Bone T Michael, Mowry Sarah E
Medical College of Georgia at Augusta University, Augusta, Georgia.
Otol Neurotol. 2016 Sep;37(8):1183-8. doi: 10.1097/MAO.0000000000001131.
Computed tomographic (CT) scans of the 3-D printed temporal bone models will be within 15% accuracy of the CT scans of the cadaveric temporal bones.
Previous studies have evaluated the face validity of 3-D-printed temporal bone models designed to train otolaryngology residents. The purpose of the study was to determine the content validity of temporal bone models printed using inexpensive printers and materials.
Four cadaveric temporal bones were randomly selected and clinical temporal bone CT scans were obtained. Models were generated using previously described methods in acrylonitrile butadiene styrene (ABS) plastic using the Makerbot Replicator 2× and Hyrel printers. Models were radiographically scanned using the same protocol as the cadaveric bones. Four images from each cadaveric CT series and four corresponding images from the model CT series were selected, and voxel values were normalized to black or white. Scan slices were compared using PixelDiff software. Gross anatomic structures were evaluated in the model scans by four board certified otolaryngologists on a 4-point scale.
Mean pixel difference between the cadaver and model scans was 14.25 ± 2.30% at the four selected CT slices. Mean cortical bone width difference and mean external auditory canal width difference were 0.58 ± 0.66 mm and 0.55 ± 0.46 mm, respectively. Expert raters felt the mastoid air cells were well represented (2.5 ± 0.5), while middle ear and otic capsule structures were not accurately rendered (all averaged <1.8).
These results suggest that these models would be sufficient adjuncts to cadaver temporal bones for training residents in cortical mastoidectomies, but less effective for middle ear procedures.
3D打印颞骨模型的计算机断层扫描（CT）与尸体颞骨的CT扫描相比，准确率在15%以内。
以往的研究评估了用于培训耳鼻喉科住院医师的3D打印颞骨模型的表面效度。本研究的目的是确定使用廉价打印机和材料打印的颞骨模型的内容效度。
随机选择四块尸体颞骨并进行临床颞骨CT扫描。使用之前描述的方法，在Makerbot Replicator 2×和Hyrel打印机上，以丙烯腈丁二烯苯乙烯（ABS）塑料制作模型。使用与尸体骨骼相同的方案对模型进行射线扫描。从每个尸体CT系列中选择四张图像，从模型CT系列中选择四张相应图像，并将体素值归一化为黑色或白色。使用PixelDiff软件比较扫描切片。由四名获得委员会认证的耳鼻喉科医生在4分制下对模型扫描中的大体解剖结构进行评估。
在四个选定的CT切片上，尸体扫描和模型扫描之间的平均像素差异为14.25±2.30%。皮质骨平均宽度差异和外耳道平均宽度差异分别为0.58±0.66毫米和0.55±0.46毫米。专家评估者认为乳突气房显示良好（2.5±0.5），而中耳和耳囊结构的呈现不准确（均平均<1.8）。
这些结果表明，这些模型对于培训住院医师进行皮质乳突切除术来说，足以作为尸体颞骨的辅助工具，但对于中耳手术则效果较差。
