Diffusion MRI

Diffusion MRI-based imaging biomarkers for monitoring therapy response in patients with breast cancer metastasized to the liver

The apparent diffusion coefficient (ADC) value of tumors estimated using diffusion MRI is emerging as an important biomarker for monitoring cancer response. In collaboration with the Arizona Cancer Center and working in an interdisciplinary research team, we investigated the problem of estimating the ADC of lesions in diffusion-weighted images of patients with breast cancer that had metastasized to the liver. We developed novel statistical methods to segment lesions in DW images and estimate a single ADC value of the lesion. The ADC estimation procedure we designed was more accurate than state-of-the-art ADC estimation procedures.  Further, we also designed methods to evaluate segmentation algorithms in diffusion MRI in the absence of ground truth segmentations. Dr. Jha’s paper on this topic was awarded the Best Student Paper award at SPIE Medical Imaging 2010. The value of the methods that we developed were assessed using a prospective clinical study. It was observed that the application of these methods was important for diffusion MRI to serve as a restrictive predictive biomarker of the therapeutic response of liver metastasis. The objective assessment of image quality (OAIQ)-based approach we used to develop these methods is summarized in figure below.

Fig. 1: Our OAIQ-based approach to measure ADC from diffusion MR images

References:

  • A. K. Jha, M. A. Kupinski, J. J. Rodriguez, R. M. Stephen and A. T. Stopeck, “Task-based evaluation of segmentation algorithms for diffusion-weighted MRI without using a gold-standard”, Phys. Med. Biol., 57(13), 4425-46, 2012. PMCID: PMC3932666
  • A. K. Jha, J. J. Rodriguez and A. T. Stopeck, “A maximum-likelihood method to estimate a single ADC value of lesions using diffusion MRI, Mag Res Med, 2016. PMCID: PMC4937834
  • A. K. Jha, J. J. Rodriguez, R. M. Stephen and A. T. Stopeck, “A clustering algorithm for liver lesion segmentation of diffusion-weighted MR images”, Proc. IEEE Southwest Symp. Image Anal. Interpret., pp. 93-6, 2010. PMCID: PMC2998770
  • R. M. Stephen, A. K. Jha, D. J. Roe et al., Diffusion MRI with Semi-Automated Segmentation Can Serve as a Restricted Predictive Biomarker of the Therapeutic Response of Liver Metastasis, Magn. Reson. Imaging, 2015, PMCID: PMC4658325