Optimization of transcostal phased-array refocusing using iterative sparse semidefinite relaxation method

McMahon, Daniel; Almekkawy, Mohamed

doi:10.1109/spmb.2017.8257038

Cited by 9 publications

(1 citation statement)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PINNs have revolutionized the approach to solving nonlinear partial differential equations (PDEs) by leveraging deep learning to approximate solutions while adhering to physical laws, reducing computational demands significantly compared to traditional methods like finite difference method (FDM) and finite element method (FEM) [25,26,27,28,29,30,31,32,33,34]. This study utilizes a 3D model based on an anatomically realistic breast phantom (ARBP) generated from T1-weighted MRIs to reflect the breast's complex anatomy accurately, enhancing the precision and clinical relevance of HIFU simulations.…”

Section: Introductionmentioning

confidence: 99%

A Holistic Physics-Informed Neural Network Solution for Precise Destruction of Breast Tumors using High-Intensity Focused Ultrasound on a Realistic Breast Model

Lari,

Rajabzadeh,

Kohandel

et al. 2024

Preprint

View full text Add to dashboard Cite

This study presents a novel approach for the precise ablation of breast tumors using high-intensity focused ultrasound (HIFU), leveraging a physics-informed neural network (PINN) integrated with a realistic breast model. The potential of high-intensity focused ultrasound (HIFU) in eliminating breast tumors has shown significant promise. This technique employs concentrated ultrasonic waves to generate intense heat, effectively destroying cancerous tissue. In previous finite element method (FEM) models, the computational demands of handling extensive datasets, multiple dimensions, and discretization pose significant challenges. Our PINN-based solution operates efficiently in a mesh-free domain, achieving remarkable accuracy with significantly reduced computational demands, compared to conventional FEM techniques. Additionally, employing PINN for estimating partial differential equations (PDE) solutions can notably decrease the enormous number of discretized elements needed. The model employs a bowlshaped acoustic transducer to focus ultrasound waves accurately on the tumor location. The simulation results offer detailed insights into each step of the HIFU process. By applying a 3.8 nm displacement amplitude of transducer input pulse at a frequency of 1.1 MHz for 1 second at the focal point, the focal point temperature reaches 38.4 °C. Subsequently, continuing the ablation process for another 90 seconds without the acoustic source leads to extensive necrosis of the tumor tissues. Validation of the PINN model’s accuracy was conducted through FEM analysis and a series of tests using ex vivo bovine liver, aligning closely with real-world HIFU therapy scenarios. This innovative platform provides physicians with a predictive tool to estimate the necrosis of tumor tissue, facilitating the customization of HIFU treatment strategies for individual breast cancer patients.

show abstract