The below text is a transcript from the webinar. Because it is a transcript, there may be oddities that arise from the process of translating speech into text. We recommend accessing the recording, above, to gain full context.
It's a true pleasure to be here today, and tell you a little bit about our story here at PierianDX, and specifically our implementation of the TSO 500 assay to enable laboratories globally to perform clinical NGS sequencing. Our mission at PierianDX is to firmly keep cancer care in the community. And we do that by democratizing access to clinical next generation sequencing through our informatics, and our services that we offer. We enable our customers to provide the same level of advanced testing and practice precision medicine at their local sites.
As Josh mentioned, we spanned out of Washington University in Saint Louis in 2011, and ... I'm sorry. We spanned out of Washington University in 2014, and the technology was developed at WashU in 2011. Our first customer at the time in 2014 was Moffitt Cancer Center, which went live in late 2014. Moffitt was also the very first customer to launch the TruSight Tumor 170 clinically, and we were instrumental in supporting their validation and their launch. Today we have over 45 plus academic medical centers, cancer centers, pediatric hospitals, commercial labs, reference labs and integrated delivery networks, that we're proud to call our customers and partners of our sharing network that we'll describe in a bit. In early 2019, we announced a multi year partnership with Illumina to support cancer research and diagnostics. That will be the focus of the presentation today.
At the very heart of our products and services is our knowledge base, which is medically powered. It's the largest opt in content sharing knowledge base that I'll describe in great detail in the latter part of this presentation. And that knowledge base is contained within a comprehensive workflow solution that we call our clinical genomics workspace. That is our informatics and reporting software. Those two core products are supported through scientific and medical services, namely assay validation, and variant and case interpretation services. And we also offer laboratory services through the CAP distributed model of testing as a clear certified CAP accredited laboratory that we established in 2018.
The agenda of the presentation today really, is to first describe the trends, and the value, and the adoption, and the increased adoption that we see with comprehensive genomic profiling or CGP, focusing on the TruSight Oncology 500 as the prototypical CGP assay that could be distributed very widely and very rapidly through our partnership with Illumina to describe the knowledge base, and the curation that's required to support an essay of the size of the TruSight Oncology 500, in order to do classification and interpretation to generate a comprehensive report very rapidly. And we'll conclude by talking about how we put all of this together within the CGW in our knowledge base in order to enable our partners both here in the US and globally.
Recent Trends and the Paradigm Shift Toward Comprehensive Genomic Profiling
First and foremost is we think about the trends in precision medicine and precision oncology. There's a rapid rise through targeted therapies as well as immunotherapies and oncology. Multiple markers that are tied to these therapies both from a therapeutic, as well as prognostic and diagnostic point of view in oncology as well as those available in pharmacogenomics. There are a number of hospitals that are now billing for large assays such as those as described by comprehensive genomic profiling. However, the majority of the volume of CGP is actually being performed at a few independent reference laboratories. And we think assays like the TSO 500, that are very broadly applicable and adaptable, will allow that cancer care to normalize back to the community.
Assays that are part of this CGP spectrum profiled biomarkers including TMB and MSI, as well as site agnostic markers such as NTRK fusions with known and other partners. These CGP assays are progressively increasing because of the need to detect these agnostic biomarkers across the number of tumor types, as well as increasing payer coverage as we described. Indeed oncology testing itself is evolving. As we think about testing even three or four years ago, and looking at single markers, single genes by single sequencing or other methodologies that were pre NGS moving now to NGS based testing to a less than 50 genes, as well as greater than 50 gene assays like the TSO 500 that do comprehensive genomic profiling. That the move to broader NGS testing is a reality today, and one that is growing through multiple modalities and testing scenarios as we'll describe in oncology.
Namely, that serial testing itself is vital to make cancer a chronic disease in that we see a vision, a future where in cancer prevention and screening through NGS is a reality. Testing at a primary diagnosis followed by each recurrence including monitoring methodologies through liquid biopsy, all being a reality. And we see all of these trends both in the precision oncology market as a whole through commercial and reference laboratories, FDA's approval path, CMS's payment approval path, as well as assays like the CGP. I'm sorry. Assays like the TSO 500 that enable a variety of these tests through a single assay.
One of the common concerns in trying to support diagnostics in oncology is that tissue is the issue. There are many tests that need to be performed, however tissue's limiting. And an essay like the TSO 500 that's capable of performing a comprehensive genomic profiling can look at a number of gene and gene targets as well as report out on large bio markers such as TMB and MSI. CGP can also identify a greater number of actionable alterations. And a number of studies have shown that this can range anywhere from 30 to 90% depending on how you classify action ability, and depending on the types of tumors that you are looking at, anywhere from pediatric tumors to advanced tumors in adults.
CGP has a great value in patient management, namely, increase use of targeted therapies and personalized therapies, much greater increase in clinical trial eligibility and that more trials are a match based on molecular criteria, enabling the patient to actually enroll in a particular trial based on the rest of the clinical criteria that are required. CGP also reduces the costs of non targeted therapy making the costs that are spent to have actual benefit, and with CGP type testing and personalized therapies reduced in adverse events.
Importantly, one of the barriers for a much more broad adoption of CGP testing has been reimbursement. And with the new NCD that was put out a little over a year ago, where a comprehensive genomic profiling for different assays was deemed to be covered by the CMS with a particular GSP code, and pricing set at 2920. Indeed, these types of CGP assays are rolling out very quickly. And these assays are initially rolling out as RUO reagents that are clinically validated at laboratories as laboratory developed tests. And many of these assays are undergoing approval and approval processes, or planned to undergo approval processes through the FDA, and parallel path to the CMS for reimbursement.
TruSight Oncology 500 Assay
The TruSight oncology 500 is one of those essays that we want to focus on today. Before I get into that, I do want to lay the foundation that PierianDX, over the last five years as a company, has supported a number of cancer and germline assays. Some of which are listed on this slide are very broad based, amplicon based, hybrid capture based, those using UMIs in cancer and in germline disease indications. And very recently, as we talked about, we have signed a partnership with Illumina to specifically support and productize the TruSight Tumor 170, and the TruSight Oncology 500 assays for very rapid and broad deployment by clinical NGS laboratories here in the US as well as globally.
The TSO 500 assay itself is a DNA and RNA based assay. The DNA component is hybrid capture. The panel size almost two mega bases, very supportive of detection of TMB, where requirements have been shown to be at least a one and a half mega bases for those types of assays. It also employs UMIs, and does have the ability to be used for liquid sample analysis, which we'll describe. The RNA component allows for the detection of known and novel fusions as well as splice variance. And as a very broad assay obviously, the content is capable of supporting a number of solid tumors as well as a very good coverage in the mileage space. Specifically, there's a hundred percent coverage of guidelines for 11 tumor types, and over 1200 clinical trials for which biomarkers have been identified that the TSO 500 supports. The TSO 500, as I alluded to, today supports solid tumor tissue based testing but is being validated through our partnership with the NCI to support liquid biopsy testing. And Illumina has partnered on companion diagnostic development for IBD use with Loxo and BMS to essentially transform the TSO 500 as that in vitro diagnostic.
What I'd like to focus on in the next few slides is our experience with the TSO 500, and the TST 170 with regard to their analytical performance based on data that were generated by Illumina. And these data were reviewed and analyzed in analyses that are analogous to analytical validation that elaborate our clinical partners, and our laboratory partners would do as part of their own validation. And thus allows us to determine the relevant components such as accuracy, analytical sensitivity, specificity, as well as components such as precision and reproducibility. And the goal here was to really optimize the pipeline, the parameters, and filters for clinical use within our platform at the CGW. Samples included multiple reference standards as well as FFPE clinical samples. Overall, as I'll show you, there was excellent analytical performance based on the datasets on the TSO 500 and the TST 170.
Starting with accuracy on the DNA as reported out here using sensitivity as the calculation measure, there was nearly 100% sensitivity for both SNVs and Indels using reference standards as well as clinical samples, and the one false-negative for SNVs that was present at a very low frequency as reviewed manually in the BAM file, and it was embedded in a homopolymer region, which is likely why it was missed. With regard to analytical sensitivity and specificity, specifically lower limit of detection on the analytical sensitivity side, our initial analyses showed that we can get down to about 3% [inaudible 00:18:17] frequencies for both SNVs and Indels, and have sensitivity and specificity at 100% using clinical samples as well as reference samples. On the other side similarly, accuracy as we measure it using sensitivity, using again reference in clinical samples, 100% of the fusions and splice variants were detected across reference standards as well as FFPE embedded cell line samples and truly clinical samples that were FFPE embedded.
Analytical specificity and sensitivity that we measured on the RNA side showed that we have a hundred percent sensitivity in reference standards and clinical samples, where there were at least five copies per nanogram of RNA input of that fusion whether known or novel. TMB and MSI calculations on the DNA, the results we're presenting here are based on Illumina analyses. We did not repeat these analyses. I didn't feel a need to do so, in that essentially, Illumina compared TMB calculations based on the TSO 500 compared to both whole exome data, as well as TSO 500 TMB reporting compared to tumor normal pairs, and showed excellent correlation demonstrating that the TSO 500 is able to accurately report out TMB at 10 mutations per mega base threshold, having greater than 90% of both positive and negative percent agreement. Similarly, for micro satellite instability, the TSO 500 MSI score was compared to the qualitative high versus stable categorization using MSI by PCR, which is standardly done using five markers, and using a threshold of greater than or equal to 40 markers sites. The specificity and the detection rate were both 100%.
Have you shown that the TSO 500 content and analytical performance is ready for clinical use? We next turned our view to the knowledge base and the curation strategies and approaches that we would need to demonstrate that the TSO 500 could be running the CGW, as well as be analyzed through our knowledge base, and have excellent performance for our clinical applications.
CAP/AMP/ASCO Classification Scheme
I bring up this slide just to remind everybody that the TSO 500 and other assays like it in somatic cancer should use the CAP and ASCO variant classification scheme, and that's the approach that we've taken for implementation of the TSO 500 in the CGW. However, the big question is with an assay of this size how can we rapidly interpret cases, and still maintain the highest level of accuracy as well as a comprehensiveness. This funnel really shows how clinically relevant variants are very rapidly identified by our knowledge base, namely, the content itself of the assay is focused on 523 genes as I described previously. The validated filter, i.e filters that pass quality criteria for each variant type ensure that clinicians, and pathologists are looking at a variance that are indeed called per the validation criteria.
We subdivide variance in different variant types for rapid review, and allow for very rapid prioritization of those different variant types and biomarkers using FDA approved labels, NCCN practice guidelines. Our prior medical interpretations and associations to clinical trials such that based on that level of information, those variants maybe rapidly either auto classified or reclassified by various scientists and medical directors as part of the interpretive review process, and report finalization.
Our knowledge base is summarized here. Our knowledge base is composed of a number of data sources. And those are listed, namely, genome builds, genome RNA protein models, minor frequency databases such as a thousand genomes NOMED, exact can ESP, cancer research databases to look at a very tumor type frequencies from COSMIC and TCGA, clinical evidence databases, such as ClinVar as well as computational frequencies. All these databases where disease or tumor type is involved or normalized to the SNOMED-CT ontology. And shared medical content is available from our medical directors as they sign out cases, namely variant classifications and medical interpretations are shared through our opted model. And overwhelmingly our customers have opted into sharing their own experiences as they review and sign out cases.
In addition, a core part of our knowledge base includes curated annotations from FDA approved labels, NCCN and ASCO guidelines for use in the US as well as active recruiting trials from clinicaltrials.gov. In the early second half of 2019, we will be adding curated information from EMA drug labels, asthma guidelines, as well as action elite recruiting trials in the EU.
An important differentiator of our knowledge base is that it is more comprehensive than a variant lookup. It is a rules based decision support engine with over six mega bases of coverage. Inferences can be made, not only on variants using HGVS syntax, or specific coordinate based syntaxes that allow us to cover codon and exon ranges, but also based on functional characteristics, such as maintaining reading frame or being truncating or things of that nature. With regard to gene fusions and copy number variants, known and novel partners can automatically be coded as part of our rules engine in order to make inferences that are relevant, similar to the end track fusion inference that's now required, as well as limiting copy number gains and losses based on ranges.
And finally, the rules engine itself can have very complex rules written such that one or more variance when they're found in the same patient context, have a different inference than when they're found a separately. Our approach to the rules engine and predicate model's really summarized in this very busy slide. Suffice it to say on the right side what you see are predicates that I described in the previous slide that then, if they are met by one or more variance in that patient case result in inferences being made. For example, [inaudible 00:26:56], a variant in melanoma with an inference being made on an FDA approves a drug label for [inaudible 00:27:04]. In addition, the other databases that I described, it's just minor little frequency databases, clinical variant databases, TCGA, COSMIC, things of that nature are coded as annotation rules. So those annotations also are assigned to variance as you review and signing up cases.
This combination of annotation rules and curated rules are when applied to a set of variants and biomarkers in a patient case, then allow those rules to fire and allow that information then to appear within the CGW, allowing our partner medical directors and various scientists to both confirm and assert that the rule indeed is applicable in the context of this patient. Allow them to then rank all those rules, and utilize that information then to finally classify and interpret a variance as part of the case review and sign up process. Well, what we describe here is the CGW, and the PierianDX knowledge basis experience already in reporting and signing out on genes that are part of the TSO 500 gene set. Mainly, what we show here is that almost 25% of the genes, about 125 genes, have already been reported in over 10000 cases within our knowledge base. And over a third of these genes have been reported, but out of over 5000 cases, over half in 2500 cases, and 100% of the genes and at least 350 cases. Mainly that the CGW, and the knowledge base has already reported out on the entire gene set in hundreds of cases, if not tens of thousands of cases.
And that demonstrates a significant experience that is part of our partner sharing network that is immediately available and leverageable as people start experiencing the TruSight Oncology 500 review and sign out process in the CGW. Important criteria that are part of our curated set are shown on the right. Almost 50 drug labels from 38 guidelines, and as I said earlier, over 1200 of all clinical trials are associated to these genes. We also wanted to benchmark the knowledge base with regard to its classification accuracy. Here, what we did is use the TruSight Tumor 170 gene set, and our experiences as our medical directors reviewed and signed out cases.
And we assessed the knowledge base's ability to classify variant as being clinically relevant, i.e tier I or II, versus either being a VUS or a benign variant i.e tiers III or IV. And calculated our specificity, sensitivity, positive predictive value and accuracy as a function of medical directors who changed the classification of a variant in a case after the knowledge base that it had assigned an initial classification. Based on that data, what we demonstrate is very, very high specificity. An excellent sensitivity of the classification that was assigned by the knowledge base compared to the signed out report.
Clinical Management Yield
The second thing we wanted to look at, with regards to the knowledge base, is what we call clinical management yield. Mainly, is there at least one actionable finding for at least one actionable variant in each case? And if that criteria were met on that case was determined actionable, and had a yield in clinical management. And what we show here is that by analyzing cases from the TruSight Tumor 170 genes set, that if we relied on FDA drug labels and guidelines alone, two thirds of the cases had at least one actionable finding. However, an additional over quarter of cases had actionable information that went beyond FDA drug labels and guidelines. And that is unique knowledge that's in the published literature, late breaking information, abstract based information, that are shared through our medical interpretations that are written by hundreds of medical directors who utilize our system today.
I want to, now, talk a little bit about evidence based reporting, and how we go about doing that within the CGW, specifically for the TruSight Oncology 500, and assays like it. The report itself, as you can see on this slide, has a clinically important, and critical front page or first page, that classifies and determines the number of variants in each of the CAP and ASCO classifications and tierings, reporting out on TMB and MSI as well as information about the number of trials that matched. As you dig deeper into the report, there are in depth interpretations for each variant or combinations of variants, and relevant clinical trials that are associated to those variants in this case.
Focusing in on the first page, variants and biomarkers are tiered based on CAP and ASCO classifications, summary of trial count and TMB results available for immediate purview by the oncologist. And evidence levels are presented for each variant or combinations of variants as well as the clinical impact, mainly therapeutic, prognostic and diagnostic information that's available from FDA approved labels guidelines as well as those that are part of medical interpretations. Within the in depth of variant interpretations are very typical structure that our variants scientists follow at PierianDX as part of the interpretation services, and that many of our medical directors have adopted including gene level information, variant level information, tied to the specific patient tumor type as well as other additional supporting evidence of that variant in that patient's tumor type, concluding then with therapy guideline information as appropriate.
Each of the variants are named according to the HBS nomenclature following both the P. and C. Nomenclature in the best practices, and publications, and as recommended by CAP checklists. Specific version transcripts, variant allele fractions, and depth are included in the report. The rest of the report then contains a variance of unknown significance as recommended by the CAP and ASCO guidelines, a description of the actual CAP and ASCO classification scheme, and concludes with methodology and disclaimer sections that are site specific and laboratory specific.
I want to conclude the Webinar by really talking about how we have deployed the TSO 500 assays like it within our CGW platform. Our primary goal with our customers is to provide a very seamless experience to be able to go from sample to answer for assays like the 500. And what I'm describing here or diagramming here, is this specific application for TSO 500. Mainly that samples are processed, typically FFPE samples, through DNA and RNA isolation, and library prep enrichment. The sample is sequenced on the next seek and analysis is performed using the Illumina trusight Oncology app. And that APP may be run locally as well as within our infrastructure. There are a variety of options the customer has in order to have secondary analysis performed, but the same secondary analysis tool and version is used. And that is critical to the prior analytical validation data that I showed.
Importantly then, that process and workflow is automated within the CGW, and then processed through our knowledge base, and the rules engine such that a draft report is generated where variants are both annotated and classified with interpretations and other information such as therapeutic, prognostic, and diagnostic information, available for professional review and sign up. That professional review and sign up itself happens within the CGW such that the final signed out report can then be pushed to the electronic medical record or alias, as appropriate.
The CGW itself is summarized on this slide for those of you very new to the CGW. A very complete molecular testing solution as I indicated in the previous slide, a full HL7 integration to receive orders as well as to push signed out reports from the LIMS to the EHR for example. We are HIPAA compliant through our high trust certification that we received last year. Undergoing GDPR compliance now that we expect to complete by the first half of 2019, we use industry best standards and security with a site to site VPN tunnels necessary for HL7 messaging as well as secure rest based APIs that allow integration with clinical, and other research based systems as appropriate with our clients' sites. We support the entire workflow to go from FASTQ to VCF as I demonstrated specifically for the TSO 500.
Whether we use a partner based secondary analysis solution or our own in other assays. It's a very flexible approach. Fully supports germline based testing as well through a germline specific databases as well as excellent base testing using a free analyzer based scoring of phenotypes associated to the patient as well as being able to filter variants based on different inheritance models. We have a number of a QA, QC reports as well that support the quality management activities of a clinical laboratory. And I've already described the clinical sharing network. Our clinical reporting capabilities are extensive. We have a number of different report templates that can be utilized by laboratories in different clinical, and disease settings to meet their needs.
I want to spend a couple of minutes on our interpretation services, which are summarized here. In a typical workflow, as I described a couple of slides ago, the simplest process and sequence data are essentially pushed into our cloud environment for a draft report that's generated within the template that a customer requires and would like. That draft report underlying annotations are fully reviewable, and transparently reviewable, and editable. By that, laboratories, variant scientists, residents, fellows, medical directors in order to finalize that report. Our own interpretation services team and variant scientist's team is available to support customers in order to assess those existing variant classifications and interpretations, and potentially update it on even newer literature, evaluate very complex pathway or co-occurrence type data to create more cohesive interpretations and refined clinical trial matching based on client specific SOPs that may exist. We're able to support customers in that. A professional review and a sign out process to get to a near final report for that customer in order to help them scale.
There are a very large number of steps that are part of clinical grade pipelines, and the adoption of secondary analysis solutions that are clinically validated is only one part of that. And I show this very complex diagram only to demonstrate that. Mainly, once pipelines are run and variance are called, there are a variety of different variant files, VCF, non VCF files, that all need to be normalized, HGVS nomenclature assigned to them, identify different variant types such as splice sites, CNVS fusions, appropriate QC, QA metrics reports that are generated based on DNA and RNA BAM files, as well as the ability to then display quality and annotation criteria within the system. That ultimately then results in that draft report, which can then be reviewed and signed out to generate the final report at the bottom right.
Assay Validation Services
Our assay validation services are very strong companion service that we provide for all assays that we support for our partners, whether customer or vendor based. That is especially true for the TruSight Tumor 170 and the TSO 500. Our essay validation services are comprised of development of a draft validation plan that defines the numbers and types of specimens that are needed to be a run on different run configurations, a sequencer instrument configurations as well as user configurations to perform a variety of different validations. Analytical and diagnostic validations, assess things like repeatability and reproducibility, assess the overall performance of the assay and coverage across exons and hotspots, including them, and resulting in then, a final validation report. And content that then maybe used to generate a CAP CLIA compliant documentation. A very prime example of this is our support of Moffit, as they validated that TruSight Tumor 170 assay. And we're the first to deploy that.
The CGW operates very nicely within the entire health IT and instrument ecosystem that exists at clinical and molecular diagnostic laboratories. Mainly API and HL7 based integration with LISs and order entry systems are shown on the bottom left. Being able to push out signed out reports to the EMR and LIS systems, including the ability to extract discrete data from our system using our rest based API. Individual variant and variant inferencing data may be pulled out using the API, and be pushed to a variety of different systems for decision support, electronic data warehouses, as well as business analytics and business intelligence applications. We have connectivity with sequencers such that we can pull both BCL files, FASTQ files, as well as downstream results. BAMS and VCF is appropriate if secondary analysis solutions are run on site.
I'd like to wrap up by telling you the different components that we described today. Mainly that CGP is a reality today. That reality is especially apparent with the TruSight Oncology 500, that enables our clinical partners to much more broadly adopted CGP in a democratized and distributed fashion, enabling our clinical partners to perform CGP in house, and partner with their oncologist to deliver the highest content, performance, and experience in their patient management processes. PierianDX is very proud to support our clinical partners. Enable those partners and being able to utilize a very comprehensive knowledge base and workflow solution. And allow that customer then, to go from sample to answer in a very streamlined fashion.
Our supporting services that I described, mainly assay validation and interpretation services, allow a customer to very rapidly validate their TSO 500 assay in their clinical laboratory, and deploy then, that essay that's available in RUO form today as a laboratory developed test at their own organization. With that, I'd like to conclude and hand it back to Josh. I believe we have some time for some questions.
All right. Thank you very much doctor Nagarajan. We've actually received quite a few questions already, which is great. But just to remind you and if you want to flip to the next slide, just to remind folks if you want to type in a question, please do so using the go to webinar control panel. I will try to get to your questions today. If we don't, we'll answer them after the Webinar's over. But please feel free to submit your questions there in the control panel. There are actually several questions about FDA approval and FDA submissions. Let me just fire these off sequentially for you, and see if there's a consistency in the theme here.
Question: But the first question is, and you touched on this early in the presentation, are you planning FDA submission like foundation medicine or Memorial Sloan Kettering did, and the ladder with the CGP assay? Maybe you can just talk about, in general doctor Nagarajan, what your understanding of the FDA submission is going to be for this assay.
Answer: Sure. My understanding is that any assay maybe pushed through as a site specific approval by any clinical laboratory. We ourselves at PierianDX do not plan to get a site specific approval of the TSO 500 or of any other assay, in that we are not a full clinical laboratory. Or any of our clinical partners or laboratories may. Having said that, the TSO 500 can go through site agnostic approval through the FDA. And that process is occurring through in vitro diagnostic approval, and through the partnerships that I previously described that Illumina's entered into.
Question: And then to continue that theme. And there's a question here about, once an assay such as TSL 500 achieves FDA approval for IBD status as a companion diagnostic assay, do you anticipate needing to implement a dual bioinformatics pipeline? In other words, one for the IBD content, another one for the non IBD content.
Answer: That's a very interesting question. Within the FDA approval process for an assay like the TSO 500, and I would imagine this is the same for the foundation one approval, one important component is the ability to call and report out analytically validated variance, and then to follow that up and report out companion diagnostics. Excuse me, there are one or more variants of biomarkers that are a part of that asset. And that is the approved component as was demonstrated by foundation and others. They limited set of biomarkers that they were calling out, and had performed appropriate clinical validity experiments as companion diagnostics. However, in the FDA tiering system, there are other tiers that the FDA called out and recommended that may be part of the report has annotation information. Mainly, clinical evidence for actionability based on guidelines, as well as variants that are part of the emerging spectrum. And those additional levels of annotation and interpretation can be applied to the CAP and ASCO classification scheme in order to put out a comprehensive report that contains both companion diagnostic information, as well as those lower tiers of evidence that the FDA has recommended.
Question: I believe you just touched on one of the next questions on FDA approval, which is, is it your understanding that once it's FDA approved, the gene content can be subsetted, sliced, and diced for generation of a draft report? The cover only tumor specific gene subsets for example.
Answer: Yeah, that's an interesting question as well. My own view there would be that I think that, that would need to be done carefully based on the FDA approvals around companion diagnostics such that at least the diagnostics that are tumor type specific are inadvertently cut out. But I think that's a possibility, and one that could be done. Although my personal opinion is that, as we've shown in the early parts of this presentation, that the CGP and the merits of performing CGP as a whole and reporting out the entire content, may outweigh the benefits of slicing and dicing, and at least in certain encounters with patients.
Question: The next one here is around something more bioinformatically. What is your average ... In your experience, what is the average turnaround time from specialist specimen receipt to report generation and finalization for the TruSight Tumor 170, as well as the TruSight Oncology 500? I believe there were a couple of questions on that, and especially in terms of generating the draft report. Yeah. A second question came in actually that, what is your expected turn around time for the secondary analysis starting from PCF to pushing a clinical draft report for all 523 genes to a client?
Answer: Sure. Going from sample in hand to a draft report, obviously, there are a number of variables to go from a sample to on sequencer. And those variables are really related to individual laboratories, and their own processes to isolate DNA and RNA. The frequency of doing that, the number of people who do that, as well as the number of times in a week that library prep is performed and the number of cycles that that library prep is performed. Where in, each sample can enter into that typical workflow. Finally, then resulting in how many times the sequencer is run on a weekly basis. Typically, however, moderate to high volume laboratory, I would expect those workflows to be completed within eight to nine business days. And then following that then, you now have sequence results going to draft report. And those processes are typically only several hours in completing from going from FASTQ to that draft report.
Question: Speaking of draft report, the next question here, you talked about the classification accuracy in one of your slides, and then you mentioned draft versus a signed out classification. And the question here is simply what does the draft classification or draft report mean? And who did that specifically?
Answer: A great question. The definition of a draft report and a draft classification is one wherein our knowledge base and our decision support engine automatically assigned a classification in each variant as well as inferred the most appropriate interpretation before that variant or combination of variance. The final classification is the classification that that variant or those set of variants had in the signed out report, i.e post sign out. We actually store the various classification as assigned by the medical director who signed out that report, and were therefore able to compare about the knowledge base set about a variant versus what the medical director ultimately think about that variant.
And I think it was a good illustration with the upside down pyramid that show the filtering strategies as well, the difference between the validated filter and say a report filter and the final classification. What is automated versus what is done manually, which you just explained. There's a number of components within the CGW software interface as well as the informatics pipeline behind it, that the idea is to get a highly accurate and fast draft report, but then to also provide, as doctor Nagarajan showed, the knowledge base components with the evidence to be able to support the final manual classifications. The next question here is on a validation.
Question: Do you have any recommendations on how the lab should go about validating tumor mutational burden, for example?
Answer: Yes, a very difficult question. I'm not sure there is really a concordance even within the leadership of the molecular pathology space. But my own opinions would be that analytical validation of the actual assay and the tool that is used has been demonstrated by Alumina in this case. And with the data that I showed, my recommendation TMB and MSI would be a clinical sample testing based on orthogonal validated TMB and MSI assays. Even if maybe PCR based for MSI, for example. A TMB based on other TMB values from reference laboratories for example. And the recommendation is to assess that bio specific tumor type to really make tumor type specific recommendations.
Question: Another question here. This one's about reporting and we showed an example report there in a couple of slides. The question is, can that report be customized to any degree or is that the report?
Answer: Wonderful question. The report itself can and is frequently customized by our clinical laboratories. There are a number of customizations that are available within that specific report template that I showed. Mainly, TMB and MSI reporting is specific to the TSO 500. So, the TST 170 or other somatic cancer assays that utilize that report template would not have those fields, and those associated detailed sections show up. Some customers may choose not to report out on clinical trials, for example. That is a component that's modular. There is modularity within order of certain components in the report, as well as options on client specific logos, addresses, patient demographic details that they choose to want to have on their report.
Question: This one is on ... Well, we talked about sharing network in the organizations and pathologists that are part of that sharing network. But what is pure indexes policy for sharing genomic data? Specifically with third party.
Answer: Wonderful question. Sharing network is predicated on enabling our customers and our laboratories, and our health systems to provide the highest quality of clinical care at their respective organizations. And it is in that vein and it is in that nature, that we enable network sharing of information. Our specific contracting language today limits that sharing to clinical use only within our partner organizations. PierianDX or other partners themselves don't have the ability or the right to utilize those data for other purposes than clinical care, and quality purposes as defined in clinical operations.
Yeah. Excellent. Excellent question. Thanks everyone. There are a number of questions here that we're not able to get to, but we've reached our time here. Thanks again Doctor Nagarajan for the great presentation. Thanks everyone for attending and taking time out of your day. We hope this Webinar was useful. I ask here at the end that once the webinar concludes, you'll be presented with a survey. We ask that you answer those questions if you could. It helps direct us in efforts for following up with folks, as well as the next webinars that we put on. I try to ascertain what folks' interests are, and how we can better serve our customers and those who are interested in this field. With that, we'll wrap it up, and want to thank you doctor Nagarajan again. And thanks everyone for attending. Have a great day.