Comprehensive Molecular Characterization of Gallbladder Carcinoma and Potential Targets for Intervention

Authors

Giraldo NA

Drill E

Satravada BA

El Dika I

Brannon AR

Dermawan J

Mohanty A

Ozcan K

Chakravarty D

Benayed R

Vakiani E

Abou-Alfa GK

Kundra R

Schultz N

Li BT

Berger MF

Harding JJ

Ladanyi M

O’Reilly EM

Jarnagin W

Vanderbilt C

Basturk O

Arcila ME

Doi

10.1158/1078-0432.CCR-22-1954

PMID: 36228155 · DOI: 10.1158/1078-0432.CCR-22-1954 · Journal: Clinical Cancer Research (2022)

TL;DR

This study provides the largest single-institution genomic characterization of gallbladder carcinoma (GBC) to date, profiling 244 samples from 233 patients using MSK-IMPACT. The most frequent oncogenic alterations were in TP53 (63%), CDKN2A (21%), SMAD4 (19%), ARID1A (18%), and ERBB2 (15%). SMAD4 and STK11 mutations independently associated with reduced overall survival in metastatic disease. Clinically actionable alterations (OncoKB levels 1, 3A, or 3B) were identified in 35% of patients, and 18% of metastatic patients received biomarker-directed therapy or enrolled in clinical trials based on molecular findings.

Cohort & data

  • 233 patients / 244 samples of GBC, collected 2014–2021 at Memorial Sloan Kettering Cancer Center.
  • 57% primary tumors, 43% metastases; 85% adenocarcinomas, 10% carcinomas with squamous differentiation, 5% neuroendocrine carcinomas.
  • Median age at collection: 66 years (range 37–90); 63% Caucasian, 12% African American, 12% Asian.
  • 67% stage IV, 18% stage III at time of collection.
  • Profiled with MSK-IMPACT (341–505 gene panel versions); dataset deposited in cBioPortal as gbc_mskcc_2022.
  • Median sequencing coverage: 634X (range 72X–1150X).
  • Microsatellite instability assessed via MSIsensor, MiMSI, and IDYLLA MSI; variants annotated with OncoKB.

Key findings

  • Most frequent oncogenic mutations: TP53 63% (all LOF), SMAD4 19% (16% LOF), ARID1A 18% (all LOF), PIK3CA 9% (all GOF), ELF3 9% (all LOF), CDKN2A 9% (8% LOF), KRAS 7% (all GOF), ERBB2 6% (all GOF), ARID2 6% (all LOF), STK11 6%, CTNNB1 6% (all GOF), KMT2C 6% (all LOF), TERT promoter 6% (all GOF), RB1 3.5% (all LOF).
  • Most frequent oncogenic CNAs: CDKN2A/CDKN2B deletions (14% each), MDM2 amplification (11%), ERBB2 amplification (10%), CCNE1 amplification (9%), MYC amplification (7%), ERBB3 amplification (5%), CDK4 amplification (5%), KRAS amplification (4%).
  • No recurrent structural variants were identified. Notable oncogenic SVs included an FGFR3::TACC3 in-frame fusion (GOF) and LMNA::NTRK1 fusion (OncoKB level 1).
  • Oncodrive clustering identified CTNNB1 (p.S45F/P), KRAS (p.G12A/C/D/R), and ERBB2 (p.S310F/Y) as potential molecular drivers in GBC.
  • Most commonly altered pathways: TP53 (74%), Cell Cycle (46%), RTK-RAS (46%), Epigenetic (40%), TGF-Beta (28%), NOTCH (9%).
  • Histology–genomics correlations: RB1 LOF enriched in small cell neuroendocrine carcinoma (q<0.001); NOTCH1 pathway alterations higher in carcinomas with squamous differentiation (26% vs. 8% in adenocarcinoma, q=0.07).
  • No significant differences in molecular landscape between primary and metastatic samples.
  • 6 tumors (3%) were MSI-High.
  • Dominant mutational signatures (in TMB-high cases): APOBEC (n=3) and MMR (n=3).
  • Microbiome: Metastatic lesions had higher Shannon diversity than primaries. No Helicobacter or Salmonella reads were detected despite their association with GBC in endemic regions.

Genes & alterations

  • TP53 – LOF mutations in 63% of GBC; most frequently altered gene. Part of the TP53 pathway altered in 74% of samples.
  • SMAD4 – LOF mutations (19%) and deletions (5%); independently associated with reduced OS in metastatic disease (HR 2.17, 95% CI 1.21–3.89, q=0.10; multivariate HR 2.11, p=0.012).
  • STK11 – LOF mutations in 6%; independently associated with reduced OS in metastatic disease (HR 3.94, 95% CI 1.62–9.55, q=0.05; multivariate HR 3.76, p=0.004).
  • ERBB2 – GOF mutations (6%) and amplifications (10%); amplification is OncoKB level 3B actionable; 10 patients received HER2-directed therapy.
  • PIK3CA – GOF mutations in 9%; actionable at OncoKB level 3B (e.g., p.E545K, p.H1047R).
  • KRAS – GOF mutations (7%) and amplifications (4%); p.G12C present in 2 patients (OncoKB level 3A).
  • CDKN2A/CDKN2B – Deletions in 14% each; mutations in CDKN2A 9%.
  • ARID1A – LOF mutations in 18%.
  • ARID2 – LOF mutations in 6%; structural variants in 2 samples.
  • CTNNB1 – GOF mutations in 6%; identified as potential driver by oncodrive clustering (p.S45F/P). Higher frequency in primaries without cholelithiasis (22% vs 4.3%).
  • ELF3 – LOF mutations in 9%.
  • MDM2 – Amplification in 11%.
  • CCNE1 – Amplification in 9%.
  • ERBB3 – Amplification in 5%; mutations more frequent in T1/T2 primaries (16% vs 0%).
  • RB1 – LOF mutations in 3.5%; enriched in small cell neuroendocrine carcinoma (q<0.001).
  • NTRK1 – LMNA::NTRK1 fusion in 1 patient (OncoKB level 1; FDA-recognized biomarker).
  • BRCA1/BRCA2 – Oncogenic mutations in 3 and 6 patients respectively (level 3B).
  • CDK12 – Structural variants in 3 patients including fusions (level 3B).
  • FGFR3 – FGFR3::TACC3 in-frame fusion (GOF) in 1 patient.
  • RET – Structural variants in 2 patients including RET fusions (level 3B).
  • ROS1 – ROS1::PARK2 fusion (likely GOF) in 1 patient.
  • MET – Amplification in 5 patients (level 3B).
  • ATM – Oncogenic mutations in 5 patients (level 3B).

Clinical implications

  • Actionable alterations: 35.2% of patients harbored at least one mutation or structural variant considered actionable in GBC or other solid tumors (OncoKB levels 1, 3A, or 3B).
  • NTRK1 fusions (level 1): FDA-recognized biomarker predictive of response to entrectinib and larotrectinib across solid tumor types.
  • KRAS G12C (level 3A): Present in 2 patients (1%); promising response rates in biliary tract tumors in phase II trials.
  • HER2-directed therapy: 10 metastatic patients received trastuzumab-based regimens targeting ERBB2 amplification/mutation (level 3B).
  • PARP inhibitors: 1 patient treated with olaparib targeting BRCA/ATM alterations.
  • Checkpoint inhibitors: 12 metastatic patients (8%) received ICIs (pembrolizumab n=9, atezolizumab n=1, nivolumab n=1, ipilimumab/nivolumab n=1); 42% (5/12) showed evidence of response (3 MSI-High, 1 MSS/TMB-high, 1 MSS/TMB-low).
  • Prognostic biomarkers: SMAD4 and STK11 mutations independently predicted reduced OS in metastatic GBC on multivariate analysis.
  • Standard first-line: 82% of patients received first-line chemotherapy for advanced disease; 143 patients received gemcitabine/platinum (mostly cisplatin).
  • 18% of metastatic patients received biomarker-directed targeted therapy or enrolled in clinical trials based on molecular findings.

Limitations & open questions

  • Single-institution cohort (MSK) with patients primarily from the continental United States, limiting geographic and environmental diversity analysis despite GBC being endemic in parts of South America and Asia.
  • Low number of neuroendocrine histology tumors (n=11) limits power for subtype-specific molecular analyses.
  • Cholelithiasis data was missing for 49% of primary tumor cases, limiting power for correlations with TERT and CTNNB1 alterations.
  • The absence of Helicobacter and Salmonella in the microbiome contrasts with reports from endemic regions, but the MSK-IMPACT microbiome pipeline is an opportunistic analysis of off-target reads rather than a dedicated metagenomic assay.
  • ERBB2/ERBB3 and TP53 alterations did not associate with clinical outcomes in this cohort, potentially confounded by HER2-directed therapy use.
  • No recurrent structural variants were found, possibly reflecting the rarity of targetable fusions in GBC compared to cholangiocarcinoma (which harbors FGFR2 fusions and IDH1 mutations at higher rates).
  • Response to targeted therapy was assessed by treating physician based on serial radiology rather than RECIST criteria.

Citations from this paper used in the wiki

  • “The most common oncogenic molecular alterations appeared in the Cell Cycle (TP53 63% and CDKN2A 21%) and RTK_RAS pathways (ERBB2 15% and KRAS 11%).” (Abstract)
  • “Variants in SMAD4 and STK11 independently associated with reduced survival in patients with metastatic disease.” (Abstract)
  • “Alterations considered clinically actionable in GBC or other solid tumor types (e.g., NTRK1 fusions or oncogenic variants in ERBB2, PIK3CA, or BRCA1/2) were identified in 35% of patients.” (Abstract)
  • “STK11: q=0.05, HR 3.94 [95% CI 1.62–9.55]; SMAD4: q=0.10, HR 2.17 [1.21–3.89]” (Results, Molecular predictors)
  • “STK11: p=0.004, HR 3.76 [95% CI 1.54–9.16]; SMAD4: p=0.012, HR 2.11 [1.18–3.80]” (Results, multivariate Cox model)
  • “One patient harbored an LMNA::NTRK1 mid-exon fusion…which is an FDA-recognized biomarker predictive of response to entrectinib and larotrectinib in all solid tumors types” (Results, Potentially actionable mutations)
  • “Five of the 12 patients (42%) treated with checkpoint inhibitors had evidence of response to treatment.” (Results)

This page was processed by crosslinker on 2026-05-06.