Cancer

Oesophageal Cancer

Reviewed September 10, 2026 11 views

Two different cancers share the same tube. The squamous type is the cancer of East Asia and its genetics lead straight to the alcohol flush gene; the adenocarcinoma type is the fastest rising cancer of high-income countries and grows out of years of reflux. This page holds variants for both and keeps them apart.

What this condition connects to

Oesophageal Cancer Variant: rs10058728 rs10058728 Variant Variant: rs2274223 rs2274223 Variant Variant: rs8030672 rs8030672 Variant Variant: rs11066015 rs11066015 Variant Variant: rs4785204 rs4785204 Variant Variant: +6 more +6 more Variant Oesophageal Cancer Oesophageal Cancer Cancer
Prevalence
A genome-wide association study of oesophageal squamous-cell carcinoma in 2,031 cases and 2,044 controls of Chinese descent, replicated in 6,276 cases and 6,165 controls, identified seven loci and found that variants on 12q24 confer risk in a gene-lifestyle manner with more pronounced risk in tobacco and alcohol users (Wu et al., Nature Genetics 2011, PMID 21642993). A 2012 analysis confirmed the ALDH2 locus and found that drinkers with both the ADH1B and ALDH2 risk alleles had a fourfold increased risk compared with drinkers without them (Wu et al., Nature Genetics 2012, PMID 22960999). For adenocarcinoma, a meta-analysis of 6,167 patients with Barrett's oesophagus, 4,112 with adenocarcinoma and 17,159 controls identified eight new loci including CFTR (Gharahkhani et al., Lancet Oncology 2016, PMID 27527254).
Inheritance
Common variants, each shifting risk slightly on its own. The squamous half is the exception on this site: at the ALDH2 and ADH1B loci the genotype combines with drinking into a risk several times larger than either alone. The adenocarcinoma half is inherited risk acting on a lining already changed by reflux.

The oesophagus is the tube that carries food from the throat to the stomach. Two different cancers arise in it, and they are different enough — in where they occur, who gets them, what causes them and what the genetics finds — that this page treats them as two halves.

Squamous cell carcinoma: the East Asian cancer, and the flush

Squamous cell carcinoma starts in the flat cells that line the upper and middle oesophagus. It is the dominant type in East Asia, where the 2011 study behind most of this half was done, and its two great risk factors are alcohol and tobacco.

That study — 2,031 cases and 2,044 controls of Chinese descent, replicated in a further 6,276 cases and 6,165 controls — found seven loci, and the one that matters most for a reader is on chromosome 12q24, the region of the ALDH2 gene. The paper states that the 12q24 variants confer their risk in a gene-lifestyle manner, with more pronounced risk enhancement in tobacco and alcohol users. A 2012 study from the same group, validated in 8,092 cases and 8,620 controls, confirmed the ALDH2 locus, found the ADH alcohol-metabolism cluster on 4q23 interacting with drinking, and put a figure on the combination: drinkers carrying both the ADH1B and ALDH2 risk alleles had a fourfold increased risk compared with drinkers carrying neither.

ALDH2 is the gene behind the alcohol flush reaction. A person who reddens after one drink is a person whose ALDH2 works slowly, so that acetaldehyde — the toxic intermediate between alcohol and vinegar — lingers. The flush is uncomfortable; the acetaldehyde is the reason this cancer's genetics keep returning to the same gene. The variant on this page in that region, labelled by the catalogue by its nearest gene ACAD10, sits 36 kilobases from ALDH2; the flush page holds rs671, the variant that actually slows the enzyme.

Also in this half: PLCE1 on 10q23, one of the earliest and best replicated squamous-cell loci; CHEK2, a DNA damage checkpoint gene that appears in the genetics of several cancers; and CSNK1A1, ANP32A and HEATR3.

Adenocarcinoma: the reflux cancer, and Barrett's

Adenocarcinoma starts in gland-forming cells near the bottom of the oesophagus, and it is a different story. It is one of the fastest rising cancers in high-income countries, and it grows out of Barrett's oesophagus — a change in the lining, caused by years of acid reflux, in which the normal flat cells are replaced by cells like those of the stomach and intestine.

The second half of this page comes from a 2016 meta-analysis of every genome-wide study of Barrett's and adenocarcinoma then published: 6,167 people with Barrett's, 4,112 with adenocarcinoma and 17,159 controls, all of European ancestry. Because Barrett's and the cancer that follows it were analysed together, the trait on those variants reads "Barrett's oesophagus or oesophageal adenocarcinoma", and this page holds them under that wording rather than pretending the study separated what it did not.

The paper's framing is the useful one. Only a few people with Barrett's ever develop adenocarcinoma, and there is no valid predictor of which — which is what makes managing Barrett's difficult, and what the study was looking for. It found eight new loci, among them CFTR, the gene of cystic fibrosis, and a locus at TPPP and CEP72. The strongest pathways it identified were muscle cell differentiation and mesenchyme development: the genetics of a lining that changes under acid points at the tissue that decides what a lining becomes.

Clinical detail

What the two halves mean in practice

They mean different things, which is why the page is divided.

For the squamous type, the message is about drinking. The flush reaction is a signal that ALDH2 is slow, and the 2012 study's fourfold figure is what that signal means for someone who drinks anyway. It is one of the few places on this site where a genotype and a habit combine into a risk large enough to act on, and the action is the obvious one.

For the adenocarcinoma type, the message is about reflux. Persistent heartburn over years is what leads to Barrett's, and Barrett's is what is watched. Whether someone is offered surveillance is decided on what an endoscopy shows, not on a genotype, and none of the variants here predicts which Barrett's will progress — the 2016 paper says that predictor does not yet exist.

The symptom that gets this cancer found. Difficulty swallowing — food sticking, or having to change what you eat because of it — especially with unexplained weight loss, is the symptom that leads to an endoscopy. It is worth a prompt appointment whatever your genotype, and nothing on this page detects the disease or should delay that appointment.

What this page cannot do

  • It cannot screen for oesophageal cancer. An endoscopy does, and it is offered on symptoms and on a known Barrett's diagnosis.
  • It cannot say which Barrett's will become cancer. That is the open question the adenocarcinoma study set out to answer, and its own conclusion is that the predictor is not there yet.
  • It cannot tell you whether you flush. You already know. The ALDH2 page explains why, and what it means for drinking.

Related here

The ALDH2 deficiency page holds rs671 and rs1229984, the two alcohol-metabolism variants the 2012 study's fourfold figure is built on. The lung cancer page is the other place tobacco does this kind of work in the genetics.

Related variants MyGeneLog checks for

What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Oesophageal Cancer comes down to these specific, well-studied positions — not a diagnosis.

Sensitive

Esophageal cancer

CSNK1A1 · rs10058728

See detailed info →
Sensitive

Esophageal cancer

PLCE1 · rs2274223

See detailed info →
Sensitive

Esophageal cancer

ANP32A · rs8030672

See detailed info →
Sensitive

Esophageal cancer

ACAD10 · rs11066015

See detailed info →
Sensitive

Esophageal cancer (squamous cell)

HEATR3 · rs4785204

See detailed info →
Sensitive

Esophageal cancer (squamous cell)

CHEK2 · rs4822983

See detailed info →
Sensitive

Barrett's esophagus or Esophageal adenocarcinoma

CFTR · rs17451754

See detailed info →
Sensitive

Barrett's esophagus or Esophageal adenocarcinoma

LOC732275 · rs1979654

See detailed info →
Sensitive

Barrett's esophagus or Esophageal adenocarcinoma

ALDH1A2 · rs2464469

See detailed info →
Sensitive

Barrett's esophagus or Esophageal adenocarcinoma

TPPP · rs9918259

See detailed info →
Sensitive

Esophageal adenocarcinoma

TPPP · rs75783973

See detailed info →

Sources

Databases, guidelines and references

Papers, with their authors

Frequently asked questions

I go red when I drink. Does that mean I will get oesophageal cancer?

No, but it is the one genotype on this page worth acting on. Flushing means the ALDH2 enzyme is slow and acetaldehyde lingers after a drink, and the 2012 study found that drinkers carrying both the ADH1B and ALDH2 risk alleles had a fourfold increased risk compared with drinkers carrying neither. The risk is in the drinking; a person who flushes and does not drink is not in that group.

I have Barrett's oesophagus. Can these variants tell me whether it will turn into cancer?

No. The 2016 meta-analysis behind those variants set out to find exactly that predictor and concluded that it does not yet exist. Only a few people with Barrett's go on to develop adenocarcinoma, and which ones is decided by what repeat endoscopy shows, not by a genotype.

Why are there two different cancers on one page?

Because they share a tube and nothing else. Squamous cell carcinoma is the cancer of East Asia, driven by alcohol and tobacco, and its genetics return the alcohol-metabolism genes. Adenocarcinoma is the cancer of high-income countries, grows out of reflux, and its genetics return tissue-development genes. A reader looking for oesophageal cancer should find one page; a reader who reads it should find two diseases.

What symptom should make me see a doctor?

Difficulty swallowing — food sticking, or changing what you eat because of it — particularly with unexplained weight loss. That is what leads to an endoscopy, which is the test that finds this cancer. No genotype changes that advice.

Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.