Reassessment of the recurrence risk of primary gastrointestinal stromal tumour after complete resection.

Objectives: The modified National Institutes of Health (NIH) risk criteria for gastrointestinal stromal tumours (GISTs) have some limitations and need to be improved.

Methods: Patients who underwent radical resection of primary GIST were retrospectively reviewed. Peripheral blood indices including the neutrophil-to-lymphocyte ratio (NLR) and prognostic nutritional index (PNI) were analysed. Recurrence-free survival (RFS) was calculated and compared. Multivariate analysis was conducted. Area under the receiver operating characteristic curve (ROC) was calculated.

Results: A total of 492 patients were enrolled. Tumour size, mitotic index (MI), tumour location and PNI were independent prognostic factors. The modified NIH criteria could not distinguish among very low-, low- and intermediate-risk patients, and PNI was the only independent prognostic factors for them. The five-year RFS rate in the high risk (HR) group was significantly lower. A further modification to the NIH risk criteria was proposed (the 'NIH-PNI stratification'). Non-high risk (NHR) patients were divided into the NHR-PNI-H group (PNI > 48.05) and the NHR-PNI-L group (PNI ≤ 48.05), respectively. HR patients were divided according to tumour size and MI: the HR1, HR2 and HR3 groups. The five-year RFS rates of the NHR-PNI-H, NHR-PNI-L, HR1, HR2 and HR3 groups were 97.3%, 93.5%, 74.1%, 61.7% and 24.4%, respectively ( < .001). The area under the curve (AUC) for the NIH-PNI stratification, modified NIH criteria, NIH criteria (2002), AFIP criteria and nomogram were 0.857, 0.807, 0.817, 0.843 and 0.831, respectively.

Conclusion: The proposed NIH-PNI stratification was able to distinguish among five groups in terms of risk of recurrence.