TCIA Multiple Delineation (Inter-Observer)

Overview

The Multiple Delineation collection comprises 21 NSCLC patients each delineated independently by five radiation oncologists. It was used by Aerts et al. 2014 to quantify inter-observer variability in the CT radiomic feature library and to identify stable features robust to delineation uncertainty. The dataset is hosted on TCIA. PMID:24892406

Composition

  • Cancer type: NSCLC, n=21 patients × 5 delineations each.
  • Modality: CT; manual tumour delineations by five independent radiation oncologists per patient.
  • Purpose: inter-observer reproducibility assessment of radiomic features.
  • Role in Aerts 2014: stability sub-study; features with high inter-observer stability ranks tend to also have higher prognostic performance. PMID:24892406

Assays / panels (linked)

  • CT radiomics — 440-feature library inter-observer variability evaluation.

Papers using this cohort

  • PMID:24892406 — Aerts et al. 2014, Nature Communications: Multiple Delineation used to quantify radiomic feature stability across delineators, complementing RIDER test-retest reproducibility.

Notable findings derived from this cohort

  • Features with higher inter-observer stability ranks tended to also have higher prognostic performance in validation cohorts, supporting delineation-robust feature selection. PMID:24892406

Sources

  • TCIA collection: NSCLC-Radiomics — https://www.cancerimagingarchive.net/collection/nsclc-radiomics/
  • PMID:24892406 — Aerts et al. 2014, Nature Communications, DOI 10.1038/ncomms5006.

This page was processed by crosslinker on 2026-04-15.