The European Society
of Human Genetics

ESHG/Eurogentest two-dimensional variant classification system

ESHG/Eurogentest two-dimensional variant classification system

An ad hoc working group in ESHG (Johan den Dunnen, Helen Firth, Nicole de Leeuw, Hans Scheffer and Gunnar Houge) has proposed a two dimensional system for variant classification, with a molecular and clinical arm. Everybody is free to test the system, and feedback is highly appreciated! The system with information and explanations use can be found below.

On the proposed two-dimensional variant classification system
By Gunnar Houge
ESHG Executive Board
Chair of the variant classification ad hoc committee

The suggested variant classification system has a molecular (A) and clinical grading (B) scale (from 0-5) that are combined in the end (A+B).

The molecular grading is based on but not identical to the ACMG/AMP system. It classifies variant consequence for gene function without considering the clinical question at hand. You can e.g. have a molecular grade 5 due to a LoF mutation that is without clinical consequence if the disease in question is likely dominant and there is no evidence suggesting haploinsufficiency as a pathogenic mechanism. You can also have a variant that clearly affects gene function in other ways (e.g. have an effect on eye colour, taste, metabolism) without any clinical consequence. Such variants can now be classified.

The main difference is that ACMG class 3 (a standard VUS) is now class 0, while a suspect VUS (sometimes called a “hot VUS” or “VUS+”) is class 3. One reason is that an ACMG class 3 may be wrongly conceived as being between classes 2 and 4, i.e. in the middle of a scale from benign to pathogenic. Although ACMG classification should rather be used without numbering (i.e. B, LB, VUS, LP and P), this misconception may still remain. A more important reason is that the current VUS group is too broad to be clinically helpful - albeit convenient if you are not exposed to the clinic and just need a group for everything you are unsure of. For these reasons the proposed system makes a distinction between a typical VUS (class 0) and a VUS that is possibly pathogenic (class 3), which is important for the clinician.

The clinical grading takes penetrance and expressivity into account. Of utmost importance is the biological likelihood of variant involvement in causing the patient’s phenotype. This applies in particular to hypomorphic and recessive alleles. To know if a variant in a gene might explain the phenotype, phenotypic details may often be needed. Why this is so important is something we need to teach the clinicians. As a follow-up project we should also consider making a standard sample referral form with a built-in HPO system that can be machine-read or transformed into barcodes.

Hopefully, the new system will make variant classification clearer for the clinician and be of general applicability. I think the move towards gene-specific variant classification systems that we increasingly see today (e.g. MYH7 specific guidelines) is a path towards chaos.

The accompanying Excel sheet shows some examples on how I have chosen to classify variants of different types from my own practice. This I have linked to nine standard variant explanations just to show how the system may be used, if desired. The variant explanations are not a part of the classification system per se.

Please feedback comments and suggestions to: gunnarhouge[a]gmail.com

Table 1 gives an overview of major variant categories and their classification in this system. Note that a variant class (e.g. D) may be linked to several reporting options, depending on the clinical question – but as mentioned this is an independent part of the system that may be used if desired – or not at all.

Variant

Clinic

A

B

A+B

Report

F5-Leiden

DVT

5

2

D

Y

BRCA1 c.5407-25T>A

Several families with CM/CO

4

5

B

Y

Trisomy 21

Down syndrome

5

5

A

Y

AQP4 c.332G>C dn

ID, transient brain edema

3

1

E

Y

Dup 1q21.1 (classic)

Tetratogy of Fallot

5

3

C

Y

JAK2 p.(R683_684insK)

Elevated platelet counts (only)

4

0

E

Y

ABCA4 p.(Asp1868Ile)

Reduc. vision. Normal macula

4

2

D

N

ABCA4 one LoF variant

Macula abnormal – Stargardt?

5

2

D

Y

CFTR Phe508del (only)

Diarrhoea

5

2

D

N

CFTR Phe508del (only)

Nasal polyposis

5

2

D

Y

SLC6A1 p.(R419H) dn

Neonatal hypotona

3

1

E

N

Extensive homozygocity

Occipital encephalocele

3

2

E

Y

Extensive homozygocity

Normal UL findings in fetus

3

2

E

N

C2orf44 p.(R628G) homoz

MCA, ID, deaf, short stature

3

0

F

N

FLVCR2 p.(W205C) homoz

Fetus with Fowler syndrome

3

5

C

Y

RET p.(Arg886Gln)

Pheochromocytoma at age 41

0

2

O

N

ATM p.(Arg3008Cys)

Family with colon cancer (only)

3

0

F

N

UMOD p.(Tyr305*)

Neonate with cyanotic spells

5

0

E

N

 

Please find further documents on the variant classification system here:

Variant classification and reporting (PDF)

ESHG Variant Classifyer EXAMPLES (Excel)

ESHG Variant Classifyer VERSION 1.1 (Excel)