New perspectives on surgical accuracy analysis of image-guided bone tumour resection surgery

Purpose: Image-guided bone tumour resection surgery has been proved in previous literatures to be more accurate than those conventional freehand ones (p < 0.001). However, in this kind of surgery, there are still many procedures depending on manual operations, which will inevitably introduce surgical errors into the surgery. In particular, the negative surgical errors (i.e., errors toward tumour) would increase the risk of tumor recurrence and metastasis. Thus, the first purpose of this study was to evaluate whether the negative surgical errors of image-guided bone tumour resection surgery were statistically significantly great, the second purpose is to evaluate whether the negative surgical errors of image-guided long-bone tumour resection surgery were statistically equivalent to those of pelvis surgery, and the last purpose is to recommend a solution for suppressing these errors when using a navigation system. Methods: Negative surgical errors of 24 osteotomies in ten pelvis tumour resection operations and 16 osteotomies in ten long-bone surgeries under the image guidance of a navigation system were statistically evaluated and compared with − 2.0 mm. The equivalence of negative surgical errors of pelvis group and those of long-bone group was statistically tested. To suppress these negative surgical errors when using a navigation system, we recommend, based on the obtained statistics, to increase the margins between cut planes and tumour boundary during pre-operatively planning cut planes, by adding an extra margin with the empirical safe margin according to the absolute lower bound of 95% CI of negative surgical errors. Results: Negative surgical errors of the pelvis group and the long-bone group were both significantly less than − 2.0 mm (p < 0.001), but not statistically equivalent (Rg > 1 mm). 95% CI of negative surgical errors were from − 3.95 to − 3.27 mm for the pelvis group, and from − 2.69 to − 2.34 mm for the long-bone group. So, the extra margin added for image-guided pelvis tumour resection surgery should be set as 3.95 mm, and the extra margin added for image-guided long-bone surgery should be set as 2.69 mm. Conclusion: The negative surgical errors of image-guided bone resection surgery were statistically significantly less than − 2.0 mm (p < 0.001), thus these errors cannot be safely ignored. Moreover, the negative surgical errors of the pelvis group were not equivalent to those of the long-bone group (Rg > 1.0 mm), thus the solution for image-guided pelvis tumour resection surgery and that for image-guided long-bone tumour resection surgery should be separately determined. In order to suppress these negative surgical errors when using a navigation system, we recommend to add extra 3.95 mm margin with the empirical safe margin for image-guided pelvis tumour resection surgery and to add extra 2.69 mm margin for image-guided long-bone tumour resection surgery during pre-operatively planning cut planes.