Retrospective Evaluation of the STONE Score and Modified Stone Score for Prediction of Renal Tract Stones

Computed tomography (CT) imaging remains the gold standard for diagnosis of renal tract stones; however, recurrence is common leading to the risk of radiation exposure with repeated scans. Predictive scoring systems have been suggested to aid decision-making and reduce unnecessary imaging; however, their accuracy has been challenged. This study aimed to assess the diagnostic performance of the STONE score and modified STONE score (MSS) in predicting ureteric calculi in an Australian population.

Retrospective cohort study of patients aged > 18 years, who were suspected of having ureteric colic and underwent CT renal tract. Patients were excluded if they had fever, recent urological surgery or trauma, active cancer or unstable vital signs. Score performance was evaluated by clinical performance analysis and receiver operating curve (ROC) analysis.

191 patients were studied; 62.3% had confirmed urolithiasis and 6.3% had a confirmed alternative diagnosis. The STONE score demonstrated an AUC of 0.806 (95% CI 0.745-0.862), and the MSS had an AUC of 0.796 (0.734-0.858) for predicting ureteric stones. If the high score (high probability urolithiasis) group patients had not undergone a CT scan, 8.9% of scans could have been avoided if the STONE score was applied compared to 54.8% with the MSS (p = 0.0004). There was poor classification agreement between the scores (Cohen's kappa -0.03; p = 1).

Both the STONE and MSS demonstrated moderate accuracy in predicting ureteric stones in an Australian emergency department (ED) population; however, the proportion of scans potentially avoided was markedly different. It is not clear whether these scores are superior to clinician gestalt.