Details

Autor(en) / Beteiligte

• Unal, Birsen
• Bagcier, Sebnem
• Simsir, Ilknur
• Bilgili, Yasemin
• Kara, Simay

Titel

Evaluation of Differences Between Observers and Automatic-Manual Measurements in Calculation of Doppler Parameters

Ist Teil von

• Journal of medical ultrasound, 2004-08, Vol.23 (8), p.1041-1048

Erscheinungsjahr

2004

Quelle

Free E-Journal (出版社公開部分のみ）

Beschreibungen/Notizen

• OBJECTIVE: We aimed to search for differences between observers and automatic and manual measurements in calculations of Doppler parameters. METHODS: The middle cerebral artery (MCA), central retinal artery, ophthalmic artery (OA), common carotid artery (CCA), vertebral artery (VA), popliteal artery (PA), interlobar renal artery (IRA), and arcuate renal artery (ARA) were evaluated in 20 healthy subjects bilaterally. Peak systolic velocity (PSV), end- diastolic velocity (EDV), time-averaged maximum velocity (TAMAX), resistive index (RI), and pulsatility index (PI) were measured from the same spectrum manually by 3 observers and automatically. Results of 4 measurements were compared by analysis of variance and Pearson tests. RESULTS: The comparison of the 4 measurements revealed significant differences for most parameters except TAMAX of the OA, VA, and ARA and PSV, EDV, and PI of the PA. An automatic calculator yielded lower PSV, RI, and PI values (except the MCA and PA) and higher EDV values compared with manual measurements. The magnitudes of difference were in the range of 1% to 16% for velocities and 4% to 14% for RI and PI. The means of difference were 3.185 cm/s for PSV of the CCA and 0.054 for RI of the IRA. Correlation was high for PSV, EDV, and TAMAX in all arteries (except TAMAX of PA) and relatively low for PI and RI in most of the arteries. CONCLUSIONS: Although our study was performed on healthy subjects, our results showed that, in most cases, readers and the automatic approach disagreed on evaluation of Doppler parameters. This may be important in preventing false diagnoses in cases with Doppler values close to upper limits and may necessitate establishment of new limits for each method.