Diagnostic agreement of bone histomorphometry for renal bone disease

An iliac bone biopsy with semi-quantitative histomorphometry is the gold standard for evaluating renal osteodystrophy (ROD). However, lack of standardization of the bone biopsy procedure, sample handling, reading and reporting represent barriers to research synthesis and clinical collaboration. This study assessed diagnostic agreement across European laboratories performing histomorphometry for evaluation of ROD.

Five-millimeter transiliac biopsy slides from 5 cases were converted to digital images for analysis by experts from 7 laboratories. Parameters included quantitative histomorphometry, bone turnover, mineralization and volume (TMV) classification, and a concluding diagnosis. Agreement was analysed using Kappa statistics for categorical and intraclass correlation coefficients (ICC) for continuous variables.

Agreement on histomorphometric variables ranged from poor to excellent. High agreement was observed for the static parameters bone area (ICC 0.91), osteoid area (ICC 0.95) and osteoid perimeter (ICC 0.84). For dynamic parameters, agreement was high for mineralizing surface (ICC 0.87) and bone formation rate (ICC 0.74), but poor for mineral apposition rate (ICC -0.13) and mineralization lag time (ICC 0.07). Agreement on the TMV classification was substantial for bone turnover (Kappa 0.77), slight for bone mineralization (Kappa 0.14), and substantial for bone volume (Kappa 0.73). The concluding diagnoses showed a high degree of consistency.

Despite considerable inter-reader variability in bone histomorphometric measures, the clinical diagnoses of ROD were consistent across laboratories. Key sources of variability included analytical methods, disagreement on reference values, and differences in the definition of mineralization disorder. Further work should be undertaken to reach consensus on these points.