Targeted NanoString Classification of Paired Gastric Tumors and Organoids

Research Applications

Assess preservation of molecular classification across tissue processing and culture.

Model & Design

A 38-patient gastric cancer cohort, with 12 paired organoid cultures. Eleven pairs were used in the principal subtype comparison and 10 in EBV/MSI reference testing.

Experimental Format

Custom 107-gene NanoString CodeSet, with RUV-III technical-replicate normalization and feature-specific quantile normalization to classifier distributions.

Model Components

• Component

  Reference tests

  Description

  EBER in situ hybridization and pentaplex PCR assessed EBV and MSI classes in 10 paired samples.

• Component

  Pharmacology

  Description

  FLOT treatment for 48 hours after 24-hour culture, with CCK-8 viability and drug sensitivity scores. Ten dose-response profiles are displayed in Fig5.

Measurements & Endpoints

• TCGA and TME subtype scores
• EBV/MSI classification versus reference tests
• Parent–organoid subtype concordance
• FLOT drug sensitivity score
• Exploratory CCLE drug correlations

Analysis

Apply previously developed classifiers to normalized targeted-expression measurements. Compare paired scores and categorical assignments, test FLOT associations and analyze 37 separate gastric cell lines.

Published Evidence

• Study Design

  Reference-test validation, pair concordance and exploratory pharmacology.

  Study Sample

  38 clinical tissue specimens; 12 organoid cultures; 11 subtype pairs; 10 EBV/MSI reference pairs. The Results report 11 malignancy-confirmed drug-tested cultures, with 10 profiles displayed in Fig5.

  Comparison

  Parent FFPE tissue; EBER/PCR reference tests; technical replicates; AGS assay reproducibility; separate CCLE cell lines.

  Finding

  EBV/MSI assignments were 100% accurate in the 10 reference-tested pairs. Parent–organoid concordance was 81.8% for TME and 63.6% for TCGA. FLOT drug sensitivity did not significantly associate with subtype scores.

  Interpretation & Limits

  • No gold-standard CIN or GS reference validation was performed.
  • Organoid subtype can differ from the parent tumor, including a shift toward CIN.
  • Individual CCLE drug correlations did not remain significant after multiple-testing correction.
  • No clinical survival or drug-response validation was available.