BACKGROUND

Systemic inflammation and malnutrition may promote tumor progression. C-reactive protein/albumin ratio (CAR) is linked to the poor long-term survival of several malignant tumors.

PURPOSE

To explore the predictive value of CAR in gastrointestinal stromal tumors (GISTs).

METHODS

A retrospective study was conducted on 325 patients with primary GIST surgically treated with curative intent from 2009 to 2018. The cut-off point of CAR was set using X-tile software. Kaplan-Meier method and multivariate Cox regression model were used to study the prognostic value of CAR. The time-dependent receiver operating characteristic curve (tROC) was drawn, and the prognostic accuracy of CAR, Glasgow prognostic score (GPS), and National Institute of Health (NIH) risk classification was compared by the area under the curve (AUC).

RESULTS

The best cut-off point of CAR was 0.55. Increased CAR was associated with the location of the lower digestive tract, larger tumor size, higher mitotic index, higher NIH risk classification, lower ALB, higher CRP, and higher GPS (all p<0.05). Multivariable analysis revealed that CAR (hazard ratio [HR] 2.598, 95% confidence interval [CI] 1.385-4.874; p=0.003) was an independent predictor of overall survival. Additionally, the AUC of CAR was lower than that of NIH risk classification at 2 years (0.601 vs. 0.775, p=0.002) and 5 years (0.629 vs 0.735, p=0.069). However, the AUC of NIH risk classification significantly increased (2-year OS 0.801, p=0.251; 5-year OS 0.777, p=0.011) when combined with CAR.

CONCLUSION

CAR is a new independent predictor of poor survival in patients with GIST. CAR combined with NIH risk classification can effectively improve the performance of prognosis prediction.