Unlocking the rNOE-CEST Contrast for Breast Cancer MRI Using the FATLESS Approach.

OBJECTIVES

Quantitative chemical exchange saturation transfer (CEST) breast imaging is limited by pronounced fat-induced artifacts. The strongest fat artifact, appearing between [-2, -4] ppm in the Z-spectrum, directly overlaps the signal of the exchange-relayed nuclear Overhauser effect (rNOE) at around -3.5 ppm, a key biomarker for protein content and cellularity, making accurate rNOE-CEST evaluation extremely challenging. The aim of this study is to evaluate rNOE-CEST contrast corrected for fat-related artifacts using a novel, fully software-based fat correction method in breast cancer patients.

MATERIALS AND METHODS

FATLESS (Fat Attenuation Technique using Lipid signal Estimation and Simulated Saturations in postprocessing) was developed for correcting fat-related artifacts across the entire Z-spectrum in CEST MRI. The FATLESS method estimates fat signals from residual signals at the direct water saturation offset (0 ppm) while accounting for partial saturation of fat resonances. FATLESS was retrospectively applied to 7T CEST data from breast cancer patients (acquired September 2018 to May 2019). Resulting fat-corrected rNOE, amide, and guanidino MTRRex contrast values were quantified from 2D snapshot GRE CEST with low saturation power (B1=0.6, 0.9 μT). Kruskal-Wallis tests and Pearson correlation analyses were used to compare MTRRex values between tumor and normal-appearing fibroglandular tissue and assess correlations with Ki-67, a tumor proliferation marker.

RESULTS

Nine biopsy-confirmed breast cancer patients [mean age, 50 y ± 10 (SD)] and 7 healthy controls [mean age, 25 y ± 4 (SD)] were included. Fat-corrected MTRRex rNOE maps were validated in phantom and in vivo data, confirming independence from fat artifacts using the FATLESS method. Tumor regions showed significantly higher fat-corrected MTRRex rNOE values than healthy tissue (+140% mean increase, P<0.001). A strong positive correlation was found between fat-corrected MTRRex rNOE values and Ki-67 (R² = 0.71).

CONCLUSIONS

The developed FATLESS fat correction method enables full utilization of all CEST MRI contrasts in the human breast. The observed significant rNOE contrast elevation and strong correlation with tumor proliferation highlight its potential as a non-invasive imaging biomarker for breast cancer characterization.