Clonal hematopoiesis is the age-related expansion of a single hematopoietic stem cell clone carrying a somatic variant, in a person with no cytopenia and no diagnosed blood cancer. It matters to the clinical genetics laboratory for one reason above all others: blood is the usual source of "germline" DNA, so a somatic clone in blood can be reported as a germline finding. Every high-stakes misinterpretation in this area follows from that single fact.
| Term | Definition |
|---|---|
| CHIP (clonal hematopoiesis of indeterminate potential) | A somatic variant in a known leukemia-associated driver gene at variant allele fraction (VAF) of at least 2%, with normal blood counts, no dysplasia, and no hematologic neoplasm |
| CCUS (clonal cytopenia of undetermined significance) | The same clone, but with an otherwise unexplained cytopenia. Higher progression risk than CHIP, and a hematology referral |
| ARCH (age-related clonal hematopoiesis) | The broader umbrella term for any clonal expansion with age, driver gene or not |
| mCA (mosaic chromosomal alterations) | The copy-number counterpart, seen on SNP array as low-level mosaic gains, losses, or copy-neutral events. Mosaic loss of Y in men and loss of X in women are the most common |
- Uncommon before age 40; present in roughly 10% of people over 70 and more than 20% over 80 at standard sequencing depth
- With ultra-deep sequencing, small clones are detectable in nearly everyone of advanced age, which is why the 2% VAF threshold exists at all
- The dominant drivers are DNMT3A, TET2, and ASXL1, together roughly two-thirds of cases, followed by PPM1D, TP53, JAK2, SF3B1, and SRSF2
- Prior cytotoxic chemotherapy or radiation selects for a distinct set of clones, enriched for TP53 and PPM1D
1. False germline calls. The variant that matters most is TP53. An apparent germline TP53 variant found in blood at a VAF of 20 to 35% may be clonal hematopoiesis rather than Li-Fraumeni syndrome, and the difference decides whether an entire family enters lifetime surveillance. Confirmation requires a non-hematopoietic tissue, and cultured skin fibroblasts are the standard. Buccal swab and saliva are not adequate, because both are substantially contaminated with leukocytes. Nail clippings and hair follicles are usable alternatives.
2. Cell-free DNA and liquid biopsy. Clonal hematopoiesis is the largest single source of false-positive "tumor" variants on plasma cfDNA panels, because lysed leukocytes contribute DNA to plasma. Paired sequencing of the buffy coat alongside the plasma is the fix, and is now expected of a well-designed assay.
3. Tumor-only sequencing. Without a paired normal sample, a clonal variant carried by tumor-infiltrating leukocytes reads as a somatic tumor driver and can misdirect therapy.
4. Prenatal cell-free DNA screening. Maternal clonal hematopoiesis, and occult maternal malignancy, are recognized causes of discordant, non-reportable, or multiple-aneuploidy results on noninvasive prenatal screening. A result reporting several aneuploidies at once is a maternal finding until proven otherwise.
5. Microarray. Low-level mosaic events on a chromosomal microarray run on blood may reflect clonal hematopoiesis rather than constitutional mosaicism, particularly mosaic loss of Y in an older man.
Features that argue the variant is clonal rather than constitutional:
- VAF meaningfully away from 50%, most often below it, though a large clone can approach or exceed 50%
- The gene is a known clonal hematopoiesis driver, especially DNMT3A, TET2, ASXL1, PPM1D, or TP53
- Older patient, and no family history consistent with the syndrome
- Prior chemotherapy or radiation
- The specific variant is a recognized hotspot in myeloid neoplasia rather than a recognized germline syndrome variant
- Absence of the variant in a fibroblast sample, which settles it
None of these is decisive alone. Confirmatory testing on non-hematopoietic tissue is the answer whenever the distinction changes management.
- Progression to a hematologic malignancy runs about 0.5 to 1% per year. Risk rises with VAF above 10%, more than one driver mutation, and spliceosome or TP53 drivers
- Cardiovascular risk is the more likely consequence. Clonal hematopoiesis roughly doubles coronary heart disease risk through an inflammatory mechanism involving interleukin-1 beta and interleukin-6 signaling in clone-derived macrophages. For most people this outranks the leukemia risk
- It is not a cancer diagnosis and should not be delivered as one. Refer to hematology for CCUS, a high VAF, or multiple drivers; otherwise the actionable advice is conventional cardiovascular risk management
A laboratory returning germline results should have a written policy for suspected somatic findings. The expected practice is to report the variant with an explicit statement that it may be somatic, and a recommendation for confirmatory testing in a non-hematopoietic tissue, rather than to either suppress it or report it as germline.
- A TP53 variant in blood at a VAF well under 50% in an older adult, with no supportive family history, is clonal hematopoiesis until fibroblasts say otherwise. This is the highest-consequence version of the error.
- Saliva and buccal swabs are not germline tissue. Both carry enough white cells to reproduce the blood result. Cultured fibroblasts are the reference sample.
- Multiple aneuploidies on prenatal cell-free DNA screening point at the mother, not the fetus. Clonal hematopoiesis and occult maternal cancer are the two explanations to work through.
- Age is the strongest predictor. Clonal hematopoiesis in a 35-year-old deserves more scrutiny than the same finding at 80.
- Chemotherapy history changes the prior. Treated cancer patients carry TP53 and PPM1D clones at much higher rates, which is exactly the population most likely to be sent for germline cancer testing.
CHIP is a chip off the marrow block: one stem cell clone, growing with age, that will masquerade as germline unless you look at skin.