Background One major advantage of ultrasound (US) is that it can be performed bedside by the clinician. In an attempt to implement US in the diagnosis of disc herniation some authors have previously described sensitivities and specificities approximating those of the conventional methods. Within the last decade, however, there have only been sparse reports on this subject. So far, the role of US in this area is unclear.

Objectives To evaluate in a blinded fashion US in the diagnosis of disc herniations at the three lower lumbar levels performed by an experienced specialist.

Methods 18 males and 19 females with a mean age of 39.7 years (SD 10.5 years) and no previous back surgery were included due to suspected disc herniation. The US was performed transabdominally with the patient in the supine position. An Acuson Sequoia with a 4 MHz vector transducer was used. MRI or CT was obtained at the three lower lumbar levels. The two modalities (US vs. CT/MRI) were performed and evaluated blinded to each other. The endpoints were existence of herniation or protrusion in contrast to normal findings and the corresponding level of the abnormal finding.

Results A total of 97 disc levels were visualised by US. Fourteen levels were not visualised. There were 45 disc abnormalities (herniations or protrusions) found by CT/MRI, and of those 12 (26.7%) were also found by US. Fifty two discs were found to be normal by CT or MRI, but of those 11 (21.2%) were suspected to be abnormal by US. A total agreement between US and CT/MRI was found in 53/97 (54.6%) of the cases. The sensitivity of US was 12/45 (26.7%) and the specificity 12/23 (52.2%). The false positive rate was 47.8% and the false negative rate 44.6%. The over all kappa value was 0.18. The findings stratified at levels are shown in the Table 1. The kappa value for level L5/S1 was -0.18 and the kappa value for the levels L3/4 and L4/5 were 0.13 and 0.25 respectively. At levels L3/4 and L4/5 a total agreement of 40/63 was found.

Conclusion At level L5/S1 US does not seem to be of any diagnostic value. At the levels L3/4 and L4/5 US was in agreement with CT/MRI in about 2/3 of the patients and could be of potential use at these levels. Further investigations are needed to establish the role of US.

References

1. Hagen A, Tertsch D, Schön R, Richwien R. Die spinale ultraschalltomographie in der diagnostik des lumbalen bandscheibenvorfalls. Zbl Neurochir 1987;48:280–4

2. Tervonen O, Lähde S, Vanharanta H. Ultrasound diagnosis of lumbar disc degeneration. Comparison with computed tomography/discography. Spine 1990;16(8):951–8