Automated data selection method to improve robustness of diffuse optical tomography for breast cancer imaging

Abstract

Imaging-guided near infrared diffuse optical tomography (DOT) has demonstrated a great potential as an adjunct modality for differentiation of malignant and benign breast lesions and for monitoring treatment response of breast cancers. However, diffused light measurements are sensitive to artifacts caused by outliers and errors in measurements due to probe-tissue coupling, patient and probe motions, and tissue heterogeneity. In general, preprocessing of the measurements is needed by experienced users to manually remove these outliers and therefore reduce imaging artifacts. An automated method of outlier removal, data selection, and filtering for diffuse optical tomography is introduced in this manuscript. This method consists of multiple steps to first combine several data sets collected from the same patient at contralateral normal breast and form a single robust reference data set using statistical tests and linear fitting of the measurements. The second step improves the perturbation measurements by filtering out outliers from the lesion site measurements using model based analysis. The results of 20 malignant and benign cases show similar performance between manual data processing and automated processing and improvement in tissue characterization of malignant to benign ratio by about 27%. Czerniecki, "Breast cancer detection based on incremental biochemical and physiological properties of breast cancers: a six-year, two-site study," Acad. Radiol. 12(8), 925-933 (2005). 9.

Correlations Among Ultrasound‐Guided Diffuse Optical Tomography, Microvessel Density, and Breast Cancer Prognosis

“…Although DOT is easy to perform, and consistency between the observers in total Hb measurements is high, 28 there are still some factors that can influence DOT measurements, such as the size, location, and depth of lesions, partial volume effects, relative movement between the transduce and breast, and heterogeneous tissue. 23,31 These may have been the causes of the outlier measurements present in the control group seen in our box plot. However, the correlation coefficient between microvessel density and total Hb was low, which may have been due to the small number of cases.…”

Mathematical and numerical challenges in optical screening of female breast

“…Remark The signals and U 0 are typically affected by noise and appropriate filtering and denoising operations must be carried out on it. In the present paper, we do not deal with this important aspect and we refer for example to Vavadi and Zhu for a discussion of useful signal preprocessing and filtering techniques in the DOT context.…”

Compact ultrasound-guided diffuse optical tomography system for breast cancer imaging

“…The third module is imaging reconstruction, which incorporates our recently developed outlier removal and data selection method before reconstruction to eliminate the need for timeconsuming data preprocessing. 27 It also includes a semiautomated method to select the region of interest (ROI) from coregistered US images and then uses the ROI for DOT image reconstruction. 28 Briefly, the overall method performs outlier removal, data selection, and data-filtering processes automated for US-guided DOT.…”

“…And finally, a filtering method was used to remove the outliers from the perturbation measurements using a model-based analysis. 27 Imaging reconstruction was performed after data preprocessing and selection of ROI from coregistered US. The reconstruction used our recently developed two-step image reconstruction method, which have shown improved reconstruction accuracy and speed compared to the previously used conjugate gradient method in US-guided DOT reconstruction.…”