External validation of the STONE score and derivation of the modified STONE score

Abstract Objective The STONE score is a clinical prediction rule for the presence of uncomplicated ureter stones with a low probability of acutely important alternative findings. This study performed an external validation of the STONE score, focusing on the Korean population, and a derivation of the modified STONE score for better specificity and sensitivity. Methods We retrospectively reviewed medical records of patients complaining of flank pain at a single emergency department from January 2013 to December 2014. Patients were categorized into 3 groups according to their STONE score. The prevalence of ureter stones and other alternative findings were calculated in each group. We derived a modified STONE score based on a multivariable analysis and performed an interval validation. Results From the 700 patients included in the analysis, 555 patients (79%) had a ureter stone. The area under the receiver operating characteristic curve of the STONE score was 0.92. The sensitivity of the high stone score was 0.56. In the modified STONE score, nausea, vomiting, and racial predictors were substituted by C-reactive protein and previous stone history. The area under the receiver operating characteristic curve and sensitivity of the modified STONE score in the internal validation group significantly increased to 0.94 and 0.80, respectively. Conclusion The STONE score can be used to predict a ureter stone with a low probability of other alternative findings. The modified STONE score might increase the diagnostic performance in suspicious urinary stone cases. Key points We performed external validation of the STONE score and derivation of the modified STONE score. This scoring system could help the clinicians with radiation reducing decision making.