Study on the consistency between CT hounsfield units and MRI evaluation of preoperative cervical paraspinal muscular fat infiltration in patients undergoing ACDF.

ABSTRACT

OBJECTIVE: To explore the feasibility of applying CT Hounsfield Units (HUs) for the assessment of preoperative paraspinal muscle fat infiltration (FI) in different segments in patients who underwent anterior cervical discectomy and fusion (ACDF). To compare the consistency of preoperative paraspinal muscle FI evaluations using MRI and those via CT HUs in patients who underwent ACDF surgery. METHODS: Ninety-five patients (45 males and 50 females, aged 37‒71 years) who received CT and MRI examinations and underwent ACDF surgery were retrospectively analyzed. In the axial T2-weighted MR images at the median level of the C3/4, C4/5, and C5/6 segments, regions of interests (ROIs) were delineated along the boundaries of the cervical multifidus (MF) and semispinalis cervicis (Scer) muscles. Using the threshold tool in ImageJ software, areas of fat tissue and intermuscular septa within the ROI were quantified. The effective cross-sectional area (ECSA) for each side was obtained by subtracting the areas of fat tissue and intermuscular septa from the total ROI area. The ratio of the fat tissue area to the CSA was then calculated to determine the initial FI value. The depth of subcutaneous fat from the midline spinous process to the epidermis at the median plane of the C4/5 intervertebral disc was measured. The initial FI values were then divided by the depth of fat to determine the post-correction FI value. Using the Picture Archiving and Communication System (PACS), at identical segments and planes, ROIs were delineated using the same method as in MRI under a standard soft tissue window (width of 500 HU, level of 60 HU). The CT HU values were measured within these defined areas. The CT HU values from both sides are summed to obtain the total HU value for the segment. According to whether the measurement results of two sets of data follow a normal distribution, Pearson's test or Sperman's test was used to analyze the correlation. RESULTS: On MRI, a statistically significant difference was observed in the post-correction FI only at the C3/4 segment compared to the other two segments (P < 0.05). No significant difference in the post-correction FI between the C4/5 and C5/6 segments was noted (P > 0.05). The CT HU results showed a substantial discrepancy between C3/4 and C4/5 segments and between C3/4 and C5/6 segments (P < 0.05), whereas no statistically significant difference was found in the CT HU value between the C4/5 and C5/6 segments (P > 0.05). The consistency analysis revealed a relatively strong correlation between the post-correction FI and CT HU values of the C3/4 and C4/5 segments. Furthermore, a strong correlation was detected in the variations in the measurement outcomes at the C5/6 segment. CONCLUSION: Patients requiring surgical treatment for the cervical spine exhibit varying degrees of FI in paraspinal muscles across different locations and segments. Evaluating the degree of FI in the paraspinal muscles of the cervical spine through CT HU values is feasible. There is considerable consistency between the post-correction FI assessed under MRI and the measurements of CT HU values in evaluating the FI of paraspinal muscles in the cervical spine.