Reading a Tumor Molecular Report: Which Findings Are Actionable, Which Are Informational, and Which Are the Laboratory's Confidence

Zetobit · Reading the Report

Reading the Report

Reading a Tumor Molecular Report: Which Findings Are Actionable, Which Are Informational, and Which Are the Laboratory's Confidence

A tumor report arrives as one list. The entries on it answer three different questions, and the most seductive line — the one naming an approved drug — is often the one that requires the most reading before it means anything for this patient.

Kanna Nandakumar, PhD · Zetobit, LLC

A molecular profiling report arrives looking like a single kind of thing: a ranked list of what was found in the tumor, with a column of drug names beside it. It reads top-down, and the eye stops at the first row with a therapy attached.

But the entries on that list are not the same kind of statement. Some describe what is physically present in the specimen. Some describe how much published evidence connects that finding to a drug. And some describe how confident the laboratory is that it saw what it says it saw. Those three claims have different failure modes, different shelf lives, and different implications for the patient in the chair — and they are printed in the same typeface, in the same table.

The tier is not a measure of importance

The most consequential misreading concerns the tier. The 2017 joint framework from the Association for Molecular Pathology, ASCO and the College of American Pathologists sorts somatic variants into four tiers: strong clinical significance, potential clinical significance, unknown clinical significance, and benign or likely benign.¹ Read quickly, that scans as a ranking of how important each variant is.

It is not. The tier measures how much clinical evidence exists for that variant in the tumor type being tested. The same alteration, in the same gene, with the same biological consequence, can sit in Tier I in one cancer and Tier III in another — because the evidence base grew up around one disease and not the other. Nothing about the tumor differs between those two reports. The literature does.

This matters most in Tier II, which is where the framework places variants supported by investigational therapies, by small published studies, or by a therapy approved in a different tumor type. That last category is the trap. A Tier II line can name a familiar, fully approved drug and still be pointing at off-label use or a trial rather than standard care. The drug name carries the authority; the tier carries the caveat; the two sit in adjacent columns.

How consistently that framework is applied is itself variable. A survey by the College of American Pathologists' Molecular Oncology Committee found adoption of the 2017 tiers has spread worldwide, but with practice differences by geography and institutional setting.⁶ The tier is a property of the evidence base, and to some degree of the laboratory reading it.

Tier III attracts the opposite error. "Unknown clinical significance" gets read as "unimportant," when what it records is that the evidence has not arrived yet.

An illustrative variants table, annotated. Gene labels are deliberately generic — the point is structural, and it survives whichever genes fill the rows. The two lines most readers skip, tumor content and matched-normal status, govern how every number above them should be read.

Actionability has a version number

If the tier reflects the evidence base rather than the tumor, then it moves when the evidence base moves — and it can move a great deal. One centre reannotated 47,271 sequenced solid tumors using two versions of the same knowledge base, deployed five years apart. The proportion carrying a standard-care predictive biomarker rose from 8.9% to 31.6%, while the proportion carrying only non-actionable drivers nearly halved, from 44.2% to 22.8%.²

Those were the same tumors and the same sequence data. What changed was the annotation layer. Which means the actionability section of a report is not a description of the cancer; it is a snapshot of what was known on the date of issue, and it is the section most likely to be wrong in the patient's favour a year later. For a patient still on treatment when their report turns two, requesting a reannotation is a cheaper and more likely intervention than resequencing.

"Actionable" is not "treatable"

The word does a lot of quiet work. On a report, actionable means a target has been identified and published evidence links it to a drug. It does not mean the drug is available to this patient, reimbursed, clinically appropriate given performance status and prior lines, or open at a nearby site.

The size of that gap is well documented. Across a decade at one large institutional precision-medicine program, detection of actionable alterations rose from 10.1% of patients tested in 2014 to 53.1% in 2024. The share of patients actually receiving a matched therapy rose too, but far less — from 1% to 14.2%, mostly through clinical trials. Taken across the whole decade, 23.5% of patients found to have an actionable alteration went on to receive biomarker-guided treatment.³

That is not a failure of the report. It is a description of what the report is for: it identifies candidates. Converting a candidate into a treatment is a separate process involving availability, eligibility and judgment, and the report has no visibility into any of it.

The confidence column, and the one absence that matters

Variant allele fraction is routinely read as clonality — the fraction of tumor cells carrying the alteration. It is not that directly. It is the fraction of sequencing reads carrying it, which is shaped by tumor content in the block the pathologist selected and by copy-number changes at that locus. At 30% tumor content, a variant present in every tumor cell appears at roughly 15%. The tumor-content estimate is therefore not a quality metric to skim past; it is the denominator for the whole table.

VAF also carries a signal most readers miss. The AMP/ASCO/CAP framework notes that apparently non-mosaic fractions — near 50% or near 100% — may be evidence of germline rather than somatic origin.¹ Without a matched normal specimen, that inference is all the assay has, and it is not enough. In a paired study of pediatric solid tumors, 308 of 434 single-nucleotide variants reported from tumor-only sequencing were in fact present in the germline, 31% of them with suggested clinical utility, and the authors concluded that allele fraction alone cannot separate the two.⁴

The consequence runs in both directions. A germline finding can be mistaken for a somatic one, and a genuine hereditary predisposition can be missed entirely: at one academic centre, only half the patients who met criteria for germline testing after tumor sequencing received it, and among those who had both tests, 43% of clinically significant germline variants were found only by the dedicated germline assay.⁵ The tumor report is not a germline test wearing a different hat, and it was never designed to be one.

What each line is, and what could change it without the tumor changing

Line on the report	What kind of claim	What moves it
Gene and variant nomenclature	Observation	Panel content and coverage at that locus
Tier and evidence level	Evidence claim, indexed to one tumor type	A guideline update, a new approval, a knowledge-base version
Named therapy	Pointer, not a prescription	Whether the approval is in this cancer or another one
Variant allele fraction	Read fraction, not cell fraction	Tumor content, copy number, which block was cut
Tumor content estimate	The denominator for every VAF above	Specimen selection by the pathologist
No alteration detected in gene X	Bounded absence	Whether gene X was on the panel and adequately covered

Three questions to ask

1. Is this tier indexed to my patient's tumor type? If a line names an approved drug, ask explicitly whether the approval is in this cancer or a different one. That single question separates standard of care from an off-label conversation, and the report frequently states it only in a footnote.
2. Was a matched normal submitted, and if not, which of these could be germline? Then ask separately whether this patient meets criteria for dedicated germline testing. A tumor-only report is not an answer to that question — it is a reason to ask it.
3. What was the tumor content, and how old is the annotation? Tumor content tells you how to read every allele fraction on the page. The report date tells you whether a reannotation is worth requesting before anything else is ordered.

A tumor molecular report is three documents printed as one: an observation about a specimen, a summary of the current literature, and a statement of the laboratory's confidence. Only the first is about the patient. The second has a version number, and the third has a denominator — and both are usually a footnote away from the line everyone reads first.

References

1. Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. Journal of Molecular Diagnostics, 2017. pubmed.ncbi.nlm.nih.gov
2. Quantifying the expanding landscape of clinical actionability for patients with cancer. Cancer Discovery, 2024. aacrjournals.org
3. Genomic actionability and matched targeted therapy in a decade-long institutional precision medicine program for solid tumors. ESMO Open, 2025. esmoopen.com
4. Germline sequencing improves tumor-only sequencing interpretation in a precision genomic study of patients with pediatric solid tumor. ncbi.nlm.nih.gov
5. Tumor-only sequencing: a story only half told. JCO Precision Oncology, 2024. ascopubs.org
6. Tiered somatic variant classification adoption has increased worldwide with some practice differences based on location and institutional setting: a study from the College of American Pathologists Molecular Oncology Committee. Archives of Pathology & Laboratory Medicine, 2022. meridian.allenpress.com

Reading the Report is Zetobit's series for the people who receive genomic output rather than produce it — clinicians, principal investigators outside genomics, and the teams who have to act on a result. Zetobit, LLC · Lexington, KY · zetobit.com