Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Optimizing Synthetic Correlated Diffusion Imaging for Breast Cancer Tumour Delineation

Chi-en Amy Tai, Alexander Wong

TL;DR

The study addresses the need for accurate breast tumour delineation by optimizing synthetic correlated diffusion imaging (CDI$^s$) exponents for task-specific performance. It applies Nelder-Mead optimization to tune the mixing coefficients $\rho$ and synthetic signal presets $\hat{S}$ on the ACRIN 6698/I-SPY2 diffusion MRI dataset, aiming to maximize AUC. The Optimized CDI$^s$ modality achieves the highest AUC, surpassing the best gold-standard by $0.0044$ and exceeding the Unoptimized CDI$^s$ by more than $0.02$, underscoring the value of dataset- and task-specific parameter optimization. However, results are contingent on dataset characteristics and masking methods, with limitations including lack of radiologist-verified masks and absence of T2w-AUC evaluation, highlighting the need for broader validation.

Abstract

Breast cancer is a significant cause of death from cancer in women globally, highlighting the need for improved diagnostic imaging to enhance patient outcomes. Accurate tumour identification is essential for diagnosis, treatment, and monitoring, emphasizing the importance of advanced imaging technologies that provide detailed views of tumour characteristics and disease. Synthetic correlated diffusion imaging (CDI$^s$) is a recent method that has shown promise for prostate cancer delineation compared to current MRI images. In this paper, we explore tuning the coefficients in the computation of CDI$^s$ for breast cancer tumour delineation by maximizing the area under the receiver operating characteristic curve (AUC) using a Nelder-Mead simplex optimization strategy. We show that the best AUC is achieved by the CDI$^s$ - Optimized modality, outperforming the best gold-standard modality by 0.0044. Notably, the optimized CDI$^s$ modality also achieves AUC values over 0.02 higher than the Unoptimized CDI$^s$ value, demonstrating the importance of optimizing the CDI$^s$ exponents for the specific cancer application.

Optimizing Synthetic Correlated Diffusion Imaging for Breast Cancer Tumour Delineation

TL;DR

The study addresses the need for accurate breast tumour delineation by optimizing synthetic correlated diffusion imaging (CDI) exponents for task-specific performance. It applies Nelder-Mead optimization to tune the mixing coefficients and synthetic signal presets on the ACRIN 6698/I-SPY2 diffusion MRI dataset, aiming to maximize AUC. The Optimized CDI modality achieves the highest AUC, surpassing the best gold-standard by and exceeding the Unoptimized CDI by more than , underscoring the value of dataset- and task-specific parameter optimization. However, results are contingent on dataset characteristics and masking methods, with limitations including lack of radiologist-verified masks and absence of T2w-AUC evaluation, highlighting the need for broader validation.

Abstract

Breast cancer is a significant cause of death from cancer in women globally, highlighting the need for improved diagnostic imaging to enhance patient outcomes. Accurate tumour identification is essential for diagnosis, treatment, and monitoring, emphasizing the importance of advanced imaging technologies that provide detailed views of tumour characteristics and disease. Synthetic correlated diffusion imaging (CDI) is a recent method that has shown promise for prostate cancer delineation compared to current MRI images. In this paper, we explore tuning the coefficients in the computation of CDI for breast cancer tumour delineation by maximizing the area under the receiver operating characteristic curve (AUC) using a Nelder-Mead simplex optimization strategy. We show that the best AUC is achieved by the CDI - Optimized modality, outperforming the best gold-standard modality by 0.0044. Notably, the optimized CDI modality also achieves AUC values over 0.02 higher than the Unoptimized CDI value, demonstrating the importance of optimizing the CDI exponents for the specific cancer application.
Paper Structure (3 sections, 2 figures, 1 table)

This paper contains 3 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methodology
  3. Results

Figures (2)

  • Figure 1: Sample breast mask from the processed DWI image.
  • Figure 2: Visual comparison of the tumour mask, ADC, DWI, ADCc, Unoptimized CDIs, and Optimized CDIs.