Purpose To evaluate the early implementation of synthesized two-dimensional (s2D) mammography in a population screened entirely with s2D and digital breast tomosynthesis (DBT) (referred to as ...s2D/DBT) and compare recall rates and cancer detection rates to historic outcomes of digital mammography combined with DBT (referred to as digital mammography/DBT) screening. Materials and Methods This was an institutional review board-approved and HIPAA-compliant retrospective interpretation of prospectively acquired data with waiver of informed consent. Compared were recall rates, biopsy rates, cancer detection rates, and radiation dose for 15 571 women screened with digital mammography/DBT from October 1, 2011, to February 28, 2013, and 5366 women screened with s2D/DBT from January 7, 2015, to June 30, 2015. Two-sample z tests of equal proportions were used to determine statistical significance. Results Recall rate for s2D/DBT versus digital mammography/DBT was 7.1% versus 8.8%, respectively (P < .001). Biopsy rate for s2D/DBT versus digital mammography/DBT decreased (1.3% vs 2.0%, respectively; P = .001). There was no significant difference in cancer detection rate for s2D/DBT versus digital mammography/DBT (5.03 of 1000 vs 5.45 of 1000, respectively; P = .72). The average glandular dose was 39% lower in s2D/DBT versus digital mammography/DBT (4.88 mGy vs 7.97 mGy, respectively; P < .001). Conclusion Screening with s2D/DBT in a large urban practice resulted in similar outcomes compared with digital mammography/DBT imaging. Screening with s2D/DBT allowed for the benefits of DBT with a decrease in radiation dose compared with digital mammography/DBT.
RSNA, 2016 An earlier incorrect version of this article appeared online. This article was corrected on August 11, 2016.
Advances in 3D imaging technology are transforming how radiologists search for cancer1,2 and how security officers scrutinize baggage for dangerous objects.3 These new 3D technologies often improve ...search over 2D images4,5 but vastly increase the image data. Here, we investigate 3D search for targets of various sizes in filtered noise and digital breast phantoms. For a Bayesian ideal observer optimally processing the filtered noise and a convolutional neural network processing the digital breast phantoms, search with 3D image stacks increases target information and improves accuracy over search with 2D images. In contrast, 3D search by humans leads to high miss rates for small targets easily detected in 2D search, but not for larger targets more visible in the visual periphery. Analyses of human eye movements, perceptual judgments, and a computational model with a foveated visual system suggest that human errors can be explained by interaction among a target’s peripheral visibility, eye movement under-exploration of the 3D images, and a perceived overestimation of the explored area. Instructing observers to extend the search reduces 75% of the small target misses without increasing false positives. Results with twelve radiologists confirm that even medical professionals reading realistic breast phantoms have high miss rates for small targets in 3D search. Thus, under-exploration represents a fundamental limitation to the efficacy with which humans search in 3D image stacks and miss targets with these prevalent image technologies.
Display omitted
•Humans searching 3D image stacks miss small targets that are salient in 2D images•Optimal observers and deep neural networks do not show the deficits for 3D images•Human eye movement under-exploration in 3D search explains the misses•New 3D-imaging technologies should consider these human visuo-cognitive bottlenecks
Will 3D imaging technologies always lead to improvements for the visual search of targets? Lago et al. show that, when humans search 3D image stacks, they under-explore with eye movements, overestimate the area they have searched, and often miss small targets that are salient in 2D images.
While handcrafted radiomic features (HRFs) have shown promise in the field of personalized medicine, many hurdles hinder its incorporation into clinical practice, including but not limited to their ...sensitivity to differences in acquisition and reconstruction parameters. In this study, we evaluated the effects of differences in in-plane spatial resolution (IPR) on HRFs, using a phantom dataset (n = 14) acquired on two scanner models. Furthermore, we assessed the effects of interpolation methods (IMs), the choice of a new unified in-plane resolution (NUIR), and ComBat harmonization on the reproducibility of HRFs. The reproducibility of HRFs was significantly affected by variations in IPR, with pairwise concordant HRFs, as measured by the concordance correlation coefficient (CCC), ranging from 42% to 95%. The number of concordant HRFs (CCC > 0.9) after resampling varied depending on (i) the scanner model, (ii) the IM, and (iii) the NUIR. The number of concordant HRFs after ComBat harmonization depended on the variations between the batches harmonized. The majority of IMs resulted in a higher number of concordant HRFs compared to ComBat harmonization, and the combination of IMs and ComBat harmonization did not yield a significant benefit. Our developed framework can be used to assess the reproducibility and harmonizability of RFs.
The reproducibility of handcrafted radiomic features (HRFs) has been reported to be affected by variations in imaging parameters, which significantly affect the generalizability of developed ...signatures and translation to clinical practice. However, the collective effect of the variations in imaging parameters on the reproducibility of HRFs remains unclear, with no objective measure to assess it in the absence of reproducibility analysis. We assessed these effects of variations in a large number of scenarios and developed the first quantitative score to assess the reproducibility of CT-based HRFs without the need for phantom or reproducibility studies. We further assessed the potential of image resampling and ComBat harmonization for removing these effects. Our findings suggest a need for radiomics-specific harmonization methods. Our developed score should be considered as a first attempt to introduce comprehensive metrics to quantify the reproducibility of CT-based handcrafted radiomic features. More research is warranted to demonstrate its validity in clinical contexts and to further improve it, possibly by the incorporation of more realistic situations, which better reflect real patients' situations.
Digital breast tomosynthesis (DBT) reconstructions introduce out-of-plane artifacts and false-tissue boundaries impacting the dense/adipose and breast outline (convex hull) segmentations. A virtual ...clinical trial method was proposed to segment both the breast tissues and the breast outline in DBT reconstructions. The DBT images of a representative population were simulated using three acquisition geometries: a left-right scan (conventional, I), a two-directional scan in the shape of a "T" (II), and an extra-wide range (XWR, III) left-right scan at a six-times higher dose than I. The nnU-Net was modified including two losses for segmentation: (1) tissues and (2) breast outline. The impact of loss (1) and the combination of loss (1) and (2) was evaluated using models trained with data simulating geometry I. The impact of the geometry was evaluated using the combined loss (1&2). The loss (1&2) improved the convex hull estimates, resolving 22.2% of the false classification of air voxels. Geometry II was superior to I and III, resolving 99.1% and 96.8% of the false classification of air voxels. Geometry III (Dice = (0.98, 0.94)) was superior to I (0.92, 0.78) and II (0.93, 0.74) for the tissue segmentation (adipose, dense, respectively). Thus, the loss (1&2) provided better segmentation, and geometries T and XWR improved the dense/adipose and breast outline segmentations relative to the conventional scan.
In breast tomosynthesis, multiple low-dose projections are acquired in a single scanning direction over a limited angular range to produce cross-sectional planes through the breast for ...three-dimensional imaging interpretation. We built a next-generation tomosynthesis system capable of multidirectional source motion with the intent to customize scanning motions around "suspicious findings". Customized acquisitions can improve the image quality in areas that require increased scrutiny, such as breast cancers, architectural distortions, and dense clusters. In this paper, virtual clinical trial techniques were used to analyze whether a finding or area at high risk of masking cancers can be detected in a single low-dose projection and thus be used for motion planning. This represents a step towards customizing the subsequent low-dose projection acquisitions autonomously, guided by the first low-dose projection; we call this technique "self-steering tomosynthesis." A U-Net was used to classify the low-dose projections into "risk classes" in simulated breasts with soft-tissue lesions; class probabilities were modified using post hoc Dirichlet calibration (DC). DC improved the multiclass segmentation (Dice = 0.43 vs. 0.28 before DC) and significantly reduced false positives (FPs) from the class of the highest risk of masking (sensitivity = 81.3% at 2 FPs per image vs. 76.0%). This simulation-based study demonstrated the feasibility of identifying suspicious areas using a single low-dose projection for self-steering tomosynthesis.
Purpose
Virtual clinical trials (VCTs) require computer simulations of representative patients and images to evaluate and compare changes in performance of imaging technologies. The simulated images ...are usually interpreted by model observers whose performance depends upon the selection of imaging cases used in training evaluation models. This work proposes an efficient method to simulate and calibrate soft tissue lesions, which matches the detectability threshold of virtual and human readings.
Methods
Anthropomorphic breast phantoms were used to evaluate the simulation of four mass models (I–IV) that vary in shape and composition of soft tissue. Ellipsoidal (I) and spiculated (II–IV) masses were simulated using composite voxels with partial volumes. Digital breast tomosynthesis projections and reconstructions of a clinical system were simulated. Channelized Hotelling observers (CHOs) were evaluated using reconstructed slices of masses that varied in shape, composition, and density of surrounded tissue. The detectability threshold of each mass model was evaluated using receiver operating characteristic (ROC) curves calculated with the CHO's scores.
Results
The area under the curve (AUC) of each calibrated mass model were within the 95% confidence interval (mean AUC 95% CI) reported in a previous reader study (0.93 0.89, 0.97). The mean AUC 95% CI obtained were 0.94 0.93, 0.96, 0.92 0.90, 0.93, 0.92 0.90, 0.94, 0.93 0.92, 0.95 for models I to IV, respectively. The mean AUC results varied substantially as a function of shape, composition, and density of surrounded tissue.
Conclusions
For successful VCTs, lesions composed of soft tissue should be calibrated to simulate imaging cases that match the case difficulty predicted by human readers. Lesion composition, shape, and size are parameters that should be carefully selected to calibrate VCTs.
Objectives
A virtual clinical trial (VCT) method is proposed to determine the limit of calcification detection in tomosynthesis.
Methods
Breast anatomy, focal findings, image acquisition, and ...interpretation (
n
= 14 readers) were simulated using screening data (
n
= 660 patients). Calcifications (0.2–0.4 mm
3
) were inserted into virtual breast phantoms. Digital breast tomosynthesis (DBT) acquisitions were simulated assuming various acquisition geometries: source motion (continuous and step-and-shoot), detector element size (140 and 70 µm), and reconstructed voxel size (35–140 µm). VCT results were estimated using multiple-reader multiple-case analyses and
d’
statistics. Signal-to-noise (SNR) analyses were also performed using BR3D phantoms.
Results
Source motion and reconstructed voxel size demonstrated significant changes in the performance of imaging systems. Acquisition geometries that use 70 µm reconstruction voxel size and step-and-shoot motion significantly improved calcification detection. Comparing 70 with 100 µm reconstructed voxel size for step-and-shoot, the ΔAUC was 0.0558 (0.0647) and
d’
ratio was 1.27 (1.29) for 140 µm (70 µm) detector element size. Comparing step-and-shoot with a continuous motion for a 70 µm reconstructed voxel size, the ΔAUC was 0.0863 (0.0434) and the
d’
ratio was 1.40 (1.19) for 140 µm (70 µm) detector element. Small detector element sizes (e.g., 70 µm) did not significantly improve detection. The SNR results with the BR3D phantom show that calcification detection is dependent upon reconstructed voxel size and detector element size, supporting VCT results with comparable agreement (ratios:
d’
= 1.16 ± 0.11,
SNR
= 1.34 ± 0.13).
Conclusion
DBT acquisition geometries that use super-resolution (smaller reconstructed voxels than the detector element size) combined with step-and-shoot motion have the potential to improve the detection of calcifications.
Clinical relevance
Calcifications may not always be discernable in tomosynthesis because of differences in acquisition and reconstruction methods. VCTs can identify strategies to optimize acquisition and reconstruction parameters for calcification detection in tomosynthesis, most notably through super-resolution in the reconstruction.
Key Points
• Super-resolution improves calcification detection and SNR in tomosynthesis; specifically, with the use of smaller reconstruction voxels.
• Calcification detection using step-and-shoot motion is superior to that using continuous tube motion.
• A detector element size of 70 µm does not provide better detection than 140 µm for small calcifications at the threshold of detectability.