Introduction

The American College of Medical Genetics and Genomics (ACMG) previously published guidance for reporting secondary findings in the context of clinical exome and genome sequencing (ES/GS) in 2013 and 2017.1,2 These recommendations were developed by the ACMG Secondary Findings Maintenance Working Group (SFWG), which was convened by the ACMG Board of Directors (BOD) to evaluate the need for a minimum list of genes that should be evaluated in individuals undergoing clinical ES/GS based on the medical actionability of the associated condition. In the past, policy recommendations concerning what types of variants to return along with lists of which genes to analyze were included. Given the increase in uptake of clinical ES/GS, the ACMG SFWG and BOD have agreed the list of recommended genes should now be updated annually. Policy updates surrounding the purpose, scope, and process for maintaining the ACMG Secondary Findings List are being published separately,3 and will be updated separately, as needed. It is important to reiterate here that use of the SF results should not be a replacement for indication-based diagnostic clinical genetic testing.

The goal of the SF gene list is to guide clinical laboratories as to which medically actionable genes unrelated to the indication for testing should be evaluated as part of clinical ES/GS, while maintaining a minimum list to balance the interests of patients with the additional burden placed on laboratories providing sequencing. The SFWG members took several aspects of the associated phenotype into consideration to evaluate genes for this list, including the actionability, severity, penetrance, and impact and/or burden of available treatment modalities or screening recommendations. The SFWG was also mindful to recommend genes where the majority of pathogenic variants are detectable by ES/GS. For instance, no gene–phenotype pairs caused by trinucleotide repeats were considered for this list. Even with these restrictions, there are still many gene–phenotype pairs that could be considered for inclusion on the ACMG SF list; however, the SFWG and BOD felt a duty to keep this list to a manageable number. Therefore, members worked toward making compromises by, for example, avoiding inclusion of disorders that would typically be diagnosed clinically, disorders where timing of the diagnosis was not critical for treatment efficacy, or disorders where a lifestyle change was the prominent intervention (e.g., avoiding tobacco use). Here, we present the ACMG SF v3.0 list, its development using the policies described in the ACMG SF Policy Statement and our rationale for and against inclusion of considered genes.

WORKING GROUP OPERATIONS

The 2018–2021 SFWG is composed of six biochemical, molecular, and/or cytogenetics clinical laboratory directors, five clinical geneticists of differing subspecialities, two genetic counselors, two cardiologists, one PhD medical geneticist, one pharmacogenomics expert, and one patient advocate. An ACMG board liaison was added to support clear communication of standards and expectations between the Board and the SFWG. The SFWG meets at least monthly via virtual web conferencing and also in-person during the ACMG and American Society of Human Genetics annual conferences to review nomination forms and vote on inclusion or exclusion of gene–phenotype topics. For all meetings, regardless of whether they are virtual or in-person, we follow established ACMG committee and working group policies for review of nominations and voting.

NOMINATION AND REVIEW PROCESS

SFWG members began the nomination and review process by evaluating genes and phenotypes from the SF v2.0 list to assess their appropriateness to remain on the SF v3.0 list. The committee also reconsidered genes that were nominated, but not included, on previous versions of the SF list. The committee then considered gene–phenotype pairs that scored a total of 10 or higher for actionability by the ClinGen Actionability Working Group as of August 2018.4 Finally, the SFWG used the actionable gene lists from the eMERGE Network and the French Society of Predictive and Personalized Medicine on hereditary cancer genes to identify genes for review.5,6

Nominations for gene–phenotype pairs to add to or remove from the SF list were accepted from ACMG members via a nomination form (ACMG Secondary Findings Panel Nomination Form) that was developed through a subcommittee of the SFWG.7 Internal nominations from SFWG members included CASQ2/catecholaminergic polymorphic ventricular tachycardia (CPVT), DICER1/DICER1-related hereditary cancer, FLNC/FLNC-related cardiomyopathy, NOTCH3/CADASIL, RPE65/RPE65-related retinopathy, TRDN/CPVT and long QT syndrome (LQTS) and TTN/cardiomyopathy. All externally submitted nominations were also considered; the committee received nominations for HNF1A/MODY3 and HNF1B/MODY5, PRKAR1A/Carney complex, SERPINA1/alpha-1-anti-trypsin deficiency and TTR/TTR-associated amyloidosis.

Based on their expertise, SFWG members were split into one of four subgroups (hereditary cancer, inborn errors of metabolism (IEM), cardiovascular, or miscellaneous) and pared down the final list of genes for review by the full SFWG. However, all nominations from the community were put forth for full review and consideration.

Genes that underwent full review were presented to the entire SFWG by a member of the corresponding subgroup. Nomination forms were circulated to the membership prior to meetings and presented by one member for consideration. After discussion, a motion to include or exclude the gene from the v3.0 list was made and seconded, which prompted a vote requiring consensus to include or exclude genes from the SF v3.0 list.

The final proposed ACMG SF v3.0 list from the SFWG was sent to the ACMG BOD for ratification with a summary of the SFWG discussion, voting outcome, and a recommendation for the suggested update to the SF minimum list. The BOD reviewed each recommendation on a gene-by-gene basis in November 2020.

RECOMMENDATIONS FOR THE ACMG SF v3.0 LIST

The overall goal of the SFWG is to recommend a minimum list of genes that places limited excess burden on patients and clinical laboratories while maximizing the potential to reduce morbidity and mortality when ES/GS is being performed. Table 1 includes the complete list of genes on the v3.0 list. A searchable, and sortable, list is available in Supplemental Table 1. No genes were removed between the v2.0 and v3.0 lists. There is a total of 73 genes on the SF v3.0 list. A list of newly added genes to the v3.0 list is shown in Table 2. A list of genes considered for inclusion, but ultimately excluded from the v3.0 list are outlined in Table 3. A number of genes have been placed on a “watchlist” to review for future versions of the SF list, particularly those that lack sufficient data as to their penetrance.

Table 1 ACMG SF v3.0 gene and associated phenotypes recommended for return as secondary findings from clinical exome and genome sequencing.
Full size table
Table 2 New gene–phenotype pairs for secondary findings (SF) list.
Full size table
Table 3 Genes not selected for secondary findings (SF) list v3.0 and reasoning.
Full size table

CONSIDERATIONS FOR SPECIFIC PHENOTYPE CATEGORIES

Genes related to cancer phenotypes

The cancer subgroup prioritized new genes for consideration by selecting 13 genes underlying seven hereditary cancer phenotypes. Relevant, recent literature on phenotype, penetrance, and actionability was curated from a gene-focused search of OMIM, GeneReviews, and PubMed, as well as the expertise of the subgroup. Technical issues of sequencing the genes were reviewed with relevant members of the SFWG.

Recommended for addition to the SF v3.0 list

A recent, international study of individuals heterozygous for a PALB2 pathogenic variant from 524 families estimated that the absolute risk of developing breast cancer by age 80 years varies from 52% (95% CI: 42–62%) for a female with an unaffected mother at age 50 years and unaffected maternal grandmother at age 70 years to 76% (95% CI: 69–83%) for a female with two affected first-degree relatives.8 Quantified risks of developing ovarian cancer and pancreatic cancer risk were much lower. Pediatric cancer (osteosarcoma, leukemia, brain tumors, and soft-tissue sarcoma) has also been reported in PALB2 heterozygotes, but absolute risk is uncertain.9 Management of risk in individuals heterozygous for pathogenic PALB2 variants is similar to that for the BRCA1 and BRCA2 genes; however, given the overall lower range of PALB2-associated risk in breast and ovarian cancer, individualized estimates are important for management decisions.10

Germline variants in MAX and TMEM127 are rare (1–2% each) causes of hereditary paraganglioma/pheochromocytoma, a well-established phenotype on the ACMG SF list.11 A large, longitudinal international investigation showed a high penetrance for pathogenic variants in both genes, although data is still limited.12

Not recommended for addition to the SF v3.0 list

As listed in Table 3, several cancer genes were reviewed and discussed but not included on the ACMG SF list for numerous reasons, even for genes with well-established phenotypes. For example, the workgroup voted not to include SDHA gene due to poor analytical specificity related to high sequence homology, although other genes that cause hereditary paraganglioma/pheochromocytoma are included on the list. Other technical difficulties were noted for genes such as EPCAM associated with Lynch syndrome and GREM1-associated polyposis, where routine detection of common deletions or duplications could be difficult at this time by ES/GS in many laboratories. Lower penetrance was also an important consideration, especially in genes such as RAD51C, RAD51D, and BRIP1 that predispose to risk for ovarian cancer, given the uncertainties in how best to manage risk, difficulty of surveillance, and morbidity of intervention. For other genes (BAP1, DICER1, POLE, POLD1), there remains uncertainty about phenotype, risk, and penetrance.

Genes related to cardiovascular phenotypes

Cardiovascular genes have been represented on the SF list since its inception, due to the morbidity and mortality of sudden cardiac death (SCD) and heart failure (HF), which can both be treated or prevented with well-established interventions.13,14

Primary arrhythmia risk, which leads to presyncope, syncope, and SCD, arises in genes encompassed by the channelopathies. With established risk, the use of antiarrhythmic medications or implantable cardioverter defibrillators (ICDs) can greatly reduce the risk of SCD and morbidity. The cardiomyopathies, classified as diseases of the myocardium, can also cause lethal arrhythmias. The cardiomyopathies also lead to heart failure, itself a morbid and mortal condition, but one that is highly amenable to medical and device therapies. With this in mind, the SFWG reviewed the evidence for nominated cardiovascular genes with a particular focus on the medical actionability of a potential SF, the penetrance and expressivity of the given gene, and the potential burden on providers and clinical laboratories should the gene be included.

Recommended for addition to the SF v3.0 list

There is strong evidence that pathogenic and likely pathogenic (P/LP) variants in FLNC significantly predispose individuals to high-risk dilated and arrhythmogenic cardiomyopathies; these often first manifest as sudden cardiac death.15,16,17 The SFWG voted to include this gene based on its high penetrance, severity of the phenotype if untreated, and the strong potential benefit of intervention based on returning P/LP variants in this gene as a SF.

TTN, the largest single gene in the human genome, has long been associated with dilated cardiomyopathy, and clinical intervention based on TTN variants that are P/LP can afford significant benefit to patients and their families. However, both its considerable length and high variant burden previously have stymied attempts to measure penetrance and made interpretation of TTN variants a challenge for clinical laboratories and clinicians alike. For these reasons, TTN had been previously considered by the SFWG, but ultimately not recommended for inclusion. Since the last iteration of the guidelines, however, new data on penetrance and expressivity derived from large population cohorts necessitated that the SFWG reconsider this gene.18 This new evidence indicated significant risk for cardiomyopathy among those with TTN truncating variants (TTNtv), specifically TTNtv in exons that are highly expressed. Further, TTNtv variants are far less frequent than missense variants in TTN (TTNtv found in 0.5–1% of the overall population) and thus identification and reporting of TTNtv variants was considered warranted and with limited burden to clinical laboratories in the assessment of this large gene. As such, the SFWG voted to include TTN on the current iteration of the list, with the critical caveat that only TTN truncating variants be returned as SF.

Pathogenic variants in the CASQ2 gene are associated with autosomal recessive catecholaminergic polymorphic ventricular tachycardia (CPVT), which commonly presents in childhood or adolescence. As with other forms of CPVT, the clinical presentation is heralded by sudden death during exercise. Patients are otherwise asymptomatic at rest and have normal structural hearts on cardiac imaging. Exercise treadmill testing provokes the typical polymorphic ventricular arrhythmia characteristic of CPVT. Treatment is highly effective, either in the form of antiarrhythmic medical therapy, or with ICD in some cases. This condition is often lethal when unrecognized, and as such the SFWG voted to include CASQ2 to the SF list for LP/P variants detected in trans or apparently homozygous variants.

TRDN is associated with autosomal recessive CPVT or an atypical form of long QT syndrome, depending on the appearance of the resting ECG. Common to all presentations is an early age of onset (<10 years) of exercise-induced sudden cardiac death. In some cases, evidence of skeletal myopathy coexists with the cardiac manifestations. Early recognition of this condition may lead to appropriate intervention in the form of antiarrhythmic therapy or ICD. In view of the early onset of disease and lethality, the SFWG voted to include TRDN to the SF list for the recessive state in which two LP/P variants are detected in trans or apparently homozygous variants.

Not recommended for addition to the SF v3.0 list

As with many other SF genes, population-based penetrance estimates are lacking for most cardiovascular genes, particularly those derived from population cohorts not ascertained for cardiovascular phenotypes. As such evidence continues to amass, we recognize that some additional “watchlist” genes not included here may meet the standard for inclusion. This includes genes associated with dilated cardiomyopathy (e.g., BAG3, DES, RBM20, TNNC1), which have evidence showing similar or greater risk of morbidity and mortality as other cardiomyopathy genes already included. Additionally, CALM1, CALM2, and CALM3, three separate genes all encoding the identical protein, have accumulated evidence supporting their cause of an atypical form of LQTS presenting in the neonatal period or early childhood, at times associated with developmental delay and seizure. As this condition usually does not escape diagnosis, and the role of variants in these three genes in adult disease presentations remains unclear, these genes have not yet been added to the SF list. The workgroup’s new policy to update the guidelines more regularly will facilitate a stringent but more agile approach to review emerging evidence for these genes and for their overall suitability for inclusion on the SF list.

Genes related to phenotypes associated with inborn errors of metabolism

When considering IEM, the SFWG first considered the broader question of whether all genes and disorders on the Recommended Uniform Screening Panel (RUSP) should be reviewed and considered for inclusion.19 Newborn screening (NBS) for disorders on the RUSP is recommended by the Department of Health and Human Services. Most states test for the majority of the recommended disorders, and some states test for additional disorders. The abundance of data associated with state screening programs, including validity of testing methodologies employed currently, are already in place and have been so for many years for many IEMs.20 Assays to measure analytes are generally more clinically sensitive to identify an IEM than molecular analysis for secondary findings, with likely limited yield for the latter if the patient had NBS. A secondary consideration noted by the SFWG would be the added cost for analysis and counseling that would be associated with the addition of more than 30 disorders to the SF list.

The SFWG, therefore, considered the following when deciding whether to review and approve an IEM for inclusion in the secondary findings list: (1) the existence of a juvenile or later-onset form of the disorder and that early or presymptomatic diagnosis of late-onset disease is unlikely for disorders recently added to the RUSP, (2) that the late-onset form should be highly medically actionable, and (3) that there appear to be a significant number of undiagnosed cases in the population.

Recommended for addition to the SF v3.0 list

Biotinidase deficiency, due to pathogenic variants in the BTD gene, was reviewed based on its high actionability score in ClinGen.21 Its addition is recommended based on the severity of clinical symptoms in a significant proportion of undiagnosed older individuals at risk for disease, ease of confirmatory diagnosis by enzyme assay, and effectiveness and ease of treatment with lifelong oral biotin.22

Pompe disease caused by recessive pathogenic variants in the acid α-glucosidase (GAA) gene was added to the RUSP in 2015. However, as of October 2020, only 23 states and Washington, DC were performing NBS for the disorder.23 Although the number of states screening is likely to increase over time, NBS may fail to diagnose later-onset, milder forms of the disorder. Given the availability of FDA-approved effective enzyme replacement therapy (ERT), we recommend adding GAA/Pompe as a SF to facilitate detection of later-onset cases and in older individuals who were not screened as newborns.24,25

While Fabry disease was included in the original SF recommendations under the disease category of cardiomyopathy, the workgroup recommends that the gene–phenotype association be broadened in affected males and females to include all P/LP variants associated with any disease manifestation(s), including significant risk for stroke and renal disease.26,27

Not recommended for addition to the SF v3.0 list

X-linked adrenoleukodystrophy (ALD) was added to the RUSP in 2016. As of October 2020, 18 states and Washington, DC perform NBS for ALD.23 The classic cerebral form of the disorder in affected males is associated with an early onset (4–8 years) and rapid progression of disease. While treatment is available in the form of hematopoietic stem cell transplantation with early stage cerebral disease, it is associated with significant morbidity and mortality and success depends upon early treatment.28,29 Therapy for later-onset cases in affected males and females is currently supportive. For these reasons, the SFWG assessed that, at the present time, NBS should be the focus, allowing presymptomatic diagnosis and the opportunity for more timely medical treatment and appropriate counseling. With NBS, it is unlikely many additional individuals would be diagnosed as a secondary finding.

The review and possible inclusion of additional lysosomal storage disorders was briefly discussed by the SFWG, particularly for forms with later onset. However, the SFWG decided that inclusion on NBS panels for some (such as Hurler syndrome), as well as the low likelihood of presymptomatic diagnosis and/or effective treatment for others, did not warrant their inclusion at this time.

For additional IEMs on the NBS list, such as organic acidemias and fatty acid oxidation disorders, the SFWG decided that insufficient numbers of additional asymptomatic patients would be secondarily diagnosed to warrant addition.

Genes related to miscellaneous phenotypes

The SFWG also reviewed nominations for genes that cause phenotypes outside of the core disease review groups. This subgroup reviewed 13 genes associated with 11 different phenotypes, and ultimately approved 4 genes to be added to the v3.0 list.

Recommended for addition to the SF v3.0 list

Hereditary hemorrhagic telangiectasia (HHT) was considered for inclusion on the ACMG SF v3.0 list, and it was ultimately decided that the ACVRL1 and ENG genes should be included. We acknowledge that the SMAD4 gene also contributes to this phenotype; however, this gene was previously placed on the list due to its association with juvenile polyposis syndrome. The HHT phenotype was added to the SF list largely due to disease severity, medical management recommendations, and disease penetrance.30 Inclusion of the GDF2 gene, which is also associated with HHT, was not considered at this time due to the small number of reported cases.31 Of note, the ACVRL1 and ENG gene have also been considered associated with hereditary pulmonary hypertension; however, review for association with the HHT phenotype only was used to include these genes on the v3.0 list.32,33,34,35

We assessed two nominated genes for maturity-onset diabetes of the young (MODY). MODY is somewhat atypical and can therefore be difficult to correctly diagnose among diabetic patients and may go undiagnosed for many years. Untreated or poorly controlled diabetes, including MODY, leads to complications including cardiovascular disease, renal disease, neuropathy, and retinopathy. Therefore, early and effective treatment is important. MODY3 is associated with pathogenic variants in HNF1A, which accounts for approximately 30–65% of MODY cases. MODY3 does not require insulin treatment and responds well to low dose oral sulfonylureas, typically lower doses than are customary for most type 2 diabetics.36 Furthermore, newborns can have transient neonatal hyperinsulinemic hypoglycemia that can lead to lifelong disabilities if hypoglycemia is not quickly recognized and treated. More than 95% of HNF1A pathogenic variants are detectable with ES. In contrast, MODY5 due to variants in HNF1B accounts for only <5% of MODY, and ~50% of pathogenic variants in HNF1B are due to deletions that are not readily detected by ES if copy-number analysis is not included.37 For this reason, only HNF1A was recommended for the SF list at this time.

The SFWG concluded that only the HFE p.Cys282Tyr variant associated with hereditary hemochromatosis should be reported from ES/GS testing and only when found in the homozygous state after deliberation. The SFWG recognized the lower penetrance levels for all other genotypes in the HFE gene, such as HFE p.His63Asp/p.Cys282Tyr compound heterozygotes and HFE p.His63Asp homozygotes. Newer studies show penetrance rates of severe iron overload to be as high as 35% and severe liver disease in 9–24% among male p.Cys282Tyr homozygotes, including larger studies without ascertainment bias.38 There is a highly effective follow-up laboratory testing (i.e., serum transferrin–iron saturation assay) that can indicate who would benefit from undergoing phlebotomy and/or iron chelation treatment.39 Additionally, this condition can easily escape detection before significant organ damage occurs, which can be prevented should treatment be initiated before significant iron overload takes place.

RPE65-associated retinopathy was nominated for inclusion on the v3.0 list due to recent availability of an FDA-approved gene replacement therapy. Individuals with biallelic pathogenic variants in RPE65 have a range of age of onset that is likely dependent on severity and combination of biallelic variants, but can be associated with symptoms, including nystagmus, at or shortly after birth.40 As the condition progresses, individuals experience a decrease in their visual field and deterioration of color vision and central visual acuity. Milder forms may present later in childhood or early adulthood, and symptoms of early retinal deterioration can be missed or overlooked. Lack of treatment over time causes devastating vision impairment or complete loss. While ongoing long-term data are still being collected, the therapy depends on viable retinal cells, and thus, may be more advantageous if administered earlier rather than later in the disease course.41,42 Therefore, the SFWG felt there was the potential for significant benefit to patients by adding this gene to the SF gene list.

Not recommended for addition to the SF v3.0 list

The SFWG received nominations for TTR-associated amyloidosis due to the availability of newer FDA-approved treatments. However, the SFWG did not ultimately recommend inclusion of this gene on the current list due to concerns of incomplete penetrance and that most patients develop recognizable disease-related symptoms allowing for diagnosis and treatment prior to late-stage disease. As part of this decision, the SFWG referenced a cardinal principle that the SF list should not be a replacement for indication-based diagnostic genetic testing.

The SFWG questioned the actionability of COL5A1-associated Ehlers–Danlos syndrome, PRKAR1A/Carney complex, and NOTCH3/CADASIL. The SFWG decided that including gene phenotypes such as HMBS-associated acute intermittent porphyria and SERPINA1/alpha-1-antitrypsin deficiency with interventions involving environmental exposures or behavior modification was beyond the scope of this list.

GCH1-associated dopa-responsive dystonia was also thought to be beyond the scope of the list using the rationale that its clinical presentation would likely prompt an individual to seek a diagnosis, and while an effective treatment is available, the timing of its initiation does not appear to compromise its effectiveness. MEFV-associated familial Mediterranean fever was ultimately not included on the list due to concerns about there being a low chance of having a SF reported out that later becomes downgraded to a new classification that would be below the threshold for reporting, which could be burdensome to patients. Finally, the HNF1B/MODY5 was not included for reasons described in detail above.

Pharmacogenomic genes/variants

The current SF list includes RYR1, and we considered several issues related to the possibility of adding additional pharmacogenetics (PGx) variants as secondary findings. The clinical validity and utility of PGx testing has been demonstrated in many studies to aid drug therapy, and guidelines for implementation are currently available from international PGx consortia such as the Clinical Pharmacogenetics Implementation Consortium (CPIC) and the Dutch Pharmacogenetics Working Group (DPWG).43,44,45 ES/GS genotypes could potentially be a cost-effective method to generate useful PGx profiles, which can then be used preemptively to guide drug dose and choice. However, several critical PGx variants or haplotypes cannot be captured through exome-based testing, and generating haplotypes requires additional data processing that is not part of a standard ES informatics pipeline. Therefore, we evaluated these technical limitations as part of the SF gene nomination process.

The difficulties for the laboratory to report clinically actionable variants in these genes arise from multiple issues: (1) many of the clinically relevant variants reside in promoter, intronic, or untranslated regions that are not captured using current methodology (e.g., the key variant for warfarin dosing [−1639G>A in VKORC1] is in the promoter region; the increased function CYP2C19*17 allele is also characterized by a promoter variant [−806C>T] that results in increased gene expression); (2) copy-number variants (CNVs), repeats, and gene hybrids have been challenging to assess with current ES technology (e.g., CYP2D6 CNVs that define the ultrarapid metabolizer phenotype); (3) for some genes and variants there is still controversy regarding genotype/phenotype correlations; and (4) as many PGx guidelines describe haplotypes, testing often requires genotyping multiple positions/regions and types of variation within the same gene, complicating the analysis and reporting, especially when phase cannot be easily determined. Some phenotypes may not be determined accurately due to a lack of coverage and missing CNV information depending on the assay design (e.g., a number of CYP2D6 alleles include SNPs at multiple positions and may also involve duplication, deletion, and large-scale gene rearrangements (hybrid). Rare CYP2D6 variants also may not be included in the genotype testing used by some laboratories, which could result in errors in diplotype/phenotype calls as well as false negative findings.

Other challenges not specific for ES include (1) lack of evidence and guidance for combining results from multiple PGx genes beyond what has been covered by existing CPIC guidelines; (2) the majority of published PGx research is conducted with European ancestry-dominant cohorts, lacking evidence from diverse patient populations (thus, the guidance based on alleles common in European ancestry–majority cohorts may not be appropriate/generalizable for other ethnicities); (3) ambiguity in PGx testing results, i.e., variants with unknown or uncertain significance; and (4) the large number of patients taking multiple medications (polypharmacy) that may have synergistic or antagonistic effects on each other, and thus affect interpretation of PGx results.

In the future, it may be possible for a workgroup to develop a universal and easily implemented method for analysis and interpretation of PGx variants that can be utilized by all diagnostic laboratories. We encourage ongoing research to document (1) the reliable identification of alleles (and proper phasing) based on standard ES/GS; (2) spectrum of PGx variants outside of European ancestry populations; (3) the time and effort required within the laboratory; (4) the time and effort required in clinics, including educational needs for clinicians who are not already familiar with this type of testing in terms of what they need to know to properly consent and return results; (5) how the results will be documented in the medical record in order to be accessible in the distant future; and (6) how often persons receiving these results will use them in medication choices.

CONCLUSIONS

With the publication of the accompanying SF policy statement, we have separated this secondary findings gene list update, which describes the rationale supporting how genes are selected for addition to or removal from the secondary findings list. This dual publication approach was done intentionally with a primary goal of providing more frequent updates to the actual SF gene list. Going forward, we foresee updates to the general policy only as needed and may be expected to occur every few years. In contrast, updates to the gene list will be targeted to occur on an annual basis, and to be published at approximately the same time each year so that all stakeholders can expect an update and be prepared to update laboratory and reporting processes. For example, we recognize that clinical laboratories must integrate updates into their workflow, and clinicians must familiarize themselves with the genes on the list for the purposes of genetic counseling and informed consent. Our intention is to publish an updated list each year in January.

The SFWG will continue to review this list of actionable genes, and new nominations, throughout the course of the year. We also wish to remind the community that ACMG members may nominate genes or variants to be added to, or removed from, the list based on an evolving evidence base and/or evolving standards in the practice of medicine. We will also consider nominations submitted through representatives of other professional organizations. Nomination forms can be found on the ACMG website.7 We hope that the detailed descriptions of our decision process during the preparation of this update will help the community to better understand the types of genes and variants that we consider appropriate for this list to guide nominations going forward.