Guided therapy selection in rheumatoid arthritis using a molecular signature response classifier: an assessment of budget impact and clinical utility

BACKGROUND: Patients with moderate to severe rheumatoid arthritis (RA) can be treated with a range of targeted therapies following inadequate response to conventional synthetic disease-modifying antirheumatic drugs such as methotrexate. Whereas clinical practice guidelines provide no formal recommendations for initial targeted therapies, the tumor necrosis factor alpha inhibitor (TNFi) class is the prevalent first-line selection based on clinician experience, its safety profile, and/or formulary requirements, while also being the costliest. Most patients do not achieve adequate clinical response with a first-line TNFi, however. A molecular signature response classifier (MSRC) test that assesses RA-related biomarkers can identify patients who are unlikely to achieve adequate response to TNFi-class therapies. OBJECTIVE: To model cost-effectiveness of MSRC-guided, first-line targeted therapy selection compared with current standard care. METHODS: This budget impact analysis used data sourced from August to September 2020. The prevalence of each first-line targeted therapy was obtained using market intelligence from Datamonitor/Informa PLC Rheumatology Dashboard Forecast 2020, and the average first-year cost of treatment for each class was calculated using wholesale acquisition costs from IBM Micromedex RED BOOK Online. Average effectiveness for each class was based on manufacturer-reported ACR50 response rates (American College of Rheumatology adequate response criteria of 50% improvement at 6 months after therapy initiation). The impact of MSRC testing on first therapy selection was predicted based on a third party-generated decision-impact study that analyzed potential alterations in rheumatologist prescribing patterns after receiving MSRC test reports. Sensitivity analysis evaluated potential impacts of variation in first-year medication cost, adherence to MSRC report, and test price on the first-year cost of treatment. Cost for response (first-year therapy cost therapy divided by probability of achieving ACR50) was compared between standard care and MSRC-guided care. RESULTS: The estimated cost for first-year, standard-care treatment was $65,117, with 80% of patients initiating treatment with a TNFi. Cost for achieving ACR50 response was $177,046. After applying MSRC-guided patient stratification and therapy selection, the first-year cost was $56,543, net of test price, with 49.0% of patients initiating with a TNFi. First-year MSRC-guided care cost, including test price, was estimated at $117,103, a 33.9% improvement over standard care. Sensitivity analysis showed a net cost improvement for guided care vs standard care across all scenarios. Patients predicted to be inadequate TNFi responders, when modeled with lower-priced alternatives, were predicted to show increased ACR50 response rates. Those with MSRC test results indicating a first-line TNFi were predicted to show an ACR50 response rate superior to that for any other class. In this model, if implemented clinically, MSRC-guided care might save the US health care system more than $850 million annually and improve ACR50 by up to 31.3%. CONCLUSIONS: Precision medicine using MSRC-guided patient stratification and therapy selection may both decrease cost and improve efficacy of targeted RA therapies. DISCLOSURES: This work was funded in full by Scipher Medicine Corporation, which participated in data analysis and interpretation and drafting, reviewing, and approving the publication. All authors contributed to data analysis and interpretation and publication preparation, maintaining control over the final content. Arnell, Withers, and Connolly-Strong are employees of and have stock ownership in Scipher Medicine Corporation. Bergman has received consulting fees from AbbVie, Gilead, GlaxoSmithKline, Novartis, Pfizer, Regeneron, Sanofi, and Scipher Medicine and owns stock or stock options in Johnson & Johnson. Kenney, Logan, and Lim-Harashima are consultants for Scipher Medicine Corporation. Basu has nothing to disclose.

Optimization of quality assurance to increase clinical utility and cost effectiveness of hereditary cancer testing

AIM

To evaluate one laboratory's hereditary cancer testing clinical quality assurance (QA) process to minimize test-ordering errors.

METHODS

The proportion of tests canceled/revised due to pre-analytic QA processes or provider consultation prior to test ordering were determined and the resulting health cost savings were estimated.

RESULTS

Over 2000 genetic test orders were canceled/revised over a 1-year period due to the laboratory QA process, saving US$5,801,832 in healthcare costs. Consultation with healthcare providers prior to submitting genetic test requests resulted in 37 canceled/revised test orders in a 2-week period, which extrapolates to a savings of US$3,049,098 over 1 year.

CONCLUSION

QA processes can contribute to the curtailment of healthcare costs through canceling or revising test orders.

Impact of Payer Constraints on Access to Genetic Testing

BACKGROUND

With increased demand for hereditary cancer genetic testing, some large national health-care insurance payers (LNHPs) have implemented policies to minimize inappropriate testing by mandating consultation with a geneticist or genetic counselor (GC). We hypothesized such a restriction would reduce access and appropriate testing.

METHODS

Test cancellation rates (ie, tests ordered that did not result in a reported test result), mutation-positive rates, and turnaround times for comprehensive BRCA1/2 testing for a study LNHP that implemented a GC-mandate policy were determined over the 12 months before and after policy implementation (excluding a 4-month transition period). Cancellation rates were evaluated based on the reason for cancellation, National Comprehensive Cancer Network testing criteria, and self-identified ancestry. A control LNHP was evaluated over the same period for comparison.

RESULTS

The study LNHP cancellation rate increased from 13.3% to 42.1% ( P < .001) after policy implementation. This increase was also observed when only individuals who met National Comprehensive Cancer Network criteria for hereditary breast and ovarian cancer testing were considered (9.5% to 37.7%; P < .001). Cancellation rates increased after policy introduction for all ancestries; however, this was more pronounced among individuals of African or Latin American ancestry, for whom cancellation rates rose to 48.9% and 49.6%, respectively, compared with 33.9% for individuals of European ancestry. Over this same time period, control LNHP cancellation rates decreased or stayed the same for all subgroups.

CONCLUSION

These findings demonstrate that a GC-mandate policy implemented by a LNHP substantially decreased access to appropriate genetic testing, disproportionately impacting minority populations without any evidence that inappropriate testing was decreased.

Erratum to: Incidence of BRCA1 and BRCA2 non-founder mutations in patients of Ashkenazi Jewish ancestry

In Table 2 of the original publication, the HGVS and legacy nomenclature were mismatched and the HGVS nomenclature did not correlate with data listed in the table. The corrected table is listed below.

Incidence of BRCA1 and BRCA2 non-founder mutations in patients of Ashkenazi Jewish ancestry

An estimated 1:40 individuals of Ashkenazi Jewish (AJ) ancestry carry one of three common founder mutations in BRCA1 or BRCA2, resulting in the inherited cancer condition, Hereditary Breast and Ovarian Cancer (HBOC) syndrome. Targeted testing for these three mutations (BRCA1 187delAG, BRCA1 5385insC, and BRCA2 6174delT) is therefore recommended for all AJ breast and ovarian cancer patients, regardless of age of diagnosis or family history. Comprehensive analysis of both genes is recommended for a subset of AJ patients in whom founder mutations are not identified, but estimates of the yield from comprehensive analysis in this population vary widely. We sought to determine the proportion of non-founder mutations as a percentage of all mutations in BRCA1 and BRCA2 among AJ patients to inform decisions about HBOC testing strategies in this population. We analyzed the genetic testing results for 37,952 AJ patients for whom clinical testing of BRCA1 and BRCA2 was performed at Myriad Genetic Laboratories from January 2006 through August 2013. Analysis was limited to AJ-only patients for whom the initial test order was either (1) comprehensive testing, or (2) founder mutation testing with instructions to automatically "reflex" to comprehensive analysis if negative. Cases were excluded if a separate follow-up order was placed to reflex to comprehensive analysis only after the founder mutation testing was reported out as negative. Among all BRCA1 and BRCA2 mutations detected in these groups, the percentage of non-founder mutations was 13 % (104/802) and 7.2 % (198/2,769). One-hundred and eighty-nine unique non-founder mutations were detected, 76 in BRCA1 and 113 in BRCA2. Non-founder mutations make up between 7.2 and 13.0 % of all BRCA1 and BRCA2 mutations in Ashkenazi Jews. A wide range of mutations are present, most of which are also seen in non-AJ individuals.

Design and validation of an oligonucleotide microarray for the detection of genomic rearrangements associated with common hereditary cancer syndromes

Background

Conventional Sanger sequencing reliably detects the majority of genetic mutations associated with hereditary cancers, such as single-base changes and small insertions or deletions. However, detection of genomic rearrangements, such as large deletions and duplications, requires special technologies. Microarray analysis has been successfully used to detect large rearrangements (LRs) in genetic disorders.

Methods

We designed and validated a high-density oligonucleotide microarray for the detection of gene-level genomic rearrangements associated with hereditary breast and ovarian cancer (HBOC), Lynch, and polyposis syndromes. The microarray consisted of probes corresponding to the exons and flanking introns of BRCA1 and BRCA2 (≈1,700) and Lynch syndrome/polyposis genes MLH1, MSH2, MSH6, APC, MUTYH, and EPCAM (≈2,200). We validated the microarray with 990 samples previously tested for LR status in BRCA1, BRCA2, MLH1, MSH2, MSH6, APC, MUTYH, or EPCAM. Microarray results were 100% concordant with previous results in the validation studies. Subsequently, clinical microarray analysis was performed on samples from patients with a high likelihood of HBOC mutations (13,124), Lynch syndrome mutations (18,498), and polyposis syndrome mutations (2,739) to determine the proportion of LRs.

Results

Our results demonstrate that LRs constitute a substantial proportion of genetic mutations found in patients referred for hereditary cancer genetic testing.

Conclusion

The use of microarray comparative genomic hybridization (CGH) for the detection of LRs is well-suited as an adjunct technology for both single syndrome (by Sanger sequencing analysis) and extended gene panel testing by next generation sequencing analysis. Genetic testing strategies using microarray analysis will help identify additional patients carrying LRs, who are predisposed to various hereditary cancers.

Analysis of current testing practices for biallelic MUTYH mutations in MUTYH-associated polyposis

MUTYH-associated polyposis (MAP) is an autosomal recessive syndrome caused by biallelic mutations in the base excision repair gene MUTYH. Owing to potential limitations in the MAP testing strategy and testing criteria, it is possible that MAP is being under-identified both genotypically and phenotypically. To determine whether full sequencing of MUTYH would increase clinical sensitivity over a founder mutation (FM) strategy, a retrospective analysis of two datasets from a commercial clinical laboratory was performed. The first cohort contained 1522 individuals who received MUTYH analysis for two FMs with subsequent full-gene sequencing. Eighty-five biallelic individuals were identified; 47 carried two FMs, 17 carried one FM and one mutation identified on full sequencing, and 21 carried biallelic mutations identified only on full sequencing. The second cohort contained 921 patients with colorectal cancer <50 years and <10 reported colorectal adenomas who had undergone MUTYH mutation testing. In this cohort, 19 of 921 (2.1%) individuals were identified as biallelic MUTYH carriers. Of these, 13 did not have a personal or family history of polyps and would not have met guidelines for MUTYH testing. These results suggest that individuals with biallelic MUTYH mutations are under-ascertained based on both genotype and phenotype under current standard testing practices.

Proffered Papers and Posters Submitted to the Fifth International Symposium on Hereditary Breast and Ovarian Cancer, BRCA: Twenty Years of Advances

Objectives: It is estimated that 1–2% of individuals of Ashkenazi Jewish (AJ) ancestry carry one of three pathogenic founder mutations in BRCA1 and BRCA2. Targeted testing for these mutations (BRCA1 187delAG and 5385insC, and BRCA2 6174delT) is therefore recommended for all AJ breast and ovarian cancer patients, regardless of age of diagnosis or family history. Comprehensive analysis of both genes is recommended for a subset of AJ patients in whom founder mutations are not identified, but estimates of the yield from comprehensive analysis in this population vary widely. Methods: We sought to establish the proportion of non-founder mutations in AJ patients undergoing clinical testing in our laboratory from January 2006 through August 2013. Analysis included AJ patients for whom: (1) comprehensive testing was ordered as the initial test, or (2) founder mutation testing was ordered with instructions to “reflex” to comprehensive analysis if negative. The latter group was limited to cases where the reflex testing was ordered on the original test request form, and not cancelled for any reason other than the detection of a founder mutation. Results: The percentage of non-founder mutations detected in these groups was 13% (104/802) and 7.2% (198/2769) respectively. We detected 189 unique non-founder mutations, 76 in BRCA1 and 113 in BRCA2. BRCA2 4075delGT was detected in 15 patients. The next most common mutations, found in 7 patients each, were BRCA1 5055delG, BRCA2 1982delA, and BRCA2 R3128X. Conclusions: Non-founder mutations make up between 13% and 7.2% of BRCA1 and BRCA2 mutations in patients reporting AJ ancestry. These numbers may represent underestimates if some patients were ascertained for testing based on the identification of a founder mutation in a relative. These numbers suggest that the prevalence of non-founder mutations in AJ individuals may be comparable to the prevalence of BRCA1/2 mutations in non-AJ individuals.

Overlapping phenotypes with AFAP and FAP in patients with Lynch syndrome gene mutations

402

Background: Lynch syndrome is a well-described cause of hereditary colon cancer. A point of the Amsterdam II criteria, which defines Lynch syndrome, is to exclude familial adenomatous polyposis (FAP). FAP is caused by mutations in the APC gene and is defined as the presence of 100+ adenomas. The attenuated form of familial adenomatous polyposis (AFAP) can be considered when an individual has 10-99 adenomas. While adenomatous polyps may be a part of the Lynch syndrome phenotype (as a precursor to carcinoma), the assumption is that Lynch syndrome and APC-associated polyposis do not have overlapping phenotypes. Methods: A retrospective analysis was performed on 8,202 individuals with a personal history of adenomatous polyps that had clinical genetic testing for Lynch syndrome (including genes MLH1, MSH2, MSH6, PMS2, and EPCAM) between January 2006 and July 2013. Patient adenoma history was collected on the test request form. The inclusion criterion was limited to personal history of adenomatous polyps and did not depend on personal or family history of cancer. Patients were excluded if the test performed was either targeted mutation testing or single gene testing, presumably based on prior immunohistochemistry tumor testing. Results: Of 8,202 patients with a personal history of adenomatous polyps that underwent Lynch syndrome testing, 610 (7.4%) were positive for a mutation. Mutations were detected in patients with a wide distribution of cumulative adenomas (Table). Seventy-five of the patients with a Lynch syndrome mutation had an adenomatous polyp phenotype suggestive of either FAP (100+ adenomas) or AFAP (10-99 adenomas). Of these 75 patients, 19 underwent APCtesting, all of whom were negative for a mutation. Conclusions: Individuals with an adenoma history, who were tested for Lynch syndrome, had an overall 7.4% mutation positive rate, indicating an overlapping phenotype between Lynch syndrome and FAP/AFAP. This overlap supports consideration of a gene panel test approach. [Table: see text]

A comprehensive laboratory-based program for classification of variants of uncertain significance in hereditary cancer genes

Genetic testing has the potential to guide the prevention and treatment of disease in a variety of settings, and recent technical advances have greatly increased our ability to acquire large amounts of genetic data. The interpretation of this data remains challenging, as the clinical significance of genetic variation detected in the laboratory is not always clear. Although regulatory agencies and professional societies provide some guidance regarding the classification, reporting, and long-term follow-up of variants, few protocols for the implementation of these guidelines have been described. Because the primary aim of clinical testing is to provide results to inform medical management, a variant classification program that offers timely, accurate, confident and cost-effective interpretation of variants should be an integral component of the laboratory process. Here we describe the components of our laboratory's current variant classification program (VCP), based on 20 years of experience and over one million samples tested, using the BRCA1/2 genes as a model. Our VCP has lowered the percentage of tests in which one or more BRCA1/2 variants of uncertain significance (VUSs) are detected to 2.1% in the absence of a pathogenic mutation, demonstrating how the coordinated application of resources toward classification and reclassification significantly impacts the clinical utility of testing.

BRCA1/2 mutation prevalence among triple-negative breast cancer patients from a large commercial testing cohort

1544

Background: BRCA1/2 deleterious mutation identification among triple-negative breast cancer (TNBC) patients has gained importance due to cancer-risk management implications for patients and their relatives, and also has an emerging role in guiding treatment selection for therapies such as PARP inhibitors. The National Comprehensive Cancer Network (NCCN) currently recommends BRCA1/2 testing for TNBC patients diagnosed at age <60. Mutation prevalence among TNBC patients has previously been studied only in small regionalized cohorts. A recent study in unselected patients using the updated definitive criteria for TNBC reported mutation prevalence as 10.6%. Methods: Following the 2011 NCCN Hereditary Breast and Ovarian Cancer (HBOC) Testing Criteria update, serial cohorts of > 5,000 Ashkenazi Jewish and > 65,000 non-Ashkenazi Jewish breast cancer patients undergoing commercial BRCA1/2 testing were analyzed. Age at diagnosis, ethnicity, and provider-reported TN status were obtained from test requisition forms completed by ordering providers, and correlated with test results. Neither the accuracy nor definitive criteria used for TN status reported was independently verified. Results: Incidence of TNBC was reported as 9.7% among non-Ashkenazi patients and 16.5% within the subset with African ancestry. Incidence of TNBC was reported as 4.5% among Ashkenazi patients, but this is likely affected by test ordering for this population. The Table displays the BRCA1/2mutation rates classified by ethnicity and age-group. Conclusions: This study provides the most robust estimate to date of BRCA1/2 mutation prevalence among TNBC patients of all ages. The mutation rates seen among TNBC patients diagnosed after age 60 also illustrate the importance of testing such patients who may not meet the current NCCN HBOC testing criteria. [Table: see text]

Founder mutation versus full sequencing of MYH for increasing clinical sensitivity

343

Background: MYH-associated polyposis is an autosomal recessive syndrome caused by biallelic mutations in the base excision repair gene MYH. Initial reports indicated that a majority of European biallelic individuals carried two founder mutations, Y165C and G382D. A common MYH analysis strategy involves evaluation of the two founder mutations (FMs) with subsequent full sequencing only if one of the FMs is identified. This study aimed to determine the sensitivity of full MYH sequencing over the strategy described above in a cohort of individuals that has undergone genetic testing by a commercial laboratory. Methods: A retrospective analysis was performed on 1,522 individuals who had clinical MYH testing ordered either independently or in conjunction with APC gene analysis. All patients underwent MYH analysis for Y165C and G382D. Subsequent full gene sequencing was performed for all patients, except for those biallellic for the two FMs. Demographic, personal, and family cancer histories were collected on the test request form. Results: 86 biallelic individuals were identified, with 47 carrying two FMs, 18 carrying one FM and one mutation identified on sequencing, and 21 carrying biallelic mutations identified only on full sequencing. 21/86 (24.4%) biallelic MYH mutation carriers would have been missed by FM analysis only. The mutation spectrum is distributed among different ethnicities (Table). The majority of individuals met clinical criteria for a polyposis syndrome, with 21 individuals reporting more than 99 adenomas (24.4%) and 50 individuals reporting 20-99 adenomas (58.1%). Conclusions: Analysis of only the two MYH FMs would miss 24% of biallelic individuals in our cohort. Due to the ethnicity distribution of mutations, it is difficult to predict who would be missed by this strategy. Consideration of MYH full sequencing is warranted to achieve highest clinical sensitivity. [Table: see text]

Mutation analysis of PALB2 in high-risk and lower-risk patients negative for BRCA1 and BRCA2 mutations

30

Background: PALB2 has been identified as a breast cancer susceptibility gene conferring ~ 2-4 fold increased risk of breast cancer. A number of studies have estimated the PALB2 mutation prevalence to range from 0.5% - 2.9% in populations of breast cancer patients. We performed an analysis to determine the PALB2 mutation prevalence in a large U.S. referral testing population. Methods: DNA samples were anonymized from two subsets of patients: 955 early onset breast cancer patients with severe family history, and 524 patients with later onset of breast cancer and/or less severe family history. All patients were negative for deleterious sequence mutations or large rearrangements in BRCA1 and BRCA2. Results: We identified 10 disease associated PALB2 mutations in the high risk group of 955 patients and 2 deleterious PALB2 mutations in the lower risk group of 524 patients. Identified PALB2 mutations included 8 nonsense, 3 frameshift mutations and a splice site mutation. The mutation prevalence for the high risk population was 1.05% (95% C.I., 0.5 -1.92) whereas that for the lower risk population was 0.38% (95% C.I., 0.05-1.37). The observed rate of PALB2 variants of unknown significance (VUS) identified in this study was ~5% (78 VUS were in 75 of the 1479 patients that were tested). Our variant classification program which successfully decreased the VUS rate in BRCA1 and BRCA2 is similarly expected to enhance mutation classification on an on-going basis for PALB2 genetic testing. Conclusions: Genetic testing for PALB2 may be indicated as a reflex test for breast cancer patients who test negative for BRCA1 and BRCA2.

Clinical significance of large rearrangements in BRCA1 and BRCA2

BACKGROUND: Current estimates of the contribution of large rearrangement (LR) mutations in the BRCA1 (breast cancer 1, early onset) and BRCA2 (breast cancer 2, early onset) genes responsible for hereditary breast and ovarian cancer are based on limited studies of relatively homogeneous patient populations. The prevalence of BRCA1/2 LRs was investigated in 48,456 patients with diverse clinical histories and ancestries, referred for clinical molecular testing for suspicion of hereditary breast and ovarian cancer. METHODS: Sanger sequencing analysis was performed for BRCA1/2 and LR testing for deletions and duplications using a quantitative multiplex polymerase chain reaction assay. Prevalence data were analyzed for patients from different risk and ethnic groups between July 2007 and April 2011. Patients were designated as “high-risk” if their clinical history predicted a high prior probability, wherein LR testing was performed automatically in conjunction with sequencing. “Elective” patients did not meet the high-risk criteria, but underwent LR testing as ordered by the referring health care provider. RESULTS: Overall BRCA1/2 mutation prevalence among high-risk patients was 23.8% versus 8.2% for the elective group. The mutation profile for high-risk patients was 90.1% sequencing mutations versus 9.9% LRs, and for elective patients, 94.1% sequencing versus 5.9% LRs. This difference may reflect the bias in high-risk patients to carry mutations in BRCA1, which has a higher penetrance and frequency of LRs compared with BRCA2. There were significant differences in the prevalence and types of LRs in patients of different ancestries. LR mutations were significantly more common in Latin American/Caribbean patients. CONCLUSIONS: Comprehensive LR testing in conjunction with full gene sequencing is an appropriate strategy for clinical BRCA1/2 analysis.