NEW YORK – Researchers in China have adapted high-resolution melting analysis (HRM) to evaluate chromosomal abnormalities that are linked to recurrent pregnancy loss (RPL) in patients.
The team believes that clinicians could eventually use the method to improve the cost-effectiveness of RPL examinations, or as a potential alternative genetic test when other methods are unavailable or cost-prohibitive.
While karyotyping, the gold-standard for RPL evaluation, can detect numerical chromosomal abnormalities and structural changes larger than 5 megabase pairs, the process is often time-consuming, labor-intensive, and requires specialists, which makes it difficult to meet the clinical demand for RPL detection. In comparison, chromosomal microarrays (CMAs) and next-generation sequencing (NGS) do not require cell culture and produce more information at the genome level. The methods are thus most often used as the standards for genetic analysis of RPL in certain clinical practices.
"Karyotyping requires cell culture and frequently yields null results due to the preferential growth of maternal cells," Qiwei Guo, lead investigator and vice director at the United Diagnostic and Research Center for Clinical Genetics in Xiamen, China, explained. "CMAs and NGS are [also] relatively expensive tests, which pose a financial challenge for patients … in underdeveloped regions in large populations."
Guo's team therefore speculated that an improved and affordable screening method for RPL integrated with a diagnostic tool — such as CMA or NGS —could improve genetic testing for product of conception samples.
In a study published earlier this week in the Journal of Molecular Diagnostics,the team evaluated the accuracy of an HRM-based assay for numerical chromosomal abnormality screening and subsequently compared the cost-effectiveness of four different strategies for evaluating RPL.
Following PCR amplification, researchers can perform HRM analysis by measuring the melt curve of a PCR product. Comparing different melting curves against each other using fluorescent dyes, the analysis provides a visual quantified representation of what's been amplified which can be used to determine if there's an excess of two different products.
In their study, Guo and his colleagues collected 860 product of conception samples from RPL patients and performed G-banded karyotyping on them, successfully analyzing 765.
The researchers then evaluated the HRM test's screening accuracy using the remaining samples from these 765 uncultured products of conception. The team extracted DNA from about 20 mg of each sample using Qiagen's QIAamp Fast DNA Tissue kit, then performed PCR amplification and HRM analysis on the extracted DNA using Roche's LightCycler platform.
Using 90 samples, Guo and his colleagues adapted the HRM method to target seven autosomes and the sex chromosomes, which he noted where the chromosomes most-frequently linked to RPL. Designing five different assays, the team compared chromosomes 13 and 15, chromosomes 18 and 21, chromosomes 16 and 22, chromosomes 2 and X, and chromosomes X and Y.
Guo's team classified samples as either low- or high-risk for numerical chromosomal abnormalities, without specification of their karyotypes. HRM data for each sample was then analyzed using Roche's Gene Scanning software. The researchers then developed a decision-analysis model to compare the cost-effectiveness of four different strategies that included different genetic tests —karyotyping, CMS/NGS, the HRM screening tests, and a combination of HRM and CMA/NGS — followed by an evidence-based workup to identify clear or potential causes of RPL.
Using the HRM assay, the researchers found that a total of 316 samples were high-risk for numerical chromosomal abnormalities, and the results from 297 of those samples correlated with karyotyping results. The remaining 449 samples were considered as low-risk for abnormalities, with the results of 437 of these samples correlating with karyotyping results.
When detecting targeted numerical chromosomal abnormalities, the team found that the HRM assay had both a sensitivity and specificity of 96 percent. In terms of screening for abnormalities in all samples however, the test's sensitivity dropped to 62 percent while maintaining a specificity of about 95 percent.
When Guo's team grouped the results based on maternal age, they found that the HRM's screening accuracy increased in the cohort of patients older than 35.
Identifying 31 discordant samples — including 17 false-positive, 12 false-negative, and two discordant positive samples — in the cohort, the team further analyzed the samples using an NGS assay, which supported HRM results in eight of the false-positive cases, eight of the 12 false-negative cases, and one of the two discordant positive cases. The researchers therefore believe that the "authentic specificity and PPV of the HRM test may be higher … and [that the] HRM test is highly accurate" when screening for numerical chromosomal abnormalities in the evaluation of RPLs.
For each genetic test strategy, the team found that when the cost of CMA and NGS dropped below $68.30 or when the cost of the HRM screening assay rose above $178.00, the addition of the HRM screening test was no longer cost-effective.
"[However,] HRM is more accurate for aneuploidy detection compare to that of relative quantification in real-time PCR," Guo said. "In comparison with digital PCR and NGS, HRM is easy to perform, rapid, and requires inexpensive equipment."
The study authors highlighted that the HRM assay is also inexpensive because its reagent costs are cheaper compared to NGS and CMA reagents. When adding HRM screening to the diagnostic strategy, the diagnostic yield fell about 1 percent, but cost-effectiveness improved by about 30 percent ($374.00 versus $531.40).
Guo also argued that HRM stands out from CMA, NGS, and karyotyping because of its ability to deliver actionable results within a shorter time frame. After DNA extraction and PCR, which takes about two hours, the workflow for HRM analysis only requires about an additional two hours compared to NGS and CMA —which take at least two days — and karyotyping — which requires up to two to three weeks —to deliver results.
The researchers also believe that changes to improve the HRM's screening sensitivity or positive predictive value could further increase its clinical utility. When conventional genetic tests are unavailable or unaffordable, the group noted that the combined use of HRM screening and an American Society for Reproductive Medicine workup provides a predicted diagnostic yield of about 77 percent for RPL evaluation.
In addition, the researchers highlighted that the cost of diagnosing a patient with RPL using the HRM is similar to that of diagnosing it through karyotyping ($208.50 vs $220.00), which is 61 percent lower than a strategy using CMA or NGS with evidence-based workup ($208.50 versus $531.40).
"Statistical analysis of the data revealed that although most patients expressed a desire for genetic testing, less than five percent of [products of conception] were examined, with the cost [of] $440 per [test] being cited as the major barrier," Guo added. "[While] the cost of CMA or NGS is decreasing, … we speculate the cost will remain high enough to hinder the acceptance of CMA and NGS as affordable detection methods for RPL in underdeveloped regions."
At the same time, Guo acknowledged that his team dealt with several limitations in the study. He noted that HRM's resolution may not be able to detect mild relative numerical changes in target chromosomes, such as numerical changes in samples with low-degree mosaicism, combined aneuploidy-polyploidy, or maternal contamination
Carl Wittwer, a researcher at the University of Utah and who first described the HRM analysis technique in 2003, said that the Chinese team's method of identifying certain PCR products using HRM and their melting temperature allows for a simplified visual representation that is potentially useful for low-resource areas. He highlighted that the shortcut of using HRM is simple and inexpensive to get targeted answers for cases like screening for recurrent miscarriages.
However, like Guo, Wittwer pointed out that there are certain limitations to the HRM assay for evaluating spontaneous abortions.
"You need the numerical chromosomal abnormalities from the specific chromosomes, which are the most frequent ones, to detect them, which then affect the sensitivities and specificities," Wittwer explained. "There are other cases of spontaneous abortions that [Guo's group] are not testing for that don't give these numerical chromosomal abnormalities, or they're not the chromosomes the group is testing for … as that takes more work and resources."
Guo also acknowledged that in rare cases, sporadic copy number variations in the target sequences may pose a risk for the HRM's screening accuracy.
However, Guo's team believes that the HRM test can be used as a screening tool — combined with other diagnostic methods —to enhance the cost-effectiveness of RPL examination, or even as an alternative genetic test when other methods are unavailable or cost-prohibitive.
Guo's team anticipates launching a prospective study that will measure the assay's cost-effectiveness in a larger patient cohort in order to minimize false positive and negative results. In order to improve the test's current technical limitations, Guo said that the researchers will either adjust or optimize the reference intervals with "additional clinical practice based on the expected sensitivity and specificity."
Because the HRM assay can provide large-scale screening due to its lower cost and higher sample capacity, Guo believes that the test could also be used for population-based screening of Klinefelter syndrome.
Guo said the researchers plan to commercialize the assay for RPL evaluation, and welcome contacts "for all kinds of commercial cooperation."
Sherin Devaskar, a pediatric professor at the University of California, Los Angeles who specializes in neonatal-perinatal research, noted that offering HRM analysis in the clinical space to determine chromosomal aberrations in the embryo — rather than having to "take on testing of the conceptus" — would be very useful. However, she pointed out that there is also cell-free DNA testing currently being done non-invasively with a mother's blood prenatally toward detecting these abnormalities in babies.
"The interesting thing [Guo's team] has done is to use paralogous sequences, which means they've searched the genome and found that the same primers will amplify two different locations on two different chromosomes," Wittwer said. "Being able to amplify equally, in a competitive way, indicates that it's a valid application of high-resolution melting in low-resource settings, where groups can't absorb sequencing costs."