NM_000314.8:c.210-2A>G is a canonical splice acceptor site variant (AG>GG at the -2 position of intron 3) in PTEN, a gene where loss of function is an established mechanism for PTEN hamartoma tumor syndrome (PHTS).¹ Per the ClinGen PTEN VCEP v3.2.0 PVS1 decision tree, this variant qualifies for PVS1 (Very Strong): it disrupts a canonical GT-AG splice site at a position 5' to p.D375 (c.1121), is predicted to cause exon 4 skipping with frameshift (p.Ala72Thrfs*5) and nonsense-mediated decay, and affects a biologically-relevant transcript (NM_000314.8). The closely related variant c.210-1G>A at the same splice junction has been experimentally confirmed to cause exon 4 skipping (PMID:28677221).² The variant is absent from gnomAD v2.1 and v4.1, meeting the PTEN VCEP threshold for PM2_Supporting (allele frequency <0.001%).³ No direct functional evidence (RNA assay, minigene, or splicing study) was identified for c.210-2A>G specifically. The Mighell et al. 2018 phosphatase activity assay (mmc2.xlsx) only covers missense variants and is not applicable to this splice variant.⁴ SpliceAI predicts a max delta score of 0.00, which is anomalous for a canonical splice acceptor variant and may represent a lookup artifact. SpliceAI predictions should be independently verified before relying on computational splicing evidence for this variant.⁵ ClinVar reports this variant as Likely pathogenic (2 clinical laboratories) and Pathogenic (1 clinical laboratory) under ClinVar ID 576440, with review status 'criteria provided, single submitter.' However, BP6 is not for use per PTEN VCEP, and ClinVar classifications alone are not used as independent evidence.⁶ This variant has been observed in somatic cancers (COSMIC COSV64302273, n=4), supporting its potential role in tumorigenesis, though somatic occurrence alone is not an ACMG/AMP criterion for germline classification. A systematic review of the five available full-text publications (PMIDs: 16199547, 28677221, 29758562, 9467011, 21194675) confirmed that none mention the exact variant NM_000314.8:c.210-2A>G. The most relevant paper (PMID:28677221) studied 34 PTEN intronic variants including c.210-1G>A at the same splice junction, but c.210-2A>G was not among the variants analyzed.⁷ Combined evidence: PVS1 (Very Strong) + PM2_Supporting. Under the PTEN VCEP combination rules v3.2.0, this specific combination (1 Very Strong + 1 Supporting without an intermediate Moderate criterion) does not match any defined pathogenic or likely pathogenic rule. Under the generic ACMG/AMP 2015 framework (PMID:25741868), 1 Very Strong + 1 Supporting likewise does not reach the threshold for Likely Pathogenic (requires >=2 Supporting or >=1 Moderate). Using a Bayesian point system (PVS1=8, PM2=1, total=9), this falls within the Likely Pathogenic range (6-9 points). A final classification should be determined by the interpreting clinical laboratory considering the strength of the PVS1 call and the anomalous SpliceAI result.⁸