Kefah Mokbel: ctDNA Identifies Patients at Risk of Breast Cancer Relapse
Kefah Mokbel/ LinkedIn

Kefah Mokbel: ctDNA Identifies Patients at Risk of Breast Cancer Relapse

Kefah Mokbel, Chair of Breast Cancer Surgery at London Breast Institute, shared on LinkedIn:

ctDNA two years after adjuvant chemotherapy identifies the patients who will relapse – but not yet what to do about it

Friedl et al., JCO (published today): a landmark analysis of 313 stage I–III patients from SUCCESS-A, tested with a tissue-free epigenomic assay (Guardant Reveal) roughly 2 years after adjuvant chemotherapy.

  •  ctDNA detected in 18/313 (5.8%)
  •  17 of 18 positives developed distant recurrence – PPV 94%
  •  Median lead time: 7.9 months
  •  DRFI: HR 33.3 (95% CI 4.12–268); OS: HR 27.3 (95% CI 1.15–647)
  •  Sensitivity 73% (11/15) within a year of recurrence; specificity 99.6%

Why it matters: tumour-informed panels need archival tissue, bespoke design and turnaround. A tissue-free assay removes all three, which is what makes population-level surveillance plausible.

Where I would be cautious:

  • The hazard ratios are directionally certain and quantitatively meaningless – 18 events cannot support a CI of 4 to 268.
  •  The 73% sensitivity applies only to samples drawn within 12 months of recurrence. Across the cohort it is lower. This is a rule-in test; a negative result reassures no one.
  •  Detection rate reflects the platform as much as the cohort. Reveal is a methylation assay with a fixed genomic footprint; tumour-informed ultra-deep panels report higher MRD positivity and longer lead times. The four missed recurrences are consistent with sub-threshold shedding rather than absent disease.
  •  SUCCESS-A recruited in the mid-2000s – no pertuzumab, T-DM1, T-DXd, CDK4/6 inhibitors, olaparib or immunotherapy. Neither the positivity rate nor the lead time can be assumed to transfer.
  •  A single draw at 2 years suits TNBC but is poorly timed for ER-positive disease, which recurs across two decades and sheds less. No subtype breakdown is given.
  •  Most importantly: prognostic is not actionable. No completed randomised evidence shows that intervening on MRD changes survival. DARE, ZEST and TRAK-ER will answer that. Until then, a positive result buys eight months of knowing.

The biology is convincing. The clinical utility question remains open.”

Title: Clinical Relevance of Post-Treatment Circulating Tumor DNA Detection in Early Breast Cancer Using a Tissue-Free Epigenomic Assay: A 2-Year Landmark Analysis

Authors: Thomas W. P. Friedl, Peter A. Fasching, Andreas D. Hartkopf, Hans Tesch, Ralf Lorenz, Georg Heinrich, Jens-Uwe Blohmer, Tanja Fehm, Volkmar Mueller, Andreas Schneeweiss, Matthias W. Beckmann, Matthias Ruebner, Nadia Harbeck, Klaus Pantel, Derek Dustin, Mingyang Cai, Brigitte Rack, Wolfgang Janni

Read the full article.

You can also read: ctDNA During Neoadjuvant Therapy: Toward Real-Time Risk Stratification in Early Breast Cancer

Kefah Mokbel: ctDNA Identifies Patients at Risk of Breast Cancer Relapse