My name is Sheetal Parmar. I'm Vice President of Medical Affairs for Natera, and I'm really excited to have the opportunity to share Natera's innovations in SNP-based NIPT technology and to introduce Panorama AI. Today we'll be answering the following questions. What is the difference between SNP-based NIPT and other NIPTs? What is Panorama AI and what is new about it? What is the SNP-based micro-deletion and aneuploidy registry, or SMART study, where Panorama AI was validated? What improvements in performance were observed in the SMART study for Panorama AI? What are the clinical implications of Panorama AI improvements in SMART study findings? So let's get started. To review, NIPT is a test that analyzes circulating cell-free DNA in the blood of a pregnant woman. All individuals have cell-free DNA that is released from cells undergoing apoptosis, or programmed cell death. These are primarily from our white blood cells and our fat cells. When pregnant, the placenta also releases cell-free DNA into the mother's bloodstream. There are two primary methods that laboratories use to evaluate cell-free DNA and determine the likelihood that the fetus has a chromosome abnormality. The first I will discuss is commonly referred to as the counting methodology. This methodology is used by all commercial NIPTs except for Natera. In the counting methodology, the laboratory determines the total amount of cell-free DNA by sequencing highly conserved chromosome regions and then comparing the relative amounts of these chromosome fragments. Highly conserved means that the mother, father, and the baby are highly likely to share the same DNA sequences in these regions. Therefore, it does not distinguish between the mother and the baby. The second method used by Panorama is commonly called SNP-based NIPT. Panorama is the only NIPT that evaluates the risk for chromosome abnormalities using single nucleotide polymorphisms, or SNPs for short. 99.9% of the DNA sequences between all of us, between humans, are shared. SNPs make up about 0.1% of our genome, and these are the DNA sequences that are different between individuals. Essentially, SNPs are our DNA fingerprint. We know that it is highly likely the baby will inherit different SNPs from its mother versus father, giving the baby a unique DNA sequence and allowing SNP-based NIPT, or Panorama, to distinguish between maternal and placental DNA sequences. By doing this, SNP-based NIPT not only screens for aneuploidies, but also has capabilities that counting-based NIPTs do not share. So, what are those unique capabilities? By using SNPs, Panorama has three main areas of differentiation that I'd like to review next. First, we have performance. It's highlighted here in blue. Panorama has the highest sensitivity for common aneuploidies in 22q11.2 deletion syndrome. It also has the highest clinical positive predictive values for common aneuploidies across all age groups and for 22q11.2 deletion syndrome. And finally, there are zero fetal sex errors in validation studies. Next, using SNP-based NIPT for twins gives the provider unique information, which you can see coming out of the green semicircle. Panorama is the only twin NIPT that can determine zygocity, that is, whether the twins have identical DNA or non-identical. This was mentioned in the updated NIPT bulletins from ACOG, SMFM, and ISPD. Panorama is the only twin NIPT that reports two fetal fractions in dizygotic twin pregnancies and is also the only test that is clinically validated to determine whether there are two girls, two boys, or one of each. The third area of differentiation comes out of the orange semicircle. NIPT is a screening test, and many of the sources of error for NIPT can be reduced when the SNP-based method is used. ACOG practice bulletin 226 points out that there are two sources of false positive NIPT results, vanishing twins, and the presence of a maternal chromosome abnormality. These are only identified when Panorama is used. Conversely, the fetal condition of triploidy cannot be detected by counting methods because the relative amounts of each chromosome is the same. This is not the case with Panorama, and triploidy of both maternal and paternal origin can be identified. In a previously published study involving over one million women, Panorama was shown to have a positive predictive value of over 90% for trisomy 21, even in women less than 30 years old. The study also showed robust PPVs in women under 35 years old for trisomies 18 and 13, and the highest published PPVs for monosomiacs. In contrast, counting-based NIPTs report significantly lower analytically-derived PPVs, especially in younger women. I will caveat that the PPVs reported by counting laboratories listed here are analytically-derived by taking measured sensitivities and specificities and applying them to population statistics on the prevalence of trisomies. The PPVs for Panorama were observed clinically, therefore represent real-world performance. So it's important to note that the comparison here is not strictly apples-to-apples in this chart. In the past, the performance of SNP-based NIPT was accompanied by a trade-off in higher rates of no-call results. This trade-off is also observed with counting NIPTs, meaning that as you reduce the false negative rate of the test, you increase the no-call rate. If you're at all familiar with Notera, you know that this is a company that thrives on innovation, pushing its technology to do more and always do better. So I'm going to close this segment with a brief overview of our innovation and how SNP-based NIPT has evolved and to highlight some of the enhancements moving forward into 2021. Here's the timeline of NIPT innovation at Notera. In 2013, we launched Panorama aneuploidy NIPT. One year later, the microdeletion panel, including 22q11.2 deletion syndrome and four additional microdeletion syndromes was added. The release of a unique NIPT for twin pregnancies was a huge milestone achieved in 2017. And fetal fraction is accepted as a key metric for NIPT. So in 2018, the ability to identify those with low fetal fraction and the highest risk for triploidy and trisomies 13 and 18 through the fetal fraction-based risk score was added to Panorama. This is a feature of Panorama that no other test offers. In 2019, Notera validated its ability to identify maternal X chromosome mosaicism and aneuploidy, which is acknowledged by ACOG and SMFM as a source of false positive sex chromosome results with NIPT. And that brings us to 2021, where Notera is introducing Panorama AI. What is Panorama AI? It is the most up-to-date version of SNP-based NIPT. After performing over 2 million Panorama tests worldwide, it made sense to evaluate whether the algorithm could use this massive set of SNP patterns to improve the ability to make calls. By combining the SNP-based technology of Panorama with artificial intelligence or AI, this version of the test can give highly accurate results in difficult to call cases, which would have previously resulted in no calls. This includes maintaining the industry-leading accuracy for aneuploidies of Panorama while significantly reducing no-call rates and improving detection of 22q11.2 deletion syndrome. Previously, the algorithm was developed to detect the most common largest deletion, the 2.5 to 3 megabase deletion, responsible for the majority of 22q11.2 deletion syndrome cases. Notera's research and development team worked diligently to enhance the algorithm to detect smaller nested deletions within the 22q11.2 region, improving the accuracy of the test. I want to thank you all for your attendance and attention. To dive into the data validating these capabilities, I want to turn things over to Dr. Kimberly Martin. Dr. Martin? Thank you, Sheetal. I have been Notera's clinical liaison to the SMART study since 2016. I am delighted to have the opportunity to share many of the key results of the SMART study with you, including the validation of Panorama AI. SMART is an acronym for SNP-Based Microdeletions and Aneuploidy Registry, or SMART study. The primary objective of the study was to measure the performance of SNP-based NIPT in a prospective study for the common trisomies, monosomy X, and microdeletions. 20,887 participants were recruited from 21 centers. 55% of patients were recruited in the U.S., and the remaining 45% were recruited from Europe and Australia. One of the things that differentiates the SMART study is the genetic truth for common chromosome abnormalities and microdeletions was obtained for almost 90% of participants. All previous large clinical studies using NIPT have generally used physical examination at birth to define pregnancy outcome. Women enrolled in the study received Panorama test results, which could be used for clinical management. After completion of the study, samples were available for reanalysis using an updated algorithm called Panorama AI. A total of 18,497 pregnancy or child samples were successfully analyzed, primarily by the Center for Advanced Genomics, which is affiliated with Children's Hospital of Philadelphia. These analyses were blind to the Panorama results and to clinical information. 13,640 pregnancies in the study were at low risk for aneuploidy. This is particularly important in view of the increasing use of NIPT as a first-tier screen for all women. In terms of test performance for the common aneuploidies, the performance of Panorama in this real-world setting was consistent with previously published sensitivity and specificity data as indicated by the dark blue bars. Similarly, Panorama AI showed equivalent performance for trisomies 21, 18, and 13, as shown in the orange bars. I would like to draw attention to the sensitivity data for trisomy 21. Not only did Panorama AI have a sensitivity of 99%, but with the reduction in no-call rate, two additional cases of trisomy 21 were correctly identified as high risk. The reduction in the no-call rate similarly resulted in two additional correct high-risk calls for trisomy 13. This graph illustrates the traditional trade-off between accuracy and no-call rates using cell-free DNA that Sheethal highlighted earlier. The reduction in no-call rate demonstrated by Panorama AI was achieved with no change in sensitivity. This is expected to have a very positive impact for patients and providers, reducing the need for second blood draws without compromising on sensitivity. The bar graphs on the right illustrate the reduction in no-call rates achieved using Panorama AI with both first draws and a single redraw in samples that did not receive a result the first time. With allowance for a single redraw, 99.4% of women will receive a highly confident result with Panorama AI. A single redraw in that small number of women further reduces the no-call with no change in sensitivity. Before I dive further into the smart study results for 22q11.2 deletion syndrome, I wanted to take a few minutes to discuss the syndrome itself and why it's a condition that is important for providers and the families they care for. It's important to think about 22q11.2 in the context of other conditions for which prenatal screening has become the standard of care. For example, trisomy 18 is accepted as a condition for which risk is evaluated and reported, although its frequency in live births is about 1 in 5,000. Carrier screening for both cystic fibrosis and spinal muscular atrophy, or SMA, are to be offered to all pregnant women, but the frequency of these two disorders are significantly lower than 22q11.2 deletion syndrome. I'm going to take just a few minutes to review the typical features observed in individuals with 22q11.2 DS. In the absence of a major congenital heart defect, this syndrome is often not recognized at birth. Diagnostic odyssey refers to the large number of tests and specialist visits required to ultimately make a specific diagnosis at an average of four years of age. This diagnostic odyssey is something that our pediatric colleagues are well aware of, and it's a source of significant stress and frustration for many families. There are many features of this syndrome that are amenable to early intervention, with the opportunity to improve outcome, and I will highlight these now. Developmental delay and learning differences are experienced by a majority of individuals with this syndrome. However, anticipatory assessment and early intervention can optimize the outcomes. Palatal abnormalities are common and result in feeding difficulties that can be corrected with targeted interventions. Not uncommonly, there is a requirement for placement of a G-tube in the first year of life. Immune deficiencies are not common, but they can lead to increases in infections, particularly in children who may require surgical procedures if the immune deficiency is not recognized. As with many genetic disorders, new features may develop with time. The increased likelihood of schizophrenia in young adulthood can have dramatic consequences for the individual and their families. Congenital heart defects may be mild, like VSDs, or unsuspected until closure of the ductus arteriosus, such as interrupted aortic arch. Discharge from hospital with an undiagnosed ductal-dependent cardiac lesion can have serious consequences, including cardiac arrest, hypoxia, and even death. Hypocalcemia is relatively common, and it may not be easy to diagnose, particularly if unsuspected. A large study of adults with 22q11.2 deletion syndrome found that untreated hypocalcemia in early life was associated with more severe neurodevelopmental phenotypes in adulthood. Almost all cases of 22q11.2 deletion syndrome are caused by missing genetic information on one copy of the 22nd chromosome. This diagram, not to scale, illustrates the region on chromosome 22 that may be affected. There are over 40 genes in this region, and there's a lot of interest in identifying the function of these genes and how absence of one copy may result in the various manifestations of this disorder. Most studies have reported that about 85% of cases of 22q11.2 deletion syndrome are caused by the absence of the entire A to D region, which is about 2.5 to 3 million base pairs in size. Most of these validation studies performed to evaluate the performance of screening for this condition using cell-free DNA have focused on this large, common deletion. In general, NIPT has better performance detecting large imbalances compared with smaller ones, with some NIPT laboratories reporting a lower threshold of about 5 million base pairs for detection. When tests such as FISH or Baxon beads are used to determine the presence or absence of 22q11.2DS, only deletions that include the A to B region will be detected. These tests do not determine whether there are deletions beyond the B region. Due to the large study cohort size in SMART, and because the majority of pregnancies enrolled underwent a complete microarray test, we were able to determine the frequency with which deletions throughout the A to D region were present, and validate the performance of PanoramaAI for these atypical nested deletions. We found that in contrast to prior reports, about 40% of the deletions were confirmed with microarray to be smaller than the full A to D deletion. One third were confirmed to be the full A to D deletion, and another quarter were confirmed via FISH or Baxon beads, so we know they involve at a minimum the A to B region. The SMART study showed that the sensitivity of PanoramaAI, which was 83%, was higher than the sensitivity in clinical reporting, which was 75%. Specificity was unchanged. One of the things that providers and patients are concerned about are positive predictive values. PanoramaAI, due to a combination of increased sensitivity and reduced false positive rates, resulted in a more than doubling of the PPV from 24% to 53%. This is even more remarkable because it includes micro-deletions in the region that are less than 2.5 megabases in size, which may be under-detected using other CFDNA methods. Of note, the incidence of 22q11.2 deletion syndrome in the study was 1 in 1,524, which is on the higher end of what has been reported in general population studies, and similar to the ranges reported in the obstetrical population. This sets a new benchmark for studies of 22q11.2DS in the prenatal population. Studies that do not obtain genetic confirmation using microarray in the entire studied population may be missing a significant number of cases. Another very interesting finding of the SMART study was that the majority of affected cases had no ultrasound findings, particularly in the first trimester. 11 of the affected pregnancies had an ultrasound between 11 and 14 weeks of pregnancy, none of which were abnormal. Specifically, there were no reports of nuchal translucency enlargement. Nuchal translucency was measured and reported in five cases and was 2.6 millimeters or below in all. Panorama was performed prior to 14 weeks in six affected cases, five of which were high risk. The cases in which panorama was performed in response to a major malformation were all performed in the mid to late second or third trimester. Identification of major malformations, particularly those that are related to a genetic abnormality later in pregnancy, significantly limits the time available for a family to receive education and counseling and still retain the full spectrum of reproductive options. This slide illustrates clearly that use of cell-free DNA for screening in the first trimester provides the opportunity for early diagnosis rather than reliance on ultrasound findings, often later in pregnancy. Finally, a very interesting result from the SMART study that I wanted to highlight was the strong correlation between adverse pregnancy outcomes and the inability of panorama to give a result. The specific factors explaining these correlations remain to be evaluated, but this is a finding that deserves attention and more research. The causes of the adverse outcome may depend in part upon the reason for the no result. For example, whether it is due to a low fetal fraction as opposed to being related to the specific SNP allele patterns and ratios that are identified. Further research is in process and we hope to be able to publish this analysis later in 2021. The implications for sample redraws and how the updated panorama AI predicts these outcomes requires further study also. Preliminary results suggest that those women who receive a second no-call are at the highest risk for these various adverse outcomes. As you can see in the chart, pregnancies with no result after a single redraw showed significantly higher rates of adverse pregnancy outcomes with a p-value of less than 0.001. SNP-based NIPT test failures had been previously established to correlate with trisomies 18 and 13 and triploidy, but a correlation with late pregnancy outcomes has not been previously demonstrated. Women who have a single redraw which is successful experience adverse outcomes at a lower frequency than women whose second draw is similarly non-reportable. As we learn more about the specific predictors for these adverse outcomes, it is likely that information which may be of use clinically in triaging these high-risk pregnancies will be able to be reported. To summarize findings from the SMART study, key takeaways are first that this is the largest prospective NIPT study to date. It is also the first general population study of this kind that had genetic testing results on all participants that were included in the data analysis. This also includes the largest cohort of low-risk pregnancies with this type of analysis used to establish performance. On a high level, the results of this study analyzed so far have demonstrated the following. The aneuploidy performance of panorama was consistent with previous validation studies using enriched cohorts, demonstrating the applicability of these performance measurements in a real-world setting. Panorama AI demonstrated improved no-call rates of less than 50 percent of that seen in the prior version of panorama. 22q11.2 deletion detection was also significantly improved, which has clinical implications for the appropriateness of screening for this syndrome, particularly in women of younger age. Finally, the association between test failures and adverse pregnancy outcomes make a compelling case that no-call results may still have clinical value. I'd like to thank you for the opportunity to share these findings from the SMART study and the validation of panorama AI. I will now hand the presentation back to Sheethal. Thank you, Dr. Martin. Now that we have gone over the results of the SMART study and the performance improvements of panorama AI, I'd like to discuss how panorama AI and the findings from SMART can be incorporated into clinical practice. Firstly, panorama AI is simply the newest version of panorama, and these improvements will be implemented during the first half of 2021. There will be no special protocol or kits necessary in order to order panorama AI for your patients. And based on the data from SMART, redraws, while rare with panorama AI, should be a routine part of a clinic's workflow to triage patients for risk of adverse pregnancy outcomes. Most patients will receive a result on a redraw, but those that do not are at significantly elevated risk for late-term adverse pregnancy outcomes, even if they do diagnostic testing and are negative for trisomies. So increased monitoring and a referral to a specialist should be considered for these high-risk patients. Another piece you should consider integrating into your practice based on the findings from the SMART study is screening for additional conditions. For example, 22Q1 1.2 deletion screening should be considered for integration into routine panorama ordering. Dr. Mary Norton of UCSF, principal investigator of the SMART study and former president of SMFM, has stated that the results of the SMART study provide exciting data supporting routine cell-free DNA screening of conditions beyond aneuploidy, such as 22Q1 1.2 deletion syndrome. Natera is committed to ensuring access to high-quality genetic tests like panorama for all patients and has developed a comprehensive suite of services to support providers and patients through the entire testing journey, including patient education through genetic information sessions, access programs like price transparency, ordering, including remote ordering, EMR integration, mobile phlebotomy, and automatic redraw programs. Results are always provided in clear and actionable reports, and additional services are available to support next steps following result reporting. I'm also very proud to share that Natera offers Natera Academy, developed by our genetic counselors and medical directors, which includes minimally branded short-form videos to help clinicians review basic scientific concepts regarding genetic testing and how to best counsel patients on this type of testing, empowering you to provide genetic education to your patients. So in summary, what questions did we answer today? We identified the SNP-based approach to NIPT as being the only method that distinguishes between maternal and placental DNA, and we discussed the unique capabilities that conveys on this type of NIPT. We introduced Panorama AI, which combines the SNP-based method for NIPT with artificial intelligence. Panorama AI was shown to have improved performance for 22p1 1.2 deletion detection and have lowered no-call rates to less than half of what was the previous version of Panorama. These improvements speak to the depth of knowledge and resources at Natera in both biostatistics and the laboratory research and development teams. We discussed results from the SMART study, the largest prospective NIPT study. The study had over 18,000 pregnancies with confirmatory testing, and over 13,000 of those were considered low-risk for aneuploidies. This constitutes the largest low-risk cohort with confirmatory testing of any NIPT study to date. With regards to screening for 22p1 1.2 deletions, Panorama AI demonstrated a sensitivity of greater than 99.9% for full deletions over 2.5 megabases and 83% for all types of deletions, including smaller deletions down to 0.5 megabases and an overall PPV of 53%. Finally, based on the results of the SMART study and the importance of perinatal interventions for this disease, 22p1 1.2 deletion screening should be considered for integration into routine Panorama ordering.

SNP-based NIPT technology

Panorama AI

SMART study

chromosomal abnormalities

pregnant women

clinical implications