Thermal Ablation Simulation Research Plan Summer 2014

Garron Deshazer with Dr. Derek Merck, Dr. Damian Dupuy, Dr. Edward Walsh, Dr. Punit Prakash, Dr. William Park


Model and validate thermal profiles produced by a physically based generative model of various microwave ablation procedures (MWA) and correlate these profiles with 3D ablation zone size and shape.


  • We are proposing that thermal models are necessary margin estimation in ablation therapy, and therefore necessary for predictive planning.

  • 4D (time driven) thermal profiles are necessary to develop and validate time dependent thermal models in order to ensure robust accuracy of proposed generative based modeling approach, as well as eventually provide clinicians with more pre-procedural planning advantages

  • Intra-operative 4D thermometry could allow physicians to directly observe when a target tissue region has been completely treated (reached a target temperature). Thus far MR thermography has not been used for MWA due to the limitations of not being able to bring the generator or metallic antenna into the room, however we now have a research MWA generator and MR safe antenna that we can bring into the room for testing. We also plan to try to perform validation of MWA ablation experiments using CT thermometry. Previous Work: Since the spring my physical based generative thermal model has been validated with homogenous ex-vivo MWA experiments. Comparisons were made by visual inspection after dissection of the ex-vivo sample, by overlaying Isothermal contours on to 3D CT data collected of the segmented ablation zone at various time points. The results of the simulation a similar ex-vivo MWTA procedure showed agreement with a 0.5 cm tolerance over all simulated time points (Deshazer AAPM 2014, Deshazer RSNA 2014) . This type of simulation-validation pipeline is ideal for optimization and simulation improvement because it allows for more data comparison as the ablation zone grows with time. The MR thermal mapping of the agarose phantom that was proposed in the spring showed agreement within less than 1 degree Celsius during the cool down period. When the initial heating of the phantom was compared to the simulation they were in agreement within 5 degrees Celsius. This initial model forms the basis of ongoing work in principled thermal simulation for MWA planning. Future goals are to incorporate heterogeneous tissue types and heat sinks into the solution domain, and validate the resulting simulation against our ongoing clinical data collection.


It is necessary to add complexity to make it more relevant in a clinical setting. Incorporating heterogeneity into my model as well as to make it completely 3D without assuming axial symmetry are the next steps in order to start doing preliminary patient specific modeling

  • Go to KSU as a visiting scholar in Dr. Punit Prakash’s lab to learn how to do 3D patient specific modeling in Comsol Multiphysics. The goal is to learn how to create a commercial antenna inside a homogeneous phantom (lung and liver) and to incorporate from patient dicom images segmented areas that we can assign different dielectric properties . The idea here is to within a 3D finite element mesh solve the heat transfer boundary conditions when there are boundaries that have a change in temperature dependent relative permittivity, electrical conductivity, ect, and observe how this thermal profile differs from a completely homogenous one over time. Since it is known that tumors in the liver and lung have different dielectric properties than normal tissue then I can use these properties to create a completely physically based generative patient specific thermal model to predict the heat transfer over the course of treatment.

  • After preliminary experiments with Dr. Walsh at the Brown MRF center we have now partnered with BSD to make a MWA MR compatible system that we can start doing experiment’s in the MR suite while the scanner is on. Once the antenna is built we will be able to compare my simulation during the ablation procedure with complete 3D thermal mapping and not just the cool down procedure. This will be ideal for further testing, however performing CT validation is also important because CT thermometry could be developed for the high temperature produced during this procedure and the mere fact that not every hospital has an MR suite.

  • Look into modeling various nanoparticle thermal enhancers that are placed close to the antenna in ex-vivo samples

Academic Plan

  • NEAAPM 2014 annual research symposium. Oral talk at Tufts University May 2014. High Honors Award for having the top 5 abstracts presented

  • AAPM 2014 annual conference abstract accepted. Poster Talk in Austin Texas in mid July

  • RSNA 2104 annual conference abstract accepted. Oral talk in Chicago in Late November

  • NCIGT MICCAI Workshop abstract accepted. Poster Talk at Harvad in September

  • SPIE conference abstract submission. Follow-up paper necessary if accepted

Goals for Fall 2014 (next 4 months)

  • Ablation Planning Workshop

  • Running simulations on Brown High performance computing center

  • MR thermography issues (coaxial filter to limit frequency interference with the signal) and creating MR safe+ compatible cable.

  • Med submission Journal Submission

    • Ex-vivo liver
    • Ex-vivo kidney
    • Absolute temperature measurements during ablation zone at the margin and further away
    • 3 patient specific liver cases
    • Statistical analysis