then CMA (if QFPCR-negative) And CMA (regardless of G-band karyotyping result) Then G-band karyotyping (if QFPCR-negative) Then FISH (if QFPCR-negative) Then G-band karyotyping (if QFPCR-negative)
– – – – – – And CMA (regardless of G-band karyotyping result) – Then FISH (if G-band karyotyping-negative)
Parental FU
Diseasepositive
pSens_Kar...
pfollowup1 No parental FU #
pPrevalence Karyo-negative
Karyotype only
Method
# Diseasenegative #
Karyo-positive
Parental FU
# Karyo-negative pSpec_karyo
Fig. 1. An example of an ‘arm’ of the decision tree for ‘karyotype
only’. pPrevalence = Prevalence of chromosomal anomalies; pSens = sensitivity of karyotyping; pSpec = specificity of karyotyping; pfollowup = probability that parental karyotyping would be required.
undetectable using standard cytogenetic analyses, to be customised and be amenable for high-volume throughput. Potential drawbacks of CMA are inability to detect balanced chromosomal rearrangements, triploidy and some instances of mosaicism. The biggest challenge presented by this new technology is the detection of chromosomal variants of unknown clinical significance (VOUS) and the potentially adverse effect reporting these variants may have on couples and their decision-making antenatally. The higher detection rate of prenatal chromosomal anomalies has been shown in multiples cohort studies [1– 3], however CMA does currently cost more and there is currently no evidence on whether its use in the prenatal 50
setting will provide value for money. This report will examine the cost-effectiveness of prenatal CMA when used because a structural anomaly has been detected on USS. It will compare CMA to standard conventional cytogenetic testing.
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
To assess the cost-effectiveness of the various tests or combination of tests, we carried out model-based economic analyses. Our analysis was conducted from the healthcare perspective and based on an outcome of cost per additional case detected. Our decision tree models were created using TreeAge software (software that is widely used in health economics to build and analyse decision trees) (fig. 1) (TreeAge Software, Inc., Williamstown, Mass., USA). The model was populated using data gained between December 2009 and April 2012 at Birmingham Women’s Hospital where we prospectively recruited women when a fetal USS showed a significant structural anomaly. Ethical approval for the present study was granted by the Staffordshire Research and Ethics Committee (Ref. No. 09/H1203/74). Women were counselled and offered fetal chromosome testing by invasive sampling. If accepted, women were also offered, counselled and consented to have CMA testing in addition to standard cytogenetic testing. 309 women were recruited and had cytogenetic testing performed [4]. In all cases they had QFPCR (quantitative fluorescence polymerase chain reaction), if this was positive for trisomy 13, 18 or 21 they did not go on to have CMA testing but for the model it was anticipated that had these cases undergone karyotype or CMA then the test would have detected aneuploidy. This cohort is described in detail in our study accepted for publication. Nine options for testing within the model were considered using combinations of the tests; QFPCR conventional G-band karyotyping, CMA, FISH (fluorescence in situ hybridization for 22q11.2 microdeletion syndrome on remaining fetal samples with cardiac abnormalities). The definition of these alternative options is presented in table 1.
Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Karyo-positive
Color version available online
FISH for 22q11.2 microdeletion syndrome on remaining fetal samples with cardiac abnormalities.
1-Mb BAC array Screening test
Probability parameter (sensitivity)
Probability parameter (specificity)
QFPCR Karyotyping CMA Karyotyping and CMA FISH
66/91 (73%) 82/91 (90%) 87/91 (96%) 91/91 (100%) 4/7 (57%)
218/218 (100%) 217/218 (99.5%) 215/218 (99%) 214/218 (98%) 34/34 (100%)
Tests accuracy data in the first model was taken using the data from our own prospective cohort. When calculating the sensitivity of CMA and karyotyping it was assumed that the overall result would be positive if either test was positive on its own. The sensitivity therefore increased as all individuals that tested positive on either test were potentially positive cases. The specificity therefore falls as there are more false positives and it is lower than the specificity of each test alone.
Some assumptions were required in order to develop a workable model: (1) In cytogenetic testing results may be positive or negative but (more so in CMA testing) it is possible to have a result or VOUS. This result shows a chromosomal variance where the evidence is lacking to link it to the phenotype with certainty but evidence does not exist to call it normal/benign with certainty either. Future research could show this to be a variant causal of a pathological phenotype or a normal but uncommon variant within the population. In these models due to the possible pathological nature of these results, we treated them as true positive results and allowed for the costs of parental follow-up to be included. However, to allow for this the model was re-run treating the VOUS as a false positive finding (there was only one VOUS in the cohort detected by CMA), the results of this sensitivity analysis are included. (2) No prior testing had been performed on the sample acquired from the fetus. (3) Triploidies (i.e. 69,XXX/69,XXY) would require no parental follow-up. (4) That a true positive or true negative result gave an effectiveness score of 1 and a false positive or false negative gave an effectiveness score of 0. However, for some patients if there was any uncertainty associated with the results they may have ranked their efficacy at 5 cases where microarray testing was performed prenatally in addition to karyotyping. Data for the model were extracted only when testing had been performed because a structural anomaly had been found on fetal USS. From the 21 primary studies, 17 had this information [1, 3, 7–20]. In total, 4,276 results were included. Five testing options were considered: karyotyping alone, CMA alone, karyotyping then CMA (if the karyotyping result is negative), CMA then karyotyping (if the CMA result is negative), karyotyping and CMA (both tests done regardless of the outcome of the other). The prevalence was calculated including VOUS as ‘test positive’ results. Cost and Resource Data In order to have cytogenetic testing a patient must be counselled by an appropriately trained healthcare professional (HCP) and have an invasive test performed (amniocentesis, chorionic villus sampling (CVS) or fetal blood sampling). The sample is then sent to an approved laboratory for testing and results are disseminated to HCPs. If negative/normal they are telephoned to the patient but if abnormal the patient re-attends for further counselling by HCPs and in some cases cytogenetic testing on parental samples is performed. Here we have only included costing from when the sample arrives at the laboratory and then subsequent follow-up costs. The initial cost of a consultation and the invasive test procedure were not included as it was assumed that these would be the same for all cases. Base costs are calculated as if the invasive sample taken was amniotic fluid by amniocentesis, however the models have been recalculated for CVS samples. The costs of resources utilised were those that were directly incurred by the NHS. The laboratory costs were calculated by the West Midlands Genetics Laboratory. Costs are reported in GBP. Costs of the cytogenetic tests included DNA extraction, cost of the base test and consumables, staffing costs, capital costs (such as the scanner service contract) and administration costs. Staffing costs for the laboratory work were calculated by timing staff performing the procedures. The average salary of those routinely doing the work was then used and the cost per hour of their time calculated. For the staffing costs for those performing the analysis work, as this can be highly variable depending on the complexity of the case, the average time was estimated. Capital costs such as the scanner service contract were calculated by taking the cost of the contract and dividing it by the number of samples processed over the time of the contract. Administration costs have been previously calculated by the laboratory per test and are included. Consumable costs were then recorded as those directly incurred by the NHS. Cost of follow-up included costs of seeing a consultant clinical geneticist and specialist midwife. This was based on a 30-min appointment. These costs were tak-
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
51
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Table 2. Test data accuracy taken from prospective cohort using a
Table 3. Unit costs and sources
Screening test
Unit cost, GBP
Source
QFPCR trisomy 13, 18 and 21 and sex chromosome aneuploidy Karyotyping amniocentesis Karyotyping chorionic villus sampling CMA (1-Mb BAC array) FISH for 22q11.2 DS Parental CMA Parental karyotype Follow-up specialist midwife per 30 min Follow-up clinical geneticist (consultant) per 30 min
129 223 265 405 186 350 382 48.50 81
West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory Unit costs of health and social care 2011 [21] Unit costs of health and social care 2011 [21]
en from unit costs of health and social care 2011 [21] (table 3). Costs used were those calculated by the laboratory for the years 2011/2012. It was assumed that follow-up with parental samples would be required for any abnormal cytogenetic test unless the results was an autosomal triploidy, trisomy or sex chromosome aneuploidy. If a trisomy 13, 18 or 21 was detected in the model by CMA or QFPCR, the cost of parental follow-up was not included (to determine if the trisomy was from a robersonian translocation) as this cost is required for future counselling regarding recurrence risk but would not change the outcome for the current fetus, as opposed to parental follow-up for VOUS which would alter the outcome for the current fetus when determining pathogenicity of the chromosomal anomaly.
by simultaneously selecting random values from each distribution. A Monte Carlo simulation repeated the process 10,000 times. This gave an indication of how variation in the test sensitivity and specificity lead to variation in the ICER. The decision uncertainty surrounding the use of CMA as a replacement for karyotyping was examined with the cost-effectiveness analysis curve (CEAC). This plots the probability that CMA will be cost-effective as compared to conventional G-band karyotyping at a given threshold of willingness to pay (WTP) that decision-makers may be willing to pay for a gain in effectiveness.
Sensitivity Analysis In addition to the base case analysis, we carried out deterministic sensitivity analysis in which the following changes were made: (1) In the CMA arm when a VOUS was found they were removed from being true positive and treated as false positive results. (2) The costings for laboratory karyotyping were changed from GBP 222 (karyotyping performed on DNA extracted from amniocytes) to GBP 265 (karyotyping performed on DNA extracted from a CVS). (3) Threshold sensitivity analyses were carried out to establish the critical value of CMA that would change the deterministic results in terms of ICERs and may affect the decision of policymakers. A probabilistic sensitivity analysis was also undertaken to determine the uncertainty of the model. A β distribution was assigned to each true positive, true negative, false positive and false negative parameter. Cost-effectiveness results were then calculated
52
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
Incremental Cost-Effectiveness Ratios Strategy 1 (presented in table 4) looks at the ICER when the 1-Mb BAC array is used and VOUS are treated as true positive results. Due to the targeted resolution of the 1-Mb BAC array and the finding of four submicroscopic deletions in the DiGeorge region (that would be detectable by FISH analysis as well as CMA) the ICER of CMA over a combination of the tests QFPCR followed by karyotype followed by FISH (for DiGeorge in known fetal cardiac anomalies) was GBP 24,600. This means that it would cost an extra GBP 24,600 for every additional case detected by CMA over the combination of tests QFPCR then karyotype (if QFPCR-negative) then FISH DiGeorge (if karyotype-negative and known cardiac anomaly). When the base case cost is changed to GBP 265 (assuming that the invasive test was CVS rather than amniocentesis, strategy 2; table 4) the ICER of CMA over QFPCR followed by karyotype followed by FISH for DiGeorge syndrome (in known fetal cardiac anomalies) decreased to GBP 14,200. The ICER of CMA alone over karyotype alone is GBP 33,300 (i.e. GBP 33,300 per additional case detected by CMA) for amniocentesis samples (strategy 1; table 4) or Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Outcome Measures In this analysis, consequences are reported in terms of the additional diagnoses provided by CMA compared to conventional G-band karyotyping. The results are reported in terms of the incremental cost-effectiveness ratio (ICER) based on the additional cost for every additional diagnosis. The ICER is universally accepted as the standard summary ratio that should be used for reporting the results of economic evaluations. The additional cost per additional unit of benefit can be applied to any evaluation as long as there is a comparison of costs and outcomes of at least two alternative pathways or strategies.
Results
Table 4. Incremental ICERs
Strategy
Cost, GBP
Incremental cost
Effectiveness
Incremental effectiveness
ICER, GBP
Strategy 1 QFPCR only QFPCR then FISH DiGeorge Karyotype only QFPCR then karyotype then FISH DiGeorge CMA only Karyotype and CMA
173 205 298 411 490 722
32 93 112 80 232
0.9159 0.9288 0.9644 0.9774 0.9806 0.9832
0.0129 0.0356 0.0129 0.0032 0.0026
2,500 2,600 8,700 24,600 90,500
Strategy 2 QFPCR only QFPCR then FISH DiGeorge Karyotype only QFPCR then karyotype then FISH DiGeorge CMA only Karyotype and CMA
173 205 341 444 490 756
32 136 103 46 266
0.9159 0.9288 0.9644 0.9774 0.9806 0.9832
0.0129 0.0356 0.0129 0.0032 0.0026
2,500 3,800 8,000 14,200 103,800
Strategy 3 QFPCR only QFPCR then FISH DiGeorge Karyotype only QFPCR then karyotype then FISH DiGeorge QFPCR then karyotype and CMA
173 205 298 411 719
32 93 112 309
0.9191 0.9320 0.9676 0.9806 0.9871
0.0129 0.0356 0.0129 0.0065
2,489 2,616 8,670 47,697
Strategy 4 Karyotype only CMA only CMA then karyo
255 469 655
214 185
0.9107 0.9925 0.9991
0.0819 0.0066
2,600 28,300
Strategy 5 Karyotype only CMA only CMA then karyo
255 469 655
214 185
0.9474 0.9558 0.9623
0.0084 0.0065
25,400 28,300
GBP 22,200 (i.e. GBP 22,200 per additional case detected by CMA) for CVS samples (strategy 2; table 4). When VOUS are treated as false positives (strategy 3; table 4), CMA alone is dominated by all other strategies. However, treating VOUS as false positives will underestimate the specificity of CMA as some VOUS will in time be determined to be benign but others will be pathogenic. The current base rate cost of CMA was GBP 405 for a 1-Mb BAC array. However the cost of CMA is rapidly decreasing and higher-resolution CMA platforms are
now being costed at GBP 350 or lower. A threshold analysis was therefore performed to see what would happen to the cost-effectiveness of the tests as CMA decreased in price. When CMA costs GBP 360 or lower and the WTP for an additional diagnosis was GBP 9,768 or greater, then CMA is cost-effective over karyotyping (fig. 2). When VOUS are treated as false positive results, then CMA has to cost GBP 302 or less and the WTP for a positive diagnosis has to be GBP 9,185 or less for CMA to be cost-effective over karyotyping (fig. 3).
Prenatal CMA: An Economic Evaluation
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
53
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Strategy 1: based on 1-Mb BAC array data from own cohort. VOUS treated as true positive results. Amniocentesis. Strategy 2: based on 1-Mb BAC array data from own cohort. VOUS treated as true positive results. Chorionic villus sampling. Strategy 3: based on 1-Mb BAC array data from own cohort. VOUS treated as false positive results. Amniocentesis. Strategy 4: based on systematic review data (indication abnormal scan). VOUS treated as true positive results. Amniocentesis. Strategy 5: based on systematic review data (indication abnormal scan). VOUS treated as false positive results. Amniocentesis. Incremental cost and incremental effectiveness are relative to the previous strategy. Effectiveness measure is cost per case of chromosomal anomaly detected. ICER GBP = Cost per additional case detected.
QFPCR + Karyo + CMA
Karyo + CMA
90,000
WTP per positive result (GBP)
80,000 70,000
CMA
60,000 50,000 40,000 30,000 20,000
QFPCR then Karyo then FISH
10,000 0 100 (45)
Fig. 2. Threshold analysis shows the effect
of decreasing the cost of fetal CMA and parental CMA (in parentheses) and the effect on the WTP for a positive result. In this analysis VOUS are treated as true positives.
Karyo 150 (95)
200 (145)
250 (195)
QFPCR 300 (245)
350 (295)
400 (345)
450 (395)
Color version available online
Cost of parental (fetal) CMA (GBP)
30,000 Karyo + CMA
QFPCR + Karyo + FISH DiGeorge
20,000 CMA 15,000
10,000
Karyo
5,000
Fig. 3. Threshold analysis shows the effect
of decreasing the cost of fetal CMA and parental CMA (in parentheses) and the effect on the WTP for detection of a chromosomal anomaly. In this analysis VOUS are treated as false positives.
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
0 100 (45)
QFPCR 150 (95)
200 (145)
250 (195)
300 (245)
350 (295)
400 (345)
Cost of fetal (parental) (GBP)
Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
WTP per positive result (GBP)
25,000
54
Color version available online
100,000
Color version available online
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
If the WTP is below GBP 9,768 and the CMA cost stayed above GBP 360, then conventional karyotyping is the test of choice. If the WTP increases but the cost of CMA stays at GBP 360 or lower, then QFPCR followed by karyotyping (if QFPCR-negative) followed by FISH for DiGeorge syndrome (if karyotype-negative and cardiac anomaly) becomes the testing strategy of choice. Cost-Effectiveness Analysis Curves A CEAC was performed for the base rate cost of GBP 405 for CMA. Using data from the Birmingham 1-Mb BAC cohort there is an 80% probability that CMA is costeffective over karyotyping when the WTP for an additional diagnosis is GBP 40,000 and a 50% probability that it is cost-effective when the WTP is GBP 13,000 (fig. 4). When the CMA cost was reduced to GBP 350, there was an 80% probability that CMA was cost-effective over karyotyping when the WTP for an additional diagnosis was GBP 25,000 (fig. 5). Model Populated Using Systematic Review Data of CMA Testing Performed for Structural Anomalies on Prenatal Ultrasound The base case deterministic results of the strategies based on the outcome of cost per additional case detected are presented in table 4. The strategies presented are undominated (if one strategy is more costly but less effective than another strategy then it is said to be dominated by another strategy and these have been removed from table 4). Prenatal CMA: An Economic Evaluation
20,000
40,000
60,000
80,000
100,000
WTP for every extra case detected (GBP)
They are at the base cost of GBP 405 for CMA and GBP 222 for karyotyping (assuming all invasive samples were amniocentesis) or GBP 265 for karyotyping (assuming all invasive samples were CVS). When systematic review data was used to populate the model, the results were very different. The ICER suggests that healthcare policy-makers would only have to spend an additional GBP 2,600 per additional case detected if they were to employ CMA testing instead of karyotyping (table 4). The probabilistic sensitivity analysis and CEAC curve shows that CMA has a 95% probability of being cost-effective when the WTP is GBP 2,830 (fig. 6). However, when CMA testing was performed for an abnormal scan result and VOUS were treated as false positives, due to the reduced effectiveness of CMA, the ICER was increased to GBP 25,400 per additional case detected by CMA over conventional karyotyping. This dramatic increase in the ICER is due to the large number of VOUS detected in the systematic review cohort. We believe that our preliminary model using our own ‘Birmingham 1-Mb BAC’ cohort [4] offers a more robust health economic analysis.
Discussion
Main Findings Using a 1-Mb BAC targeted array platform prenatally (at a cost of GBP 405) does not presently offer good value for money, as it would cost GBP 24,600 for Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
55
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
CMA is cost-effective over karyotyping as the WTP for every extra case detected by CMA increases. Here the case rate cost of CMA is GBP 405. There is an 80% probability that CMA is cost-effective over karyotyping when the WTP for a positive diagnosis is GBP 40,000 and a 50% probability that it is cost-effective when the WTP is GBP 13,000.
Probability that CMA is cost-effective over Karyotyping
Fig. 4. CEAC showing the probability that
1.0
CMA is cost-effective over karyotyping as the WTP for every extra case detected by CMA increases. Here the base rate cost of CMA is GBP 405. This curve is populated by data taken from systematic review analysis when the indication for chromosomal testing if an abnormal fetal USS finding. CMA has a 95% probability of being costeffective over karyotyping when the WTP is GBP 2,830.
Color version available online
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
40,000
60,000
80,000
100,000
WTP for every extra case detected (GBP)
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
CMA to detect an additional case over using the combination of QFPCR followed by karyotype (if QFPCRnegative) followed by FISH for DiGeorge syndrome (if karyotype-negative and the fetal anomaly is cardiac). The reason for this is the higher price of this array 56
20,000
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
500
1,000
1,500
2,000
2,500
3,000
WTP for every extra case detected (GBP)
platform, its targeted nature and therefore lower detection of pathogenic findings. It offers less value for money when comparing CMA alone over karyotyping alone, costing an extra GBP 33,000 per additional case detected. Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Fig. 6. CEAC showing the probability that
0.9
Color version available online
CMA is cost-effective over karyotyping as the WTP for every extra case detected by CMA increases. Here the base rate cost of CMA is GBP 350. There is an 80% probability that CMA was cost-effective over karyotyping when the WTP for a positive diagnosis was GBP 25,000.
Probability of CMA being cost-effective over Karyotyping
Fig. 5. CEAC showing the probability that
Probability that CMA is cost-effective over Karyotyping
1.0
Strengths and Limitations The model’s strength was that only one array platform was used and therefore there was no heterogeneity in the data that would have existed had various arrays been used at differing resolutions. It was primary data taken from a complete dataset so we were able to perform complex modelling using various testing strategies. This was in contrast to the sub-analysis model using systematically reviewed data that used heterogeneous CMA platforms and lacked the information to provide more detailed modelling of tests other than CMA and karyotype. There were limitations to our study. Firstly, that the outcome was measured as effectiveness per additional case detected. In economic evaluation this is seen as an intermediate outcome because the final pathway that is followed based on the detection of a fetal chromosomal abnormality is not explored in terms of the additional costs and effects that are likely to be incurred. Furthermore, the perspective of the analysis is limited to that of the healthcare sector and a societal perspective has not been considered. Clearly the societal cost of a false negative result could be considerable: Doran et al. [22] recently reported the costs of caring for a child with intellectual disability as ranging from AUD 36,681 to 58,129 every 6 months. However, the effect of a false positive (a VOUS that later turns out to be benign) is also not known: a false positive might lead to the pregnancy being terminated which may also impose a loss to society, through parental guilt or foregone productivity, which cannot be quantified. Thus the overall balance of these two potentially opPrenatal CMA: An Economic Evaluation
posing effects on the overall cost-effectiveness cannot be predicted. Furthermore, no numerical figure could be placed on the benefit that women and their partners would state for a positive result, particularly if that positive result had some uncertainty attached (such as a VOUS). Quality adjusted life years (QALYs) are not appropriate as the resulting conditions are heterogeneous and there is no ‘treatment’ for a positive result at present, the only action that can be taken is to terminate the pregnancy. So whilst there are limitations of adopting both an intermediate outcome and an NHS perspective, a wider, societal approach to the analysis, that includes the impact of the pathway taken as a result of the test, is not feasible at the current time and is beyond the scope of the current study. The authors acknowledge that the 1-Mb BAC array is more expensive than other prenatal arrays used. For this reason the analysis includes reduction in the cost of the CMA test which is also included within the sensitivity analysis. The targeted nature and conservative resolution of the array (and hence its decreased detection rate) is also acknowledged, but this platform was chosen in the knowledge that the rate of VOUS would also be significantly decreased. This would therefore not create potentially unnecessary uncertainty/worry for women and their partners. Interpretation Prenatal targeted BAC CMA platforms may not presently offer good value for money in terms of cost per additional case detected. However, increasing resolution and decreasing costs of CMA mean that it is likely to become a cost-effective option in the future. Before this can be ascertained, VOUS must be awarded an effectiveness score. Current qualitative data seems to suggest that parents have mixed feelings about their worth, depending upon whether they are given pre- or postnatally and that prenatally they may be given a limited value, or even a negative value [23]. If awarded a negative value it may be better to not report them to parents at all. If VOUS continue to be reported to patients they should continue to be modelled as we have done (as true positive results) but with an adjusted effectiveness score. If an international decision is made to not report VOUS, they could be re-modelled as true negative results but with associated parental follow-up costs. Modelling VOUS as false positives (as we have done in our sensitivity analysis) may be inaccurate as many will turn out to be pathological findings. Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
57
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
However, the threshold analysis shows that if the cost of CMA can be reduced to GBP 360 then the WTP would have to be only GBP 9,768 (for every additional case detected) to make CMA cost-effective over conventional karyotyping. Even when VOUS were treated as false positives the cost of CMA would have to be GBP 302 and the WTP for every additional result GBP 9,185 for CMA to be cost-effective over conventional karyotyping. This is within the realms of possibility as the current cost of a higher-resolution 60K array in the West Midlands Genetics Laboratory is GBP 350. It should be noted that this analysis refers to the costeffectiveness of prenatal CMA when an anomaly is found on USS only. Prenatal CMA can also be employed due to other reasons such as a ‘high-risk’ result on Down’s screening or advanced maternal age. As prenatal CMA has a lower detection rate in these circumstances [2] it is also likely to be less cost-effective.
Acknowledgement
Conclusion
The use of a targeted prenatal BAC CMA when an anomaly is found on scan is currently not a cost-effective option when compared to other cytogenetic testing strategies. However, as CMA costs decrease and resolution (and hence detection rates) increase, it is likely to become the cost-effective option in the future. Future research should include developing an effectiveness score for VOUS.
Sarah Hillman was funded by the children’s charity SPARKS.
Disclosure Statement The authors have no conflicts of interest to disclose.
References
58
8 Bi W, Breman AM, Venable SF, Eng PA, Sahoo T, Lu XY, Patel A, Beaudet AL, Cheung SW, White LD: Rapid prenatal diagnosis using uncultured amniocytes and oligonucleotide array CGH. Prenat Diagn 2008;28:943–949. 9 Tyreman M, Abbott KM, Willatt LR, Nash R, Lees C, Whittaker J, Simonic I: High-resolution array analysis: diagnosing pregnancies with abnormal ultrasound findings. J Med Genet 2009;46:531–541. 10 Kleeman L, Bianchi DW, Shaffer LG, Rorem E, Cowan J, Craigo SD, Tighiouart H, WilkinsHaug LE: Use of array comparative genomic hybridization for prenatal diagnosis of fetuses with sonographic anomalies and normal metaphase karyotype. Prenat Diagn 2009; 29: 1213–1217. 11 Vialard F, Molina GD, Leroy B, Quarello E, Escalona A, Le SC, Serazin V, Roume J, Ville Y, de MP, Selva J: Array comparative genomic hybridization in prenatal diagnosis: another experience. Fetal Diagn Ther 2009;25:277– 284. 12 Faas BH, van der Burgt, I, Kooper AJ, Pfundt R, Hehir-Kwa JY, Smits AP, de Leeuw N: Identification of clinically significant, submicroscopic chromosome alterations and UPD in fetuses with ultrasound anomalies using genome-wide 250K SNP array analysis. J Med Genet 2010;47:586–594. 13 Evangelidou P, Sismani C, Ioannides M, Christodoulou C, Koumbaris G, Kallikas I, Georgiou I, Velissariou V, Patsalis PC: Clinical application of whole-genome array CGH during prenatal diagnosis: study of 25 selected pregnancies with abnormal ultrasound findings or apparently balanced structural aberrations. Mol Cytogenet 2010;3:24. 14 Valduga M, Philippe C, Bach SP, Thiebaugeorges O, Miton A, Beri M, Bonnet C, Nemos C, Foliguet B, Jonveaux P: A retrospective study by oligonucleotide array-CGH analysis in 50 fetuses with multiple malformations. Prenat Diagn 2010;30:333–341. 15 Fiorentino F, Caiazzo F, Napolitano S, Spizzichino L, Bono S, Sessa M, Nuccitelli A, Biricik A, Gordon A, Rizzo G, Baldi M: Introducing array comparative genomic hybridization into
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
16
17
18
19
20
21 22
23
routine prenatal diagnosis practice: a prospective study on over 1,000 consecutive clinical cases. Prenat Diagn 2011;31:1270–1282. Leung TY, Vogel I, Lau TK, Chong W, Hyett JA, Petersen OB, Choy KW: Identification of submicroscopic chromosomal aberrations in fetuses with increased nuchal translucency and apparently normal karyotype. Ultrasound Obstet Gynecol 2011;38:314–319. D’Amours G, Kibar Z, Mathonnet G, Fetni R, Tihy F, Desilets V, Nizard S, Michaud J, Lemyre E: Whole-genome array CGH identifies pathogenic copy number variations in fetuses with major malformations and a normal karyotype. Clin Genet 2011;81:128–141. Schmid M, Stary S, Blaicher W, Gollinger M, Husslein P, Streubel B: Prenatal genetic diagnosis using microarray analysis in fetuses with congenital heart defects. Prenat Diagn 2012;32:376–382. Srebniak MI, Boter M, Oudesluijs GO, Cohen-Overbeek T, Govaerts LC, Diderich KE, Oegema R, Knapen MF, van de Laar IM, Joosten M, Van OD, Galjaard RJ: Genomic SNP array as a gold standard for prenatal diagnosis of foetal ultrasound abnormalities. Mol Cytogenet 2012;5:14. Armengol L, Nevado J, Serra-Juhe C, Plaja A, Mediano C, Garcia-Santiago FA, Garcia-Aragones M, Villa O, Mansilla E, Preciado C, Fernandez L, Angeles MM, Garcia-Perez L, Lapunzina PD, Perez-Jurado LA: Clinical utility of chromosomal microarray analysis in invasive prenatal diagnosis. Hum Genet 2012;131: 513–523. Curtis L (ed): Unit Costs of Health and Social Care 2011. Canterbury, Personal Social Services Research Unit, 2011, pp 1–237. Doran CM, Einfeld SL, Madden RH, Otim M, Horstead SK, Ellis LA, Emerson E: How much does intellectual disability really cost? First estimates for Australia. J Intellect Dev Disabil 2012;37:42–49. Bernhardt BA, Soucier D, Hanson K, Savage MS, Jackson L, Wapner RJ: Women’s experiences receiving abnormal prenatal chromosomal microarray testing results. Genet Med 2013;15:139–145.
Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
1 Wapner RJ, Martin CL, Levy B, Ballif BC, Eng CM, Zachary JM, Savage M, Platt LD, Saltzman D, Grobman WA, Klugman S, Scholl T, Simpson JL, McCall K, Aggarwal VS, Bunke B, Nahum O, Patel A, Lamb AN, Thom EA, Beaudet AL, Ledbetter DH, Shaffer LG, Jackson L: Chromosomal microarray versus karyotyping for prenatal diagnosis. N Engl J Med 2012;367:2175–2184. 2 Hillman SC, Pretlove S, Coomarasamy A, McMullan DJ, Davison EV, Maher ER, Kilby MD: Additional information from array comparative genomic hybridization technology over conventional karyotyping in prenatal diagnosis: a systematic review and meta-analysis. Ultrasound Obstet Gynecol 2011;37:6–14. 3 Shaffer LG, Dabell MP, Fisher AJ, Coppinger J, Bandholz AM, Ellison JW, Ravnan JB, Torchia BS, Ballif BC, Rosenfeld JA: Experience with microarray-based comparative genomic hybridization for prenatal diagnosis in over 5,000 pregnancies. Prenat Diagn 2012; 32: 976–985. 4 Hillman SC, McMullan DJ, Hall G, Togneri FS, James N, Maher EJ, Meller CH, Williams D, Wapner RJ, Maher ER, Kilby MD: Prenatal chromosomal microarray use: a prospective cohort of fetuses and a systematic review and meta-analysis. Ultrasound Obstet Gynecol 2013;41:610–620. 5 Van den Veyver IB, Patel A, Shaw CA, Pursley AN, Kang SH, Simovich MJ, Ward PA, Darilek S, Johnson A, Neill SE, Bi W, White LD, Eng CM, Lupski JR, Cheung SW, Beaudet AL: Clinical use of array comparative genomic hybridization for prenatal diagnosis in 300 cases. Prenat Diagn 2009;29:29–39. 6 Lee CN, Lin SY, Lin CH, Shih JC, Lin TH, Su YN: Clinical utility of array comparative genomic hybridisation for prenatal diagnosis: a cohort study of 3,171 pregnancies. BJOG 2012;119:614–625. 7 Le Caignec C, Boceno M, Saugier-Veber P, Jacquemont S, Joubert M, David A, Frebourg T, Rival JM: Detection of genomic imbalances by array based comparative genomic hybridisation in fetuses with multiple malformations. J Med Genet 2005;42:121–128.
Received: July 31, 2013 Accepted after revision: December 27, 2013 Published online: June 13, 2014
BAC Chromosomal Microarray for Prenatal Detection of Chromosome Anomalies in Fetal Ultrasound Anomalies: An Economic Evaluation Sarah C. Hillman a, e Pelham M. Barton b Tracy E. Roberts b Eamonn R. Maher a Dominic M. McMullan c Mark D. Kilby a, d, e a
School of Clinical and Experimental Medicine, College of Medicine and Dentistry, b Health Economics Unit, School of Health and Populations Science, c West Midlands Genetics Laboratory and d Fetal Medicine Centre, Birmingham Women’s Foundation Trust, and e Centre for Women’s and Children’s Health, University of Birmingham, Birmingham, UK
Abstract Introduction: To determine the cost-effectiveness of prenatal chromosomal microarray (CMA) when performed for structural anomalies on fetal ultrasound scan over conventional techniques. Method: A decision tree was populated using data from a prospective cohort of women undergoing testing when a fetal ultrasound scan showed a structural abnormality. Nine strategies of testing were modeled including combinations of the tests: QFPCR, G-band karyotyping, CMA and FISH for DiGeorge (22q) microdeletion. Results: When CMA costs GBP 405 and using a 1-Mb BAC array it would cost GBP 24,600 for every additional case detected by CMA over a combination of QFPCR, followed by G-band karyotype, followed lastly by FISH (for DiGeorge syndrome). If CMA is performed instead of conventional karyotyping alone it costs GBP 33,000 for every additional case detected. However, if the cost of CMA is reduced to GBP 360 than when CMA is performed instead of conventional karyotyping alone it would cost GBP 9,768 for every additional case detected. Discussion: The use of a prenatal BAC CMA is not currently cost-effective when compared to other testing strategies.
© 2014 S. Karger AG, Basel 1015–3837/14/0361–0049$39.50/0 E-Mail [email protected] www.karger.com/fdt
However, as CMA costs decrease and resolution (and detection rates) increase it is likely to become the cost-effective option of the future. © 2014 S. Karger AG, Basel
Introduction
Prenatal analysis of fetal chromosomes has been available since the 1960s. Full G-band karyotyping has classically been used to detect chromosomal anomalies at a resolution of between 5 and 10 Mb. This technology is now being supplemented, and in some instances replaced, by chromosomal microarray (CMA) which is capable of examining chromosomes to a resolution of 1 kb, smaller than the average gene. CMA would be appropriate when there is a ‘high-risk’ result found on aneuploidy screening, there is a previous history of chromosomal anomalies, advanced maternal age or for parental anxiety. But its highest detection rate and arguably its most effective use in the prenatal setting is for detecting submicroscopic chromosome deletions and duplications when a congenital anomaly is detected on fetal ultrasound scan (USS) [1, 2]. The advantages of CMA lie in its ability to detect smaller potentially pathological chromosomal variants that are Prof. Mark D. Kilby School of Clinical and Experimental Medicine College of Medical and Dental Sciences University of Birmingham, Birmingham B15 2TT (UK) E-Mail m.d.kilby @ bham.ac.uk
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Key Words Chromosomal microarray · Fetal anomaly · Ultrasound · Economic evaluation
Table 1. Different strategies of chromosomal testing that were analysed within the model
Option
Test 1
Followed by test 2
Followed by test 3
1 2 3 4 5 6 7 8 9
QFPCR G-band karyotyping CMA QFPCR QFPCR G-band karyotyping QFPCR QFPCR QFPCR
– – – Then G-band karyotyping (if QFPCR-negative) Then CMA (if QFPCR-negative) And CMA (regardless of G-band karyotyping result) Then G-band karyotyping (if QFPCR-negative) Then FISH (if QFPCR-negative) Then G-band karyotyping (if QFPCR-negative)
– – – – – – And CMA (regardless of G-band karyotyping result) – Then FISH (if G-band karyotyping-negative)
Parental FU
Diseasepositive
pSens_Kar...
pfollowup1 No parental FU #
pPrevalence Karyo-negative
Karyotype only
Method
# Diseasenegative #
Karyo-positive
Parental FU
# Karyo-negative pSpec_karyo
Fig. 1. An example of an ‘arm’ of the decision tree for ‘karyotype
only’. pPrevalence = Prevalence of chromosomal anomalies; pSens = sensitivity of karyotyping; pSpec = specificity of karyotyping; pfollowup = probability that parental karyotyping would be required.
undetectable using standard cytogenetic analyses, to be customised and be amenable for high-volume throughput. Potential drawbacks of CMA are inability to detect balanced chromosomal rearrangements, triploidy and some instances of mosaicism. The biggest challenge presented by this new technology is the detection of chromosomal variants of unknown clinical significance (VOUS) and the potentially adverse effect reporting these variants may have on couples and their decision-making antenatally. The higher detection rate of prenatal chromosomal anomalies has been shown in multiples cohort studies [1– 3], however CMA does currently cost more and there is currently no evidence on whether its use in the prenatal 50
setting will provide value for money. This report will examine the cost-effectiveness of prenatal CMA when used because a structural anomaly has been detected on USS. It will compare CMA to standard conventional cytogenetic testing.
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
To assess the cost-effectiveness of the various tests or combination of tests, we carried out model-based economic analyses. Our analysis was conducted from the healthcare perspective and based on an outcome of cost per additional case detected. Our decision tree models were created using TreeAge software (software that is widely used in health economics to build and analyse decision trees) (fig. 1) (TreeAge Software, Inc., Williamstown, Mass., USA). The model was populated using data gained between December 2009 and April 2012 at Birmingham Women’s Hospital where we prospectively recruited women when a fetal USS showed a significant structural anomaly. Ethical approval for the present study was granted by the Staffordshire Research and Ethics Committee (Ref. No. 09/H1203/74). Women were counselled and offered fetal chromosome testing by invasive sampling. If accepted, women were also offered, counselled and consented to have CMA testing in addition to standard cytogenetic testing. 309 women were recruited and had cytogenetic testing performed [4]. In all cases they had QFPCR (quantitative fluorescence polymerase chain reaction), if this was positive for trisomy 13, 18 or 21 they did not go on to have CMA testing but for the model it was anticipated that had these cases undergone karyotype or CMA then the test would have detected aneuploidy. This cohort is described in detail in our study accepted for publication. Nine options for testing within the model were considered using combinations of the tests; QFPCR conventional G-band karyotyping, CMA, FISH (fluorescence in situ hybridization for 22q11.2 microdeletion syndrome on remaining fetal samples with cardiac abnormalities). The definition of these alternative options is presented in table 1.
Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Karyo-positive
Color version available online
FISH for 22q11.2 microdeletion syndrome on remaining fetal samples with cardiac abnormalities.
1-Mb BAC array Screening test
Probability parameter (sensitivity)
Probability parameter (specificity)
QFPCR Karyotyping CMA Karyotyping and CMA FISH
66/91 (73%) 82/91 (90%) 87/91 (96%) 91/91 (100%) 4/7 (57%)
218/218 (100%) 217/218 (99.5%) 215/218 (99%) 214/218 (98%) 34/34 (100%)
Tests accuracy data in the first model was taken using the data from our own prospective cohort. When calculating the sensitivity of CMA and karyotyping it was assumed that the overall result would be positive if either test was positive on its own. The sensitivity therefore increased as all individuals that tested positive on either test were potentially positive cases. The specificity therefore falls as there are more false positives and it is lower than the specificity of each test alone.
Some assumptions were required in order to develop a workable model: (1) In cytogenetic testing results may be positive or negative but (more so in CMA testing) it is possible to have a result or VOUS. This result shows a chromosomal variance where the evidence is lacking to link it to the phenotype with certainty but evidence does not exist to call it normal/benign with certainty either. Future research could show this to be a variant causal of a pathological phenotype or a normal but uncommon variant within the population. In these models due to the possible pathological nature of these results, we treated them as true positive results and allowed for the costs of parental follow-up to be included. However, to allow for this the model was re-run treating the VOUS as a false positive finding (there was only one VOUS in the cohort detected by CMA), the results of this sensitivity analysis are included. (2) No prior testing had been performed on the sample acquired from the fetus. (3) Triploidies (i.e. 69,XXX/69,XXY) would require no parental follow-up. (4) That a true positive or true negative result gave an effectiveness score of 1 and a false positive or false negative gave an effectiveness score of 0. However, for some patients if there was any uncertainty associated with the results they may have ranked their efficacy at 5 cases where microarray testing was performed prenatally in addition to karyotyping. Data for the model were extracted only when testing had been performed because a structural anomaly had been found on fetal USS. From the 21 primary studies, 17 had this information [1, 3, 7–20]. In total, 4,276 results were included. Five testing options were considered: karyotyping alone, CMA alone, karyotyping then CMA (if the karyotyping result is negative), CMA then karyotyping (if the CMA result is negative), karyotyping and CMA (both tests done regardless of the outcome of the other). The prevalence was calculated including VOUS as ‘test positive’ results. Cost and Resource Data In order to have cytogenetic testing a patient must be counselled by an appropriately trained healthcare professional (HCP) and have an invasive test performed (amniocentesis, chorionic villus sampling (CVS) or fetal blood sampling). The sample is then sent to an approved laboratory for testing and results are disseminated to HCPs. If negative/normal they are telephoned to the patient but if abnormal the patient re-attends for further counselling by HCPs and in some cases cytogenetic testing on parental samples is performed. Here we have only included costing from when the sample arrives at the laboratory and then subsequent follow-up costs. The initial cost of a consultation and the invasive test procedure were not included as it was assumed that these would be the same for all cases. Base costs are calculated as if the invasive sample taken was amniotic fluid by amniocentesis, however the models have been recalculated for CVS samples. The costs of resources utilised were those that were directly incurred by the NHS. The laboratory costs were calculated by the West Midlands Genetics Laboratory. Costs are reported in GBP. Costs of the cytogenetic tests included DNA extraction, cost of the base test and consumables, staffing costs, capital costs (such as the scanner service contract) and administration costs. Staffing costs for the laboratory work were calculated by timing staff performing the procedures. The average salary of those routinely doing the work was then used and the cost per hour of their time calculated. For the staffing costs for those performing the analysis work, as this can be highly variable depending on the complexity of the case, the average time was estimated. Capital costs such as the scanner service contract were calculated by taking the cost of the contract and dividing it by the number of samples processed over the time of the contract. Administration costs have been previously calculated by the laboratory per test and are included. Consumable costs were then recorded as those directly incurred by the NHS. Cost of follow-up included costs of seeing a consultant clinical geneticist and specialist midwife. This was based on a 30-min appointment. These costs were tak-
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
51
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Table 2. Test data accuracy taken from prospective cohort using a
Table 3. Unit costs and sources
Screening test
Unit cost, GBP
Source
QFPCR trisomy 13, 18 and 21 and sex chromosome aneuploidy Karyotyping amniocentesis Karyotyping chorionic villus sampling CMA (1-Mb BAC array) FISH for 22q11.2 DS Parental CMA Parental karyotype Follow-up specialist midwife per 30 min Follow-up clinical geneticist (consultant) per 30 min
129 223 265 405 186 350 382 48.50 81
West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory West Midlands Genetics Laboratory Unit costs of health and social care 2011 [21] Unit costs of health and social care 2011 [21]
en from unit costs of health and social care 2011 [21] (table 3). Costs used were those calculated by the laboratory for the years 2011/2012. It was assumed that follow-up with parental samples would be required for any abnormal cytogenetic test unless the results was an autosomal triploidy, trisomy or sex chromosome aneuploidy. If a trisomy 13, 18 or 21 was detected in the model by CMA or QFPCR, the cost of parental follow-up was not included (to determine if the trisomy was from a robersonian translocation) as this cost is required for future counselling regarding recurrence risk but would not change the outcome for the current fetus, as opposed to parental follow-up for VOUS which would alter the outcome for the current fetus when determining pathogenicity of the chromosomal anomaly.
by simultaneously selecting random values from each distribution. A Monte Carlo simulation repeated the process 10,000 times. This gave an indication of how variation in the test sensitivity and specificity lead to variation in the ICER. The decision uncertainty surrounding the use of CMA as a replacement for karyotyping was examined with the cost-effectiveness analysis curve (CEAC). This plots the probability that CMA will be cost-effective as compared to conventional G-band karyotyping at a given threshold of willingness to pay (WTP) that decision-makers may be willing to pay for a gain in effectiveness.
Sensitivity Analysis In addition to the base case analysis, we carried out deterministic sensitivity analysis in which the following changes were made: (1) In the CMA arm when a VOUS was found they were removed from being true positive and treated as false positive results. (2) The costings for laboratory karyotyping were changed from GBP 222 (karyotyping performed on DNA extracted from amniocytes) to GBP 265 (karyotyping performed on DNA extracted from a CVS). (3) Threshold sensitivity analyses were carried out to establish the critical value of CMA that would change the deterministic results in terms of ICERs and may affect the decision of policymakers. A probabilistic sensitivity analysis was also undertaken to determine the uncertainty of the model. A β distribution was assigned to each true positive, true negative, false positive and false negative parameter. Cost-effectiveness results were then calculated
52
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
Incremental Cost-Effectiveness Ratios Strategy 1 (presented in table 4) looks at the ICER when the 1-Mb BAC array is used and VOUS are treated as true positive results. Due to the targeted resolution of the 1-Mb BAC array and the finding of four submicroscopic deletions in the DiGeorge region (that would be detectable by FISH analysis as well as CMA) the ICER of CMA over a combination of the tests QFPCR followed by karyotype followed by FISH (for DiGeorge in known fetal cardiac anomalies) was GBP 24,600. This means that it would cost an extra GBP 24,600 for every additional case detected by CMA over the combination of tests QFPCR then karyotype (if QFPCR-negative) then FISH DiGeorge (if karyotype-negative and known cardiac anomaly). When the base case cost is changed to GBP 265 (assuming that the invasive test was CVS rather than amniocentesis, strategy 2; table 4) the ICER of CMA over QFPCR followed by karyotype followed by FISH for DiGeorge syndrome (in known fetal cardiac anomalies) decreased to GBP 14,200. The ICER of CMA alone over karyotype alone is GBP 33,300 (i.e. GBP 33,300 per additional case detected by CMA) for amniocentesis samples (strategy 1; table 4) or Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Outcome Measures In this analysis, consequences are reported in terms of the additional diagnoses provided by CMA compared to conventional G-band karyotyping. The results are reported in terms of the incremental cost-effectiveness ratio (ICER) based on the additional cost for every additional diagnosis. The ICER is universally accepted as the standard summary ratio that should be used for reporting the results of economic evaluations. The additional cost per additional unit of benefit can be applied to any evaluation as long as there is a comparison of costs and outcomes of at least two alternative pathways or strategies.
Results
Table 4. Incremental ICERs
Strategy
Cost, GBP
Incremental cost
Effectiveness
Incremental effectiveness
ICER, GBP
Strategy 1 QFPCR only QFPCR then FISH DiGeorge Karyotype only QFPCR then karyotype then FISH DiGeorge CMA only Karyotype and CMA
173 205 298 411 490 722
32 93 112 80 232
0.9159 0.9288 0.9644 0.9774 0.9806 0.9832
0.0129 0.0356 0.0129 0.0032 0.0026
2,500 2,600 8,700 24,600 90,500
Strategy 2 QFPCR only QFPCR then FISH DiGeorge Karyotype only QFPCR then karyotype then FISH DiGeorge CMA only Karyotype and CMA
173 205 341 444 490 756
32 136 103 46 266
0.9159 0.9288 0.9644 0.9774 0.9806 0.9832
0.0129 0.0356 0.0129 0.0032 0.0026
2,500 3,800 8,000 14,200 103,800
Strategy 3 QFPCR only QFPCR then FISH DiGeorge Karyotype only QFPCR then karyotype then FISH DiGeorge QFPCR then karyotype and CMA
173 205 298 411 719
32 93 112 309
0.9191 0.9320 0.9676 0.9806 0.9871
0.0129 0.0356 0.0129 0.0065
2,489 2,616 8,670 47,697
Strategy 4 Karyotype only CMA only CMA then karyo
255 469 655
214 185
0.9107 0.9925 0.9991
0.0819 0.0066
2,600 28,300
Strategy 5 Karyotype only CMA only CMA then karyo
255 469 655
214 185
0.9474 0.9558 0.9623
0.0084 0.0065
25,400 28,300
GBP 22,200 (i.e. GBP 22,200 per additional case detected by CMA) for CVS samples (strategy 2; table 4). When VOUS are treated as false positives (strategy 3; table 4), CMA alone is dominated by all other strategies. However, treating VOUS as false positives will underestimate the specificity of CMA as some VOUS will in time be determined to be benign but others will be pathogenic. The current base rate cost of CMA was GBP 405 for a 1-Mb BAC array. However the cost of CMA is rapidly decreasing and higher-resolution CMA platforms are
now being costed at GBP 350 or lower. A threshold analysis was therefore performed to see what would happen to the cost-effectiveness of the tests as CMA decreased in price. When CMA costs GBP 360 or lower and the WTP for an additional diagnosis was GBP 9,768 or greater, then CMA is cost-effective over karyotyping (fig. 2). When VOUS are treated as false positive results, then CMA has to cost GBP 302 or less and the WTP for a positive diagnosis has to be GBP 9,185 or less for CMA to be cost-effective over karyotyping (fig. 3).
Prenatal CMA: An Economic Evaluation
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
53
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Strategy 1: based on 1-Mb BAC array data from own cohort. VOUS treated as true positive results. Amniocentesis. Strategy 2: based on 1-Mb BAC array data from own cohort. VOUS treated as true positive results. Chorionic villus sampling. Strategy 3: based on 1-Mb BAC array data from own cohort. VOUS treated as false positive results. Amniocentesis. Strategy 4: based on systematic review data (indication abnormal scan). VOUS treated as true positive results. Amniocentesis. Strategy 5: based on systematic review data (indication abnormal scan). VOUS treated as false positive results. Amniocentesis. Incremental cost and incremental effectiveness are relative to the previous strategy. Effectiveness measure is cost per case of chromosomal anomaly detected. ICER GBP = Cost per additional case detected.
QFPCR + Karyo + CMA
Karyo + CMA
90,000
WTP per positive result (GBP)
80,000 70,000
CMA
60,000 50,000 40,000 30,000 20,000
QFPCR then Karyo then FISH
10,000 0 100 (45)
Fig. 2. Threshold analysis shows the effect
of decreasing the cost of fetal CMA and parental CMA (in parentheses) and the effect on the WTP for a positive result. In this analysis VOUS are treated as true positives.
Karyo 150 (95)
200 (145)
250 (195)
QFPCR 300 (245)
350 (295)
400 (345)
450 (395)
Color version available online
Cost of parental (fetal) CMA (GBP)
30,000 Karyo + CMA
QFPCR + Karyo + FISH DiGeorge
20,000 CMA 15,000
10,000
Karyo
5,000
Fig. 3. Threshold analysis shows the effect
of decreasing the cost of fetal CMA and parental CMA (in parentheses) and the effect on the WTP for detection of a chromosomal anomaly. In this analysis VOUS are treated as false positives.
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
0 100 (45)
QFPCR 150 (95)
200 (145)
250 (195)
300 (245)
350 (295)
400 (345)
Cost of fetal (parental) (GBP)
Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
WTP per positive result (GBP)
25,000
54
Color version available online
100,000
Color version available online
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
If the WTP is below GBP 9,768 and the CMA cost stayed above GBP 360, then conventional karyotyping is the test of choice. If the WTP increases but the cost of CMA stays at GBP 360 or lower, then QFPCR followed by karyotyping (if QFPCR-negative) followed by FISH for DiGeorge syndrome (if karyotype-negative and cardiac anomaly) becomes the testing strategy of choice. Cost-Effectiveness Analysis Curves A CEAC was performed for the base rate cost of GBP 405 for CMA. Using data from the Birmingham 1-Mb BAC cohort there is an 80% probability that CMA is costeffective over karyotyping when the WTP for an additional diagnosis is GBP 40,000 and a 50% probability that it is cost-effective when the WTP is GBP 13,000 (fig. 4). When the CMA cost was reduced to GBP 350, there was an 80% probability that CMA was cost-effective over karyotyping when the WTP for an additional diagnosis was GBP 25,000 (fig. 5). Model Populated Using Systematic Review Data of CMA Testing Performed for Structural Anomalies on Prenatal Ultrasound The base case deterministic results of the strategies based on the outcome of cost per additional case detected are presented in table 4. The strategies presented are undominated (if one strategy is more costly but less effective than another strategy then it is said to be dominated by another strategy and these have been removed from table 4). Prenatal CMA: An Economic Evaluation
20,000
40,000
60,000
80,000
100,000
WTP for every extra case detected (GBP)
They are at the base cost of GBP 405 for CMA and GBP 222 for karyotyping (assuming all invasive samples were amniocentesis) or GBP 265 for karyotyping (assuming all invasive samples were CVS). When systematic review data was used to populate the model, the results were very different. The ICER suggests that healthcare policy-makers would only have to spend an additional GBP 2,600 per additional case detected if they were to employ CMA testing instead of karyotyping (table 4). The probabilistic sensitivity analysis and CEAC curve shows that CMA has a 95% probability of being cost-effective when the WTP is GBP 2,830 (fig. 6). However, when CMA testing was performed for an abnormal scan result and VOUS were treated as false positives, due to the reduced effectiveness of CMA, the ICER was increased to GBP 25,400 per additional case detected by CMA over conventional karyotyping. This dramatic increase in the ICER is due to the large number of VOUS detected in the systematic review cohort. We believe that our preliminary model using our own ‘Birmingham 1-Mb BAC’ cohort [4] offers a more robust health economic analysis.
Discussion
Main Findings Using a 1-Mb BAC targeted array platform prenatally (at a cost of GBP 405) does not presently offer good value for money, as it would cost GBP 24,600 for Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
55
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
CMA is cost-effective over karyotyping as the WTP for every extra case detected by CMA increases. Here the case rate cost of CMA is GBP 405. There is an 80% probability that CMA is cost-effective over karyotyping when the WTP for a positive diagnosis is GBP 40,000 and a 50% probability that it is cost-effective when the WTP is GBP 13,000.
Probability that CMA is cost-effective over Karyotyping
Fig. 4. CEAC showing the probability that
1.0
CMA is cost-effective over karyotyping as the WTP for every extra case detected by CMA increases. Here the base rate cost of CMA is GBP 405. This curve is populated by data taken from systematic review analysis when the indication for chromosomal testing if an abnormal fetal USS finding. CMA has a 95% probability of being costeffective over karyotyping when the WTP is GBP 2,830.
Color version available online
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
40,000
60,000
80,000
100,000
WTP for every extra case detected (GBP)
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
CMA to detect an additional case over using the combination of QFPCR followed by karyotype (if QFPCRnegative) followed by FISH for DiGeorge syndrome (if karyotype-negative and the fetal anomaly is cardiac). The reason for this is the higher price of this array 56
20,000
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
500
1,000
1,500
2,000
2,500
3,000
WTP for every extra case detected (GBP)
platform, its targeted nature and therefore lower detection of pathogenic findings. It offers less value for money when comparing CMA alone over karyotyping alone, costing an extra GBP 33,000 per additional case detected. Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
Fig. 6. CEAC showing the probability that
0.9
Color version available online
CMA is cost-effective over karyotyping as the WTP for every extra case detected by CMA increases. Here the base rate cost of CMA is GBP 350. There is an 80% probability that CMA was cost-effective over karyotyping when the WTP for a positive diagnosis was GBP 25,000.
Probability of CMA being cost-effective over Karyotyping
Fig. 5. CEAC showing the probability that
Probability that CMA is cost-effective over Karyotyping
1.0
Strengths and Limitations The model’s strength was that only one array platform was used and therefore there was no heterogeneity in the data that would have existed had various arrays been used at differing resolutions. It was primary data taken from a complete dataset so we were able to perform complex modelling using various testing strategies. This was in contrast to the sub-analysis model using systematically reviewed data that used heterogeneous CMA platforms and lacked the information to provide more detailed modelling of tests other than CMA and karyotype. There were limitations to our study. Firstly, that the outcome was measured as effectiveness per additional case detected. In economic evaluation this is seen as an intermediate outcome because the final pathway that is followed based on the detection of a fetal chromosomal abnormality is not explored in terms of the additional costs and effects that are likely to be incurred. Furthermore, the perspective of the analysis is limited to that of the healthcare sector and a societal perspective has not been considered. Clearly the societal cost of a false negative result could be considerable: Doran et al. [22] recently reported the costs of caring for a child with intellectual disability as ranging from AUD 36,681 to 58,129 every 6 months. However, the effect of a false positive (a VOUS that later turns out to be benign) is also not known: a false positive might lead to the pregnancy being terminated which may also impose a loss to society, through parental guilt or foregone productivity, which cannot be quantified. Thus the overall balance of these two potentially opPrenatal CMA: An Economic Evaluation
posing effects on the overall cost-effectiveness cannot be predicted. Furthermore, no numerical figure could be placed on the benefit that women and their partners would state for a positive result, particularly if that positive result had some uncertainty attached (such as a VOUS). Quality adjusted life years (QALYs) are not appropriate as the resulting conditions are heterogeneous and there is no ‘treatment’ for a positive result at present, the only action that can be taken is to terminate the pregnancy. So whilst there are limitations of adopting both an intermediate outcome and an NHS perspective, a wider, societal approach to the analysis, that includes the impact of the pathway taken as a result of the test, is not feasible at the current time and is beyond the scope of the current study. The authors acknowledge that the 1-Mb BAC array is more expensive than other prenatal arrays used. For this reason the analysis includes reduction in the cost of the CMA test which is also included within the sensitivity analysis. The targeted nature and conservative resolution of the array (and hence its decreased detection rate) is also acknowledged, but this platform was chosen in the knowledge that the rate of VOUS would also be significantly decreased. This would therefore not create potentially unnecessary uncertainty/worry for women and their partners. Interpretation Prenatal targeted BAC CMA platforms may not presently offer good value for money in terms of cost per additional case detected. However, increasing resolution and decreasing costs of CMA mean that it is likely to become a cost-effective option in the future. Before this can be ascertained, VOUS must be awarded an effectiveness score. Current qualitative data seems to suggest that parents have mixed feelings about their worth, depending upon whether they are given pre- or postnatally and that prenatally they may be given a limited value, or even a negative value [23]. If awarded a negative value it may be better to not report them to parents at all. If VOUS continue to be reported to patients they should continue to be modelled as we have done (as true positive results) but with an adjusted effectiveness score. If an international decision is made to not report VOUS, they could be re-modelled as true negative results but with associated parental follow-up costs. Modelling VOUS as false positives (as we have done in our sensitivity analysis) may be inaccurate as many will turn out to be pathological findings. Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
57
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
However, the threshold analysis shows that if the cost of CMA can be reduced to GBP 360 then the WTP would have to be only GBP 9,768 (for every additional case detected) to make CMA cost-effective over conventional karyotyping. Even when VOUS were treated as false positives the cost of CMA would have to be GBP 302 and the WTP for every additional result GBP 9,185 for CMA to be cost-effective over conventional karyotyping. This is within the realms of possibility as the current cost of a higher-resolution 60K array in the West Midlands Genetics Laboratory is GBP 350. It should be noted that this analysis refers to the costeffectiveness of prenatal CMA when an anomaly is found on USS only. Prenatal CMA can also be employed due to other reasons such as a ‘high-risk’ result on Down’s screening or advanced maternal age. As prenatal CMA has a lower detection rate in these circumstances [2] it is also likely to be less cost-effective.
Acknowledgement
Conclusion
The use of a targeted prenatal BAC CMA when an anomaly is found on scan is currently not a cost-effective option when compared to other cytogenetic testing strategies. However, as CMA costs decrease and resolution (and hence detection rates) increase, it is likely to become the cost-effective option in the future. Future research should include developing an effectiveness score for VOUS.
Sarah Hillman was funded by the children’s charity SPARKS.
Disclosure Statement The authors have no conflicts of interest to disclose.
References
58
8 Bi W, Breman AM, Venable SF, Eng PA, Sahoo T, Lu XY, Patel A, Beaudet AL, Cheung SW, White LD: Rapid prenatal diagnosis using uncultured amniocytes and oligonucleotide array CGH. Prenat Diagn 2008;28:943–949. 9 Tyreman M, Abbott KM, Willatt LR, Nash R, Lees C, Whittaker J, Simonic I: High-resolution array analysis: diagnosing pregnancies with abnormal ultrasound findings. J Med Genet 2009;46:531–541. 10 Kleeman L, Bianchi DW, Shaffer LG, Rorem E, Cowan J, Craigo SD, Tighiouart H, WilkinsHaug LE: Use of array comparative genomic hybridization for prenatal diagnosis of fetuses with sonographic anomalies and normal metaphase karyotype. Prenat Diagn 2009; 29: 1213–1217. 11 Vialard F, Molina GD, Leroy B, Quarello E, Escalona A, Le SC, Serazin V, Roume J, Ville Y, de MP, Selva J: Array comparative genomic hybridization in prenatal diagnosis: another experience. Fetal Diagn Ther 2009;25:277– 284. 12 Faas BH, van der Burgt, I, Kooper AJ, Pfundt R, Hehir-Kwa JY, Smits AP, de Leeuw N: Identification of clinically significant, submicroscopic chromosome alterations and UPD in fetuses with ultrasound anomalies using genome-wide 250K SNP array analysis. J Med Genet 2010;47:586–594. 13 Evangelidou P, Sismani C, Ioannides M, Christodoulou C, Koumbaris G, Kallikas I, Georgiou I, Velissariou V, Patsalis PC: Clinical application of whole-genome array CGH during prenatal diagnosis: study of 25 selected pregnancies with abnormal ultrasound findings or apparently balanced structural aberrations. Mol Cytogenet 2010;3:24. 14 Valduga M, Philippe C, Bach SP, Thiebaugeorges O, Miton A, Beri M, Bonnet C, Nemos C, Foliguet B, Jonveaux P: A retrospective study by oligonucleotide array-CGH analysis in 50 fetuses with multiple malformations. Prenat Diagn 2010;30:333–341. 15 Fiorentino F, Caiazzo F, Napolitano S, Spizzichino L, Bono S, Sessa M, Nuccitelli A, Biricik A, Gordon A, Rizzo G, Baldi M: Introducing array comparative genomic hybridization into
Fetal Diagn Ther 2014;36:49–58 DOI: 10.1159/000358387
16
17
18
19
20
21 22
23
routine prenatal diagnosis practice: a prospective study on over 1,000 consecutive clinical cases. Prenat Diagn 2011;31:1270–1282. Leung TY, Vogel I, Lau TK, Chong W, Hyett JA, Petersen OB, Choy KW: Identification of submicroscopic chromosomal aberrations in fetuses with increased nuchal translucency and apparently normal karyotype. Ultrasound Obstet Gynecol 2011;38:314–319. D’Amours G, Kibar Z, Mathonnet G, Fetni R, Tihy F, Desilets V, Nizard S, Michaud J, Lemyre E: Whole-genome array CGH identifies pathogenic copy number variations in fetuses with major malformations and a normal karyotype. Clin Genet 2011;81:128–141. Schmid M, Stary S, Blaicher W, Gollinger M, Husslein P, Streubel B: Prenatal genetic diagnosis using microarray analysis in fetuses with congenital heart defects. Prenat Diagn 2012;32:376–382. Srebniak MI, Boter M, Oudesluijs GO, Cohen-Overbeek T, Govaerts LC, Diderich KE, Oegema R, Knapen MF, van de Laar IM, Joosten M, Van OD, Galjaard RJ: Genomic SNP array as a gold standard for prenatal diagnosis of foetal ultrasound abnormalities. Mol Cytogenet 2012;5:14. Armengol L, Nevado J, Serra-Juhe C, Plaja A, Mediano C, Garcia-Santiago FA, Garcia-Aragones M, Villa O, Mansilla E, Preciado C, Fernandez L, Angeles MM, Garcia-Perez L, Lapunzina PD, Perez-Jurado LA: Clinical utility of chromosomal microarray analysis in invasive prenatal diagnosis. Hum Genet 2012;131: 513–523. Curtis L (ed): Unit Costs of Health and Social Care 2011. Canterbury, Personal Social Services Research Unit, 2011, pp 1–237. Doran CM, Einfeld SL, Madden RH, Otim M, Horstead SK, Ellis LA, Emerson E: How much does intellectual disability really cost? First estimates for Australia. J Intellect Dev Disabil 2012;37:42–49. Bernhardt BA, Soucier D, Hanson K, Savage MS, Jackson L, Wapner RJ: Women’s experiences receiving abnormal prenatal chromosomal microarray testing results. Genet Med 2013;15:139–145.
Hillman/Barton/Roberts/Maher/ McMullan/Kilby
Downloaded by: John Rylands Library 130.88.90.140 - 5/13/2015 6:09:26 PM
1 Wapner RJ, Martin CL, Levy B, Ballif BC, Eng CM, Zachary JM, Savage M, Platt LD, Saltzman D, Grobman WA, Klugman S, Scholl T, Simpson JL, McCall K, Aggarwal VS, Bunke B, Nahum O, Patel A, Lamb AN, Thom EA, Beaudet AL, Ledbetter DH, Shaffer LG, Jackson L: Chromosomal microarray versus karyotyping for prenatal diagnosis. N Engl J Med 2012;367:2175–2184. 2 Hillman SC, Pretlove S, Coomarasamy A, McMullan DJ, Davison EV, Maher ER, Kilby MD: Additional information from array comparative genomic hybridization technology over conventional karyotyping in prenatal diagnosis: a systematic review and meta-analysis. Ultrasound Obstet Gynecol 2011;37:6–14. 3 Shaffer LG, Dabell MP, Fisher AJ, Coppinger J, Bandholz AM, Ellison JW, Ravnan JB, Torchia BS, Ballif BC, Rosenfeld JA: Experience with microarray-based comparative genomic hybridization for prenatal diagnosis in over 5,000 pregnancies. Prenat Diagn 2012; 32: 976–985. 4 Hillman SC, McMullan DJ, Hall G, Togneri FS, James N, Maher EJ, Meller CH, Williams D, Wapner RJ, Maher ER, Kilby MD: Prenatal chromosomal microarray use: a prospective cohort of fetuses and a systematic review and meta-analysis. Ultrasound Obstet Gynecol 2013;41:610–620. 5 Van den Veyver IB, Patel A, Shaw CA, Pursley AN, Kang SH, Simovich MJ, Ward PA, Darilek S, Johnson A, Neill SE, Bi W, White LD, Eng CM, Lupski JR, Cheung SW, Beaudet AL: Clinical use of array comparative genomic hybridization for prenatal diagnosis in 300 cases. Prenat Diagn 2009;29:29–39. 6 Lee CN, Lin SY, Lin CH, Shih JC, Lin TH, Su YN: Clinical utility of array comparative genomic hybridisation for prenatal diagnosis: a cohort study of 3,171 pregnancies. BJOG 2012;119:614–625. 7 Le Caignec C, Boceno M, Saugier-Veber P, Jacquemont S, Joubert M, David A, Frebourg T, Rival JM: Detection of genomic imbalances by array based comparative genomic hybridisation in fetuses with multiple malformations. J Med Genet 2005;42:121–128.
