Three-dimensional meniscus allograft sizing—a study of 280 healthy menisci

Silvan Beeler, Lukas Jud, Marco von Atzigen, Reto Sutter, Philipp Fürnstahl, Sandro F. Fucentese, Lazaros VlachopoulosJournal of Orthopaedic Surgery and Research 2020 · DOI

TL;DR

This study introduces and validates a novel 3D-MRI based method for sizing meniscus allografts, demonstrating its superior accuracy over traditional 2D methods. By leveraging the contralateral meniscus as a precise 3D template, this approach significantly reduces sizing errors and the number of mismatched allografts, which is crucial for improving patient outcomes in meniscus transplantation.

What Problem Does This Paper Address?

Meniscus transplantation is a valuable procedure for patients suffering from chronic knee pain due to meniscectomy. However, a significant challenge in this field is accurately sizing the allograft to match the patient's native meniscus. Both undersized and oversized grafts can lead to poor joint congruity, increased load, extrusion, and ultimately, poorer biomechanical outcomes and accelerated degenerative changes. Current sizing methods, primarily based on conventional radiography (2D-RX) or 2D MRI, have limitations and inaccuracies, leading to a relevant problem in allograft surgery. This paper aims to address this by evaluating a new three-dimensional (3D) meniscus-sizing method that utilizes the contralateral knee as a template to improve allograft selection accuracy.

Methods

The study developed and compared three distinct meniscus sizing methods:

  1. 3D-MRI Sizing: This novel method uses a 3D surface model of the healthy contralateral meniscus as a template. It calculates the "mean surface distance" (MeSD) between the template and potential allografts. The allograft with the smallest MeSD is selected, aiming for the closest geometric match in three dimensions.
  2. 2D-MRI Sizing: This method uses 2D measurements of meniscal width and length from MRI scans of the contralateral meniscus. Allografts are selected based on the smallest difference in these 2D dimensions.
  3. 2D-RX Sizing (Pollard method): This conventional method uses measurements derived from plain radiography of the ipsilateral knee, specifically the tibial plateau width and length, to indirectly calculate meniscal dimensions. Allografts are selected based on the smallest difference between these calculated dimensions and the allograft's dimensions.

The core of the validation involved creating 3D surface models of 280 healthy menisci from MRI scans. An "imaginary tissue bank" of 138 different menisci was then constructed. For each of the 100 menisci in the "validation group," the three sizing methods were simulated to select the "best fitting" allograft from the tissue bank. The selected allografts were then compared to the original "validation group" menisci using surface distance and dimensional differences (width, length, height) as metrics.

Datasets

The study utilized data from:

  • 50 patients with bilateral knee imaging: This group served as the primary "validation group" for comparing the sizing methods. It included 34 men and 56 women, with a mean age of 26.7 years.
  • 40 patients with unilateral knee imaging: These patients contributed additional menisci to create a larger "allograft pool."

The total dataset comprised 280 healthy menisci (140 medial, 140 lateral) from which 3D surface models were generated. The "allograft pool" consisted of 138 different menisci. The study did not introduce a new benchmark dataset but rather used existing patient data to simulate and validate sizing methods against a simulated allograft pool.

Key Insights and Ideas

  • 3D Shape Matters: The study strongly suggests that matching the overall 3D shape of the meniscus is more critical than just matching width and length. The MeSD metric effectively captures this 3D similarity.
  • Contralateral Template is Key: Using the contralateral meniscus as a template for 3D sizing provides a highly accurate and reliable reference, as it represents the patient's native anatomy.
  • Limitations of 2D Methods: Traditional 2D methods, whether from MRI or radiography, are shown to be less precise and lead to a higher number of "outliers" (significantly mismatched grafts).
  • Outlier Reduction is Crucial: The primary benefit of the 3D-MRI method lies in its ability to significantly reduce the number of mismatched grafts, which is a major factor in clinical success.
  • Potential for Improved Distribution: By improving sizing accuracy, this method could lead to more effective distribution of the limited supply of meniscus allografts.

Key Results

  • Superiority of 3D-MRI Sizing: The 3D-MRI sizing method using MeSD was significantly superior to both 2D-MRI and 2D-RX sizing (p < 0.001) across all measurement parameters.
  • Outlier Reduction:
    • For the medial meniscus, 3D-MRI sizing reduced outliers by up to 55% compared to 2D-MRI and 83% compared to 2D-RX.
  • For the lateral meniscus, outliers were reduced by 39% compared to 2D-MRI and 56% compared to 2D-RX.
  • Dimensional Differences: While 3D shape matching (MeSD) was significantly improved, the mean differences in meniscal width, length, and height were only slightly better with 3D-MRI compared to the other methods, highlighting the importance of the overall 3D geometry.
  • Inter-rater/Intra-rater Reliability: The reliability of the measurements for all three methods was good to excellent, indicating that the measurements themselves are robust.

Limitations and Future Work

  • Direct Anatomical Comparison: The study acknowledges the lack of a direct comparison between the segmented 3D models and the actual anatomical shapes, though it notes MRI's established reliability in identifying meniscus tissue.
  • Slide Thickness Influence: The 3 mm slide thickness used in MRI could influence results, especially given that a 3-4 mm mismatch is considered clinically relevant. Biplanar segmentation was used to mitigate this.
  • Tissue Bank Size: While the study used a relatively large "allograft pool" of 138 menisci, a larger pool could potentially yield even better results.
  • Cost of 3D Sizing: The authors note that 3D sizing involves additional costs for MRI scans and 3D model generation. However, they suggest that future advancements in automated segmentation could reduce these costs.
  • Future Directions: The authors propose that 3D templates could be used for more advanced applications, such as guiding the creation of biomimetic 3D-printed scaffolds or improving bone plug/bridge fixation techniques. They also suggest that semiautomated or fully automated meniscus segmentation could further streamline the process.

Summary updated

Abstract

Kantonale Ethikkommission Zürich had given the approval for the study (BASEC-No. 2018-00856).
BibTeX
@article{beeler2020three,
  author = {Silvan Beeler and Lukas Jud and Marco von Atzigen and Reto Sutter and Philipp Fürnstahl and Sandro F. Fucentese and Lazaros Vlachopoulos},
  title = {{Three-dimensional meniscus allograft sizing—a study of 280 healthy menisci}},
  journal = {Journal of Orthopaedic Surgery and Research},
  year = {2020},
  doi = {10.1186/s13018-020-01591-z},
}
Figure