Article Text

Download PDFPDF
Regression grading in neoadjuvant treated pancreatic cancer: an interobserver study
  1. Sangeetha N Kalimuthu1,
  2. Stefano Serra1,
  3. Neesha Dhani2,
  4. Sara Hafezi-Bakhtiari1,
  5. Eva Szentgyorgyi1,
  6. Rajkumar Vajpeyi1,
  7. Runjan Chetty1
  1. 1Laboratory Medicine Program, Department of Pathology, University Health Network, University of Toronto, Toronto, Ontario, Canada
  2. 2Laboratory Medicine Program, Department of Medical Oncology, University Health Network, University of Toronto, Toronto, Ontario, Canada
  1. Correspondence to Professor Runjan Chetty, Department of Pathology, University Health Network, Toronto General Hospital, 200 Elizabeth Street, 11th Floor, Eaton Wing, Toronto, Ontario, Canada M5G 2C4; runjan.chetty{at}


Aim Several regression grading systems have been proposed for neoadjuvant chemoradiation-treated pancreatic ductal adenocarcinoma (PDAC). This study aimed to examine the utility, reproducibility and level of concordance of three most frequently used grading systems.

Methods Four gastrointestinal pathologists used the College of American Pathologists (CAP), Evans, MD Anderson Cancer Centre (MDA) regression grading systems to grade 14 selected cases (7–20 slides from each case) of neoadjuvant chemoradiation-treated PDAC. A postscoring discussion with each pathologist was conducted. The results were entered into a standardised data collection form and statistical analyses were performed.

Results There was little concordance across the three systems. The Kendall coefficient of concordance agreement scores were: CAP: 2-poor, 2-fair; Evans: 1-fair, 1-moderate, 2-good; MDA: 1-poor, 2-moderate, 1-good. Interpretation in all three grades in the CAP grading system was a source of discrepancy. Furthermore, using fibrosis as a criterion to assess regression was contentious. In the Evans system, quantifying tumour destruction using arbitrary percentage cut-offs (ie, 9% vs 10%; 50% vs 51%, etc) was imprecise and subjective. Although the MDA system generated greatest concordance, this was due to ‘oversimplification’ surrounding wide, arbitrarily assigned thresholds of </> 5% of tumour.

Conclusions All systems lacked precision and clarity for accurate regression grading. Presently the clinical utility and impact of histological regression grading in patient management is questionable. There is a need to re-evaluate regression grading in the pancreas and establish a reproducible, clinically relevant grading system.


Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • This manuscript was presented as a platform presentation at the US and Canadian Academy of Pathology meeting, Seattle, March 2016.

  • Handling editor Cheok Soon Lee

  • Contributors All authors participated and contributed to the manuscript.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.