Didem DERİCİ YILDIRIM, Bahar TAŞDELEN, Filiz ÇAYAN, Özlem İZCİ AY, Mustafa Ertan AY

Kromozomal bozuklukların referans test ile doğrulanamadığı durumda iki aşamalı Bayes yöntemiyle performans ölçülerinin değerlendirilmesi

Amaç: Kromozomal anormallik prevalansını ve diagnostik testlerin doğruluğunu tahmin etmek kromozomal bozukluklar altın Standard test ile doğrulanamadığı durumda ciddi bir problemdir. Bu çalışmanın amacı tanı testlerinin performasını ve kromozomal anomali prevalansını bayes analizi ile tahmin etmektir. Gereç ve Yöntem: Ağustos 2014 ve Ağustos 2015 tarihleri arasında Mersin Üniversitesi Hastanesi Kadın Hastalıkları ve Doğum polikliniğine başvurmuş gebeler çalışmaya dahil edilmiştir. Başvuran gebeler içerisinden, birinci trimester dönemde ultrasonografi ve ikili tarama test sonucu olan 1759 gebeye ait kayıtlardan yararlanılmıştır. İki Aşamalı Bayes yöntemi kullanılmıştır. İlk aşamada, bireyler hasta ve sağlıklı olarak sınıflandırılmaktadır. İkinci aşamada ise ilk aşamada hasta ya da sağlıklı olduğuna karar verilen bireylere altın standart test uygulanarak durumları doğrulanmaktadır. Bulgular: Önsel dağılım bilgisi olduğu durumda, birlikte duyarlılık ve seçicilik değerleri %77 ve %99 olarak tahmin edilmiştir. Önsel dağılım bilgisi olmadığı durumda ise, birlikte duyarlılık ve seçicilik değerleri %50 ve %97 olarak tahmin edilmiştir. Sapma Bilgi Ölçütü (DIC) ve Monte Carlo Hata (MC) değerleri birlikte incelendiğinde, önsel bilgi olduğu durumda modele uyum daha iyidir. Sonuç: Doğrulanamama problemi ile gebelikte uygulanan rutin tarama testleri için sıklıkla karşılaşılmaktadır. Tarama testlerinin doğruluğu bu yöntemden yararlanılarak daha doğru şekilde tahmin edilmektedir.

Anahtar Kelimeler:

Kromozomal bozukluklar, iki aşamalı bayes yöntemi, doğrulama problemi

Evaluation of diagnostic performance measures with two-stage Bayesian method when chromosomal disorders can not be verified with gold standard test

Purpose: Predicting chromosomal abnormality prevalance and accuracy of diagnostic tests represents a considerable problem when chromosomal disorders could not be verified with gold standard test. Aim of the study was to predict performance of screening tests and chromosomal abnormality prevalance within a bayesian analysis.Materials and Methods: Retrospective study at the Hospital of Mersin University, including pregnants who were admitted to gynecology and obstetrics clinic between August 2014 and August 2015. Within applicants, 1759 gestational’s records were utilized who had ultrasound examination and double screening test result in first trimester period. Two stage Bayesian approach was used. In the first stage, people were classified as patient and healthy. In the second stage, people suspected of having disease were examined by a gold standard test. Results: In the situation of known prior information, joint sensitivity and joint specificity were estimated as 77% and 99%. In the case of no prior information, joint sensitivity and joint specificity were estimated as 50% and 97%. Considering Deviation Information Criterion and Monte Carlo Error values together, model under prior information was better than the other. Conclusion: Verification problem will be frequently encountered for screening tests in pregnancy. Performance measures of the tests may be calculated more accurately by utilizing this method.

Keywords:

Chromosomal abnormalities, two-stage bayesian method, verification problem.,

PDF

___

1. Antonarakis SE, Lyle R, Chrast R, Scott HS. Differential gene expression studies to explore the molecular pathophysiology of Down syndrome. Brain Res Rev. 2001;36:265-74.
2. Gersen SL, Keagle MB. The Principles of Clinical Cytogenetics. 2nd ed. USA, Humana Press, 2005.
3. Gardner McKinlay RJ, Sutherland GR. Chromosome Abnormalities and Genetic Counseling. 3th ed. England, Oxford University Press. 2004.
4. Kuskucu AC. Fetal kromozom anomalisi tarama testleri. Jjinekoloji Obstetri Pediyatri ve Pediyatrik Cerrahi Dergisi. 2010;2:55-60.
5. Khalil A, Pandya P. Screening for Down syndrome. J Obstet Gynaecol India. 2006;56:205-11.
6. Baer RJ, Currier RJ, Norton ME, Flessel MC, Goldman S, Towner D et al. Outcomes of pregnancies with more than one positive prenatal screening result in the first and second trimester. Prenat Diagn. 2015;35:1223-31.
7. Maxim LD, Niebo R, Utell MJ. Screening test:a review with examples. Inhal Toxicol. 2014;26:811-28.
8. Dunson DB. Commentary: practical advantages of Bayesian analysis of epidemiology data. Am J Epidemiol. 2001;153:1222-6.
9. Berkvens D, Speybroeck N, Praet N, Adel A, Lesaffre E. Estimating disease prevalance in a Bayesian framework using probabilistic constraints. Epidemiology. 2006;17:145-53.
10. Dendukuri N, Joseph L. Bayesian approaches to modeling the conditional dependence between multiple diagnostic tests. Biometrics. 2001;57:158-67.
11. Geurden T, Claerebout E, Vercruysse J, Berkvens D. Estimation of diagnostic test characteristics and prevalance of Giardia duodenalis in dairy calves in Belgium using a Bayesian approach. Int J Parasitol. 2004;34:1121-7.
12. Geurden T, Berkvens D, Geldhof P, Vercruysse J, Claerebout E. A Bayesian approach for the evaluation of six diagnostic assays and estimation of cryptosporidium prevalance in dairy calves. Vet Res. 2006;37:671-82.
13. Liu P, Yang H, Qiang L, Xiao S, Shi ZX. Estimation of the sensitivty and specificity of assays for screening antibodies to HIV: a comparison between the frequentiest and Bayesian approaches. J Virol Methods. 2012;186: 89-93.
14. Liu J, Chen F, Yu H, Zeng P, Liu L. A two-stage bayesian method for estimating accuracy and disease prevalance for two dependent dichtomous screening tests when the status of individuals who are negative on both tests is unverified. BMC Med Res Methodol. 2014;14:1-11.
15. Li F, Chu H, Nie L. A two-stage estimation for screening studies using two diagnostic tests with binary disease status verified in test positives only. Stat Methods Med Res. 2015;24:635-56.
16. Joseph L, Gyorkos TW, Coupal L. Bayesian estimation of disease prevalance and the parameters of diagnostic tests in the absence of a gold standard. Am J Epidemiol. 1995;141:263-72.
17. Spiegelhalter D, Thromas A, Best N, Lunn D. 2003. WinBUGS User Manual (Version 1.4)
18. Toft N, Innocent GT, Gettinby G, Stuart WJR. Assessing the convergence of Markov Chain Monte Carlo methods: An example from evaluation of diagnostic tests in absence of a gold standard. Prev Vet Med. 2007;79:244-56.
19. Benn PA, Ying J, Beazoglou T, Egan JF. Estimates for the sensitivity and false-positive rates for second trimester serum screening for Down syndrome and trisomy 18 with adjustment for cross-identification and double-positive results. Prenat Diagn. 2001;21:46-51.
20. Abele H, Wagner P, Sonek J, Hoopmann M, Brucker S, Artunc-Ulkumen B et al. First trimester ultrasound screening for Down syndrome based on maternal age, fethal nuchal translucency and different combinations of the additional markers nasal bone, tricuspid and ductus venosus flow. Prenat Diagn. 2015;35:1182-6.
21. Xiao H, Yang YL, Zhang CY, Liao EJ, Zhao HR, Liao SX. Karyotype analysis with amniotic fluid in 12365 pregnant women with indications for genetic amniocentesis and strategies of prenatal diagnosis. J Obstet Gynaecol. 2016;36:293-6.
22. Walter SD. Estimation of test sensitivity and specificity when disease confirmation is limited to positive results. Epidemiology. 1999;10:67-72.
23. Chu H, Zhou Y, Cole SR, Ibrahim JG. On the estimation of disease prevalance by latent class models for screening studies using two screening tests with categorical disease status verified in test positives only. Stat Med. 2010;29:1206-18.
24. Pepe MS, Alonzo TA. Comparing disease screening tests when true disease status is ascertained only for screen positives. Biostatistics. 2001;2:249-60.
25. Chock C, Irwig L, Berry G, Glasziou P. Comparing dichotomous screening tests when individuals negative on both tests are not verified. J Clin Epidemiol. 1997;50:1211-7.
26. Berry G, Smith CL, Macaskill P, Irwig L. Analytic methods for comparing two dichotomous screening or diagnostic tests applied to two populations of differing disease prevalance when individuals negative on both tests are unverified. Stat Med. 2002;21:853-62.
27. Bohning D, Patılea V. A capture-recapture approach for screening using two diagnostic tests with availability of disease status for the test positivies only. J Am Stat Assoc. 2008;103:212-20.
28. Genc Y, Tuccar E. Effect of verification bias on sensitivity and specificity of diagnostic tests. Journal of Ankara Medical School. 2003;25:107-12.
29. Groot JA, Dendukuri N, Janssen KJM, Reitsma JB, Bossuyt PM, Moons KG. Adjusting for differential-verification bias in diagnostic- accuracy studies: a Bayesian approach. Epidemiology. 2011;22:234-41.
30. Alonzo TA, Brinton JT, Ringham BM, Glueck DH. Bias in estimating accuracy of a binary screening test with differential disease verification. Stat Med. 2011;30:1852-64.
31. Collins J, Huynh M. Estimation of diagnostic test accuracy without full verification: a review of latent class methods. Stat Med. 2014;33:4141-69.
32. Kosinski AS, Barnhart HX. Accounting for nonignorable verification bias in assessment of diagnostic tests. Biometrics. 2003;59:163-71.
33. Buzoianu M, Kadane JB. Adjusting for verification bias in diagnostic test evaluation: a Bayesian approach. Stat Med. 2008;27:2453-73.