A points-based framework steadily cut uncertain variant calls

Design

Retrospective analysis of a variant-classification framework across its versions

Population

2.6 million variants from 5.5 million individuals referred for genetic testing

Primary outcome

Change in variant-of-uncertain-significance rate and reclassification

Effect

73% of tracked reclassifications used post-baseline criteria; score predicted reclassification direction (r-squared 0.95)

Variants of uncertain significance are the bane of clinical genetic testing, and how classification rules evolve directly changes how many a laboratory issues. This analysis evaluated a decade of a points-based refinement of the 2015 ACMG/AMP guidelines, applied to 2.6 million variants across 5.5 million people referred for testing.

Tracking a set of 32,241 variants first classified under an early version, 3,834 uncertain variants were later reclassified; 73% of those reclassifications relied on evidence criteria added after that early version — showing that the framework updates, not just new data, drove the gains. In a simulation removing later artificial-intelligence-related criteria, 21% of recent classifications changed, and the numerical score predicted the direction of reclassification well.

The lesson for a reporting laboratory is that structured, points-based classification systems materially reduce uncertain calls over time, and that a variant's score carries quantitative information about how likely it is to be reclassified. This is one commercial framework's experience, so the specific numbers are not a universal benchmark.

• Points-based refinement of the 2015 ACMG/AMP guidelines applied to 2.6 million variants across 5.5 million people.
• Of 32,241 tracked variants, 3,834 uncertain ones were reclassified; 73% relied on criteria added after the early version.
• Removing later AI-related criteria changed 21% of recent classifications in simulation.
• The numerical score predicted the direction of reclassification well.
• Structured scoring reduces uncertain calls, but these figures are one framework's experience.

Why it matters

It shows that how a laboratory structures its classification rules, not only the evidence it gathers, changes how many uncertain results patients receive.

Don't overread it

This reflects one commercial classification framework; the specific reclassification rates are not a universal benchmark for other laboratories.

The statistics, in plain English

These are operational metrics from one laboratory's framework, not a trial; they show the direction and scale of change in classification, but the exact percentages depend on that system and its caseload.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.