RESUMO
Newborn screening for congenital hypothyroidism remains challenging decades after broad implementation worldwide. Testing protocols are not uniform in terms of targets (TSH and/or T4) and protocols (parallel vs. sequential testing; one or two specimen collection times), and specificity (with or without collection of a second specimen) is overall poor. The purpose of this retrospective study is to investigate the potential impact of multivariate pattern recognition software (CLIR) to improve the post-analytical interpretation of screening results. Seven programs contributed reference data (N = 1,970,536) and two sets of true (TP, N = 1369 combined) and false (FP, N = 15,201) positive cases for validation and verification purposes, respectively. Data were adjusted for age at collection, birth weight, and location using polynomial regression models of the fifth degree to create three-dimensional regression surfaces. Customized Single Condition Tools and Dual Scatter Plots were created using CLIR to optimize the differential diagnosis between TP and FP cases in the validation set. Verification testing correctly identified 446/454 (98%) of the TP cases, and could have prevented 1931/5447 (35%) of the FP cases, with variable impact among locations (range 4% to 50%). CLIR tools either as made here or preferably standardized to the recommended uniform screening panel could improve performance of newborn screening for congenital hypothyroidism.
RESUMO
PURPOSE: The purpose of this study was to compare performance metrics of postanalytical interpretive tools of the Region 4 Stork collaborative project to the actual outcome based on cutoff values for amino acids and acylcarnitines selected by the California newborn screening program. METHODS: This study was a retrospective review of the outcome of 176,186 subjects born in California between 1 January and 30 June 2012. Raw data were uploaded to the Region 4 Stork Web portal as .csv files to calculate tool scores for 48 conditions simultaneously using a previously unpublished functionality, the tool runner. Scores for individual target conditions were deemed informative when equal or greater to the value representing the first percentile rank of known true-positive cases (17,099 cases in total). RESULTS: In the study period, the actual false-positive rate and positive predictive value were 0.26 and 10%, respectively. Utilization of the Region 4 Stork tools, simple interpretation rules, and second-tier tests could have achieved a false-positive rate as low as 0.02% and a positive predictive value >50% by replacing the cutoff system with Region 4 Stork tools as the primary method for postanalytical interpretation. CONCLUSION: Region 4 Stork interpretive tools, second-tier tests, and other evidence-based interpretation rules could have reduced false-positive cases by up to 90% in California.
Assuntos
Triagem Neonatal/métodos , Espectrometria de Massas em Tandem/métodos , Aminoácidos/sangue , California , Carnitina/análogos & derivados , Carnitina/sangue , Biologia Computacional , Reações Falso-Positivas , Feminino , Humanos , Recém-Nascido , Internet , Masculino , Minnesota , Valor Preditivo dos Testes , Estudos Retrospectivos , SoftwareRESUMO
PURPOSE: To improve quality of newborn screening by tandem mass spectrometry with a novel approach made possible by the collaboration of 154 laboratories in 49 countries. METHODS: A database of 767,464 results from 12,721 cases affected with 60 conditions was used to build multivariate pattern recognition software that generates tools integrating multiple clinically significant results into a single score. This score is determined by the overlap between normal and disease ranges, penetration within the disease range, differences between conditions, and weighted correction factors. RESULTS: Ninety tools target either a single condition or the differential diagnosis between multiple conditions. Scores are expressed as the percentile rank among all cases with the same condition and are compared to interpretation guidelines. Retrospective evaluation of past cases suggests that these tools could have avoided at least half of 279 false-positive outcomes caused by carrier status for fatty-acid oxidation disorders and could have prevented 88% of known false-negative events. CONCLUSION: Application of this computational approach to raw data is independent from single analyte cutoff values. In Minnesota, the tools have been a major contributing factor to the sustained achievement of a false-positive rate below 0.1% and a positive predictive value above 60%.