Introduction:: Breast cancer is the most common cancer in women worldwide. The conventional dynamic contrast enhanced (DCE) MRI has the highest sensitivity for cancer detection among all breast imaging modalities. However, it is limited by moderate specificity resulting in many unnecessary biopsies. Diffusion weighted MRI (DWI) has emerged as a promising adjunct to DCE-MRI to reduce false positives. DWI measures the diffusion of water molecules in tissue using motion sensitizing gradients and provides quantitative information on cellularity and microstructure. The apparent diffusion coefficient (ADC) derived from DWI reflects cellular density and has demonstrated usefulness for distinguishing benign and malignant lesions [1]. Approaches to measure lesion ADC values vary and emphasize different aspects of tumor microenvironment and biology. Choice of diffusion sensitization (or b-value) affects contributions from microcirculation and restriction, and region-of-interest (ROI) sampling methods capture different aspects of the lesion e.g., whole volume of the lesion for describing heterogeneity or ‘hotspot’ for peak cellularity). Here, we explored different approaches for measuring lesion ADC and potential effect on reducing false positives and unnecessary biopsies of breast MRI.


Materials and Methods:: In this IRB approved study, consecutive breast screening MRI examinations (August 2014 - January 2019) that had a suspicious breast lesion followed by definitive biopsy outcome were retrospectively evaluated. Generation of ADC maps and segmentation of lesion region of interest (ROI) were performed using custom software developed in MATLAB (MathWorks, Natick, MA). ADC maps were calculated by the standard mono exponential decay model and linear least squares fitting of signal intensities at different b value [2]. Guided by the radiologist, lesion ROIs were segmented on b=800 s/mm² images blinded to biopsy outcomes, using a semi-automated threshold-based tool to avoid fat and fibroglandular tissue. Nine ADC measures were computed for each lesion using three methods for generating ADC maps and three segmentation techniques (ref). ADC maps were generated using different b value combinations: b=0/800 s/mm² (ADC_0-800), b=100/800 s/mm² (ADC_100-800) images and b =0/100/600/800 s/mm² (ADC_all). Lesion ROIs were defined for a single representative slice (‘2D’), whole tumor volume (‘3D’) and darkest region with lowest ADC values (‘hotspot’) [3]. Diagnostic performance was evaluated by calculating area under the receiver operating characteristic curves (AUC). ADC differences between benign and malignant lesions were assessed by Wilcoxon rank sum test. The highest malignant ADC value by each method was identified to determine the optimal ADC threshold to reduce benign biopsies without reducing sensitivity, where lesions with ADC values above this threshold could hypothetically be considered benign without biopsy. [4]


Results, Conclusions, and Discussions:: Over the study period, 142 suspicious lesions (in 121 women, median age 44 years [range, 20-75yrs]) were detected on DCE-MRI and recommended for biopsy. Of those, 31 (21.8 %) were malignant and 111 (78.2 %) were benign. Lesion ADC characterization for representative benign and malignant breast lesions are shown in Fig 1. Mean ADC values of benign lesions were significantly higher than malignancies in all nine measurement approaches (Table 1). The area under the ROC curves for predicting malignancy ranged 0.63–0.69 (95% CI range, 0.55-0.75) (Fig 2). The optimal ADC cut-off identified for each of the ADC measurement methods ranged from 1.38-1.74 (Fig 3). Applying optimal cutoff values resulted in a maximum reduction of benign biopsies of nearly 20% (21/111) using ADC_0-800 and 2D ROI (cutoff = 1.53x10^-3 mm²/s). Hotspot lesion ADC values resulted in biopsy reduction very close to the 2D approach for all ADC maps. The hotspot ROI approach on ADC_100-800 map had the highest AUC (0.69) for predicting malignancy and the resulting cut-off could potentially reduce 18% (20/111) benign biopsies (Table 2). The other alternate ADC calculation techniques provided 7-17.11% reduction in benign biopsies.

Our findings demonstrate that lesion ADC measurements computed from a variety of approaches can help reduce unnecessary biopsies of breast MRI. The approach using mean ADC calculated with b values 0 and 800 s/mm² for a representative slice (2D) had the highest biopsy reduction rate. Measuring the hotspot or the darkest region of the lesion could be easier than defining an ROI contour of the lesion on an ADC map for a radiologist in the clinic without access to custom segmentation tools, which might facilitate integration of DW imaging into the clinical workflow to reduce unnecessary biopsies.


Acknowledgements (Optional): : We would like to thank our funding sources NIH/NCI Grant R01 CA207290

References (Optional): :

1. Partridge S.C., Diffusion-weighted breast MRI: Clinical applications and emerging techniques J. Magn. Reson. Imaging 2017;45 (2)

2. Stejskal E.O. and Tanner J.E., Spin Diffusion Measurements: Spin Echoes in the presence of a time dependent field gradient J. Chem Physics 42 (288)

3. McDonald E.E., et al Mean Apparent Diffusion Coefficient Is a Sufficient Conventional Diffusion-weighted MRI Metric to Improve Breast MRI Diagnostic Performance: Results from the ECOG-ACRIN Cancer Research Group A6702 Diffusion Imaging Trial, Radiology 2021 Jan;298(1)

4. Rahbar H et al Utility of Diffusion-weighted Imaging to Decrease Unnecessary Biopsies Prompted by Breast MRI: A Trial of the ECOG-ACRIN Cancer Research Group (A6702) Clinical Cancer Research 2019 Mar; 25(6)