Accuracy of multimodal image fusion for oral and maxillofacial tumors: A revised evaluation method and its application

Lei-Hao Hu,Wen-Bo Zhang,Yao Yu,Xin Peng
DOI: https://doi.org/10.1016/j.jcms.2020.05.009
2020-08-01
Abstract:<h2 class="section-title u-h3 u-margin-l-top u-margin-xs-bottom">Summary</h2><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Objectives</h3><p>To develop a revised evaluation method for accuracy of multimodal image fusion for oral and maxillofacial tumors and explore its application for comparing the accuracy of three commonly used fusion algorithms, automatic fusion, manual fusion, and registration point-based fusion.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Materials and Methods</h3><p>Image sets of patients with oral and maxillofacial tumor were fused using the iPlan 3.0 navigation system. Fusion accuracy included two aspects: (1) overall fusion accuracy: represented by the mean value of the coordinate differences along the x-, y-, and z-axes (<span class="math"><math>Δx</math></span>, <span class="math"><math>Δy</math></span>, and <span class="math"><math>Δz</math></span>), mean deviation (<em>MD</em>), and root mean square (<em>RMS</em>) of six pairs of landmarks on the two image sets; (2) tumor volume fusion accuracy: represented by Fusion Index (<em>FI</em>), which was calculated based on the volume of tumor delineated on the two image sets.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Results</h3><p>Eighteen pairs of image sets of 17 patients were enrolled in this study. The <span class="math"><math>Δx</math></span> and <span class="math"><math>Δy</math></span> values for the three algorithms were less than 1.5 mm. The <span class="math"><math>Δz</math></span> values for automatic fusion, manual fusion and registration point-based fusion was 1.049 mm, 1.864 mm and 1.254 mm. The <em>MD</em> for automatic fusion, manual fusion and registration point-based fusion was 1.978 mm, 2.788 mm and 1.926 mm. Significant differences existed in <span class="math"><math>Δz</math></span> for manual fusion and that for automatic fusion (<em>P</em>=0.058), in <em>MD</em> for manual fusion and that for automatic fusion (<em>P</em>=0.087), and in <em>MD</em> for manual fusion and that for registration point-based fusion (<em>P</em>=0.069). The <em>FI</em> for automatic fusion, manual fusion, and registration point-based fusion was 0.594, 0.520, and 0.549; the inter-algorithm differences were not significant (<em>P</em>=0.290).</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Conclusion</h3><p>The automatic fusion and the registration point-based fusion were more accurate than manual fusion, and therefore were recommended to be used in multimodal image fusion for oral and maxillofacial tumors.</p>
dentistry, oral surgery & medicine,surgery
What problem does this paper attempt to address?