TRACERx 100 NSCLC Cohort (UCL/CRUK, Nature 2017)

Overview

The TRACERx (TRAcking Cancer Evolution through therapy (Rx)) study is a prospective longitudinal cohort study of non-small cell lung cancer (NSCLC) led by the Cancer Research UK Lung Cancer Centre of Excellence at University College London. The nsclc_tracerx_2017 dataset encompasses the first 100 surgically resected early-stage NSCLC patients enrolled in TRACERx (NCT01888601), paired with bespoke multiplex-PCR ctDNA profiling. Multi-region whole-exome sequencing of 327 spatially separated primary tumor regions (plus 4 metastatic biopsies) enables reconstruction of clonal and subclonal tumor architecture, which is then tracked longitudinally in plasma via patient-specific ctDNA panels (precursor to the Signatera assay, designed by Natera). This foundational dataset established that ctDNA can detect NSCLC relapse with a median 70-day lead-time over CT imaging.

Composition

  • 100 patients, 96 evaluable for pre-operative ctDNA; 24-patient longitudinal sub-cohort with pre- and post-surgical plasma profiling.
  • Histologies: 58 LUAD, 31 LUSC, 7 other NSCLC.
  • Tumor genomics: 327 spatially separated primary tumor regions + 4 metastatic biopsies; whole-exome sequencing (Illumina); copy number / purity / ploidy via ASCAT; subclonal deconvolution via PyClone; phylogenetic tree reconstruction with CITUP.
  • ctDNA: Bespoke patient-specific multiplex-PCR NGS panels (median 18 SNVs pre-operative; expanded to median 28 SNVs for LUAD longitudinal phase), sequenced to ~40,000× target depth on Illumina HiSeq 2500. Analytical sensitivity >99% at VAF ≥0.1%; specificity 99.6% per SNV.
  • Imaging: Pre-operative PET/CT in 92/96 patients; 3D tumor volume measured via CT.
  • Data deposition: EGA accessions EGAS00001002247 (primary) and EGAS00001002415 (metastatic).

Assays / panels (linked)

Papers using this cohort

  • PMID:28445469 — Abbosh et al., Nature 2017: First ctDNA analysis from TRACERx; bespoke multiplex-PCR plasma profiling of 96 early-stage NSCLC patients detected relapse with 93% sensitivity and a median 70-day lead-time over imaging; established histology (LUSC vs LUAD) as the dominant predictor of pre-operative ctDNA detectability. PMID:28445469

Notable findings derived from this cohort

  • Pre-operative ctDNA detection rate is strongly histology-dependent: 97% (30/31) of LUSC vs 19% (11/58) of LUAD; independent predictors include non-adenocarcinoma histology, lymphovascular invasion, and high Ki67. PMID:28445469
  • Tumor volume correlates linearly with mean clonal plasma VAF (Spearman’s ρ = 0.63, P<0.001, n=37); a 10 cm³ primary tumor predicts a mean clonal VAF of 0.1%. PMID:28445469
  • Post-operative ctDNA detected confirmed NSCLC relapse in 13/14 (93%) patients, with a median lead-time of 70 days (range 10–346 days) over CT-confirmed relapse. PMID:28445469
  • Phylogenetic ctDNA tracking resolved clonal architecture of metastatic seeding; a subclone harboring an ERBB2 amplification (>15 copies) was identified as driving relapse in patient CRUK0004. PMID:28445469
  • An expanded 103-SNV panel detected ctDNA in a post-mortem case 189 days earlier than the 19-SNV panel, demonstrating that panel breadth substantially improves MRD sensitivity. PMID:28445469

Sources

  • cBioPortal study nsclc_tracerx_2017.
  • Abbosh C, et al. Phylogenetic ctDNA analysis depicts early-stage lung cancer evolution. Nature. 2017;545:446–451. PMID:28445469

This page was processed by crosslinker on 2026-05-15.