Magnetic Resonance Temperature Imaging Research Articles

Experimental assessment of phase aberration correction for breast MRgFUS therapy

Purpose: This study validates that phase aberrations in breast magnetic resonance-guided focussed ultrasound (MRgFUS) therapies can be corrected in a clinically relevant time frame to generate more intense, smaller and more spatially accurate foci.Materials and methods: Hybrid angular spectrum (HAS) ultrasound calculations in an magnetic resonance imaging (MRI)-based tissue model, were used to compute phase aberration corrections for improved experimental MRgFUS heating in four heterogeneous breast-mimicking phantoms (n = 18 total locations). Magnetic resonance(MR) temperature imaging was used to evaluate the maximum temperature rise, focus volume and focus accuracy for uncorrected and phase aberration-corrected sonications. Thermal simulations assessed the effectiveness of the phase aberration correction implementation.Results: In 13 of 18 locations, the maximum temperature rise increased by an average of 30%, focus volume was reduced by 40% and focus accuracy improved from 4.6 to 3.6 mm. Mixed results were observed in five of the 18 locations, with focus accuracy improving from 6.1 to 2.5 mm and the maximum temperature rise decreasing by 8% and focus volume increasing by 10%. Overall, the study demonstrated significant improvements (p < 0.005) in maximum temperature rise, focus volume and focus accuracy. Simulations predicted greater improvements than observed experimentally, suggesting potential for improvement in implementing the technique. The complete phase aberration correction procedure, including model generation, segmentation and phase aberration computations, required less than 45 min per sonication location.Conclusion: The significant improvements demonstrated in this study i.e., focus intensity, size and accuracy from phase aberration correction have the potential to improve the efficacy, time-efficiency and safety of breast MRgFUS therapies.

A spiral-based volumetric acquisition for MR temperature imaging.

To develop a rapid pulse sequence for volumetric MR thermometry. Simulations were carried out to assess temperature deviation, focal spot distortion/blurring, and focal spot shift across a range of readout durations and maximum temperatures for Cartesian, spiral-out, and retraced spiral-in/out (RIO) trajectories. The RIO trajectory was applied for stack-of-spirals 3D imaging on a real-time imaging platform and preliminary evaluation was carried out compared to a standard 2D sequence in vivo using a swine brain model, comparing maximum and mean temperatures measured between the two methods, as well as the temporal standard deviation measured by the two methods. In simulations, low-bandwidth Cartesian trajectories showed substantial shift of the focal spot, whereas both spiral trajectories showed no shift while maintaining focal spot geometry. In vivo, the 3D sequence achieved real-time 4D monitoring of thermometry, with an update time of 2.9-3.3 s. Spiral imaging, and RIO imaging in particular, is an effective way to speed up volumetric MR thermometry. Magn Reson Med 79:3122-3127, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

A methodology for thermal dose model parameter development using perioperative MRI

Post-treatment imaging is the principal method for evaluating thermal lesions following image-guided thermal ablation procedures. While real-time temperature feedback using magnetic resonance temperature imaging (MRTI) is a complementary tool that can be used to optimise lesion size throughout the procedure, a thermal dose model is needed to convert temperature–time histories to estimates of thermal damage. However, existing models rely on empirical parameters derived from laboratory experiments that are not direct indicators of post-treatment radiologic appearance. In this work, we investigate a technique that uses perioperative MR data to find novel thermal dose model parameters that are tailored to the appearance of the thermal lesion on post-treatment contrast-enhanced imaging. Perioperative MR data were analysed for five patients receiving magnetic resonance-guided laser-induced thermal therapy (MRgLITT) for brain metastases. The characteristic enhancing ring was manually segmented on post-treatment T1-weighted imaging and registered into the MRTI geometry. Post-treatment appearance was modelled using a coupled Arrhenius-logistic model and non-linear optimisation techniques were used to find the maximum-likelihood kinetic parameters and dose thresholds that characterise the inner and outer boundary of the enhancing ring. The parameter values and thresholds were consistent with previous investigations, while the average difference between the predicted and segmented boundaries was on the order of one pixel (1 mm). The areas predicted using the optimised model parameters were also within 1 mm of those predicted by clinically utilised dose models. This technique makes clinically acquired data available for investigating new thermal dose model parameters driven by clinically relevant endpoints.

Effect of applied energy in renal sympathetic denervation with magnetic resonance guided focused ultrasound in a porcine model

BackgroundPast catheter-based and focused ultrasound renal denervation studies have indicated that procedure efficacy is related to the number of ablations performed or the amount of energy used for the ablation. This study extends those prior results and investigates energy level effects on the efficacy of MR guided focused ultrasound renal denervation performed in a porcine model.MethodsTwenty-four normotensive pigs underwent unilateral denervation at three intensity levels. The applied intensity level was retrospectively de-rated to account for variability in animal size. Efficacy was assessed through evaluating the norepinephrine present in the kidney medulla and through histological analysis. The treatment was performed under MRI guidance including pre- and post-procedure T1-weighted and quantitative T1 and T2 imaging. During treatment, the temperature in the near field of the ultrasound beam was monitored in real time with MR temperature imaging. Energy delivery in the regions surrounding the renal artery was independently confirmed through an invasive fiberoptic temperature probe placed in the right renal artery.ResultsAnimals that underwent denervation at a de-rated acoustic intensity of greater than 1.2 kW/cm2 had a significantly lower norepinephrine concentration in the kidney indicating successful denervation. Images obtained during the treatment indicated no tissue changes in the kidneys as a function of the procedure but there were significant T1 changes present in the right lumbar muscles, although only one animal had indication of muscle damage at the time of necropsy.ConclusionsWhile MR guided focused ultrasound renal denervation was found to be safe and effective in this normotensive animal model, the results indicated the need to incorporate patient-specific details in the treatment planning of MRgFUS renal denervation procedure.

Referenceless magnetic resonance temperature imaging using Gaussian process modeling

During magnetic resonance (MR)-guided thermal therapies, water proton resonance frequency shift (PRFS)-based MR temperature imaging can quantitatively monitor tissue temperature changes. It is widely known that the PRFS technique is easily perturbed by tissue motion, tissue susceptibility changes, magnetic field drift, and modality-dependent applicator-induced artifacts. Here, a referenceless Gaussian process modeling (GPM)-based estimation of the PRFS is investigated as a methodology to mitigate unwanted background field changes. The GPM offers a complementary trade-off between data fitting and smoothing and allows prior information to be used. The end result being the GPM provides a full probabilistic prediction and an estimate of the uncertainty. GPM was employed to estimate the covariance between the spatial position and MR phase measurements. The mean and variance provided by the statistical model extrapolated background phase values from nonheated neighboring voxels used to train the model. MR phase predictions in the heating ROI are computed using the spatial coordinates as the test input. The method is demonstrated in ex vivo rabbit liver tissue during focused ultrasound heating with manually introduced perturbations (n = 6) and in vivo during laser-induced interstitial thermal therapy to treat the human brain (n = 1) and liver (n = 1). Temperature maps estimated using the GPM referenceless method demonstrated a RMS error of <0.8°C with artifact-induced reference-based MR thermometry during ex vivo heating using focused ultrasound. Nonheated surrounding areas were <0.5°C from the artifact-free MR measurements. The GPM referenceless MR temperature values and thermally damaged regions were within the 95% confidence interval during in vivo laser ablations. A new approach to estimation for referenceless PRFS temperature imaging is introduced that allows for an accurate probabilistic extrapolation of the background phase. The technique demonstrated reliable temperature estimates in the presence of the background phase changes and was demonstrated useful in the in vivo brain and liver ablation scenarios presented.

Spatially-segmented undersampled MRI temperature reconstruction for transcranial MR-guided focused ultrasound

BackgroundVolumetric thermometry with fine spatiotemporal resolution is desirable to monitor MR-guided focused ultrasound (MRgFUS) procedures in the brain, but requires some form of accelerated imaging. Accelerated MR temperature imaging methods have been developed that undersample k-space and leverage signal correlations over time to suppress the resulting undersampling artifacts. However, in transcranial MRgFUS treatments, the water bath surrounding the skull creates signal variations that do not follow those correlations, leading to temperature errors in the brain due to signal aliasing.MethodsTo eliminate temperature errors due to the water bath, a spatially-segmented iterative reconstruction method was developed. The method fits a k-space hybrid signal model to reconstruct temperature changes in the brain, and a conventional MR signal model in the water bath. It was evaluated using single-channel 2DFT Cartesian, golden angle radial, and spiral data from gel phantom heating, and in vivo 8-channel 2DFT data from a FUS thalamotomy. Water bath signal intensity in phantom heating images was scaled between 0-100% to investigate its effect on temperature error. Temperature reconstructions of retrospectively undersampled data were performed using the spatially-segmented method, and compared to conventional whole-image k-space hybrid (phantom) and SENSE (in vivo) reconstructions.ResultsAt 100% water bath signal intensity, 3 ×-undersampled spatially-segmented temperature reconstruction error was nearly 5-fold lower than the whole-image k-space hybrid method. Temperature root-mean square error in the hot spot was reduced on average by 27 × (2DFT), 5 × (radial), and 12 × (spiral) using the proposed method. It reduced in vivo error 2 × in the brain for all acceleration factors, and between 2 × and 3 × in the temperature hot spot for 2-4 × undersampling compared to SENSE.ConclusionsSeparate reconstruction of brain and water bath signals enables accelerated MR temperature imaging during MRgFUS procedures with low errors due to undersampling using Cartesian and non-Cartesian trajectories. The spatially-segmented method benefits from multiple coils, and reconstructs temperature with lower error compared to measurements from SENSE-reconstructed images. The acceleration can be applied to increase volumetric coverage and spatiotemporal resolution.

A printed Yagi–Uda antenna for application in magnetic resonance thermometry guided microwave hyperthermia applicators

Biological studies and clinical trials show that addition of hyperthermia stimulates conventional cancer treatment modalities and significantly improves treatment outcome. This supra-additive stimulation can be optimized by adaptive hyperthermia to counteract strong and dynamic thermoregulation. The only clinically proven method for the 3D non-invasive temperature monitoring required is by magnetic resonance (MR) temperature imaging, but the currently available set of MR compatible hyperthermia applicators lack the degree of heat control required. In this work, we present the design and validation of a high-frequency (433 MHz ISM band) printed circuit board antenna with a very low MR-footprint. This design is ideally suited for use in a range of hyperthermia applicator configurations. Experiments emulating the clinical situation show excellent matching properties of the antenna over a 7.2% bandwidth (S11 < −15 dB). Its strongly directional radiation properties minimize inter-element coupling for typical array configurations (S21 < −23 dB). MR imaging distortion by the antenna was found negligible and MR temperature imaging in a homogeneous muscle phantom was highly correlated with gold-standard probe measurements (root mean square error: RMSE = 0.51 °C and R2 = 0.99). This work paves the way for tailored MR imaging guided hyperthermia devices ranging from single antenna or incoherent antenna-arrays, to real-time adaptive hyperthermia with phased-arrays.

MR thermometry for focused ultrasound monitoring utilizing model predictive filtering and ultrasound beam modeling.

BackgroundA major challenge in using magnetic resonance temperature imaging (MRTI) to monitor focused ultrasound (FUS) applications is achieving high spatio-temporal resolution over a large field of view (FOV). This is important to accurately monitor all ultrasound (US) power depositions. Magnetic resonance (MR) subsampling in conjunction with thermal model-based reconstruction of the MRTI utilizing Pennes bioheat transfer equation (PBTE) is one promising approach. The thermal properties used in the thermal model are often estimated from a pre-treatment, low-power sonication.MethodsIn this proof-of-concept study we investigate the use of US simulations computed using the hybrid angular spectrum (HAS) method to estimate the US power deposition density Q, thereby avoiding the pre-treatment sonication and any potential tissue damage. MRTI reconstructions are performed using a thermal model-based reconstruction method called model predictive filtering (MPF). Experiments are performed in a homogeneous gelatin phantom and in a gelatin phantom with embedded plastic skull. MPF reconstructions are compared to separate sonications imaged with fully sampled data over a smaller FOV. Temperature root-mean-square errors (RMSE) and focal spot positions and shapes are evaluated.ResultsHAS simulations accurately predict the location of the focal spot (to within 1 mm) in both phantoms. Accurate temperature maps (RMSE below 1 °C), where the location of the focal spot agrees well with fully sampled “truth” (to within 1 mm), are also achieved in both phantoms.ConclusionsHAS simulations can be used to accurately predict the focal spot location in homogeneous media and when focusing through an aberrating plastic skull. The HAS simulated power deposition (Q) patterns can be used in the MPF thermal model-based reconstruction to obtain accurate temperature maps with high spatio-temporal resolution over large FOVs.

Development and validation of a MRgHIFU non-invasive tissue acoustic property estimation technique

MR-guided high-intensity focussed ultrasound (MRgHIFU) non-invasive ablative surgeries have advanced into clinical trials for treating many pathologies and cancers. A remaining challenge of these surgeries is accurately planning and monitoring tissue heating in the face of patient-specific and dynamic acoustic properties of tissues. Currently, non-invasive measurements of acoustic properties have not been implemented in MRgHIFU treatment planning and monitoring procedures. This methods-driven study presents a technique using MR temperature imaging (MRTI) during low-temperature HIFU sonications to non-invasively estimate sample-specific acoustic absorption and speed of sound values in tissue-mimicking phantoms. Using measured thermal properties, specific absorption rate (SAR) patterns are calculated from the MRTI data and compared to simulated SAR patterns iteratively generated via the Hybrid Angular Spectrum (HAS) method. Once the error between the simulated and measured patterns is minimised, the estimated acoustic property values are compared to the true phantom values obtained via an independent technique. The estimated values are then used to simulate temperature profiles in the phantoms, and compared to experimental temperature profiles. This study demonstrates that trends in acoustic absorption and speed of sound can be non-invasively estimated with average errors of 21% and 1%, respectively. Additionally, temperature predictions using the estimated properties on average match within 1.2 °C of the experimental peak temperature rises in the phantoms. The positive results achieved in tissue-mimicking phantoms presented in this study indicate that this technique may be extended to in vivo applications, improving HIFU sonication temperature rise predictions and treatment assessment.

Heat equation inversion framework for average SAR calculation from magnetic resonance thermal imaging.

Deposition of radiofrequency (RF) energy can be quantified via electric field or temperature change measurements. Magnetic resonance imaging has been used as a tool to measure three dimensional small temperature changes associated with RF radiation exposure. When duration of RF exposure is long, conversion from temperature change to specific absorption rate (SAR) is nontrivial due to prominent heat-diffusion and conduction effects. In this work, we demonstrated a method for calculation of SAR via an inversion of the heat equation including heat-diffusion and conduction effects. This method utilizes high-resolution three dimensional magnetic resonance temperature images and measured thermal properties of the phantom to achieve accurate calculation of SAR. Accuracy of the proposed method was analyzed with respect to operating frequency of a dipole antenna and parameters used in heat equation inversion. Bioelectromagnetics. 37:493-503, 2016. © 2016 Wiley Periodicals, Inc.

Precision Nanomedicine Using Dual PET and MR Temperature Imaging-Guided Photothermal Therapy.

We investigated the feasibility of PTT for ovarian cancer under the guidance of PET and MR temperature imaging using copper sulfide nanoparticles (CuS NPs). The tumor distribution of the CuS NPs after systemic administration was assessed using highly sensitive, quantifiable PET imaging. Two wavelengths of near-infrared (NIR) lasers-808 and 980 nm-were tested for PTT using noninvasive MR temperature imaging real-time monitoring. The in vivo studies revealed that the 980-nm NIR laser had better photothermal effects than the 808-nm NIR laser. These results were in accord with the histologic findings. In vivo PTT using CuS NPs combined with 980-nm laser irradiation achieved significant tumor ablation compared with no treatment control in both subcutaneous (P = 0.007) and orthotopic (P < 0.001) models of ovarian cancer with regard to the percentage of necrotic damage. Our results indicate that real-time monitoring of the accuracy of PTT is a promising approach for future clinical translation of this emerging thermal ablation technique.

SU‐F‐J‐02: Flexible Training of MR‐Guided Laser Ablation Models Via Global Optimization

Purpose:MR‐guided laser ablation is currying interest in neurosurgery as a minimally invasive procedure. The procedure is planned via an assumed lesion overlaid on various anatomic and functional images. Computational models of heat transfer may further inform planning the intervention, reducing surgical morbidity and ensuring therapeutic effect. These models require physical parameter data, via less relevant literature values. Using commercialized laser ablation equipment combined with MR temperature imaging (MRTI) feedback, recent work has demonstrated the use of inverse problems as a means to gain the necessary information via retrospective analysis. Here, performance of a global inverse problem solution is explored that allows for simultaneous optimization of multiple objective functions.Methods:A steady state bioheat model is trained on MRTI patient datasets (n=20); the MRTI regions exceeding a 57°C threshold are considered lethally damaged. Optical and blood perfusion parameters are densely sampled, allowing for global optimization of the model. There are two objective functions to optimize with respect to damage prediction: Dice similarity coefficient (DSC) and Hausdorff distance (HD). DSC is an overlap metric; HD is the distance‐to‐agreement in physical distance. Predictive performance is simulated via leave‐one‐out cross‐validation (LOOCV) and is compared to the same model using naive literature values.Results:Post‐optimization and during LOOCV, mean and median DSC are 0.824 and 0.850 respectively; mean and median HD are 4.64mm and 4.03mm respectively. These compare favorably to the model using literature values. Mean and median DSC are 0.6685 and 0.6564 respectively; mean and median HD are 5.85mm and 5.58mm respectively. A two‐tailed paired Student's t‐test for the mean DSC reported p=0.0002. For mean HD, p=0.0064.Conclusion:Clinical MRTI feedback provides useful information for designing optimized laser ablation models. Global optimization facilitates simultaneous optimization of multiple objective functions. Trained models are considerably more accurate than naive literature based models as measured by DSC and HD.

WE‐AB‐BRA‐04: Investigation of MRI Derived Thermal Dose Models

Purpose:To develop and investigate a novel thermal dose model methodology using MR temperature imaging (MRTI) and post‐treatment MRI changes as a surrogate of tissue damage in clinical laser ablation of the brain.Methods:Post‐treatment contrast enhanced T1‐weighted images were registered to MRTI acquired during two MR‐guided thermal ablation procedures. The non‐enhancing central region and hyperintense ring of vascular disruption characteristic of the thermal lesion were segmented manually and resampled into the 2D geometry of the MRTI. Regions immediately inside and outside the ring were classified as nonviable and viable tissue, respectively (N=1860). Logistic regression was performed on these classifications using thermal dose (Ω) calculated via four Arrhenius dose models from the literature (Henriques (H), Weaver &amp; Stoll (WS), Diller &amp; Klutke (DK), and Brown (B)). The optimal (O) Arrhenius dose model was found by iteratively changing the parameters of the Arrhenius model, Ea and A, until logistic model deviance was minimized.Results:Deviance of the Henriques model was closest to the optimized model (360 (H), 630 (WS), 438 (DK), 803(B), 330(O)). The thermal dose required to achieve a 90% probability of nonviability was 1.4 (H), 2.65 (WS), 1.8 (D), 4.5 (B), 1.1 (O) for each model. Agreement between the area of &gt;90% nonviability between each model and the optimal model was 0.97 (H), 0.89(WS), 0.94(DK), and 0.88(B) using the Dice Similarity Coefficient.Conclusion:A methodology for deriving new thermal dose models from clinically used surrogates of tissue damage has been developed and tested. The logistic model suggests that the common dose threshold of Ω=1 may overestimate tissue damage. Good agreement is observed between the nonviable regions predicted by the optimized model and literature models. However, agreement between models is expected to decrease for longer ablation procedures designed to create larger thermal lesions and is an opportunity for further study.

Fast temperature estimation from undersampled k-space with fully-sampled center for MR guided microwave ablation

PurposeThis study aims to accelerate MR temperature imaging using the proton resonance frequency (PRF) shift method for real time temperature monitoring during thermal ablation. Materials and methodsThe proposed method estimates temperature changes from undersampled k-space with a fully sampled center. This proposed algorithm is based on the hybrid multi-baseline and referenceless treatment image model and can be seen as an extension of the conventional k-space-based hybrid thermometry. The parameters of hybrid model are acquired by utilizing information from low resolution images which are obtained from fully-sampled centers of k-space. Registration is used to correct temperature errors due to the displacement of the subject. Phantom heating simulations, motion simulations, phantom heating and in-vivo experiments were performed to investigate the efficiency of the proposed method. SPIRiT and the conventional k-space estimation reconstruction thermometry were implemented for comparison using the same sampling pattern. ResultsThe phantom heating simulations showed that the proposed method results in lower RMSEs than the conventional k-space hybrid thermometry and SPIRiT at various reduction factors tested. The motion simulations indicated the robustness of the proposed method to displacement of the subject. Phantom heating experiment further demonstrated the ability of the method to reconstruct temperature maps with less computation time and higher accuracy (RMSEs lower than 0.4°C) at a net reduction factor of 3.5 in the presence of large noise caused by a microwave needle. In-vivo experiments validated the feasibility of the proposed method to estimate temperature changes from undersampled k-space (net reduction factor 4.3) in presence of respiratory motion and complicated anatomical structure, while reducing computation time as much as 10-fold compared with the conventional k-space method. ConclusionThe proposed method accelerates the PRF-shift MR thermometry and provides more accurate temperature maps in presence of motion with relatively short computation time, which may make real time imaging for MR-guided microwave ablation possible.

238 - Magnetic resonance temperature imaging in clinical hyperthermia: past experience and prospects

238 - Magnetic resonance temperature imaging in clinical hyperthermia: past experience and prospects

MR temperature imaging using PRF phase difference and a geometric model-based fat suppression method.

Because protons in fat do not exhibit a temperature-dependent frequency shift, proton resonance frequency shift (PRFS)-based MR thermometry always suffers from disturbances due to the presence of fats or lipids. A new fat suppression method for PRFS-based MR thermometry is proposed to obtain accurate variation of phase angle. Similar to the approach of separating fat and water with the two-point Dixon technique, we first scan a complex MR image for reference and then scan another complex image varying with temperature at the same TE point. Based on the conventional PRFS method, we use geometric relationships to remove the effect of fat on the variation of the phase angle. Two phantoms with different water-to-fat ratios are involved in the temperature mapping test. Experimental results show that the temperature images of two phantoms are approximated under the same conditions. The proposed fat suppression method is simple and effective for PRFS-based MR thermometry.

MO‐FG‐BRA‐09: Quantification of Nanoparticle Heating and Concentration for MR‐Guided Laser Interstitial Thermal Therapy

Purpose:Nanoparticle Mediated Laser Interstitial Thermal Therapy (npLITT) is a technique that utilizes tumor localized optically activated nanoparticles to increase the conformality of laser ablation procedures. Temperatures in these procedures are dependent on the particle concentration which generally cannot be measured noninvasively prior to therapy. In this work we attempt to quantify particle concentration in vivo by estimating the increase in R2* relaxation induced by bifunctional magnetic resonance (MR)‐visible gold‐based nanoparticles (SPIO@Au) and relate it to the temperature increase observed during real time MR temperature imaging (MRTI) of laser ablation.Methods:SPIO@Au nanoparticles (90nm) were synthesized containing a silica‐iron core (for MR visibility via R2*) and gold shell (for near‐infrared absorption). High resolution R2* maps were acquired before and after injecting four different particle concentrations (saline,1e10, 5e10, and 10e10 particles/mL) into HN5 flank xenografts. Tumors were monitored using MRTI during treatment with an interstitial fiber. (1 watt, 808 nm, 3 minutes)Results:The maximum temperature within the tumors increased linearly with concentration of injected particles, reaching 34.0, 37.6, 45.8, and 55.4 ⁰C for saline, 1e10, 5e10 and 10e10 particles/mL injections, respectively (R2=.994). The highest temperatures occur at the injection site rather than the fiber, confirming that SPIO@Au nanoparticles are the primary absorber. The differences between the median R2* measured at the injection site and the rest of the tumor were −6, 134, 111, 156 s‐1 for the saline,1e10,5e10 and 10e10 particles/mL injections, respectively. This R2* change is consistent with the measured relaxivity for the 1e10 particles/mL injection but does not maintain linearity at higher concentrations.Conclusion:Bifunctional SPIO@Au nanoparticles are a promising technology for providing noninvasive estimates of particle concentration via MRI and temperature increase in npLITT procedures. Future experiments will focus on lower, physiologically relevant particle concentrations and spin echo R2 mapping to better quantify the particle concentration.This research was supported by the National Institutes of Health and National Cancer Institute under Award Numbers TL1TR000369 and P30CA016672 and was conducted at the MD Anderson Center for Advanced Biomedical Imaging in‐part with equipment support from General Electric Healthcare.

SU‐C‐BRA‐03: Prediction of Laser Induced Thermal Therapy: Results of Model Training and Cross Validation

Purpose:MR‐guided laser induced thermal therapy (MRgLITT) is a minimally invasive surgery with applications in the brain, among other sites. In especially precise interventions, like neurosurgery, accurate planning may behoove surgical planning by aiding in the decision of where and how many laser ablations are required. Previous models of tissue heating have relied on literature values extrapolated primarily from normal brain animal research and ex vivo data. In this abstract, an inverse problem provides model parameter data from retrospective analysis of MR temperature imaging data in patient tumor tissue, which represent a training cohort. Within the same cohort, leave‐one‐out cross validation (LOOCV) estimates the predictive accuracy of the trained model.Methods:The training has three parts: MR temperature datasets (n=20), a relatively simple steady state bioheat model, and a global optimization algorithm maximizing the Dice similarity coefficient (DSC). DSC ranges from 0 to one; &gt;0.7 is considered a ‘successful’ prediction. DSC compares the regions exceeding 57 °C—i.e. ablated tissue regions—and measures accuracy. Blood perfusion and optical parameters are optimized according to DSC, creating a library of 20 pairs of parameters that are used in prediction. The predictive accuracy is estimated by using LOOCV. LOOCV begins by dropping an optimal parameter pair from the library and makes a model prediction, given the average from the remaining 19 pairs. This procedure is permuted so that every dataset is predicted using the other parameter pairs.Results:The distribution of DSC during predictive LOOCV is described by: mean=0.822; median=0.849; standard deviation=0.0872; minimum=0.583; maximum=0.930; 19/20 datasets pass (i.e. DSC&gt;0.7). Best available literature values perform comparably worse: mean=0.668; median=0.656; standard deviation=0.115; minimum=0.455; maximum=0.857; 8/20 pass.Conclusion:Data strongly indicate population‐based parameter optimization is potentially useful in treatment planning, as the model far outperforms available literature values in this preliminary study.Research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number TL1TR000369. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Magnetic resonance temperature imaging-based quantification of blood flow-related energy losses.

This study presents a new approach for evaluating bioheat transfer equation (BHTE) models used in treatment planning, control and evaluation of all thermal therapies. First, 3D magnetic resonance temperature imaging (MRTI) data are used to quantify blood flow-related energy losses, including the effects of perfusion and convection. Second, this information is used to calculate parameters of a BHTE model: in this paper the widely used Pennes BHTE. As a self-consistency check, the BHTE parameters are utilized to predict the temperatures from which they were initially derived. The approach is evaluated with finite-difference simulations and implemented experimentally with focused ultrasound heating of an ex vivo porcine kidney perfused at 0, 20 and 40 ml/min (n = 4 each). The simulation results demonstrate accurate quantification of blood flow-related energy losses, except in regions of sharp blood flow discontinuities, where the transitions are spatially smoothed. The smoothed transitions propagate into estimates of the Pennes perfusion parameter but have limited effect on the accuracy of temperature predictions using these estimates. Longer acquisition time periods mitigate the effects of MRTI noise, but worsen the effect of flow discontinuities. For the no-flow kidney experiments the estimates of a uniform, constant Pennes perfusion parameter are approximately zero, and at 20 and 40 ml/min the average estimates increase with flow rate to 3.0 and 4.2 kg/m(3) /s, respectively. When Pennes perfusion parameter values are allowed to vary spatially, but remain temporally constant, BHTE temperature predictions are more accurate than when using spatially uniform, constant Pennes perfusion values, with reductions in RMSE values of up to 79%. Locations with large estimated perfusion values correspond to high flow regions of the kidney observed in T1 -weighted MR images. This novel, MRTI-based technique holds promise for improving understanding of thermal therapy biophysics and for evaluating biothermal models.

Development of an endoluminal high-intensity ultrasound applicator for image-guided thermal therapy of pancreatic tumors.

An ultrasound applicator for endoluminal thermal therapy of pancreatic tumors has been introduced and evaluated through acoustic/biothermal simulations and ex vivo experimental investigations. Endoluminal therapeutic ultrasound constitutes a minimally invinvasive conformal therapy and is compatible with ultrasound or MR-based image guidance. The applicator would be placed in the stomach or duodenal lumen, and sonication would be performed through the luminal wall into the tumor, with concurrent water cooling of the wall tissue to prevent its thermal injury. A finite-element (FEM) 3D acoustic and biothermal model was implemented for theoretical analysis of the approach. Parametric studies over transducer geometries and frequencies revealed that operating frequencies within 1-3 MHz maximize penetration depth and lesion volume while sparing damage to the luminal wall. Patient-specific FEM models of pancreatic head tumors were generated and used to assess the feasibility of performing endoluminal ultrasound thermal ablation and hyperthermia of pancreatic tumors. Results indicated over 80% of the volume of small tumors (~2 cm diameter) within 35 mm of the duodenum could be safely ablated in under 30 minutes or elevated to hyperthermic temperatures at steady-state. Approximately 60% of a large tumor (~5 cm diameter) model could be safely ablated by considering multiple positions of the applicator along the length of the duodenum to increase coverage. Prototype applicators containing two 3.2 MHz planar transducers were fabricated and evaluated in ex vivo porcine carcass heating experiments under MR temperature imaging (MRTI) guidance. The applicator was positioned in the stomach adjacent to the pancreas, and sonications were performed for 10 min at 5 W/cm2 applied intensity. MRTI indicated over 40°C temperature rise in pancreatic tissue with heating penetration extending 3 cm from the luminal wall.