Three stains on a lung carcinoid biopsy predicted relapse better than WHO typing

Design

Retrospective registry cohort with matched biopsy and resection

Population

98 patients with resected stage I-III lung neuroendocrine tumours, Netherlands

Primary outcome

Relapse prediction from biopsy

Effect

Panel flagged 17/19 relapses vs 2/19 by WHO; NPV 0.96 vs 0.82

This Dutch registry study took 98 patients with stage I to III lung neuroendocrine tumours resected between 2003 and 2021, with matched preoperative biopsies. Three pathologists reclassified every case by WHO 2021 criteria, and biopsies and resections were stained for OTP, CD44 and Ki-67. Low risk was defined as OTP H-score 50 or more, CD44 H-score 30 or more and Ki-67 below 5%.

Nineteen patients relapsed over a median of 83 months. On the biopsy, the panel classed 17 of the 19 (89%) as high risk; WHO typing called only 2 of 19 atypical. The panel's negative predictive value was 0.96, against 0.82 for WHO. Interobserver agreement was much better (kappa 0.67 vs 0.28), and the biopsy result agreed better with the resection (kappa 0.58 vs 0.17).

Typical versus atypical carcinoid depends on mitotic count and necrosis, which small biopsies often cannot assess. This panel may help surgeons decide when a limited resection is reasonable. It is a small retrospective series with 19 events; the thresholds need external validation before they drive surgery.

• Avoid firm typical versus atypical calls on small carcinoid biopsies; mitoses and necrosis are often unassessable.
• Consider adding Ki-67 with a stated hot-spot method to every lung carcinoid biopsy report.
• Note OTP and CD44 where your laboratory can stain them, as research-stage prognostic markers.
• Flag in the report when biopsy sampling limits grading, so the surgeon does not over-rely on it.

Why it matters

Biopsy grading of lung carcinoids is unreliable, and surgeons increasingly use it to choose sublobar resection.

Don't overread it

A retrospective series of 98 patients with 19 relapses; the cut-offs have not been validated elsewhere.

The statistics, in plain English

A negative predictive value of 0.96 means 96% of tumours called low risk did not relapse. Kappa measures agreement beyond chance: 0.67 is substantial, 0.28 fair. With only 19 relapses, one or two cases more or less would shift these figures noticeably.

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.