An In-Silico Testbed for Fast and Accurate MR Labeling of Orthopaedic Implants

  1. Electrical & Computer Eng. Dept. Worcester Polytechnic Institute, Worcester MA 01609-2280 USA
  2. Ansys, Inc., 2600 Ansys Drive, Canonsburg, PA 15317 USA
  3. Dassault Systèmes Deutschland GmbH, Bad Nauheimer Str. 19, 64289 Darmstadt, Germany
  4. Neva Electromagnetics, LLC, 1010 Main St., Holden, MA 01520
  5. GE HealthCare, 500 W Monroe Street, Chicago, IL, 60661 USA
  6. Micro Systems Engineering, Inc., an affiliate of Biotronik, 6024 Jean Road, Lake Oswego, OR 97035 USA
  7. Musculoskeletal Translational Innovation Initiative, Department of Orthopedic Surgery, Beth Israel Deaconess Medical Center and Harvard Medical School, 330 Brookline Ave, RN123, Boston, MA, 02215 USA
  8. Harvard Medical School, Boston, MA 02115 USA
  9. Athinoula A Martinos Center for Biomed. Imaging, Massachusetts General Hospital 149 13, St. Charlestown, MA 02129 USA


  • Reviewing Editor
    Peng Liu
    Icahn School of Medicine at Mount Sinai, United States of America
  • Senior Editor
    Aleksandra Walczak
    École Normale Supérieure - PSL, France

Reviewer #1 (Public Review):

In this work, the authors are trying to satisfy a real need in MR safety, when concerns can arise about the thermal increase due to metallic materials in patients carrying orthopedic implants. The "MR conditional" labeling of the implant obtained by ASTM in-vitro tests may help to plan the MR scan, but it is normally limited to a single specific MR sequence and a B0 value, and it is not always available. The adoption of an in-silico simulation testbed overcomes this limitation, providing a fast and reliable prediction of temperature increase from RF, in real-life scan conditions on human-like digital models. The FDA is pushing this approach.

The presented in-silico testbed looks valuable and validated. It is based on the widely available Visible Human Project (VHP) datasets, and the testbed is available online. The approval of the testbed by the FDA as a medical device development tool (MDDT) is a good premise for the large-scale adoption of this kind of solution.

There are a couple of limitations in the study that must be clearly highlighted to the readers.

While the RF-related heating is very well modeled, the gradients-related heating is out of the scope of this paper and not considered. Readers must be warned that RF causes only a part of the heating, and literature is reporting cases where also gradient switching can contribute, as correctly mentioned in this work. A cautious attitude should consider this as a significant limitation of the study.

Moreover, the way the implant is embedded in the VHP model is shortly documented in the materials and methods and mostly focuses on implant registration on bone tissue. It is not clear how to manage the empty space and the soft tissue stretching/reshaping generated by the simulated surgery (for example, by the cut of the femoral head in total hip arthroplasty). It is reported by literature that the level of accuracy in the simulated surgery can impact in some cases (RF vs. gradients heating, massive vs. thin or elongated implants) on temperature predictions.

Reviewer #2 (Public Review):

In this article, the authors provide a method of evaluating the safety of orthopedic implants in relation to radiofrequency-induced heating issues. The authors provide an open-source computational heterogeneous human model and explain computational techniques in a finite element method solver to predict the RF-induced temperature increase due to an orthopedic implant while being exposed to MRI RF fields at 1.5 T.

The open-access computational human model along with their semiautomatic algorithm to position the implant can help realistically model the implant RF exposure in patients avoiding over- or under-estimation of RF heating measured using rectangular box phantoms such as ASTM phantom. Additionally, using numerical simulation to predict radiofrequency-induced heating will be much easier compared to the experimental measurements in an MRI scanner, especially when the scanner availability is limited.

The proposed method only used radiofrequency (RF) field exposure to evaluate the heating around the implant. However, in the case of bulky implants, the rapidly changing gradient field can also produce significant heating due to large eddy currents. So the gradient-induced heating still remains an issue to be evaluated to decide on the safety of the patient. Moreover, the method is limited to a single human model and might not be representative of patients with different age, sex, and body weights. Additionally, the authors compare the temperature rise predicted by their method to an earlier study. However, there is no information about how they controlled the input power in their simulation testbed compared to the earlier study in showing validation of the method.

  1. Howard Hughes Medical Institute
  2. Wellcome Trust
  3. Max-Planck-Gesellschaft
  4. Knut and Alice Wallenberg Foundation