Diagnostic accuracy of the Mirels score at the humerus in metastatic bone disease: a narrative review.

purposeTo critically evaluate the diagnostic performance of the Mirels score for pathological fracture risk prediction at the humerus, and to determine whether site-specific factors warrant a revised approach to fracture risk stratification at this location.

methodsNarrative review. PubMed was searched without date restriction for studies addressing Mirels score validation at the humerus, alternative fracture risk assessment tools, and clinical outcomes of completed versus impending pathological humeral fracture. Additional studies were identified by manual reference screening.

resultsThree dedicated humeral validation studies report sensitivity of approximately 14% at the standard Mirels threshold of nine or more, equivalent to chance (area under the curve [AUC] 0.51-0.55). A larger cohort supports a revised threshold of eight or more with improved performance (AUC 0.82), identifying cortical breach as the dominant independent predictor of fracture (odds ratio 21.0). Humeral lesions receive lower Mirels scores and less frequent staging computed tomography than femoral lesions despite a higher fracture rate. Completed pathological humeral fracture is associated with increased one-year mortality (hazard ratio 1.28), higher reoperation risk (odds ratio 2.50), and a clinically significant rate of iatrogenic radial nerve palsy (up to 13.9%). Alternative quantitative tools, including computed tomography-based rigidity analysis and finite element analysis, have not been validated at the humerus.

conclusionThe Mirels score performs poorly at its standard threshold when applied to humeral lesions. A revised threshold incorporating cortical breach assessment, combined with systematic staging imaging of the humerus, may reduce avoidable morbidity. Prospective validation of a humerus-specific risk stratification tool is needed.