Reproducibility of sonographic measurement of the substantia nigra

The aim of this study was to evaluate inter-reader, intra-investigator and inter-investigator reproducibility and correlations in the assessment of substantia nigra (SN) echogenicity and area measurement by a physician-sonographer (PS), a sonographic laboratory assistant (SLA) and a physician without sonographic experience (PN). A total of 22 patients with extrapyramidal symptoms were examined using transcranial sonography (TCS). SN images were encoded and evaluated by the three readers. A second TCS examination was performed after 7+/-2 d. A second investigator performed TCS examination 1 mo later. Spearman rank correlation and Pearson's correlation coefficient were used when assessing the agreement between readers. All three readers identified the same 15 patients with SN echogenicity III or more. Inter-reader SN echogenicity and area measurement correlations were r=0.55 to 0.82 and r=0.31 to 0.74 between PS and SLA and r=0.55 to 0.77 and 0.49 to 0.62 between PS and PN, respectively (p<0.05 in all cases). Intra-reader echogenicity and area measurement correlations (r=0.85 to 0.96 and r=0.51 to 0.69) were statistically significant only for PS (p<0.001). All intra- and inter-investigator correlations of SN area measurement (r=0.69 to 0.88 and r=0.5 to 0.61) and SN echogenicity (r=0.64 to 0.92 and r=0.51 to 0.69) were statistically significant (p<0.05). Semiquantitative evaluation of SN echogenicity and area using TCS is highly dependent on the experience of the sonographer. Only an experienced sonographer was able to produce very reproducible results with statistically significant correlations; SLA and PN intra-reader correlations were poor.