Colorectal cancer grade depends more on the system than the pathologist

Design

Interobserver agreement study on scanned slides

Population

100 colorectal adenocarcinomas graded by five gastrointestinal pathologists

Primary outcome

Fleiss kappa and number of high-grade calls under each system

Effect

Kappa 0.730 (WHO6) vs 0.731 (AJCC8); high grade 34.4 vs 11.8 tumours

Five gastrointestinal pathologists graded 100 scanned colorectal adenocarcinomas under four approaches: WHO 5th edition, WHO 6th edition, two-tier AJCC 8th edition, and their usual practice. AJCC grades by the proportion of gland formation; WHO grades by the least differentiated component, and the 6th edition requires that component to cover at least 0.196 mm² away from the invasive front.

Agreement was substantial with all systems: Fleiss kappa 0.730 for WHO6, 0.731 for AJCC8 and 0.655 for WHO5. But the WHO systems called many more tumours high grade — an average of 34.4 under WHO6 and 33.6 under WHO5, against 11.8 under AJCC8. Pathologists agreed on discounting tumour buds at the invasive front, but not on how to handle stroma-rich areas.

Grade can influence adjuvant treatment decisions in stage II disease. A report that says 'high grade' without naming the system may mean different things to different readers. Until outcome studies settle which threshold matters, name the system every time.

• State the grading system used — WHO 6th edition or AJCC 8th edition — in every colorectal cancer report.
• Apply the WHO 6th edition minimum area of 0.196 mm² for the least differentiated component, away from the invasive front.
• Do not count invasive front tumour budding towards the grade; report budding separately.
• Agree a departmental approach to stroma-rich areas, where pathologists disagreed most.
• Tell oncologists that WHO criteria classify about three times as many tumours as high grade as AJCC does.

Why it matters

A grade that triples depending on the system can tip an adjuvant treatment decision in stage II disease.

Don't overread it

This measured agreement and grade distribution, not which system better predicts outcome.

The statistics, in plain English

Fleiss kappa measures agreement beyond chance among several raters; 0.61 to 0.80 is usually called substantial. Similar kappa values with very different high-grade counts mean pathologists apply each system consistently — the systems themselves disagree.

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.