Dipartimento di Ingegneria, University of Ferrara, Via Saragat 1, 44122, Ferrara, Italy.
Int J Comput Assist Radiol Surg. 2013 Mar;8(2):247-57. doi: 10.1007/s11548-012-0773-7. Epub 2012 Jun 17.
Measurements accomplished on most oral implantology software are often affected by some systematic effects, of which those related to the CT dataset anisotropy are the most relevant. In fact, most of these commercial systems do not manage possible anisotropy in input datasets, leaving the responsibility to users and radiologists. Therefore, in order to achieve a better knowledge of the patient's anatomy before inserting the implants, and thus reducing the risk of damaging the surrounding structures, the implementation of a complete and precise anisotropy management system is required.
We present an anisotropy management algorithm for pre-operative planning software that is able to handle any anisotropic CT dataset, and, as a result, provides a very precise isotropic equivalent. The developed algorithm exploits two interpolation passes to correct anisotropy and is characterised by linear complexity, needing just a few seconds to accomplish the tasks. The first pass concerns the integer-filling of possible intra-slice void spaces of the original slices, having the responsibility of a correct spreading of the radiographic details along the volume height axis. The second pass, instead, reformats its input dataset under isotropic conditions exploiting a contribution-based interpolation sub-algorithm.
The algorithm has been evaluated by comparing the anisotropy implied systematic effects for both anisotropic and interpolation-reconstructed radiographic volumes of five different scans. The proposed system demonstrated to be able to successfully handle any dataset interslice-pixel-size ratio. Moreover, the precision achieved proved to be even better than that of another precise algorithm that we previously developed and published, validating the proposed approach as a consequence.
The proposed algorithm makes it possible to handle and correct anisotropy in input CT datasets, helping to avoid anisotropy implied systematic effects on related measurements, and consequently supporting pre-operative planning software by providing a precise and isotropic equivalent volume on which to work.
大多数口腔种植学软件上进行的测量结果往往会受到一些系统效应的影响，其中与 CT 数据集各向异性相关的影响最为显著。事实上，大多数商业系统都无法处理输入数据集可能存在的各向异性，而将这一责任留给了用户和放射科医生。因此，为了在植入种植体之前更好地了解患者的解剖结构，从而降低损坏周围结构的风险，需要实现一个完整且精确的各向异性管理系统。
我们提出了一种用于术前规划软件的各向异性管理算法，该算法能够处理任何各向异性的 CT 数据集，并提供非常精确的各向同性等效体。所开发的算法利用两次插值传递来纠正各向异性，其特点是线性复杂度，仅需几秒钟即可完成任务。第一次传递涉及原始切片中可能存在的切片内空隙的整数填充，负责正确沿体积高度轴扩展射线细节。第二次传递则根据基于贡献的插值子算法在各向同性条件下重新格式化其输入数据集。
通过比较五个不同扫描的各向异性和插值重建射线照相体积的各向异性所带来的系统影响，对算法进行了评估。该系统证明能够成功处理任何切片间像素尺寸比的数据集。此外，所达到的精度甚至优于我们之前开发和发布的另一种精确算法，从而验证了所提出方法的有效性。
所提出的算法使得处理和纠正输入 CT 数据集中的各向异性成为可能，有助于避免各向异性对相关测量结果产生的系统影响，并通过提供精确的各向同性等效体积来支持术前规划软件，从而在该体积上进行工作。
