Clinical application of HEDI for biomechanical evaluation and visualisation in incisional hernia repair

1. Problem Statement

Incisional hernia repair is frequently associated with high recurrence rates (22–32%) and chronic postoperative pain. Current limitations in surgical planning include:

• Oversimplified Mesh Sizing: Traditional approaches rely on fixed overlap margins (5–7 cm) based solely on the hernia defect size, ignoring the broader area of abdominal wall instability.
• Neglect of Biomechanics: Standard methods fail to account for cyclic loading (e.g., coughing, jumping), tissue elasticity, and internal pressure, which can cause mesh sliding, "shakedown," or failure at the mesh-tissue interface.
• Manual Assessment Limitations: Preoperative evaluation of hernia size and wall instability via manual CT review is time-consuming, observer-dependent, and lacks quantitative precision.
• Misalignment of Defect and Instability: The hernia orifice is often not centered within the unstable abdominal wall region; covering only the defect with a fixed margin leaves "weak spots" prone to recurrence.

2. Methodology

The study introduces HEDI (Hernia Evaluation, Detection, and Imaging), a computational tool designed to automate the assessment of abdominal wall biomechanics using standard clinical data.

A. Data Acquisition

• Input: Dual-phase Computed Tomography (CT) scans for each patient:
  1. Rest State: During shallow expiration.
  2. Strain State: During a forced Valsalva maneuver (increasing intra-abdominal pressure).
• Cohort:
  • Development: 141 patients used to train and validate the tool.
  • Clinical Application: 31 patients underwent surgery with HEDI-guided planning and were followed for three years.

B. Technical Workflow

1. Automatic Segmentation:
   • Utilized a deep neural network (Biomedisa) trained on manually annotated CT scans.
   • Segmented four key regions: abdominal cavity volume, lateral muscle structure, abdominal muscles, and hernia volume.
   • Performance: Achieved Dice scores of 0.93 (cavity), 0.79 (lateral muscle), and 0.55 (hernia volume). Hernia volume segmentation was more accurate during the Valsalva maneuver (Dice 0.64) than at rest (Dice 0.49).
2. Symmetric Diffeomorphic Registration:
   • Used the DIPY library (Python) to align the "Rest" and "Valsalva" CT scans.
   • Generated a 3D displacement field transforming the rest state to the strain state.
   • Calculated the Green-Lagrange strain tensor to quantify local tissue deformation.
3. Visualization & Thresholding:
   • Defined the "unstable abdominal wall" as areas with displacement >15 mm during the Valsalva maneuver.
   • Generated color-coded 3D surface meshes: Red/Yellow/White for high displacement (>15 mm) and Cyan/Blue for stable areas (<15 mm).
   • Visualized strain "hotspots" to guide fixation strategies.

C. Surgical Integration (GRIP Concept)

HEDI was used as an adjunct to the GRIP concept (a biomechanical framework for hernia repair).

• Mesh Selection: Mesh size was determined to cover the entire unstable area (not just the defect) and all strain hotspots.
• Fixation Strategy: Targeted fixation points were placed in areas of low displacement, while high-strain areas received reinforced fixation (e.g., continuous running sutures).
• Mesh-to-Defect Ratio: Aimed for a ratio of ~6:1 to ensure stability in elastic tissues.

3. Key Contributions

• Novel Tool (HEDI): The first automated tool to quantify abdominal wall instability and hernia volume by comparing rest and strain-state CT scans.
• Biomechanical Visualization: Provides surgeons with a visual map of tissue displacement and strain, moving beyond static anatomical measurements to dynamic functional assessment.
• Refined Surgical Strategy: Demonstrates that successful repair requires covering the entire unstable wall (often much larger than the defect) rather than just the hernia orifice.
• Open-Source Implementation: The source code and a test dataset are publicly available via GitHub and Zenodo, promoting reproducibility.

4. Results

• Clinical Outcomes (31 Patients):
  • Recurrence: 0% recurrence rate over a 3-year follow-up.
  • Pain: 100% of patients were pain-free (defined as no pain or minor discomfort relieved by rest without analgesics).
  • Mesh Usage: Average mesh size was 1009 cm² (significantly larger than the average defect area of 210 cm²), with an average mesh-to-defect ratio of 6:1.
• Tool Performance:
  • Processing Time: Average computation time of ~5.5 minutes per patient.
  • Accuracy: Registration errors were within the range of human landmark placement error (mean absolute error ~1.6 mm).
  • Limitations: 27 out of 168 total cases (141 development + 31 clinical) were excluded due to scanning artifacts (e.g., truncation, misalignment, different slice thickness), highlighting the need for strict acquisition protocols.

5. Significance and Future Directions

• Paradigm Shift: HEDI shifts hernia repair from a "defect-centric" approach to a "biomechanical stability-centric" approach, acknowledging that the unstable tissue extends beyond the visible hernia.
• Clinical Utility: While not a deterministic decision-maker, HEDI serves as a powerful visual aid that complements surgeon expertise, leading to tailored mesh sizing and fixation.
• Caveats:
  • The study lacks a formal negative control arm (comparing HEDI-guided vs. standard care in a randomized trial).
  • The 15 mm displacement threshold is currently absolute; future work aims to establish patient-specific relative thresholds.
  • The sample size is small, and the tool's development occurred concurrently with its clinical use.
• Future Work: Plans include larger cohort studies, propensity score matching with the HERNIAMED registry, and exploring non-invasive alternatives (MRI/Laser) to reduce radiation exposure.

Conclusion: The study validates the feasibility of integrating biomechanical imaging into routine clinical workflows. By visualizing abdominal wall dynamics, HEDI supports the creation of durable, patient-specific repairs, potentially addressing the high recurrence and pain rates associated with traditional incisional hernia surgery.