Embedded Files
Proposed Solution
Since typical MRI sequences are not suitable as a screening tool for high-risk ovarian patients, a solution would be to develop a new screening protocol that would have high in-plane and through-plane resolution, a reasonable acquisition time, and the ability to discriminate between benign and malignant lesions. In-plane resolution refers to the clarity of details within the imaging slice, while through-plane resolution refers to the clarity of details from one slice to another along the body. Our proposed solution that meets those requirements is high resolution imaging with high b-value DWI. The three main features of this protocol are 3D isotropic T2 Fast Spin Echo, multiband DWI, and high b-value DWI. Essentially, the voxel size, which is the three-dimensional equivalent of a pixel, and slice thickness are reduced to improve in-plane and through-plane resolution. DWI images are also taken at higher b-values with multiple slices excited simultaneously to enable greater differentiation between benign and malignant lesions, while reducing acquisition time.
Protocol Design
These are the imaging parameters developed by the Women's Imaging Laboratory (WIL) at UC San Diego Health to satisfy the proposed solution. Using these imaging parameters and a 3.0T (GE Healthcare, USA) MRI scanner with a 32-channel body coil, 14 patients were scanned. Among the 14 patients, 3 were high-risk BRCA 1 or 2 scheduled for prophylactic oopherectomy and 11 were suspected to have cancer based on prior imaging. Upon scanning, signal-to-noise (SNR) and contrast-to-noise (CNR) ratios were calculated according to the equations below:
Protocol Optimization: Noise Correction
Protocol Optimization: Noise Correction
A consequence of high b-values is decreased SNR and CNR. With greater noise, image quality and potentially confidence in image interpretation is reduced. To optimize the protocol and improve image quality, a noise correction method is necessary. A collaborator of the WIL developed a noise correction algorithm. We applied that algorithm to resulting DWI images and assessed its impact on apparent diffusion coefficient (ADC), a clinical biomarker.
Summary of Noise Correction Assessment
For each patient:
Generate T2 anatomical, no correction, correction, and noise map images
Generate figures of DWI at b-values of 0 - 3000
Generate ADC maps before and after noise correction to quantitatively assess the effect of noise on a clinical biomarker
Draw ROIs on normal ovary and lesion to get ADC value for each ADC map
Compare ADC values to expected ADC values in literature by calculating percentage error according to the equation below:
Page Author: Hillary Tran
Page updated
Report abuse