Article
Diagnostic value of three-dimensional ultrasound to detect tumor remnants in glioma surgery
Search Medline for
Authors
-
Dorothea Miller
- Klinik für Neurochirurgie, Universitätsklinikum Essen
-
Marc Caspary
- Klinik für Neurochirurgie, Universitätsklinikum Essen
-
Klaus-Peter Stein
- Klinik für Neurochirurgie, Universitätsklinikum Essen
-
Neriman Özkan
- Klinik für Neurochirurgie, Universitätsklinikum Essen
-
Ulrich Sure
- Klinik für Neurochirurgie, Universitätsklinikum Essen
| Published: | May 21, 2013 |
|---|
Outline
Text
Objective: Three-dimensional ultrasound (3D US) has gained increasing interest in neurosurgery as a low-cost alternative to intraoperative MRI. However, 3D image quality depends on the way of 3D reconstruction. Recently, we showed that distance weighted interpolation algorithms (DWIA) improve the utility of 3D US during intraoperative navigation. The aim of the present study was to determine the diagnostic value of 3D US using DWIA to detect tumor remnants as compared to postoperative MRI.
Method: Eighteen patients with neuroepithelial tumors were included in the study. A navigated 3D US volume was acquired prior and after resection by moving a tracked 2D probe over the area of interest. A 3D volume was reconstructed using a DWIA. Ten points were selected at the resection margins in each case. Points were judged as tumor-free, suspicious for tumor or including residual tumor according to 3D US images. US imaging results were compared to post-operative MRI at the specific points by fusing post-resection US and post-surgical MRI. Sensitivity, specificity as well as predictive values were calculated.
Results: Tumors could be detected and delineated well from surrounding brain tissue in all cases prior to resection. 3D US with DWIA was very helpful in determining the extent of resection in 16 of 18 cases, but was less helpful in two cases of small temporomesial low-grade lesions. The sensitivity to detect residual tumor after resection was 67.2%, specificity was 66.5%, the positive predictive value was 32.8% and the negative predictive value was 66.5%. The diagnostic value was significantly improved as compared to 3D volumes reconstructed without a DWIA, where artifact formation and gaps within the 3D volume lead to inconclusive results in one third of cases.
Conclusions: 3D US is a suitable and cost-effective tool during tumor surgery. DWIA improve 3D image quality, increase its diagnostic value and allow the detection of tumor remnants with a good sensitivity even at the end of resection.
