WHO criteria called three times as many colorectal cancers high grade as AJCC

Design

Interobserver agreement study on scanned slides

Population

100 colorectal adenocarcinomas graded by 5 gastrointestinal pathologists

Primary outcome

Agreement (Fleiss kappa) and frequency of high-grade calls by scheme

Effect

Kappa 0.730 (WHO6), 0.731 (AJCC8), 0.655 (WHO5); high grade 34.4 (WHO6) vs 11.8 (AJCC8) per 100

Five gastrointestinal pathologists graded the same 100 scanned colorectal adenocarcinoma slides under four schemes: WHO 5th edition, WHO 6th edition, the two-tier AJCC 8th edition, and their usual practice. AJCC grades by the percentage of gland formation; WHO grades by the least differentiated component, and the sixth edition sets a minimum area for that component away from the invasive front.

Agreement among the five was substantial under all schemes, and about equal for WHO 6th edition and AJCC (kappa about 0.73), both better than WHO 5th edition. But the two schemes did not call the same tumours. On average, about 34 of 100 tumours were high grade under WHO criteria against about 12 under AJCC.

The pathologists agreed on some rules, such as ignoring tumour buds at the invasive front when applying WHO grading, and disagreed on others, notably how to grade stroma-rich areas.

Grade feeds into decisions such as adjuvant therapy in stage II disease. A tripling of high-grade calls depending on the scheme means a reader of the report needs to know which scheme was used. Until the systems are aligned and tested against outcomes, reports should name the grading system explicitly.

• State which grading system was used on every colorectal cancer report.
• Expect WHO criteria to label many more tumours high grade than AJCC.
• Exclude invasive-front tumour buds when grading by the least differentiated component.
• Agree a department approach to grading stroma-rich areas.
• Flag grade discrepancies at the multidisciplinary meeting when they affect stage II decisions.

Why it matters

The same tumour can be low or high grade depending on the rulebook, and treatment decisions downstream may not know which was used.

Don't overread it

Five expert reviewers and no outcome data; it does not show which scheme better predicts prognosis.

The statistics, in plain English

A kappa of 0.73 means substantial agreement beyond chance among the five pathologists; 1.0 is perfect. High agreement within a scheme does not mean the schemes agree with each other: they reliably sort tumours differently.

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.