Caresio Cristina, Salvi Massimo, Molinari Filippo, Meiburger Kristen M, Minetto Marco Alessandro
Biolab, Department of Electronics and Telecommunications, Politecnico di Torino, Turin, Italy.
Biolab, Department of Electronics and Telecommunications, Politecnico di Torino, Turin, Italy.
Ultrasound Med Biol. 2017 Jan;43(1):195-205. doi: 10.1016/j.ultrasmedbio.2016.08.032. Epub 2016 Oct 6.
Musculoskeletal ultrasound imaging allows non-invasive measurement of skeletal muscle thickness. Current techniques generally suffer from manual operator dependency, while all the computer-aided approaches are limited to be semi-automatic or specifically optimized for a single muscle. The aim of this study was to develop and validate a fully automatic method, named MUSA (Muscle UltraSound Analysis), for measurement of muscle thickness on longitudinal ultrasound images acquired from different skeletal muscles. The MUSA algorithm was tested on a database of 200 B-mode ultrasound images of rectus femoris, vastus lateralis, tibialis anterior and medial gastrocnemius. The automatic muscle thickness measurements were compared to the manual measurements obtained by three operators. The MUSA algorithm achieved a 100% segmentation success rate, with mean differences between the automatic and manual measurements in the range of 0.06-0.45 mm. MUSA performance was statistically equal to the operators and its measurement accuracy was independent of the muscle thickness value.
肌肉骨骼超声成像可实现对骨骼肌厚度的无创测量。当前技术通常依赖人工操作,而所有计算机辅助方法都局限于半自动或仅针对单一肌肉进行特定优化。本研究的目的是开发并验证一种名为MUSA(肌肉超声分析)的全自动方法,用于测量从不同骨骼肌获取的纵向超声图像上的肌肉厚度。MUSA算法在包含200张股直肌、股外侧肌、胫骨前肌和腓肠肌B模式超声图像的数据库上进行了测试。将自动测量的肌肉厚度与三名操作人员手动测量的结果进行了比较。MUSA算法的分割成功率达到了100%,自动测量与手动测量之间的平均差异在0.06 - 0.45毫米范围内。MUSA的性能在统计学上与操作人员相当,其测量精度与肌肉厚度值无关。