Clarification on Segment Statistics: Surface Area vs. Cross-Sectional Area (CSA) for Muscle Analysis

Support
, , ,
1

Hi everyone,

I am performing a quantitative body composition analysis on single-slice CT scans (at the D10 level) using the Segment Statistics module in 3D Slicer. I would like to clarify a methodological doubt to ensure the accuracy of my data:

  1. Surface Area vs. CSA: Can you confirm that the ‘Surface Area’ (mm²) metric provided by the module refers to the total 3D surface area of the generated mesh (the “envelope” or “shell” of the segment)? In my single-slice analysis, this value is significantly higher than the expected planar area. Is it correct to assume that ‘Surface Area’ should not be used as a proxy for the Cross-Sectional Area (CSA) or Skeletal Muscle Area (SMA)?

  2. SMA Calculation: To derive the Skeletal Muscle Area (SMA) in cm^2, I am currently taking the ‘Volume’ (mm³) of the segment and dividing it by the slice thickness (e.g., 1 mm). Is this “Volume / Thickness” conversion considered a standard and reliable method in Slicer to obtain the 2D planar area from a single-slice segmentation?

I want to make sure I am not misinterpreting the geometric outputs of the software, as using the wrong metric would lead to incorrect Skeletal Muscle Index (SMI) calculations.

Thank you in advance for your help!

2

You are correct that Segment Statistics returns the 3D surface area of the generated mesh, the sum of the areas of all the triangles which make up the surface. To your second question, no, you can’t derive cross sectional areas by dividing the total surface area by a slice thickness. Such a calculation would give you an average volume per slice, but would not tell you about the CSA. To get a series of CSA values, you should cut across the muscle perpendicular to the axis you want to measure relative to (presumably the long axis of the muscle or perhaps a centerline of the muscle), and measure the enclosed area. This discussion will likely be helpful for you: Extracting muscle cross-sectional area and its centroid locatation from MRI/CT scans