In-vivo abdominal wall strain determination in complex ventral hernias (cVH)

Hazel Marecki, BS, BS, Khashayar Rafatzand, MD, FRCPC, Richard Perugini, MD. University of Massachusetts Medical School

Objective:

Static computed tomography (CT) is the clinical standard for assessment of cVH, yielding limited anatomical & no physiologic/functional data. No established criteria exist for assessing cVH repair impact on abdominal wall function preoperatively. Moreover, postoperative assessment of outcomes is limited to hernia recurrence.

We utilized dynamic-MRI (dMRI) at our tertiary care center for cVH preoperative planning. dMRI assesses hernia metrics & muscle mass/atrophy similar to CT; it also evaluates muscle contraction, component geometry, deformation & core muscle interactions.

dMRI also allows strain-measurement along abdominal wall surfaces by tracking anatomically matched points throughout deformation induced by valsalva. Mechanical properties of abdominal wall & meshes have been tested ex-vivo. Currently no method exists for in-vivo strain analysis of cVH components.

Methods:

IRB-approved, HIPPAA-compliant study.
MRI: Axial T2-weighted SSFSE images pre- & post-Valsalva, Xiphoid-pubis
Point selection: Minimum 8 point-pairs per slice along sub-rectus plane & linea alba (Fig-1) using anatomical landmarks.
Strain calculation: A mesh of 3d points was imported into MatLAB. Point correlation was used for finite element analysis of each constant strain triangle; deformation for each triangle was computed individually & independently. Each two-dimensional triangle was transformed by inverse Euler rotation to a primary coordinate frame to align normal vectors before & after deformation. Surface strain was calculated as a system of two-vector deformation in relation to initial length. Strain was transformed to principal axes, representing axes of maximal & minimal strain. Average surface strain (Von Mises) was computed.

Results:

Patient:

70 year-old male with OSA
Weight=233 lbs, BMI=33.4 kg/m2
Three separate defects, Xiphoid to umbilicus (Fig-2).
Recti/lateral wall muscles: No atrophy, normal contraction.
Hernia properties presented in Fig-3.

Strain measurement:

Figures 4-5 show vector & net strain, respectively, for each triangular component of abdominal wall, coupled to coronal MRI images at rest.
- Fig-4: Red & blue vectors represent axes of maximal & minimal strain, respectively.
- Fig-5: Numerical strain corresponds to:
  - =1: no change compared to initial size
  - >1: stretched compared to initial size
  - <1: contracted compared to initial size
Impaired contraction (magenta) at larger hernia defects #1 & 3
Preserved contraction (cyan) at defect#2
No correlation between strain & hernia volume
Direct correlation between strain & hernia defect size & post-Valsalva increase in hernia volume.

Limitations:

Single patient, narrow range of hernia volumes & defect area.

Conclusion & Future Directions:

We demonstrated feasibility of in-vivo strain analysis using dMRIs as a new technique to assess functional properties of abdominal wall.

We aim to use this technique for standardized & multi-parametric pre- and post-operative assessment of cVH. We hope to better stratify the physiologic abnormality induced by cVH, and to better assess functional improvement of the abdominal wall following repair. Strain measurement may identify recurrence-prone areas following cVH repair as high strain regions.

Future technical applications include patient-specific, compliance-matched mesh synthesis & automation of point-selection & matching using machine algorithms.

Figure 1.