FOR DOCTORS. Complete tumour-board submission of 14.10.2026. Ukrainian version: tumour-board-submission-2026-10-14-UA.html; one-page print versions: English · Ukrainian.

Superscript numbers link either to a numbered case document (documents) or to a numbered literature reference on the case page (Surgical and local-treatment literature for the board).

Tumour board submission: operation planning for imatinib-resistant liver metastases of gastric GIST

Question to the board

Which lesions should be treated now, in which order and by which technique, and how are imatinib and regorafenib managed around the procedures?

Key facts

1. Progression is metabolic, inside pre-existing lesions: L1 21 → 33 → 54 mL (05/2025 → 06/2026 → 09/2026), a new 14 × 23 mm focus in the 79 mL S8 dome lesion, FDG-active plaques on the wall of the size-stable 1.2 L mass R1 INTERPRETATION ^[187].
2. L1 is biopsy-proven viable GIST with low proliferation (Ki-67 2 %) ^[195]; it is 0–0.8 mm from the left portal branch (vein patent) and touches a low-confidence LHV outline INTERPRETATION; thermal ablation was declined for this reason OPINION.
3. ctDNA shows two resistance clones (exon 13 and exon 18); no single approved second-line drug covers both well ^[1].
4. Volumetry (AI-assisted, inferred segments): non-tumour parenchyma 1.41 L; remnant after right hemihepatectomy + S1 ≈ 0.69–0.78 L (49–53 %); after both targets in one stage ≈ S4 only, 0.12–0.26 L (9–20 %) INTERPRETATION.
5. The surgeon’s stage order changed from right-first (18.09.2026) ^[190] to left-first (09.10.2026) OPINION; the reason is not documented.

Lesions (labels used throughout)

Questions for the board (in order of dependency)

1. Operate now? Does the unresolved cytology (category V; AUS on review) stop surgery, or is it settled by peritoneal inspection and frozen-section washings at the start of surgery? [alternatives within days: repeat tap with cell block at CSD if fluid reaccumulates; second-look laparoscopy. Note: frozen section applies to nodules; DOG1/CD117 on a cell block is a post-operative result] ^[2,3]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): he considers the earlier category V finding over-diagnosis against the background of the cancer diagnosis (“they overdid it”; presumably the first report); malignant cells would have been on all slides.
2. Order and staging. Right hemihepatectomy (R1 with the S8 lesion) first or L1 first; one stage or two; and how is R1 — the 1.2 L right-lobe mass with FDG-active wall plaques — handled: resection only, or is drainage/aspiration an option? [surgeon’s order changed 18.09 ^[190] → 09.10 OPINION; remnant estimates in the collapsed block; no published GIST series on drainage] ^[4,14]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): right hemihepatectomy first: the right-lobe lesion has a fluid component that could rupture (an emergency, then carcinomatosis, after which no surgery would be radical); most of the right lobe is non-functioning tumour, so little healthy liver is lost and he sees no high risk of liver failure; the left-lobe lesion at a second stage, because one stage would leave too little functional parenchyma. Drainage or aspiration of R1 was not mentioned.
3. L1 technique and operator. For L1 — 54 mL on the 23.09.2026 consensus model (4.9 × 3.9 × 5.9 cm), biopsy-proven viable GIST (05.10.2026) ^[195], 0–1 mm from the left portal branch and the left bile duct — which option is feasible here, where and by whom: resection of the left lateral section (segments II–III), irreversible electroporation (NanoKnife), histotripsy (HistoSonics), SBRT, or transarterial treatment (DEB-TACE/TARE) of the segment II/III feeding artery? [thermal ablation declined by Dr Korshak, 10–11.09.2026 OPINION; written: thermal ablation of the left-lobe focus “judged technically inadvisable” (discharge summary, 13.10.2026 ^[201]) FACT; conditional per ULIS ^[188]; June 2026: 42 × 36 mm (Israeli reading ^[77]). Access / cost: TARE from about EUR 40,000, not performed in Ukraine; histotripsy at HCA Healthcare UK (London) costs from £37,500 for the treatment itself, plus £500 initial review (oncologist + MDT), about £250 per clinician per visit and about £100 for the radiologist’s imaging review (e-mail reply of 11.09.2026; no quote for her case issued)] ^[5,6,7,8]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): no local treatment of L1 now: for embolisation it is unclear what to embolise; ablation works for small lesions and here would be only partial; if L1 grows before stage two, palliative ablation or chemoembolisation, most likely not radical. (background: case page)
4. Same-session targets. Are the S8 dome lesion (79 mL) with its new 14 × 23 mm active focus, the R1 wall plaques and the calcified gastric primary ^[195] addressed in the same operation, or left or treated separately? ^[9,10]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): the tumour board recommended removing the gastric tumour; in the same operation if a local resection is enough, the extent decided at intra-operative revision (20.10.2026: otherwise left for the second stage).
5. Operator, venue, window. Who operates, where, and in which window (from 19.10.2026? OPINION), and is the contrast CT with volumetry done before the first operation (last contrast study 19.06.2026 ^[70])? ^[14]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): the operation is possible “this Wednesday” if the patient agrees (INTERPRETATION 21.10.2026, computed from the call date); the written tumour-board conclusion was due on 19.10.2026.
6. Imatinib and regorafenib. Imatinib stopped 20.10.2026 before surgery: confirm the restart date after the operation; regorafenib not allowed now: timing of regorafenib or an alternative after the local step? [imatinib stopped 20.10.2026 before surgery, restart a few days after the operation, timing to be discussed; regorafenib dispensed 30.09.2026 ^[76], not started, not allowed currently (family report, 20.10.2026); label: regorafenib ≥ 2 weeks after major surgery and wound healing] ^[11,12,13]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): imatinib “not now”; after the operation she returns to it once her condition allows — consistent with the stop on 20.10.2026 before surgery (family report). Regorafenib was not discussed.
7. PRES safety. After haemorrhagic PRES (2023) ^[22]: peri-operative blood-pressure target, anaesthesia plan, and which VEGFR-active drugs (sunitinib, regorafenib) to avoid around surgery and SBRT? ^[12]
8. Response check. Which check after each step — contrast CT at 4–6 weeks, PET-CT at 3 months, repeat ctDNA — and who orders and reads it? ^[15]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): one CT about two weeks after surgery to assess liver regeneration; no more frequent imaging because of the contrast and radiation load.
9. Next systemic line or trial. Which drug or trial would cover both resistance clones (KIT exon 13 V654A and exon 18 A829P) at the lowest neurovascular risk after her PRES — sunitinib, regorafenib, ripretinib, velzatinib (StrateGIST 3), bezuclastinib + sunitinib or pimitespib — and at what point: now, or only after local control fails? [regorafenib dispensed 30.09.2026, not started, held because surgery is expected soon (family report, 14.10.2026); sunitinib prescribed 30.07 and 08.09.2026, not started; each later TKI line gives a shorter time to the next progression with more toxicity; blood-pressure safeguards needed; full comparison in the drug table] ^[1]
   OPINION Dr Fedorov’s position (16.10.2026, opinion): no systemic therapy other than imatinib was discussed. (background: case page)
10. Immune- and antibody-based options. Is any immune- or antibody-based approach a candidate for her now or alongside local treatment — the KIT-targeted antibody–drug conjugate trial (NN3201, NCT06805825), a checkpoint inhibitor with regorafenib, or an intratumoral or cell-therapy trial — and under what conditions? [evidence weak and mostly preclinical; TMB 3–4.8 Muts/Mb, MSS; OpenEvidence review, sections 13–16]
    OPINION Dr Fedorov’s position (16.10.2026, opinion): immunotherapy is not indicated in her case; immunomodulators are of unproven efficacy; on in-vivo CAR-T: “it would be great if something like that appeared”. (background: case page)

Needed within days (owner): operation protocol of 05.10.2026 (Shalimov) · contrast CT or MRI with volumetry and arterial phase (surgeon/radiology) · coagulation, CBC, biochemistry, ECOG, height/weight (treating team) · anaesthesia review incl. blood-pressure plan and SVC/port status (anaesthesia) · cell-block cytology if fluid reaccumulates (CSD) · EGD done 02.10.2026 (discharge summary ^[201]: erythematous gastropathy, duodeno-gastric bile reflux; the oesophagus is not described; family report 20.10.2026: a cold with a mild sore throat in the days of the 23.09.2026 PET-CT) · histology of the colon neoplasias seen at colonoscopy 02.10.2026 (Shalimov Centre).

Prepared as a discussion document; not a treatment recommendation.

Supporting material (Priority 2 and 3, decision map, extended questions, appendices)

Priority 2: supporting questions (for the same meeting if time allows; otherwise within the month)

1. Intent: radical two-stage clearance or control of the FDG-active clones with imatinib continued; is debulking without radical intent acceptable? ^[11,14,16]
2. Remnant: what functional remnant is required at each stage, and is portal-vein embolisation or ALPPS-type augmentation applicable if S4 (± S1) is the final remnant? ^[16]
3. L1 anatomy: enucleation versus left lateral sectionectomy — depends on invasion of the LHV and segment III pedicle; who reads the contrast CT/MRI for this? ^[5]
4. Boundary nodules: does the board accept the R1 wall as the origin of the dome focus and wall plaques (“nodule within a mass”), and does that change the order? ^[17]
5. Small foci (S1 1.6 mL, P7, P12): included in the resection field, surveillance, or ablation if they grow?
6. Gastric primary: endoscopic/EUS surveillance versus wedge resection at a liver stage, given that one hyalinised fragment does not exclude viable cells? ^[9,10,18]
7. If the frozen section at surgery is positive: abort, limited procedure, or proceed for liver-dominant disease? ^[19,20,21]
8. Second-line alternative if regorafenib is unsuitable after PRES review: sunitinib (covers V654A only), ripretinib (self-pay), or trial (velzatinib versus sunitinib)? ^[1,22,23]
9. Transarterial option for R1 wall plaques if the right lobe is not resected first: TACE/DEB-TACE available in Kyiv; TARE abroad (centre, cost, lung-shunt study)? ^[24,25]
10. Extrahepatic clearance before major surgery: EGD for the oesophageal uptake; SVC/port status; is anything else required? ^[187][77]

Key to tags

FACT FACT: what a document or official written conclusion states (record number [Dn] on the case site). OPINION OPINION: a doctor’s informal view given in a call, chat or teleconsultation, never upgraded to fact. INTERPRETATION INTERPRETATION: inference by the case site (including AI-assisted measurements) or by a doctor, not an official conclusion. COMMENTARY COMMENTARY: an explicitly labelled written commentary. In the Markdown versions these appear as superscript letters F, O, I, C.

Decision map (which prerequisite gates which decision; numbers refer to the board questions 1–10, P2 to Priority 2)

1. Peritoneal status (Q1) gates everything surgical: confirmed malignant fluid → the surgeon has said surgery is not indicated OPINION; indeterminate → the board decides whether to proceed with peritoneal inspection and frozen-section washings at the start of surgery, or to re-sample first; negative → operability assessment. Inputs: operation protocol; cell block if fluid reaccumulates; the AUS category’s pooled risk of malignancy (about 50 %, carcinoma-dominated data) ^[2].
2. Current imaging (Q5) gates target definition: the only contrast study is 19.06.2026; the 23.09.2026 CT is non-contrast and cannot show vessels, ducts or small isodense lesions. Procedure planning on imaging older than about 30 days is outside usual practice (case-site evidence review INTERPRETATION).
3. Intent (P2-1) gates the sequence: radical intent requires that L1 be resectable with a margin after stage 1 (unknown: invasion of the LHV/segment III pedicle OPINION); control intent allows single-target procedures (L1 alone, S8 alone) with imatinib continued.
4. Remnant volume and function (Q2, P2-2) gates which stage can be first: by the inferred map the remnant after right hemihepatectomy + S1 is 49–53 % of non-tumour parenchyma; the remnant after removing both targets in one stage is S4 (± S1) only, 9–20 % INTERPRETATION. Surgical CT volumetry must replace these estimates.
5. R1 handling (Q2) depends on 3 and 4: resection of R1 is the right-sided stage; drainage/sclerosis has no GIST-specific evidence base (topic 2) and raises spillage/infection questions; transarterial therapy targets only the perfused wall.
6. Modality for L1 and the S8 focus (Q3, Q4) depends on 2 (vessel/duct relation) and 3: thermal ablation declined for L1 OPINION; non-thermal options (IRE, SBRT, histotripsy) and transarterial options have little or no GIST-specific data (topics 5, 6); the S8 focus is subcapsular and away from the hilum.
7. Gastric primary (Q4) is independent of the liver sequence except for the combined-procedure risk raised by the surgeon OPINION.
8. Systemic timing and PRES safety (Q6, Q7) wrap around every step: imatinib hold only peri-procedurally (guideline practice ^[11]); regorafenib label hold ≥ 2 weeks before elective surgery and ≥ 2 weeks after major surgery until wound healing ^[12] (regorafenib not started, held pending surgery per family report, so only the post-operative condition applies); PRES history argues for blood-pressure control and specialist review before any VEGFR-pathway drug.
9. Response check (Q8) closes the loop after each step: contrast CT (Choi criteria), PET-CT at 3 months, ctDNA (exploratory).

Extended questions

Format for each item: question; why it matters now; facts on file (tags, records); options the board can weigh (pros/cons as reported by the sources, not a recommendation); what would change the answer; who should answer; one literature pointer (details in the topic sections below).

A. Prerequisites and staging

A1. Peritoneal cytology. Should the ascitic finding be treated as peritoneal metastasis? - Why now: the surgeon’s conditional statement of 13.10.2026 (“if confirmed, surgery is not indicated”) OPINION has not been commented on after the CSD review did not confirm it (14.10.2026). - Facts: Shalimov 06.10.2026, 3 smears, Romanowsky–Giemsa: clusters with marked nuclear polymorphism, category V FACT ^[196]. CSD LAB 14.10.2026, 2 of those slides, Pappenheim: low-quality sample, dense groups of small relatively monomorphic cells, no pleomorphism, AUS, origin undeterminable without a cell block FACT ^[197]. Ascites volume and peritoneal inspection: not documented (protocol not received). 2022 anamnesis: “sarcomatosis of the abdominal cavity, greater omentum” FACT ^[5]; 2025: ametabolic cystic peritoneal node over the uterus FACT ^[67]; 06/2026: small fluid in the pouch of Douglas FACT ^[70]; imatinib itself causes fluid retention with ascites in a minority of patients ^[26]. - Options: (i) repeat cytology on new fluid with cell block and CD117/DOG1/calretinin (needs a puncture or re-laparoscopy; a small-volume effusion may not be reachable); (ii) treat as indeterminate and proceed with planning (risk: operating on peritoneal disease); (iii) treat as positive (risk: withholding local therapy on a non-reproduced reading). Pooled risk of malignancy for AUS is 49 % (95 % CI 37–61 %) in carcinoma-dominated series, with poor inter-observer agreement for the atypical categories ^[2,27]; a cell-block immunopanel re-classified 86 % of AUS fluids in one series ^[28]; low-grade sarcomas rarely exfoliate into effusions ^[29]; in the published GIST effusion cases the tumour cells were large/pleomorphic and DOG1/CD117 positive on cell block ^[3,30]. - What changes the answer: the operation protocol (nodules seen or not); a cell-block result; interval imaging of the peritoneum. - Intra-operative practicality (parallel search, lit_peritoneal_staging/SOURCES.md; carcinoma data): laparoscopic inspection agrees with laparotomy on the presence of any peritoneal disease in about 96 % but under-estimates the extent; frozen section of peritoneal nodules is discordant in about 2.4 % (gastric cancer); rapid stains on washings take 21–54 minutes with sensitivity about 63–79 %; the definitive DOG1/CD117 cell block is a paraffin result. Washings should be taken into fixative for a cell block whatever the frozen-section result. - Who: cytopathology (both laboratories), the operating surgeon, medical oncology. - Literature: topic 4 — AUS pooled risk of malignancy 49 % ^[2].

A2. Laparoscopy protocol and peritoneal staging. What did the 05.10.2026 laparoscopy show, and is a second look needed before any resection? - Facts: the surgeon’s stated aim was to exclude extrahepatic lesions and biopsy anything suspicious OPINION (chat 04.10.2026); only the pathology and cytology forms have been received FACT ^[195][196]. Peritoneal GIST surgery series describe outcomes for macroscopic disease, not for cytology-only findings ^[31,32]. - Options: request the protocol; re-laparoscopy with peritoneal biopsies and washing at the start of any resection; PET-CT review of the peritoneum (23.09.2026: no peritoneal uptake described) FACT ^[187]. - Who: operating team. - Literature: topic 4 — peritoneal versus liver-only disease: HR for death 2.67 (peritoneum only versus liver only) in one multivariable analysis ^[19], no OS difference by site in an oligometastatic cohort ^[20].

A3. Imaging currency. Is a contrast-enhanced CT or MRI of the liver with volumetry required before any decision? - Facts: last contrast study 19.06.2026 FACT ^[70]; 23.09.2026 CT non-contrast, 3.27 mm slices; vessels and ducts not assessable FACT ^[187]; the three-date model transfers the June portal vein onto the September study by registration (approximation of a few millimetres) INTERPRETATION. - Options: multiphase CT with volumetry; MRI with hepatobiliary contrast (small lesions, biliary relation); MRCP; ICG clearance or hepatobiliary scintigraphy for function. - Who: liver surgeon with radiology. - Literature: topic 3 — remnant volume thresholds are set on CT volumetry, not on inferred segment maps (case-site estimate INTERPRETATION).

A4. Tissue genotype of the progressing lesion. Should block 4222/26 be sequenced now? - Facts: ctDNA carries V654A (8.8 %) and A829P (4.0 %) alongside the exon 11 driver (17 %) FACT ^[75]; the biopsied L1 was not genotyped FACT ^[195]; the Shalimov pathologist offered a 40-gene panel including KIT (about 3 weeks) OPINION; the treating oncologist saw no fundamental point in further genetics before surgery OPINION (13.10.2026). - Options: sequence the cores now (information on which clone L1 carries; may guide the second-line choice); defer to resection material (more tumour cells, per the pathologist OPINION); rely on ctDNA (cannot assign clones to lesions). - Literature: topic 8 — secondary KIT mutations differ between metastases within one patient (inter-lesion heterogeneity) ^[33]; exon 13 and exon 17/18 co-mutation defines a subgroup where sunitinib and ripretinib performed alike in INTRIGUE ctDNA analyses ^[1].

A5. ctDNA as a monitoring tool. Should a repeat FoundationOne Liquid (or equivalent) be drawn before and 4–8 weeks after any local procedure? - Facts: one ctDNA time point exists (20.08.2026) FACT ^[75]. - Options: paired samples to document clone clearance after removal of L1/S8 (exploratory; no GIST standard); PET-CT at 3 months (Tahover OPINION). - Literature: topic 8 — ctDNA tumour fraction tracks active disease in GIST; no validated post-local-therapy endpoint identified.

A6. Extrahepatic clearance. Oesophageal uptake, femoral-head lucency, uterine masses and lung nodules: is anything unresolved that changes the intent? - Facts: oesophageal SUVmax 2.0–3.2, EGD advised, status undocumented FACT ^[187]; femoral lucency reported once (LISOD 07/2026 ^[71]) and not confirmed by two later readings FACT ^[77][187]; uterine lesions stable with low uptake for years FACT; lung nodules tiny and stable FACT ^[77]. - Who: gastroenterology (EGD); radiology.

B. Goals

B1. Radical versus control intent. What does “oligoprogression” mean in this patient and does it justify a two-stage operation with risk at both stages? - Facts: active tissue is multifocal within the liver (L1, S8 focus, several R1 wall plaques, P7) {F/I} ^[187]; all macroscopic disease is hepatic plus a FDG-cold primary and an unresolved peritoneum; the surgeon framed two-stage surgery as justified mainly with radical intent and questioned oncologists’ endorsement of debulking alone OPINION (18.09.2026). - Options as reported: radical two-stage resection (surgeon OPINION); clone-directed local therapy of the active foci with imatinib continued (NCCN/ESMO option for limited progression ^[11,34]); systemic switch without local therapy (Czarnecka OPINION; Tahover favours systemic plus IR OPINION). - What changes the answer: peritoneal status; whether the pattern is judged unifocal or multifocal progression (post-surgical PFS 11 versus 6 months and OS 59 versus 24 months in the two-institution series ^[14]); R0 feasibility for L1 (median OS 8.7 years after R0/R1 versus 5.3 years after R2 in the EORTC series; debulking did not appear to prolong survival ^[16]). - Who: the whole board. - Literature: topic 1 — multifocal progression predicted worse PFS after surgery, HR 3.08 ^[14].

B2. Definition of success. Which endpoint should the plan be judged by: PET/ctDNA clearance of the resistant clones, time on imatinib gained, or R0 status? - Options: FDG-PET response at 3 months (used by the consultants OPINION); ctDNA clearance (exploratory); imaging per Choi criteria rather than RECIST (size-stable disease here).

C. Sequencing

C1. Which target first. Right-sided stage (R1 + dome lesion + S1) first versus L1 first? - Facts: 18.09.2026: right hemihepatectomy + S1 first, L1 second after regeneration, “never in one stage” OPINION; 09.10.2026: “most likely two-stage: first removal of the focus from the left lobe, then 2–3 weeks of recovery; after regeneration the second stage, complete removal of the right-lobe focus; first operation not before 19.10” OPINION; the reason for the change is not documented. Volumetry INTERPRETATION: non-tumour parenchyma 1.41 L; remnant after right hemihepatectomy + S1 ≈ 0.69–0.78 L (49–53 %); remnant after left lateral sectionectomy ≈ 0.84–0.95 L (60–66 %), but that remnant contains R1 and is mostly right-sided parenchyma compressed around a 1.2 L mass; final remnant after both ≈ S4 (± S1) ≈ 0.12–0.26 L (9–20 %) before regeneration. Standardised liver volume for 65 kg ≈ 1.2–1.4 L (height unknown) INTERPRETATION. - Considerations reported by sources: right first removes the largest tumour load and the dome focus in one operation and leaves the left lobe to regenerate OPINION; left first treats the biopsy-proven, fastest-growing, FDG-dominant lesion and is a smaller operation, but leaves R1 and the S8 focus for a later major hepatectomy whose remnant would then be S4 ± S1 after interval regeneration {O/I}; one-stage removal of both is excluded by remnant size {O/I}. - Risks created by each order for the next step: right first → the second stage operates on a regenerated but tumour-bearing left lobe; invasion of the LHV/segment III pedicle may force a larger resection OPINION; left first → delays removal of R1 and the S8 focus; the remaining liver is dominated by the mass; a complication (bile leak, collection) could delay stage 2 while regorafenib is withheld. - What changes the answer: surgical volumetry and function tests; L1’s relation to the LHV and portal pedicle on contrast imaging; peritoneal status; the board’s intent. - Who: liver surgery with anaesthesia. - Literature: topic 3 — no published series addresses lesion order in bilobar GIST; staged hepatectomy data come from colorectal metastases (see topic 3 table).

C2. R1 first by non-resective means. Can the mass be drained, aspirated or sclerosed to “remove the dangerous right lobe” before anything else? - Facts: R1 is 1.2 L, 97 % fluid/necrotic by density partition (core 1050 mL, wall 153 mL at the September HU split) INTERPRETATION; active tissue is confined to wall plaques (anterolateral strip, base, medial wall) and to the dome lesion above the septum INTERPRETATION; the May 2025 11 mm mural nodule maps to the anterolateral strip and not to the dome focus INTERPRETATION; the mass contacts the IVC, MHV/RHV and right portal branches and abuts the gallbladder INTERPRETATION. - What the user’s question implies: that the enhancing wall “generates” boundary nodules. Imaging supports growth of activity along the wall and in the adjacent dome lesion FACT ^[187], but not that drainage of the fluid would remove the wall, which is the viable compartment. - Options and the concerns reported by sources: aspiration/drainage (no GIST-specific series found; concerns raised in the literature search: infection of a necrotic cavity, spillage/seeding along a tract, haemorrhage; the wall remains); ethanol sclerosis (no GIST data); resection (right hemihepatectomy; surgeon: most of the removed volume is non-functional tumour, risk “not high” OPINION); transarterial therapy of the wall (TACE/DEB-TACE/TARE; surgeon sceptical: hypovascular tumour, “resorption of a large amount of tumour parenchyma, which may turn into an abscess” OPINION; radioembolisation is not available in Ukraine OPINION). - Who: liver surgery and interventional radiology together. - Literature: topic 2 — rupture/haemorrhage of GIST liver metastases is reported as case reports; tumour rupture is defined by spillage and piecemeal resection criteria ^[35].

C3. L1 before or after R1 when the intent is control only. If the board does not pursue radical surgery, which single procedure gives the most benefit per risk? - Facts: L1 holds about half of the active tissue (36–61 mL of 51–111 mL depending on threshold) INTERPRETATION; the S8 focus is 3.6–10.8 mL; R1 wall plaques 10–39 mL INTERPRETATION. - Options: L1 alone (resection or non-thermal ablation or SBRT) with imatinib continued; L1 plus S8 focus in one session (combined resection + ablation/SBRT); R1 wall plaques by transarterial therapy only. - Literature: topic 1 — local treatment in metastatic GIST was associated with better OS in the Dutch registry (HR 0.56), with the benefit concentrated in liver-confined, responding disease ^[36].

C4. Boundary foci and small lesions. How should S1 (1.6 mL, ametabolic), P7 (0.4 mL, PET-defined) and P12 (0.7 mL PET-only) be handled in each sequence? - Facts: S1 caudate focus 7–16 mm from the IVC and portal vein INTERPRETATION; the surgeon would remove S1 with the right-sided stage OPINION; P7 touches the R1 wall INTERPRETATION. - Options: include in the resection field; surveillance; ablation if they grow.

C5. Timing between stages and with systemic therapy. What interval between stages, and when does regorafenib (or an alternative) start? - Facts: surgeon: 2–3 weeks recovery then regeneration before stage 2 (09.10) OPINION; earlier: “a month or two” (18.09) OPINION; regorafenib label: hold ≥ 2 weeks before elective surgery, do not give for ≥ 2 weeks after major surgery and until wound healing ^[12]; imatinib: continue until surgery and restart with oral intake (ESMO practice ^[11]); imatinib interruption in responding GIST leads to early progression in most patients (BFR14) ^[13]. - Options: imatinib throughout both stages, regorafenib after the second stage; regorafenib between stages (requires two label holds); no second-line drug while the disease is controlled locally.

D. Modality per target

D1. L1 modalities. - Resection: left lateral sectionectomy (anatomical; takes 0.45–0.57 L of parenchyma by the inferred map INTERPRETATION) or enucleation leaving a cavity (surgeon: technically feasible, limited by invasion of the LHV/segment III branches OPINION). Contact with the left portal branch is 0–0.8 mm with a contact patch of 0.25–0.34 mL and a patent vein INTERPRETATION; the LHV relation is unresolved INTERPRETATION. - Thermal ablation (RFA/MWA): contraindicated per Dr Korshak (portal branch and left bile duct adjacency) OPINION; ULIS: only with a safe ultrasound path, bile ducts being the main obstacle OPINION; heat-sink and biliary injury are the reported risks for perihilar ablation. - IRE: non-thermal, used for tumours within 1 cm of major bile ducts in small series ^[5]; no GIST series; lesion size 5 cm is above the typical ≤ 3 cm indication. - SBRT/CyberKnife: composition-independent; in the only prospective GIST radiotherapy study 80 % of irradiated lesions stabilised for a median 16 months ^[7]; dose constraints for the hilum and stomach (15–20 mm away) apply; combination with VEGFR inhibitors is a reported toxicity concern. - Transarterial (TACE/DEB-TACE/TARE): treats perfused tissue; the surgeon doubts an arterial supply (“hypovascular, dark on CT”) OPINION; GIST series exist for TACE and SIRT (topic 6); TARE not available in Ukraine OPINION. - Histotripsy: approved in the EU in 2026 per the case site (not verified from a primary source here); pivotal trial treated tumours ≤ 3 cm with 95 % technical success ^[8]; no GIST patients identified; HCA London: initial review £500, treatment from £37,500 FACT ^[191]. - Electrochemotherapy: unsuitable (electrode depth) OPINION; an open study in Padova accepts visceral lesions of any histotype (NCT06753136; parallel search lit_local_therapy/TRIALS.md). - HDR brachytherapy: 10 imatinib-refractory GIST patients, local control 97.5 %, median PFS 6.8 months (parallel search lit_local_therapy/SOURCES.md, R7); not available locally as far as known. - Literature: topic 5 — IRE within 1 cm of major ducts: 11 patients, 22 metastases, no definite duct injury ^[5].

D2. S8 focus modalities. Subcapsular dome location, 3.9–8.2 mm from the hepatic-vein outline, inside a 79 mL cystic lesion INTERPRETATION. - Options: percutaneous MWA/RFA (ultrasound visibility of a 14 × 23 mm focus inside a cystic lesion is the practical question OPINION ULIS); laparoscopic ablation; SBRT; wedge/segmentectomy; part of a right hemihepatectomy; transarterial therapy. - Literature: topic 5 — thermal ablation of small GIST metastases has mature data for solid lesions, not for foci inside cystic masses (OpenEvidence review on the case site COMMENTARY).

D3. R1 wall plaques. Options: transarterial therapy; SBRT to the dominant anterolateral plaque (6.8–15 mL) INTERPRETATION; ablation of individual plaques (subcapsular lateral wall); resection with R1.

D4. Gastric primary. FDG-negative since 08/2023, calcifying, 4 cm, no viable cells in one fragment FACT ^[195]; exophytic, abutting the pancreatic tail INTERPRETATION; “technically easy to remove”, but the surgeon would rather not combine it with an extensive liver resection (healing, anastomotic leak) OPINION. - Options: surveillance (EGD/EUS, PET); wedge resection at a liver stage; later separate resection; pathological complete response is uncommon in resected GIST after imatinib and single-fragment sampling cannot exclude viable cells ^[9]; resection of the primary in metastatic GIST before imatinib did not improve outcome in one series ^[37].

D5. Centre and access. Which procedures are available in Kyiv (Shalimov NSCST: resection, transplantation, systemic therapy, ablation, TACE OPINION), in Zaporizhzhia (RFA OPINION), and abroad (SIRT Gliwice; CyberKnife Warsaw; liver surgery Warsaw OPINION Czarnecka; histotripsy London/Cambridge FACT case site)? Centres with published GIST liver local-therapy experience and verified contacts (parallel search, lit_trials_centres/CENTRES.md): NIO-PIB Warsaw (surgery on imatinib); WTZ Essen (metastasectomy on TKI); Mannheim (SIRT for GIST liver metastases, 2015 and 2025 series); INT Milan (surgery of residual/limited-progressive disease); Gustave Roussy (RFA of GIST metastases); Leeds (liver resection; histotripsy site); Royal Marsden (liver surgery). No Israeli centre has a published GIST liver local-therapy series.

E. Systemic therapy interplay

E1. Imatinib continuation. Continue imatinib through all local therapy? - Facts: guideline practice is to continue the TKI with local treatment of limited progression ^[11,34]; the ULIS doctor suggested pausing targeted therapy around RFA for 1–2 weeks OPINION; the surgeon noted impaired wound healing with imatinib and regorafenib and advised waiting, but judged the laparoscopy risk moderate OPINION; imatinib was interrupted around the colonoscopy (02.10) and the laparoscopy (05.10) per chat OPINION, resumed on 06.10.2026, currently taken and to be continued until the operation (family report, 14.10.2026). - Literature: topic 7 — imatinib interruption after response: most patients progress within months (BFR14) ^[13]. Practice statements on the pre-operative pause: ESMO “a few days or even 1 day” ^[11]; GEIS 24 h ^[38]; British Sarcoma Group 2024 “usually stopped a week before surgery; much shorter intervals appear to be safe” (parallel search, lit_periop_tki/SOURCES.md A1).

E2. Regorafenib start. Before, between or after the local procedures? - Facts: dispensed 30.09.2026 FACT ^[76]; not started, held because surgery is expected soon (family report, 14.10.2026), so only the post-operative label condition applies at present; label: wound-healing hold ≥ 2 weeks pre-, ≥ 2 weeks post-surgery ^[12]; grade ≥ 3 hypertension and PRES are label warnings; Czarnecka: “not a bad choice” OPINION; Tahover: acceptable but harder to tolerate OPINION; TARGET clinic: sceptical of both sunitinib and regorafenib, proposes immunotherapy off-label OPINION (hypothesis only). - Options: start after the last planned procedure (≥ 2 weeks after major surgery and after wound healing); start between stages (would require a ≥ 2-week hold before the next operation); alternative agents (sunitinib — covers V654A only; ripretinib — covers A829P, weak for V654A, lowest neurovascular risk, self-pay; trial). - Literature: topic 8 — INTRIGUE ctDNA: exon 11 + 13/14 favoured sunitinib, exon 11 + 17/18 favoured ripretinib; the co-mutated subgroup showed no difference ^[1].

E3. PRES history. What does a haemorrhagic PRES on pazopanib (2023) require before sunitinib/regorafenib/SBRT or major surgery? - Facts: PRES with left occipital haemorrhage and multifocal vasoconstriction, BP 150–160 mmHg, resolved by 06/2023 FACT ^[22][54]; no neurology review since 2023 on file. - Options: baseline neurology review and BP plan; home BP monitoring; reduced starting dose (consultants’ suggestion on the case site INTERPRETATION); avoidance of VEGFR inhibitors around SBRT. - Literature: topic 7 — rechallenge data after TKI-associated PRES are limited to case reports (see topic 7).

E4. Monitoring. PET-CT at 3 months (Tahover OPINION) versus contrast CT with Choi criteria; ctDNA kinetics.

F. Risk and function

F1. Remnant volume. Estimates INTERPRETATION (inferred segments; to be replaced by surgical volumetry): see Appendix 1. Conventional thresholds (25–30 % of a healthy liver, 40 % after chemotherapy/steatosis) are from colorectal/HCC practice, not GIST-specific; the parenchyma here has hypertrophied around a 1.2 L mass, so volume may overstate function. F2. Vascular and biliary anatomy. L1: left portal branch contact, LHV unresolved; R1: IVC contact, MHV/RHV ≤ 2 mm, right portal branches ≤ 1 mm; the S8 focus: 4–8 mm from the hepatic-vein outline INTERPRETATION. Hepatic artery and bile ducts were not segmentable on the available studies INTERPRETATION. F3. Bleeding and haemostasis. Platelets 216, bilirubin normal, albumin 42.8 g/L FACT ^[193]; coagulation not documented since 2023; SVC occlusion suspected (port) FACT ^[77]. F4. PRES and anaesthesia. BP control intra- and postoperatively; HCA’s eligibility question on general anaesthesia FACT (case site, “ablation” section). F5. Infection and spillage. Necrotic 1.2 L cavity (abscess risk after embolisation/drainage OPINION surgeon; literature topic 2); tumour rupture criteria include spillage and piecemeal resection ^[35]; peritoneal recurrence after rupture is high even with adjuvant imatinib in gastric GIST cohorts ^[39]. F6. Diabetes, anaemia, SVC. Glucose 6.9 mmol/L; Hb 121 g/L, MCV 101 fL (B12/folate not recent) FACT ^[192][193].

G. Information still missing (who could provide it)

See the ‘Needed within days’ list in the main document; in addition: height and weight (volumetry ratios); ECOG; whether L1 was biopsied under laparoscopic or percutaneous ultrasound (sampling path matters for seeding discussions) (Shalimov); the exact fluid volume aspirated at laparoscopy (Shalimov); whether frozen or additional fluid remains (Shalimov cytology); and Dr Korshak’s pending answer on whether the target of any local treatment would be the new focus or the old lesion (case site, 11.09.2026).

H. Disagreements between consultants (exact statements, level, what would resolve them)

I. Evidence summary (short; details in the topic sections)

• Local treatment of limited progression on imatinib with the TKI continued is an option in ESMO and NCCN guidance ^[11,34]; outcomes after surgery depend on the response pattern at operation (responsive ≫ unifocal > multifocal progression) ^[14] and on completeness of resection ^[16]; a registry analysis associated local treatment with better OS in selected patients ^[36].
• No published series was identified for drainage, aspiration or sclerosis of a cystic GIST liver metastasis; the related literature concerns rupture, haemorrhage and the definition of tumour rupture ^[35].
• Ablation adjacent to the portal pedicle: thermal techniques carry heat-sink and biliary-injury risk; IRE has small series near major ducts ^[5]; histotripsy has no GIST data and a ≤ 3 cm pivotal-trial population ^[8].
• Transarterial therapy and radiotherapy have small GIST series (topics 5–6); the only prospective GIST radiotherapy study stabilised 80 % of irradiated lesions ^[7].
• Peritoneal cytology: the AUS category carries a pooled risk of malignancy near 50 % with poor reproducibility; cell block with DOG1/CD117 is how GIST cells are identified in fluid ^[2,3].
• Systemic therapy: imatinib interruption after response leads to early progression ^[13]; regorafenib must be withheld ≥ 2 weeks before elective surgery ^[12]; the exon 13 + 18 co-mutation is a subgroup without a clear winner between sunitinib and ripretinib ^[1].

Appendix 1. Volumes (consensus masks; segments inferred; label as estimates)

Percentages are the remnant’s share of the total non-tumour parenchyma on that date. These are arithmetic on inferred labels and say nothing about inflow, outflow, biliary drainage or function.

Appendix 2. Lesion-to-structure distances (mm; consensus masks)

Appendix 3. Statements by consultants (dates, level)

Appendix 4. Claims that could not be verified from a primary source in this preparation

• The 15.10.2026 consilium date and the earliest operation date (19.10.2026): chat messages only OPINION.
• Imatinib interruption 02–05.10.2026, resumption on 06.10.2026 and continuation until the operation: chat messages and a family report (14.10.2026), no document.
• Regorafenib dispensed 30.09.2026 and deliberately not started because surgery is expected soon: family report (14.10.2026), no document.
• “2-year PFS 75 % with continued TKI versus 29 % without after RFA”: quoted in the case site’s evidence review without a traceable primary source; not used here.
• Histotripsy CE mark in 07/2026 and HOPE4LIVER site status: case-site statements from a foundation newsletter and company material; not verified against a regulator document in this preparation (see topic 5 for what was retrieved).
• Peritoneal findings at the 05.10.2026 laparoscopy: no document.
• Standardised liver volume: height not on file; the 1.2–1.4 L figure assumes 160–170 cm.
• Segment labels on all three dates are automatic inferences; the September S1/S4 volumes are implausible and the surgeon’s “segment I” remark about a left-lobe lesion OPINION could not be reconciled with the masks.

Priority 3: research-type and longer-horizon items (not for this month’s decision)

1. KIT/PDGFRA genotyping of the liver cores (block 4222/26) or of the resection specimen: which clone does L1 carry? ^[33,40]
2. Repeat ctDNA before and 4–8 weeks after local therapy to document clone clearance (exploratory; no validated endpoint) ^[41].
3. Cell block and CD117/DOG1 on any archived fluid or a third slide of 996/26 (confirmatory only; not a reason to delay) ^[28,30].
4. Trial screening for StrateGIST 3 (velzatinib vs sunitinib; Budapest, Pécs and German sites; the Polish sites were listed as recruiting on ClinicalTrials.gov on 05.10.2026 but a parallel registry check reports Polish recruitment ended 12.10.2026 — confirm with GSK) if no local step is taken or after recovery; starting regorafenib may remove eligibility ^[23].
5. Histotripsy candidacy opinion (Cambridge/London) for L1: no GIST data, pivotal trials ≤ 3 cm, hilar structures excluded ^[8,42,43].
6. TARE enquiry (Gliwice, London): largest GIST series median hepatic PFS 16 months in TKI-exhausted patients ^[24].
7. Tissue biobanking of any resection specimen (CSD has a biobank; family’s note). OPINION Dr Fedorov’s position (16.10.2026, opinion): the specimen goes unaltered to the Centre’s pathology laboratory under the standard protocol; any special preservation (e.g. freezing) must be arranged with the pathologist or laboratory; biobanking does not affect treatment (16.10 and 20.10.2026).
8. Literature review for the long term: see the topic sections below.