EGFR c.2390G>C, p.Cys797Ser

Reasoning

PS3 is met at moderate strength. The queried EGFR NM_005228.5:c.2390G>C (p.Cys797Ser, C797S) variant was directly evaluated in multiple independent, biologically relevant cell-based functional studies. PMID:25939061 used a G>C-specific ddPCR assay to identify acquired C797S and showed in Ba/F3 cell assays that C797S constructs were markedly less sensitive to AZD9291 and resistant to CO-1686, with reduced drug-associated inhibition of cell growth and EGFR phosphorylation. PMID:25964297 identified C797S in a selected resistant cell line, compared it with the parental line, and showed that engineered C797S prevented suppression of EGFR phosphorylation and downstream signaling and conferred resistance to EGFR tyrosine-kinase inhibitors. PMID:25948633 independently showed C797S-associated resistance to WZ4002, CO-1686, and AZD9291 in engineered Ba/F3, PC9, and NSCLC models, with context-dependent inhibitor-response controls. PMID:29410323 further evaluated C797S in Ba/F3 models and reported altered inhibitor sensitivity, including moderate dacomitinib resistance. The convergence of variant-specific resistance, signaling, and cell-growth results supports a damaging functional effect. Moderate rather than strong strength is assigned because the studies are research cell assays and the available evidence does not establish reproducibility and robustness in a clinical diagnostic laboratory setting.

Rule

Well-established in vitro or in vivo functional studies supportive of a damaging effect on the gene or gene product. ... Functional studies that have been validated and shown to be reproducible and robust in a clinical diagnostic laboratory setting are considered the most well-established.

Evidence considered

• PMID:25939061: Variant-specific G>C ddPCR and serial plasma/tumor analyses identified acquired EGFR C797S; Ba/F3 growth and EGFR-phosphorylation assays showed markedly reduced AZD9291 sensitivity and resistance to CO-1686.
• PMID:25964297: C797S was identified in a WZ4002-resistant line but not its parental line; engineered models showed failure to suppress EGFR phosphorylation and downstream signaling and conferred inhibitor resistance.
• PMID:25948633: Engineered Ba/F3, PC9, and NSCLC models showed C797S resistance to WZ4002, CO-1686, and AZD9291, with responses dependent on the accompanying EGFR genotype.
• PMID:29410323: Ba/F3 models expressing EGFR L858R plus C797S or G719A plus C797S showed altered inhibitor sensitivity and moderate dacomitinib resistance, with reported IC50 values of 23.8 nM for dacomitinib in the L858R+C797S model.

Still missing

• No evidence was identified that these assays were validated and shown to be reproducible and robust in a clinical diagnostic laboratory setting.
• The available extracts do not provide a standardized quantitative calibration linking assay effect size to PS3 strong-level evidence.

Sources

Reasoning

PM2 is met at supporting strength. The variant is reported absent from gnomAD v2.1 and gnomAD v4.1, the default all-comers population sources because no governing VCEP specifies a non-cancer or exome-only subset. Thus its observed allele frequency is 0 in the reported datasets and is below the supplied PM2 supporting threshold of <=0.0001. No non-cancer subset was required by the governing generic rules, and no population-frequency artifact or indel-calling caveat applies to this single-nucleotide substitution.

Rule

Absent from controls (or at extremely low frequency if recessive) in Exome Sequencing Project, 1000 Genomes or Exome Aggregation Consortium.

Evidence considered

• gnomAD v2.1: variant search status absent in the default all-comers dataset.
• gnomAD v4.1: variant search status absent in the default all-comers dataset.
• The case has no EGFR VCEP or gene-specific population rule; therefore the prompt's default all-comers sources apply rather than a non-cancer or exome-only subset.
• Generic PM2 threshold supplied for this case: <=0.0001, with PM2 downgraded to supporting strength under the supplied ClinGen SVI recommendation.

Sources

Reasoning

The governing rule applies when the variant is found in a case with an alternate molecular basis for disease. Multiple exact-variant reports identify C797S in tumors that already carry other EGFR alterations, including exon 19 deletion and T790M, and describe C797S as an acquired treatment-resistance mutation. Those pre-existing driver alterations provide an alternate molecular basis for the underlying lung cancer, supporting BP5 at supporting strength.

Rule

Variant found in a case with an alternate molecular basis for disease.

Evidence considered

• PMID:25939061 reports newly detected C797S at progression in a tumor with pre-existing EGFR exon 19 deletion and T790M; the confirmatory cohort comprised EGFR-mutant NSCLC cases.
• PMID:26181354 reports acquired C797S in a patient whose lung adenocarcinoma already had EGFR exon 19 deletion and T790M.
• PMID:26749488 reports C797S together with T790M and a pre-existing exon 19 mutation in a post-progression tumor biopsy.
• PMID:25948633 shows C797S resistance effects depend on the accompanying EGFR activating mutation and T790M context, consistent with an alternate molecular basis for the underlying cancer.

Sources

Reasoning

This is a missense variant, NP_005219.2:p.(Cys797Ser), so the governing generic rule requires the missense REVEL path only; the SpliceAI result is not used for PP3/BP4 in this situation. REVEL is 0.516, below the generic PP3 supporting threshold of >=0.644 (ClinGen SVI REVEL calibration, Pejaver et al. 2022, PMID:36413997), and therefore PP3 is not met. BayesDel is not used because no generic ACMG-strength calibration is established.

Rule

Multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.). ... Missense variant -> REVEL only. Its SpliceAI score says nothing about whether the amino-acid substitution is tolerated, so 'no predicted splice impact' is never BP4 support for a missense variant. ... PP3 (supporting): >= 0.644 — ClinGen SVI REVEL calibration (Pejaver et al. 2022, Am J Hum Genet) (PMID:36413997)

Evidence considered

• The variant is a missense substitution, NP_005219.2:p.(Cys797Ser), in EGFR exon 20; therefore REVEL is the applicable computational path and SpliceAI is excluded.
• Local REVEL lookup reports 0.516 for NM_005228.5:c.2390G>C (source_registry key: revel).
• The supplied generic calibration assigns PP3 supporting at REVEL >= 0.644: ClinGen SVI REVEL calibration, Pejaver et al. 2022, PMID:36413997.
• The supplied generic rules state that BayesDel has no established generic ACMG-strength calibration, so the BayesDel score is not_available for PP3/BP4.

Sources

Reasoning

The governing rule requires a phenotype or family history highly specific for a disease with a single genetic etiology. The available reports describe EGFR C797S in non-small-cell lung cancer acquired-resistance settings, often with pre-existing EGFR driver mutations and treatment exposure; this phenotype is not highly specific for a single genetic etiology attributable to this variant.

Rule

Patient’s phenotype or family history is highly specific for a disease with a single genetic etiology.

Evidence considered

• PMID:25939061 describes acquired C797S at progression in EGFR-mutant non-small-cell lung cancer after AZD9291 treatment, with pre-existing exon 19 deletion and T790M.
• PMID:26181354 describes acquired C797S in a patient with EGFR exon 19 deletion and T790M lung adenocarcinoma after AZD9291 treatment.
• PMID:26749488 describes C797S in a post-progression tumor biopsy after HM61713 treatment, together with T790M and a pre-existing exon 19 mutation.

Sources

Reasoning

The governing rule concerns a reputable source reporting the variant as pathogenic when the evidence is unavailable for independent evaluation. The exact ClinVar record is classified as uncertain significance with one single-submitter, criteria-provided submission, no expert-panel submission, and no usable expert-panel pathogenic or likely-pathogenic classification. OncoKB labels the variant oncogenic, but it is not an exact-variant ClinVar expert-panel classification and therefore does not trigger this criterion under the specified rule.

Rule

Reputable source recently reports variant as pathogenic, but the evidence is not available to the laboratory to perform an independent evaluation.

Evidence considered

• ClinVar variation 3257908 exactly matches NM_005228.5(EGFR):c.2390G>C (p.Cys797Ser) but is classified as uncertain significance with criteria provided by a single submitter; expert-panel submissions are absent.
• OncoKB reports a variant-specific oncogenic label, but this is not an exact-variant ClinVar expert-panel classification and does not satisfy the required PP5 trigger.

Sources

Reasoning

BA1 is not met. The queried EGFR NM_005228.5:c.2390G>C (p.Cys797Ser) variant is reported absent from both gnomAD v2.1 and gnomAD v4.1, so no filtering allele frequency or grpmax FAF reaches the generic stand-alone threshold of >=0.05. No population-frequency evidence supports a BA1 call.

Rule

Allele frequency is above 5% in Exome Sequencing Project, 1000 Genomes or Exome Aggregation Consortium.

Evidence considered

• gnomAD v2.1: variant search status absent; no allele count, allele number, point AF, or filtering grpmax FAF was reported.
• gnomAD v4.1: variant search status absent; no allele count, allele number, point AF, or filtering grpmax FAF was reported.
• Generic BA1 default threshold supplied for this case: >=0.05, with filtering allele frequency/grpmax FAF required rather than a popmax point estimate.

Sources

Reasoning

BS1 is not met. The variant is absent from gnomAD v2.1 and v4.1, so its filtering allele frequency/grpmax FAF is not greater than expected for the disorder and does not reach the generic 0.01 default. There is therefore no elevated population frequency requiring a maximum-credible-frequency adjustment for prevalence, penetrance, or allelic heterogeneity.

Rule

Allele frequency is greater than expected for the disorder.

Evidence considered

• gnomAD v2.1: variant search status absent; no allele count, allele number, point AF, or filtering grpmax FAF was reported.
• gnomAD v4.1: variant search status absent; no allele count, allele number, point AF, or filtering grpmax FAF was reported.
• Generic BS1 default supplied for this case: >=0.01; the prompt requires filtering allele frequency/grpmax FAF and consideration of prevalence, penetrance, and allelic heterogeneity when an elevated frequency exists.
• No elevated ancestry-specific frequency was observed, so a Whiffin et al. 2017 maximum-credible-frequency calculation was not needed.

Sources

Reasoning

BS3 is not met. The available variant-specific functional evidence does not show no damaging effect on protein function or splicing; instead, independent cell-based assays consistently report altered EGFR inhibitor response, acquired drug resistance, and reduced inhibitor-associated EGFR phosphorylation for C797S. No benign functional study with appropriate controls and a demonstrated normal protein or splicing result was identified.

Rule

Well-established in vitro or in vivo functional studies show no damaging effect on protein function or splicing.

Evidence considered

• PMID:25939061: C797S-containing cells showed reduced AZD9291-associated inhibition of cell growth and EGFR phosphorylation.
• PMID:25964297: C797S prevented suppression of EGFR phosphorylation and downstream signaling in engineered models and conferred resistance.
• PMID:25948633: C797S caused resistance to multiple irreversible EGFR inhibitors in engineered cellular models.
• PMID:29410323: C797S altered EGFR inhibitor sensitivity and produced moderate dacomitinib resistance in Ba/F3 models.

Still missing

• No well-established benign functional assay demonstrating normal EGFR protein function or splicing was identified.

Sources

Reasoning

This is a missense variant, NP_005219.2:p.(Cys797Ser), so the governing generic rule requires the missense REVEL path only; the SpliceAI result is not used for PP3/BP4 in this situation. REVEL is 0.516, above the generic BP4 supporting threshold of <=0.29 (ClinGen SVI REVEL calibration, Pejaver et al. 2022, PMID:36413997), so BP4 is not met. BayesDel is not used because no generic ACMG-strength calibration is established.

Rule

Multiple lines of computational evidence suggest no impact on gene or gene product (conservation, evolutionary, splicing impact, etc). ... Missense variant -> REVEL only. Its SpliceAI score says nothing about whether the amino-acid substitution is tolerated, so 'no predicted splice impact' is never BP4 support for a missense variant. ... BP4 (supporting): <= 0.29 — ClinGen SVI REVEL calibration (Pejaver et al. 2022, Am J Hum Genet) (PMID:36413997)

Evidence considered

• The variant is a missense substitution, NP_005219.2:p.(Cys797Ser), in EGFR exon 20; therefore REVEL is the applicable computational path and SpliceAI is excluded.
• Local REVEL lookup reports 0.516 for NM_005228.5:c.2390G>C (source_registry key: revel).
• The supplied generic calibration assigns BP4 supporting at REVEL <= 0.29: ClinGen SVI REVEL calibration, Pejaver et al. 2022, PMID:36413997.
• The supplied generic rules state that BayesDel has no established generic ACMG-strength calibration, so the BayesDel score is not_available for PP3/BP4.

Sources

Reasoning

The governing rule concerns a reputable source reporting the variant as benign when the evidence is unavailable for independent evaluation. The exact ClinVar record has one single-submitter uncertain-significance classification and no expert-panel benign or likely-benign assertion. No qualifying exact-variant ClinVar expert-panel benign classification is present.

Rule

Reputable source recently reports variant as benign, but the evidence is not available to the laboratory to perform an independent evaluation.

Evidence considered

• ClinVar variation 3257908 exactly matches the variant but reports uncertain significance, criteria provided by a single submitter, with no expert-panel submission.

Sources

Reasoning

The variant is EGFR p.Cys797Ser (c.2390G>C), and multiple papers report the same amino-acid change in acquired resistance or functional studies. However, the available reports do not establish a previously pathogenic germline or ACMG-classified comparator variant; OncoKB's somatic oncogenic label and resistance reports are not sufficient to satisfy the PS1 requirement. The variant is protein-level missense rather than a splice-impacting change, and SpliceAI's maximum delta score is 0.004, but that does not resolve the missing established-pathogenic comparator.

Rule

Same amino acid change as a previously established pathogenic variant regardless of nucleotide change.

Evidence considered

• EGFR p.Cys797Ser is the normalized protein consequence of NM_005228.5:c.2390G>C.
• Ercan et al. (PMID:25948633) and Niederst et al. (PMID:25964297) report C797S in engineered or acquired-resistance models; these are reports of the queried variant, not an established pathogenic comparator classification.
• The ClinVar record is a single-submitter somatic record classified as uncertain significance, while OncoKB labels the variant oncogenic; neither is an ACMG pathogenic classification for PS1.
• SpliceAI maximum delta score is 0.004, supporting a protein-level interpretation rather than a splice-impact caveat.

Still missing

• A previously established pathogenic variant with the same amino-acid change, supported by an applicable germline or ACMG/AMP classification, is not available.

Sources

Reasoning

The available evidence does not document a proband with a relevant disease phenotype, absence of family history, or confirmed maternity and paternity. Therefore, the required de novo evidence for PS2 cannot be assessed.

Rule

De novo (both maternity and paternity confirmed) in a patient with the disease and no family history.

Evidence considered

• No proband phenotype, family history, or parental maternity and paternity testing results are documented in the available case evidence.

Still missing

• Proband disease phenotype and clinical diagnosis
• Family history
• Confirmed maternity
• Confirmed paternity
• Trio-level variant testing results

Reasoning

The governing rule requires a significantly increased prevalence in affected individuals compared with controls and specifies an OR or RR >5.0 with a confidence interval excluding 1.0. The literature-pass PS4 leads were checked against the available full-text extracts: PMID:25948633 reports engineered-cell drug resistance, PMID:29410323 reports in-vitro resistant clones, and PMID:25939061 reports C797S in 6/15 T790M-positive treated cases (40%) without an unaffected or population control group. PMID:26181354 and PMID:26749488 are single-patient acquired-resistance reports. No qualifying case-control counts, OR/RR, or confidence interval were identified, so no case-control comparison can be computed and PS4 cannot be assessed.

Rule

The prevalence of the variant in affected individuals is significantly increased compared to the prevalence in controls. Note 1: Relative risk (RR) or odds ratio (OR), as obtained from case-control studies, is >5.0 and the confidence interval around the estimate of RR or OR does not include 1.0.

Evidence considered

• PMID:25939061 reports C797S in 6 of 15 T790M-positive cases at progression (40%) and describes serial treated cases, but provides no unaffected or population control comparison.
• PMID:25948633 and PMID:29410323 report engineered-cell or resistant-clone experiments rather than affected-person versus control prevalence.
• PMID:26181354 and PMID:26749488 are single-patient acquired-resistance reports without control groups.

Still missing

• A case-control cohort with affected-case and unaffected or matched-population control counts for exact EGFR c.2390G>C is needed to calculate OR/RR and its confidence interval.

Sources

Reasoning

EGFR p.Cys797Ser lies in the EGFR tyrosine-kinase domain and C797 is described as part of the ATP-binding pocket that forms the covalent bond targeted by third-generation EGFR inhibitors. The case evidence also reports a statistically significant C797 hotspot. Nevertheless, the generic PM1 definition requires a critical and well-established functional domain or hotspot without benign variation, and the supplied evidence does not establish the absence of benign variation across the relevant domain or hotspot. No EGFR VCEP domain table is available to provide a complete approved region and benign-variation assessment.

Rule

Located in a mutational hot spot and/or critical and well-established functional domain (e.g. active site of an enzyme) without benign variation.

Evidence considered

• Niederst et al. (PMID:25964297) state that C797 is located opposite T790 in the ATP-binding pocket and forms a covalent bond with third-generation TKIs.
• The C797S case report (PMID:26181354) places the variant in EGFR's tyrosine-kinase functional context and reports acquired resistance.
• Cancer Hotspots reports EGFR C797 as a statistically significant hotspot, but the parsed record supplies no benign-variation assessment.
• The vcep_materials domain_tables object is empty for EGFR, so no authoritative gene-specific domain list is available.

Still missing

• A documented benign-variation assessment for the EGFR kinase-domain or C797 hotspot region is unavailable.
• No EGFR VCEP domain table or approved PM1 region specification is available.

Sources

Reasoning

A different missense substitution at the same residue, EGFR C797G, was experimentally studied and altered EGFR activity, while C797S is reported in several resistance studies. However, the supplied evidence does not establish C797G or another alternate C797 missense variant as pathogenic under the governing ACMG/AMP framework. Therefore the same-residue observation does not satisfy PM5.

Rule

Novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before.

Evidence considered

• Zhang et al. (PMID:26378037) studied the alternate same-residue substitution EGFR C797G and reported reduced palmitoylation, phosphorylation, dimerization, and EGF-induced activation, but did not classify C797G as pathogenic.
• PMID:25964297 and PMID:25948633 report the queried C797S substitution in resistance models, not a different pathogenic C797 missense comparator.

Still missing

• A different missense variant at EGFR residue 797 with an established pathogenic classification is not available.

Sources

Reasoning

The available evidence does not document an assumed de novo occurrence in a patient with the relevant disease, nor does it provide parental testing information showing that maternity or paternity was unconfirmed. Therefore, PM6 cannot be assessed.

Rule

Assumed de novo, but without confirmation of paternity and maternity.

Evidence considered

• No assumed de novo finding, parental testing results, or parental identity-confirmation status is documented in the available case evidence.

Still missing

• Proband disease phenotype and clinical diagnosis
• Variant testing in both biological parents
• Documentation that maternity and paternity were unconfirmed
• Family history

Reasoning

The available evidence does not identify affected family members carrying the variant, provide phenotype concordance, or report informative meioses. Consequently, co-segregation and any possible increased strength cannot be assessed.

Rule

Co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease.

Evidence considered

• No family pedigree, affected-relative testing, phenotype-segregation data, or informative meioses are documented in the available case evidence.

Still missing

• Affected family members
• Phenotype consistency among relatives
• Variant testing results for relatives
• Number of informative meioses

Reasoning

The available evidence concerns EGFR somatic cancer alterations and acquired drug resistance, not a gene-level analysis establishing both a low rate of benign missense variation and missense as a common disease mechanism under the generic germline ACMG/AMP criterion. Absence of this individual variant from gnomAD does not establish either gene-level requirement.

Rule

Missense variant in a gene that has a low rate of benign missense variation and where missense variants are a common mechanism of disease.

Evidence considered

• The supplied literature and database evidence is focused on somatic EGFR cancer mutations and acquired resistance.
• The queried variant is absent from gnomAD v2.1 and v4.1, but single-variant absence does not establish the PP2 gene-level requirements.

Still missing

• A gene-level benign-missense-rate assessment for EGFR is unavailable.
• A disease-context assessment showing missense variants are a common mechanism of EGFR disease is unavailable.

Reasoning

BS2 cannot be assessed from the available evidence. gnomAD absence is population-frequency evidence, not documentation that the variant was observed in a healthy adult, and the available literature concerns acquired somatic EGFR C797S in cancer or in-vitro resistance models. No validated healthy-adult heterozygote, homozygote, or hemizygote observation with the relevant inheritance and expected early full penetrance was supplied.

Rule

Observed in a healthy adult individual for a recessive (homozygous), dominant (heterozygous), or X-linked (hemizygous) disorder, with full penetrance expected at an early age.

Evidence considered

• gnomAD v2.1 and v4.1 report the variant as absent; they do not provide a healthy-carrier observation.
• The available variant-specific papers describe acquired C797S in treated cancer or engineered cell models, not a healthy adult constitutional carrier.

Still missing

• A validated observation of the variant in a healthy adult, with zygosity, inheritance model, age, phenotype ascertainment, and penetrance context, is unavailable.

Sources

Reasoning

The available evidence does not identify affected family members who were tested and found not to carry the variant. Because no non-segregation observations or phenocopy assessment are documented, BS4 cannot be assessed.

Rule

Lack of segregation in affected members of a family.

Evidence considered

• No affected relatives who tested negative for the variant, pedigree-based non-segregation, or phenocopy assessment is documented in the available case evidence.

Still missing

• Affected relatives with comparable disease phenotypes
• Variant testing results for affected relatives
• Pedigree-based non-segregation analysis
• Phenocopy assessment

Reasoning

The supplied evidence does not establish that EGFR disease is caused primarily by truncating variants. The available EGFR reports concern somatic missense resistance mutations, and no curated gene-level variant-spectrum evidence demonstrating a predominantly truncating disease mechanism is provided.

Rule

Missense variant in a gene for which primarily truncating variants are known to cause disease.

Evidence considered

• The supplied EGFR literature describes somatic missense substitutions and drug resistance, not a predominantly truncating disease mechanism.
• No gene-level source in the supplied evidence establishes that truncating variants primarily cause the relevant EGFR disease.

Still missing

• A curated EGFR disease variant-spectrum assessment comparing truncating and missense mechanisms is unavailable.

Reasoning

BP2 requires either observation in trans with a pathogenic variant for a fully penetrant dominant gene/disorder or observation in cis with a pathogenic variant in any inheritance pattern. PMID:26181354 reports somatic acquired EGFR C797S in cis with T790M in a resistant lung-cancer biopsy, but the record does not establish T790M as a germline pathogenic variant for a fully penetrant dominant EGFR disorder, and no germline inheritance or affected-proband phase data are provided. The cis observation is therefore insufficient to assess BP2 under the stated ACMG/AMP definition.

Rule

Observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder or observed in cis with a pathogenic variant in any inheritance pattern.

Evidence considered

• PMID:26181354, Report of a Case, biopsy 3: NGS showed C797S acting in cis with T790M, with all C797S alleles also carrying T790M and vice versa; the report describes acquired resistance after AZD9291 treatment.
• ClinVar variation 3257908 identifies the available submission as somatic and related to neoplasm, without a germline pathogenic-variant designation or inheritance data.
• COSMIC COSV51766509 records the assessed variant in somatic cancers and does not establish a qualifying germline pathogenic partner.

Still missing

• The phase observation is from somatic acquired-resistance material rather than a germline family or proband study.
• T790M is not established in this case as a pathogenic germline variant for a fully penetrant dominant EGFR disorder.
• No germline affected-proband or parental testing confirms a qualifying BP2 configuration.

Sources