Implications of Genetic Testing on Radiation Therapy Decisions
TL;DRASTRO think tank consensus: heterozygous germline DNA-repair variants (BRCA1/2, PALB2, CHEK2, RAD50) and VUS should not alter indicated RT; young deleterious ATM carriers are the exception.
REQUITE
The ATM carve-out is smaller than it sounds: an RR of 1.3 moves 5-yr contralateral breast cancer risk from 2.9% to an estimated 3.8%, and the panel calls toxicity data evenly split. Indicated RT in young ATM carriers still goes ahead with close CBC screening. Optional RT, such as omission-eligible adjuvant breast RT, is the decision this sharpens.
In a BRCA1/2, PALB2, CHEK2 or RAD50 carrier, or a pt with only a VUS, this consensus supports offering indicated RT unchanged; it does not cover recognized radiosensitivity syndromes such as ataxia telangiectasia or Li-Fraumeni.
The ATM carve-out is smaller than it sounds: an RR of 1.3 moves 5-yr contralateral breast cancer risk from 2.9% to an estimated 3.8%, and toxicity data are evenly split. Indicated RT still goes ahead with close CBC screening; optional RT in young ATM carriers is the decision this sharpens.
13 details
ASTRO-convened think tank, February 2019, of physicians and scientists in genetic screening, genomics, radiosensitivity, risk estimation and clinical oncology. Narrative review with consensus points; no voting or agreement metrics reported.
Addresses pts with germline or somatic variants found on hereditary screening, direct-to-consumer CLIA testing, or tumor genomic panels. Panel recommends only CLIA-certified tests inform RT discussion and prefers paired tumor-normal sequencing.
VUS = normal. BRCA1/2: no excess RT toxicity or RT-induced second cancers, extended to PALB2, CHEK2, RAD50. Deleterious heterozygous ATM: avoid optional RT in young pts, close CBC screening after indicated RT. RB1, NF1, TP53, PTCH1 syndromes carry elevated in-field second malignancy risk.
| Germline finding | RT recommendation | Basis in source |
|---|---|---|
| VUS | Treat as normal | Not actionable; RT risks similar to normal |
| BRCA1/2 | No change to indicated RT | Consistent data: no excess toxicity or RT-induced second cancers |
| PALB2, CHEK2, RAD50 | No change to indicated RT | Extended by analogy from BRCA |
| Deleterious heterozygous ATM | Avoid optional RT in young pts; close CBC screening after indicated RT | Possible RT-attributable CBC excess; toxicity data mixed |
| RB1, NF1, TP53, PTCH1 syndromes | Elevated in-field second malignancy risk | No known excess acute or late toxicity |
Somatic ATM truncation tracked with improved local control after palliative RT (P < .001) without OS difference, read as an RT-specific effect. No tumor mutation is yet clinically actionable for dose.
NCCN guidance covers counseling for common germline mutations but gives little RT-specific risk information. REQUITE (~4400 pts, 8 countries) and the Radiogenomics Consortium are named as the routes to validated polygenic toxicity predictors.
The ATM absolute risk is a rough estimate built from a SEER baseline and one cohort's RR, not a measured carrier-specific rate. Most cited studies lack a no-RT control, so RT-attributable risk is often inferred.
Panel proposes future genomic-adjusted RT trials include a no-RT arm to separate prognostic from predictive value. Polygenic, not single-gene, predictors are expected to be what eventually becomes actionable.
- Does heterozygous ATM truly raise RT toxicity risk?
- Validated polygenic risk score for radiation toxicity
- Genomic-adjusted RT dose tested against a no-RT arm
📚 Sources · 📄 1 paper
The longer read
The document's main contribution is a default. A heterozygous germline hit in a DNA-repair gene is not, on its own, a reason to change an indicated RT plan. That default rests on uneven footing. For BRCA1/2 the panel cites consistent clinical data showing no excess toxicity or RT-induced second cancers with standard adjuvant regimens, and this is the firmest ground in the paper. PALB2, CHEK2 and RAD50 are then brought under the same umbrella as phenotypically similar heterozygous repair variants, by analogy rather than from direct outcome data. A reader should keep those two tiers separate: the first is measured, the second is reasoned.
ATM is where the paper earns its carve-out, and where its evidence is thinnest. The panel concedes that data on heterozygous ATM and toxicity cut both ways about equally, yet still recommends avoiding optional RT in young carriers. The justification is the contralateral breast cancer signal plus the age gradient of radiation carcinogenesis, not toxicity. The Box 3 arithmetic is the most usable counseling tool in the paper: a relative risk of 1.3 becomes a 5-year move from 2.9% to roughly 3.8%. That figure joins a population baseline to one cohort's relative risk, so it is an order-of-magnitude estimate, not a carrier-specific measurement, and should be presented to pts that way.
The second load-bearing idea is the line between syndromic and incidental findings. The rare syndromes in Table 1 carry real in-field second malignancy risk; hereditary retinoblastoma, with 30-year second cancers of 35% after EBRT vs 6% without RT, shows the stakes. The panel argues these syndromes usually declare themselves in childhood, so an adult panel result is unlikely to be one. The exception it flags is paired tumor-normal sequencing, which surfaces occult TP53 carriers in otherwise low-risk pts. That is the setting where an adult finding deserves a closer look before RT.
The somatic material is more speculative but more relevant to dose. Truncating somatic ATM mutations tracked with better local control after palliative RT with no OS difference, which reads as an RT-specific effect. Still, the panel itself states no tumor mutation is clinically actionable, and its proposed trial design, with a no-RT arm to separate prognostic from predictive value, is the right bar for any future genomic dose-adjustment claim.
What would overturn this? A large prospective toxicity dataset such as REQUITE showing a clinically meaningful excess for a specific heterozygous variant. The consensus process is also undocumented: no votes, agreement rates or evidence grades are reported, so this carries the weight of expert opinion from a 2019 convening. Of its conclusions, the BRCA position is least likely to move; the ATM recommendation is most exposed to new data in either direction.