A new ultrasound parameter for osteoporosis diagnosis: Clinical validation on normal- and under-weight women

Abstract

Aim of this work was to evaluate the effectiveness of a recently introduced ultrasound (US) method for osteoporosis diagnosis, when extensively used in a clinical context to investigate adult women of variable age. A total of 384 female patients (46-65 years; body mass index <; 25 kg/m2) underwent a spinal dual X-ray absorptiometry (DXA) and an abdominal US scan of lumbar spine, acquiring both echographic images and unprocessed radiofrequency signals. US data were analyzed through a new fully automatic algorithm, which performed a series of spectral and statistical analyses to calculate the parameter called Osteoporosis Score (O.S.). Diagnostic effectiveness of O.S. was assessed through a direct comparison with DXA measurements (assumed as the gold standard reference), quantifying the agreement between the two methods through accuracy calculation, Cohen's kappa (k) and Pearson correlation coefficient (r). The overall accuracy of O.S.-based diagnoses resulted 84.6%, ranging from a minimum of 81.7% for the oldest patients (aged in 61-65 y) to a maximum of 87.2% for the youngest patients (aged in 46-50 y). Cohen's kappa showed an analogous trend, confirming a significant agreement between DXA and US-based diagnoses along the whole considered age interval (k=0.758, p<;0.0001). A good correlation was also found between O.S.-derived BMD values and corresponding DXA measurements (r=0.72, p<;0.001). These results demonstrated that US-measured O.S. is significantly correlated with spinal BMD in normal- and under-weight adult women belonging to a wide age interval. Therefore, the routine clinical application of this innovative approach to osteoporosis diagnosis can be envisioned.