Publication detail

Accelerated Design of HIFU Treatment Plans Using Island-Based Evolutionary Strategy

KUKLIŠ, F. JAROŠ, M. JAROŠ, J.

Original Title

Accelerated Design of HIFU Treatment Plans Using Island-Based Evolutionary Strategy

Type

conference paper

Language

English

Original Abstract

High Intensity Focused Ultrasound (HIFU) is an emerging technique for non-invasive cancer treatment where malignant tissue is destroyed by thermal ablation. Such a treatment consists of a series of short sonications destroying small volumes of tissue. High-quality treatment plans allow to precisely target malignant tissue and protect surrounding healthy tissue. Recently, we developed an evolutionary strategy to design such HIFU treatment plans using patient-specific material properties and a realistic thermal model. Unfortunately, the execution time was prohibitive for routine use. Here, we present an optimized evolutionary strategy based on island model parallelization and a revised fitness function implementation. The proposed improvements allow to develop a good treatment plan 4 times faster and with 5% higher success rate.

Keywords

Evolutionary strategy, Island model, HIFU, Treatment planning, k-Wave toolbox 

Authors

KUKLIŠ, F.; JAROŠ, M.; JAROŠ, J.

Released

9. 4. 2020

Publisher

Springer International Publishing

Location

Cham

ISBN

978-3-030-43721-3

Book

Applications of Evolutionary Computation

Edition

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Pages from

463

Pages to

478

Pages count

16

URL

BibTex

@inproceedings{BUT162282,
  author="Filip {Kukliš} and Marta {Jaroš} and Jiří {Jaroš}",
  title="Accelerated Design of HIFU Treatment Plans Using Island-Based Evolutionary Strategy",
  booktitle="Applications of Evolutionary Computation",
  year="2020",
  series="Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
  volume="12104",
  pages="463--478",
  publisher="Springer International Publishing",
  address="Cham",
  doi="10.1007/978-3-030-43722-0\{_}30",
  isbn="978-3-030-43721-3",
  url="https://link.springer.com/chapter/10.1007/978-3-030-43722-0_30"
}