Retrospective analysis of resected gastrointestinal stromal tumours (GISTs) comparing existing risk stratification models and assessing the influence of GIST genotypes on recurrence and mortality  — ASN Events

Retrospective analysis of resected gastrointestinal stromal tumours (GISTs) comparing existing risk stratification models and assessing the influence of GIST genotypes on recurrence and mortality  (#308)

Yada Kanjanapan 1 , Mohammad Cheik-Hussein 1 , Christine Hemmings 1 , Desmond Yip 1
  1. Department of Medical Oncology, Canberra Hospital, Canberra, ACT

Aims

GISTs are rare gastrointestinal mesenchymal tumours. Imatinib, a tyrosine kinase inhibitor against c-KIT, is indicated in tumours with a high risk of recurrence following resection. Various models combine prognostic factors such as the tumour’s size, mitotic count, site and presence of rupture; for risk-stratification. The study aims to compare the National Institute of Health (NIH)1 , Joensuu2  classifications and the Gold3  nomogram in their predictive power for GIST recurrence. We also aim to test the influence of GIST genotypes on recurrence risk.

Methods

A retrospective review of 100 GIST cases diagnosed between 1998 and 2012 treated in ACT/NSW is conducted. Tumours are stratified per the NIH and Joensuu classifications and predicted 2-year recurrence-free survival (RFS) calculated from the Gold nomogram. Actual RFS and overall survival is calculated from the date of GIST diagnosis to the date of tumour recurrence or death respectively.  Patients alive at last follow-up are censored. Survival functions will be estimated using the Kaplan-Meier method. Log rank tests will compare survival between risk groups, and between predicted and actual 2yr RFS. Cox proportional hazards model is used assess the impact of genotype as a prognostic factor on survival.

Results

Interim results of 70 patients show 7%, 29%, 25%, 39% fall in the very low, low, intermediate and high NIH risk categories respectively. The rate of agreement with the Joensuu classification is 91% (63 of 69 cases). In the other 6 cases, Joensuu’s classification upstaged the tumour from intermediate to high risk. There is a trend for differences in RFS between NIH groups which has not reached statistical significance. The results will be updated with final recurrence and mortality data, and analysis on genotype will be presented.

Conclusion

Interim results suggest good agreement between the NIH and Joensuu classifications, but refinement of their predictive power with other prognostic factors such as GIST genotype warrants further study.

  1. Fletcher CDM, Berman JJ, Corless C, et al. Diagnosis of 
gastrointestinal stromal tumors: a consensus approach. Hum Pathol. 
2002;33:459–465.
  2. Joensuu H. Risk stratification of patients diagnosed with gastro-intestinal stromal tumor. Hum Pathol. 2008;39:1411–1419.
  3. Gold JS, Gonen M, Gutierrez A, et al. Development and validation of a prognostic nomogram for recurrence-free survival after complete surgical resection of localised primary gastrointestinal stromal tumour: a retrospective analysis. Lancet Oncol. 2009;10:1045–1052.