Gastric Cancer

Primary malignancy of the stomach, often a highly invasive disease associated with high mortality.

top D I A B M I M home


Common worldwide (4th biggest cancer, 2nd biggest killer)
Reducing in West; used to be highest killer; now 1/3 what it was 35 yrs ago (poss. related to H. Pylori).
(Trend is decreasing distal, increasing cardia/GOJ, increase in diffuse type, represents 30% of neoplasms.


Relatively low in West, high in SE Asia, Japan (decreasing with Westernization), Russia, S America.
2-4x in Polynesians.



Mean 60s at discovery
Rare under 40 except in familial forms.

top D I A B M I M home


Gastric CA = 90% adenocarcinomas.
10% others (mainly lymphoma, GIST).
Occasional others: SCC, oat cell carcinoma, carcinoid, mesodermal tumours etc.

AdenoCa Predisposing factors

Familial Gastric cancer
1. Hereditary diffuse gastric cancer
- 2 or more cases in first/second degree relatives; one diagnosed before 50
- 3 or more independent of age
- E-cadherin gene inactivating mutation
- esp common in Maori.
2. Familial diffuse gastric cancer
3. Familial intestinal gastric cancer

Chronic gastritis usual background.
H. pylori main cause of this (3-4x risk if infected).
- risk proportionate to H pylori serum levels.
Atrophy of gastric mucosa (decreased acidity, more bacteria, produce toxins), assoc. with pernicious anaemia.
Also chronic peptic ulcers.
- not significantly (h. pylori is), but 1% associated by virtue of chronic inflammatory change.
Vitamin C deficiency.
Low fruit and veg, high starch.
Dietary toxins (nitrosamines, ?polycyclic hydrocarbons, dried, smoked and salted food).
Previous gastric surgery.
- 2% risk of cancer at a gastroduodenal anastomosis
- typically 15-20y after surgery; young surgical at risk.
- chronic gastritis and atrophy, fundic gland polyps, hyperplastic polyps also more common in the remnant.
- both intestinal and diffuse types, 40% early,
Also in blood group A people; ?why.
Menetrier's Disease
- rare; rugal hypertrophy and hyperplasia, hypochlorhydria and protein-losing enteropathy.
- cancer in 10%, often diagnosed within 12mo of each other.

top D I A B M I M home




Lauren Classification
- Intestinal (well differentiated; mucosal origin)
- or diffuse (poorly differentiated; lamina propria origin)

Intestinal (20% 5-yr survival)
- associated with chronic gastritis and intestinal metaplasia, H. pylori.
- tend to be well differentiated and expansile
- high risk populations, M>F 2:1; older pts; declining
- tends to form glands and met via blood.

Diffuse are ominous (10% 5-yr survival)
- no elements secrete mucous; unclear cause; ?genetic, ?h. pylori
- tend to be poorly differentiated and infiltrative
- tend to affect F>M, younger pts and may be more genetically determined.
- spreads via submucosal infiltration

Classical development theorized as:
-> H pylori infection
-> Acute gastritis
-> chronic active gastritis
-> atrophic gastritis
-> intestinal metaplasia
-> dysplasia
-> gastric cancer
This is not fixed; 60-70% of mild-mod dysplasia may revert to normal; severe = much more likely to progress.
This only applies to intestinal type; diffuse tend not to have association with chronic gastritis / atrophy.

Natural History
Arises in the mucosa.

Early Gastric Ca
When confined to the mucosa or submucosa, regardless of lymph involvement.
5 yr survival of such cancer good (90%), diagnosis in West very poor (10% early)
- in Japan, screening with double contrast barium meal and endoscopy means more (e.g. 30%) detected early.
- better when confined to submucosa (90%) vs submucosa (75%) - 15 yr survivals.
Predominantly distal stomach, most 2-5cm; if <5mm then 'minute'.
LN risk <5% with mucosal disease; 20% with submucosal invasion.
- higher risk with ulceration, larger tumours

Early Gastric Ca Macroscopic Classification:

Type I: protuberant
- sessile, smooth nodules with a broad stalk, often pale
Type II: superficial (a=elevated; b=flat; c=depressed)
- slightly
Type III: ulcerating
Exophytic tumours = better prognosis (more intestinal type); Ulcerating = diffuse.

Commonest in the antrum (40%), but diffuse cancers of prox stomach (30%) becoming more common in West, unclear why.
Spread into the duodenum is a poor prognostic sign (5-yr survival approaching 5%).

Intramural, direct extraluminal growth, lymph, and mets.
- involves regional nodes at least in 3/4 at presentation.
Intragastric spread often proximal > distal
- pylorus is a partial barrier; tumor in 25% past bulb still.
Later liver, bone, brain, lung.
5 yr survival rate for advanced Ca is below 15%.

Subtypes : Borrman Classification
Four subdivisions correlate loosely with natural history and outcome
In order of nastiness:

I. Polypoid carcinomas
- large, bulky intraluminal growths that tend to metastasize late.
- often soft and red in colour with a broad base

II. Ulcerating / excavated carcinoma
- deep penetrating ulcer-tumor; all layers deep
- flat edges instead of rolled edges

III. Ulcerative with elevated margins, infiltrated

IV. Linitis plastica type
- all layers thick, spreads with a marked desmoplastic reaction.
- cure is rare, spread advanced

*There is also a lymphoepithelioma-like carcinoma, often associated with EBV, with a relatively good prognosis.


Disruption of the genome, participation of oncogenes, mutation of tumor-suppressor genes.
p53 is most common abnormality; tumour suppressor gene involved in preserving stability during DNA turnover
- if not, then usually microsatellite instability; different pathway for collecting genetic abnormalities.
APC and DCC genes also commonly affected.
E-cadherin (cellular adhesion molecules) reduced expression in 90%, esp diffuse types.
Familial clustering in HNPCC and familial gastric cancer.


Late cancers risk perforation.
Local perf with abscess, internal or external fistula possible, especially in the ulcerating form.

top D I A B M I M home



No symptoms when superficial and curable.
Later varies from:
- vague non-specific symptoms e.g. early satiety, heaviness, not identified as pain.
- to severe steady abdo pain (though this often slight and can be relieved by food).
May have had dyspepsia for years, but persistence is new.

Anorexia (often w meat), dyspepsia,
Nausea, vomiting (severe if near pylorus), and altered bowel habit possible.

Less frequently dysphagia (if involves cardia)
Haemorrhage (haematemesis / malaena), perforation

Loss of weight and strength, malaise.
Frequently anaemic symptoms.

Local spread.
Symptoms from lymph, liver, bones, brain, lung.


May be jaundice, cachexia, anaemia.
If metastatis - appropriate signs to liver, bone, brain, lung.
Associated with dermatomyositis (non-suppurative inflammation of skin) and acanthosis nigricans (thickened, hyperpigmented skin at flexures).

Palpate; Advanced Disease
Palpable tumour usually means diagnosis too late.
Virchow's Node (left supraclavicular fossa); Sister Mary Joseph Nodule
?Large liver
?Malignant ascites.
Rectal or PV may show Blumer's shelf (pre-rectal mets) in advanced disease.

Consider also lungs, bone, brain.
top D I A B M I M home



Gastroscopic biopsy is standard, with biopsy for all pts with ulcer (6+ bxs)
- GC ulcers are suspect.

Role of EUS

Can show depth of tumor (80% accuracy), and lymph nodes (only 50% accuracy!)
But specifically useful for the showing depth of invasion
- differentiating T2 tumours, which need neoadjuvant chemo and may be amenable to EMR in specialist centres.

Staging depends on CT, EUS and laparoscopy
- CT shows metastatic disease
AXR - stomach rigid if 'leather bottle'.
Ba meal study (double contrast)

Staging Laparoscopy
Highly advisable; done as a day stay procedure.
Shows peritoneal metastatic disease in 30% of pts with normal cross-sectional imaging.
Collect peritoneal washings
- if +ve, predicts early peritoneal recurrence.
- in absence of visible mets, decreases survival from 98.5 months to 14.8 months,

Role of PET
Less utility in gastric cancer as many are not FDG avid.
- only 60%; vs ~90% of esophageal cancers.
More likely in intestinal-type tumours vs diffuse signet-ring type cancers.


Leucocytosis possible.
May be picked up on FOB (50% +ve)

CEA +ve in 65%, usually with extensive disease

Dysplasia in general is divided into low-grade and high-grade
- on basis of cytological vs architectural
Spread correlates with degree of differentiation.
- glandular = better than 'diffuse'
- 'diffuse' type often associated with substrantial stromal component.
- 'Signet ring' = poorly differentiated; bad; more in women, younger patients
Lymphovascular invasion is also prognostic.


Tx = unassessable.
T0 = no primary tumour evident.
Tis = in-situ (intra-epithelial; not invading lamina propria).
T1 = Invades lamina propria or submucosa.
T2 = Invades muscularis (T2a) or subserosa (T2b)
T3 = Invades serosa / visceral peritoneum, but not adjacent structures
T4 = Invades adjacent structures.
Nx = unassessable
N0 = no regional nodes
N1 = 1-6 regional nodes
N2 = 7-15 nodes
N3 = >15 nodes
[need 15 nodes in specimen for accurate staging]
Mx = unassessable
M0 = no mets
M1 = distant mets

0      Tis
IA    T1N0M0
IB    T1N1M0, or T2a/bN0M0
II     T1N2M0, T2N1M0, T3N0M0
IIIA T2N2M0, T3N1M0, T4N0M0
IV    T4N1-3M0, T1-3N3M0, anyTanyNM1
- ie T4 or N3 or M1 is very bad news. 

top D I A B M I M home

Multidisciplinary; gastroenterologists, surgeons, oncologists, radiologists.

Done in Japan  with mobile X-ray unit and double contrast study, increasing early diagnosis and survival rates.
However, not justified as overall benefit is minimal, and at present even in dyspeptic sufferers, not sufficiently feasible.


Surgical excision remains the cornerstone of curative therapy.
- possible in <25%
Want an R0 resection, radical lymphadenectomy, 5cm margin (due to submucosal spread).

What do do with R1 resection?
Controversial.  Decide based on patient's overall condition and treatment goals
- makes little difference when there is advanced disease.
- but if early tumour / good lymph status and fit for surgery, consider it.

See gastrectomy

Role of proximal partial gastrectomy?
Intractable reflux limits utility.
Avoided generally.

Extent of Lymphadenectomy?
Remains controversial.
D0 = no nodes, i.e. palliative resection
D1 = excision of perigastric nodes
D2 = nodes along trunks of coeliac axis.
Radical in Japan, where evidence supports D2 clearance.
Most Western surgeons are skeptical of radical lymph resections.
- Dutch Gastric Cancer Group failed to show benefit of D2>D1 in survival or recurrence-free survival, but did show higher morbidity.
- Similar results from British Medical Research Council Trial.
--> distal pancreatectomy and splenectomy were included; contributed to morbidity.
--> when not included, non-randomized trials show comparable morbidity D1=D2 and improved survival.
Aggressive lymphadenectomy allows adequate staging.
- No. +ve nodes, regardless of location determines survival.
Need at least 15 nodes for accurate staging and improved survival at all disease stages.
--> likely due to less understaging
However, new long-term follow-up data emerging that may state very-long-term survival benefit weighs in favour of D2, unclear.

Frozen Section?
Good practice to send routine frozen section to ensure R0 resection.
- before constructing anastomosis; may need extension.

Emerging role, currently mainly in Japan where tumours diagnosed very early.
Little role in West where most are found late
Also require highly specialist center and some may not be amenable, such as GOJ tumours or at high curvatures.

Palliative Surgery?
No role at present; no evidence of survival and cannot be justified for QOL.
Significant morbidity (50% complications)
Observational studies show only 10% of those who have operations aborted will undergo subsequent surgical intervention.
If obstructs, stent is best option.

Adjuvant Therapy
Many R0 patients have recurrent disease in <2 yrs, hence adjuvant focus.
Survival benefit was initially not shown on RCTs for chemo/rad.
Intergroup 116 trial showed both overall and relapse-free survival better with 5-FU and lecovorin with external beam radRx.
--> now standard of care in US
--> however, muddied by poor quality and standardization of medical therapy; 54% had R0, only 10% R2.
Newer trials from Japan beginning to show promise using different agents.
- however, relevance to non-D2 population in West unclear.
Bottom line:
When accurately staged and standardized lymph resection, adjuvant chemorad not known to clearly surpass that of surgery alone.
This is the focus of the Top Gear study, contributing out of Australia

Neoadjuvant Therapy
>1/3 patients cannot tolerate post-op chemo; better tolerated pre-op.
Better tolerance, can assess disease response in-vivo, downstaging potential.
MAGIC trial (British Medical Council Adjuvant Gastric Infusional Chemo) = periop epirubicin, cisplatin, 5-FU, vs surgery alone = better overall and progression free survival, and better resectability.
Bottom line:
--> Give neoadjuvant chemo for patients with T2 or higher tumours.

Nomogram on p84 of Cameron.
- Age (the young and very old do worse)
- Sex (Men worse)
- Primary site (GEJ bad)
- Lauren (Diffuse worse)
- Size (20 = bad)
- Nodes (5 = bad, >10 = very bad)
- Depth (deeper = very bad)

Many elderly with lymph spread, contra-indicating surgery.
Palliation, pain relief, relief of dysphagia, referral to med onc.
In advanced metastatic disease, mild chemo may help; 5-fluorouacil and cisplatin.
Generally gastric CA radioresistive, though it has some role in palliation.

top D I A B M I M home