Evaluation of a variable dose acquisition technique for microcalcification and mass detection in digital breast tomosynthesis

Abstract

In this article the authors evaluate a recently proposed variable dose ͑VD͒-digital breast tomosynthesis ͑DBT͒ acquisition technique in terms of the detection accuracy for breast masses and microcalcification ͑MC͒ clusters. With this technique, approximately half of the total dose is used for one center projection and the remaining dose is split among the other tomosynthesis projection views. This acquisition method would yield both a projection view and a reconstruction view. One of the aims of this study was to evaluate whether the center projection alone of the VD acquisition can provide equal or superior MC detection in comparison to the 3D images from uniform dose ͑UD͒-DBT. Another aim was to compare the mass-detection capabilities of 3D reconstructions from VD-DBT and UD-DBT. In a localization receiver operating characteristic ͑LROC͒ observer study of MC detection, the authors compared the center projection of a VD acquisition scheme ͑at 2 mGy dose͒ with detector pixel size of 100 m with the UD-DBT reconstruction ͑at 4 mGy dose͒ obtained with a voxel size of 100 m. MCs with sizes of 150 and 180 m were used in the study, with each cluster consisting of seven MCs distributed randomly within a small volume. Reconstructed images in UD-DBT were obtained from a projection set that had a total of 4 mGy dose. The current study shows that for MC detection, using the center projection alone of VD acquisition scheme performs worse with area under the LROC curve ͑A L ͒ of 0.76 than when using the 3D reconstructed image using the UD acquisition scheme ͑A L = 0.84͒. A 2D ANOVA found a statistically significant difference ͑p = 0.038͒ at a significance level of 0.05. In the current study, although a reconstructed image was also available using the VD acquisition scheme, it was not used to assist the MC detection task which was done using the center projection alone. In the case of evaluation of detection accuracy of masses, the reconstruction with VD-DBT ͑A L = 0.71͒ was compared to that obtained from the UD-DBT ͑A L = 0.78͒. The authors found no statistically significant difference between the two ͑p-value= 0.22͒, although all the observers performed better for UD-DBT.

The TOMMY trial: a comparison of TOMosynthesis with digital MammographY in the UK NHS Breast Screening Programme – a multicentre retrospective reading study comparing the diagnostic performance of digital breast tomosynthesis and digital mammography with digital mammography alone

“…These observations suggested that DBT was unlikely to be used as a stand-alone imaging modality if a 2D mammogram was required for optimal microcalcification assessment. 25,52,53 Cancer detection and recall rate…”

Digital mammography and tomosynthesis for breast cancer diagnosis

“…Dose levels between 1.6 mGy [53] and 4 mGy [54] have been reported. The total tube loading (mAs value) is often divided equally among the projection images [55], but recently the effect of spending a larger fraction of the total dose on the center projection in order to increase detection of microcalcifications have been investigated [56][57][58]. Spangler et al [59] did not find any difference in area under the receiver operating characteristic (ROC) curve [60] for BI-RADS (breast imaging reporting and data system) scores of calcifications for DM and BT, although they found a higher sensitivity for calcification detection.…”

Patchwork reconstruction with resolution modeling for digital breast tomosynthesis

“…The contributions of the homogeneous parts of the volume are then placed outside of the smoothing kernel in equation (5), with b sθ the total blank value for angle θ and w θϕ the normalized waveform of the exposure at angle θ. The key approximation is then applied in (6) where the sum over the sparsely sampled angles with kernel w θϕ is replaced by a sum over the densely sampled detector pixels with kernel w P nξ , which is combined with kernel A sn to form A P sn in (7). At this point we can assume that there are more high contrast lesions in the volume, but that they don't overlap in the projections.…”

“…One path to improve visualization of microcalcifications is to optimize the dose and angular distribution of the limited angle set of projections [7][8][9]. These projections are then usually reconstructed with filtered backprojection (FBP) although this reconstruction by FBP is not always optimal [10,11] because of the limited angular range and low dose acquisitions, * koen.michielsen@uzleuven.be especially for small angular range [12].…”