23GIST — gastrointestinal stromal tumour
Snapshot. The commonest mesenchymal (sarcoma-family) tumour of the GI tract — NOT an adenocarcinoma, and none of the epithelial-cancer logic applies. Driven by a single kinase: KIT (~80%) or PDGFRA (~10%); the rest are KIT/PDGFRA wild-type (mostly SDH-deficient — young, gastric, multifocal, indolent — or NF1-associated). Stomach ~60% · small bowel ~30%; spreads to liver + peritoneum (nodes and lungs are rare). 🔴 THE MUTATION DECIDES THE DRUG AND THE DOSE → MUTATIONAL TESTING BEFORE ANY TKI, ALWAYS — the one unforgivable miss: PDGFRA D842V does NOT respond to imatinib. Risk of recurrence = size + mitotic rate + site (gastric behaves better than small bowel/rectal) + rupture (rupture ≈ treat as high risk).
Workup
CT abdomen/pelvis with contrast (+ baseline chest)
liver + peritoneum are the metastatic sites · EUS (endoscopic ultrasound) characterises the primary
Biopsy → EUS-guided core, NOT percutaneous if resectable
GISTs are soft and vascular — a percutaneous track risks rupture/seeding; if imaging is classic and it is going to theatre anyway, tissue can wait for the specimen
IHC → KIT (CD117) + DOG1
the diagnostic pair · a KIT-negative GIST still exists (DOG1 catches it, often PDGFRA)
🔴 Mutational analysis — MANDATORY before any TKI
KIT exon 11 (imatinib 400) · KIT exon 9 (imatinib 800) · PDGFRA D842V → imatinib-RESISTANT → avapritinib · wild-type → SDHB IHC + germline SDHx referral, and NGS for NTRK fusions (the rare wild-type with its own drug)
Risk-stratify every resected primary → size · mitotic count · site · rupture
this is what decides adjuvant imatinib
Incidental gastric <2 cm, asymptomatic → EUS surveillance is an option
| Setting | Treatment |
|---|
| Localized, resectable | Surgery — R0, NO routine lymphadenectomy, do NOT rupture it
nodal spread is rare (skip the nodal surgery an adenocarcinoma would get) · intact-tumour handling is oncologic technique, rupture converts the case to high-risk HIGH-risk resected + imatinib-sensitive mutation → adjuvant IMATINIB ×3 years (SSGXVIII)
3 yr beat 1 yr on recurrence-free AND overall survival · do NOT give adjuvant imatinib to D842V / most SDH-deficient (insensitive) |
Localized, morbid surgery
GEJ, duodenum, rectum, or bulky | NEOADJUVANT imatinib → downsize → function-preserving surgery
the whole point is a smaller operation (sphincter, stomach, pancreas preserved) · ⚠ test the mutation FIRST — D842V will not shrink on imatinib (avapritinib or straight to surgery instead) |
Metastatic / unresectable
1st line | IMATINIB 400 mg — and continue UNTIL PROGRESSION; never interrupt a responder (BFR14)
interruption → progression, even in long responders · KIT exon 9 → 800 mg Limited/focal progression on TKI → treat the focal site (surgery/ablation) + CONTINUE the TKI
one resistant clone does not mean the drug has failed everywhere |
On progression
lines 2–4 | 2nd → SUNITINIB · 3rd → REGORAFENIB (GRID) · 4th → RIPRETINIB (INVICTUS)
fixed ladder, all kinase inhibitors · re-biopsy/molecular review where available — resistance mutations differ NTRK fusion (rare, wild-type) → larotrectinib / entrectinib |
SDH-deficient
young, gastric, multifocal | Surgery-led + surveillance — often indolent, TKI-poor
germline SDHx testing + family cascade · avoid reflex adjuvant imatinib |
Watch
Imatinib → periorbital/peripheral OEDEMA, cytopenias, nausea, muscle cramps — and GI-bleed risk from a necrosing tumour early on. Sunitinib → HYPERTENSION, hand-foot syndrome, HYPOTHYROIDISM, cardiotoxicity (check TSH + BP each visit). Regorafenib → hand-foot, hypertension, hepatotoxicity (LFTs). Ripretinib → alopecia, hand-foot. Avapritinib → cognitive/memory effects, rare intracranial haemorrhage. Never stop a responding TKI; manage toxicity with dose holds, not abandonment.