Realistic simulation of breast mass appearance using random walk

Abstract

The aim of the present work was to develop a method for simulating breast lesions in digital mammographic images. Based on the visual appearance of real masses, three dimensional masses were created using a 3D random walk method where the choice of parameters (number of walks and number of steps) enables one to control the appearance of the simulated structure. This work is the first occasion that the random walk results have been combined with a model of digital mammographic imaging systems. This model takes into account appropriate physical image acquisition processes representing a particular digital X-ray mammography system. The X-ray spectrum, local glandularity above the insertion site and scatter were all taken account during the insertion procedure. A preliminary observer study was used to validate the realism of the masses. Seven expert readers each viewed 60 full field mammograms and rated the realism of the masses they contained. Half of the images contained real, histologically-confirmed masses, and half contained simulated lesions. The ROC analysis of the study (average AUC of 0.58±0.06) suggests that, on the average, there is evidence that the radiologists could distinguish, somewhat, between real and simulated masses.

Simulation of 3D DLA masses in digital breast tomosynthesis

“…The depth of insertion was arbitrarily selected within the breast thickness. The simulated DLA mass was inserted in a 2D mammogram using the method developed by Rashidnasab et al 5,7 which used a degradation model of the imaging system, taking into account the polychromatic nature of the X-ray spectra, local glandularity and the scatter around the insertion site. The corresponding insertion location and local glandularity estimation were then calculated for the DBT 0° projection.…”

“…Simulated masses were inserted into 2D mammograms using the method developed by Rashidnasab et al 5,7 as shown in Figure 1(left). The primary photon flux was calculated at each pixel before and after insertion using polychromatic X-ray spectra model (block 3).…”

A Modelling Framework for Evaluation of 2D-Mammography and Breast Tomosynthesis Systems

“…An initial experiment was conducted whereby the masses [6][7] and microcalcifications [8] were inserted into the breast phantom [5] by replacing the tissue voxels with lesion/micro-calcification voxels at appropriate locations. One mammogram and 15 tomosynthesis primary projections were acquired using the ray tracing tool [9], Beer's law and attenuation coefficients [10] corresponding to the appropriate X-ray spectrum [11], at different angles in accord with the Hologic specification.…”

“…The method of insertion into 2D-mammogram was adopted from Rashidnasab et al [6][7], and the method of insertion into tomosynthesis projections was adopted from Shaheen et al [8]. A series of templates were created from the projections of the lesion or microcalcifications.…”

“…We used lesion simulation tool proposed by Rashidnasab et al [6][7], which uses fractal growth methods such as DLA (diffusion limited aggregation) and Random walk, to grow lesions in a 3D space. The model parameters can be varied to control the simulated mass structure and size.…”