Abstract
Ultrasound (US)-guided diffuse optical tomography (DOT) has demonstrated potential value for breast cancer diagnosis and treatment response assessment. However, in clinical use, the chest wall, poor probe-tissue contact, and tissue heterogeneity can all cause image artifacts. These image artifacts, appearing commonly as hot spots in the non-lesion regions (edge artifacts), can decrease the reconstruction accuracy and cause misinterpretation of lesion images. Here we introduce an iterative, connected component analysis-based image artifact reduction algorithm. A convolutional neural network (CNN) is used to segment co-registered US images to extract the lesion location and size to guide the artifact reduction. We demonstrate its performance using Monte Carlo simulations on VICTRE digital breast phantoms and breast patient images. In simulated tissue mismatch models, this algorithm successfully reduces edge artifacts without significantly changing the reconstructed target absorption coefficients. With clinical data it improves the optical contrast between malignant and benign groups, from 1.55 without artifact reduction to 1.91 with artifact reduction. The proposed algorithm has a broad range of applications in other modality-guided DOT imaging.
Highlights
Near-infrared diffuse optical tomography (DOT) has demonstrated its great potential in breast cancer diagnosis and treatment response assessment [1,2,3,4,5,6]
We adopt VICTRE breast phantoms in DOT for the first time to account for tissue heterogeneity, and we demonstrate the effectiveness of the artifact reduction algorithm
We introduced an iterative edge artifact reduction algorithm based on connected component analysis
Summary
Near-infrared diffuse optical tomography (DOT) has demonstrated its great potential in breast cancer diagnosis and treatment response assessment [1,2,3,4,5,6]. Perturbation is sensitive to probe positions due to tissue heterogeneity differences (tissue compositions, and chest wall positions) underneath the probe, and to poor probe-tissue contact, which can generate image artifacts, appearing commonly as hot spots in the non-lesion regions (edge artifacts). Reducing these artifacts can more accurately quantify breast lesions and improve diagnosis. Deng et al extracted realistic digital breast phantoms from dual-energy X-ray mammographic imaging of human breasts and used them in DOT in-silico tests [30,31]. We adopt VICTRE breast phantoms in DOT for the first time to account for tissue heterogeneity, and we demonstrate the effectiveness of the artifact reduction algorithm. The algorithm, combined with an interactive convolutional neural network (CNN) that segments coregistered US images to extract the lesions, is evaluated with clinical data that includes 14 patients of seven benign and seven malignant lesions
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have