p53 immunohistochemistry in Barrett's oesophagus was reproducible and changed diagnoses

Design

Nationwide consultation cohort with multi-observer review

Population

1146 patients with Barrett's oesophagus biopsies referred to a Dutch expert panel

Primary outcome

Agreement on p53 patterns and diagnostic reclassification

Effect

Hospital–panel κ 0.61 (95% CI 0.58 to 0.64); interobserver κ 0.65 (0.61 to 0.69)

Published in June, this study used real-world consultation cases sent to the Dutch Esophageal Pathology Panel: 1146 consecutive patients and 1704 biopsy levels, each assessed by a median of 12 panel pathologists. They classified H&E findings by the Vienna system and p53 patterns as wild-type, overexpression, double clone, null or equivocal.

Agreement on p53 between referring hospitals and the panel was κ 0.61 (95% CI 0.58 to 0.64), and 24% of referring p53 assessments were reclassified. Among 13 panel pathologists reading 316 biopsies, agreement was κ 0.65 (0.61 to 0.69). Aberrant p53 moved 9% of non-dysplastic biopsies to indefinite and 5% to low-grade dysplasia. In indefinite biopsies, p53 shifted 17% to non-dysplastic and 37% to low-grade dysplasia.

Indefinite for dysplasia is an uncomfortable diagnosis for the endoscopist and the patient, and p53 resolved more than half of those cases here. The one-in-four reclassification rate also shows that interpretation outside expert centres needs standardised criteria. Guidelines differ on routine p53 use in Barrett's, so laboratories should align with their national guidance while adopting consistent pattern definitions.

• Consider adding p53 immunohistochemistry to Barrett's biopsies that are indefinite for dysplasia or suspicious for low-grade dysplasia.
• Use defined pattern categories: wild-type, overexpression, null, double clone and equivocal.
• Interpret p53 together with H&E morphology, not on its own.
• Seek expert or second review for low-grade dysplasia diagnoses, which still carry the most disagreement.

Why it matters

It turns a frequent indeterminate diagnosis into an actionable one in many cases, and shows where interpretation drifts outside expert centres.

Don't overread it

The study shows reproducibility and diagnostic shifts, not that p53-guided management reduces progression to cancer.

The statistics, in plain English

Kappa measures agreement beyond chance: 0 is chance level and 1 is perfect. Values around 0.6 are usually described as moderate to substantial, which is better than the agreement typically reported for low-grade dysplasia on H&E alone. The study measured agreement and reclassification, not whether patients' outcomes improved.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.