Prevalence threshold of screening tests in obstetrics and gynecology

Prevalence threshold and positive predictive value of noninvasive prenatal testing

OBJECTIVE

Noninvasive prenatal testing (NIPT) has increased the number of conditions that can be screened. However, the prevalence of conditions assessed by NIPT has remained stable. The "prevalence threshold," a novel epidemiological concept, uses a test's sensitivity and specificity to determine the prevalence below which a test's positive predictive value declines most sharply relative to disease prevalence. In this article, we calculated the prevalence threshold for common conditions assessed through NIPT and compared the value with the actual prevalence of each condition to best ascertain the reliability of NIPT results.

METHODS

Six databases and PubMed were searched from January 2010 to March 2023 for sensitivity and specificity parameters of common conditions tested through NIPT. Using an equation previously derived by the authors of the current paper, the prevalence threshold for each condition was calculated. The theoretical number of test iterations required to reach the prevalence threshold was also reported.

RESULTS

None of the conditions tested through the NIPT had a prevalence rate that met or exceeded the calculated prevalence threshold. Trisomy 21 had the greatest concordance between the prevalence rate and the prevalence threshold. In contrast, Angelman, Cri-du-chat, and Prader-Willi syndromes had the most significant discordance. Apart from trisomy 21 and XXY, all remaining conditions required more than one test iteration to reach their respective prevalence threshold.

CONCLUSION

We conclude that at the current prevalence levels, the positive predictive value of NIPT remains low, with the prevalence of disease levels significantly lower than the prevalence threshold for each condition tested.

A statistical comparison between Matthews correlation coefficient (MCC), prevalence threshold, and Fowlkes-Mallows index

Even if assessing binary classifications is a common task in scientific research, no consensus on a single statistic summarizing the confusion matrix has been reached so far. In recent studies, we demonstrated the advantages of the Matthews correlation coefficient (MCC) over other popular rates such as cross-entropy error, F1 score, accuracy, balanced accuracy, bookmaker informedness, diagnostic odds ratio, Brier score, and Cohen's kappa. In this study, we compared the MCC to other two statistics: prevalence threshold (PT), frequently used in obstetrics and gynecology, and Fowlkes-Mallows index, a metric employed in fuzzy logic and drug discovery. Through the investigation of the mutual relations among three metrics and the study of some relevant use cases, we show that, when positive data elements and negative data elements have the same importance, the Matthews correlation coefficient can be more informative than its two competitors, even this time.

False-positives and false-negatives in non-invasive prenatal testing (NIPT): what can we learn from a meta-analyses on > 750,000 tests?

Background

Non-invasive prenatal testing (NIPT) has had an incomparable triumph in prenatal diagnostics in the last decade. Over 1400 research articles have been published, predominantly praising the advantages of this test.

Methods

The present study identified among the 1400 papers 24 original and one review paper, which were suited to re-evaluate the efficacy of > 750,000 published NIPT-results. Special attention was given to false-positive and false-negative result-rates. Those were discussed under different aspects—mainly from a patient-perspective.

Results

A 27: 1 rate of false-positive compared to false-negative NIPT results was found. Besides, according to all reported, real-positive, chromosomally aberrant NIPT cases, 90% of those would have been aborted spontaneously before birth. These findings are here discussed under aspects like (i) How efficient is NIPT compared to first trimester screening? (ii) What are the differences in expectations towards NIPT from specialists and the public? and (iii) There should also be children born suffering from not by NIPT tested chromosomal aberrations; why are those never reported in all available NIPT studies?

Conclusions

Even though much research has been published on NIPT, unbiased figures concerning NIPT and first trimester screening efficacy are yet not available. While false positive rates of different NIPT tests maybe halfway accurate, reported false-negative rates are most likely too low. The latter is as NIPT-cases with negative results for tested conditions are yet not in detail followed up for cases with other genetic or teratogenic caused disorders. This promotes an image in public, that NIPT is suited to replace all invasive tests, and also to solve the problem of inborn errors in humans, if not now then in near future. Overall, it is worth discussing the usefulness of NIPT in practical clinical application. Particularly, asking for unbiased figures concerning the efficacy of first trimester-screening compared to NIPT, and for really comprehensive data on false-positive and false-negative NIPT results.

Time to reconsider: Have the 2015 FIGO and 2017 Swedish intrapartum cardiotocogram classifications led us from Charybdis to Scylla?

In 2015, FIGO revised the 1987 intrapartum cardiotocography (CTG) classification (FIGO1987). A less radical FIGO2015 version was introduced in Sweden 2017 (SWE2017). Now, post hoc simulation studies show that FIGO2015 and SWE2017 are less reliable than (a modified) FIGO1987. FIGO2015 shows significantly better interobserver agreement for normal CTG traces than FIGO1987, but significantly worse for pathological traces. Agreements between templates are moderate to good, but different classifications of mainly variable decelerations and tachycardia cause significant heterogeneities. FIGO2015 shows insufficient sensitivity to identify fetal acidemia compared with FIGO1987. In connection with fetal electrocardiogram ST analysis, one study showed no template was superior in identifying fetal acidemia, but in a series of only academia, FIGO1987 had significantly higher sensitivity than FIGO2015 (73% vs. 43%) and set of an alarm for fetal acidemia considerably earlier. With SWE2017, operative interventions declined significantly in Sweden but several adverse neonatal outcomes increased significantly. It remains to investigate the development with FIGO2015.