分级(工程)
医学
结直肠癌
肿瘤科
癌症
内科学
工程类
土木工程
作者
Dipti M. Karamchandani,Raul S. González,Hwajeong Lee,Maria Westerhoff,Brian Cox,Rish K. Pai
出处
期刊:Histopathology
[Wiley]
日期:2025-01-28
卷期号:86 (7): 1101-1111
被引量:14
摘要
Aims The current American Joint Committee on Cancer (AJCC) staging manual and the College of American Pathologists (CAP) colorectal carcinoma (CRC) protocol specify use of a four‐tiered grading system (i.e. grades 1–4; well‐differentiated–undifferentiated) for CRC, based on percentage of gland formation. The World Health Organization (WHO) 5th edition grades CRC into low‐grade (well‐ and moderately differentiated) and high‐grade (poorly and undifferentiated), based on the least differentiated component. We studied interobserver agreement and practice patterns among pathologists when grading CRC by these two grading systems. Methods and results Five gastrointestinal pathologists reviewed 100 scanned CRC slides and graded the tumour on each slide, per provided criteria in (a) WHO 5th edition book, (b) AJCC manual/CAP CRC protocol and (c) their clinical practice. A questionnaire for grading selected CRC subtypes was also provided. Statistical analysis was performed using Pearson's χ 2 test and Fleiss multi‐rater kappa analyses. Overall, agreement among the five reviewers when grading via WHO and AJCC criteria for low‐grade and high‐grade CRC was moderate ( κ = 0.568, P < 0.001) and good ( κ = 0.611, P < 0.001), respectively. All reviewers graded significantly more tumours as high‐grade when using WHO (median = 46) versus AJCC/CAP criteria (median = 20). Conclusions Interobserver agreement was higher using the AJCC grading criteria as a two‐tiered system. Significantly more tumours were called high‐grade using the WHO criteria. This raises concerns regarding upgrading tumours, as well as potential differences in grading tumours among pathologists worldwide, based on regional preferred grading systems. Synchronisation of these two grading systems is necessary for uniform grading of CRCs throughout institutions.
科研通智能强力驱动
Strongly Powered by AbleSci AI