Loren Niklason

Comparison of Digital Mammography Alone and Digital Mammography Plus Tomosynthesis in a Population-based Screening Program

PURPOSE To assess cancer detection rates, false-positive rates before arbitration, positive predictive values for women recalled after arbitration, and the type of cancers detected with use of digital mammography alone and combined with tomosynthesis in a large prospective screening trial. MATERIALS AND METHODS A prospective, reader- and modality-balanced screening study of participants undergoing combined mammography plus tomosynthesis, the results of which were read independently by four different radiologists, is under way. The study was approved by a regional ethics committee, and all participants provided written informed consent. The authors performed a preplanned interim analysis of results from 12,631 examinations interpreted by using mammography alone and mammography plus tomosynthesis from November 22, 2010, to December 31, 2011. Analyses were based on marginal log-linear models for binary data, accounting for correlated interpretations and adjusting for reader-specific performance levels by using a two-sided significance level of .0294. RESULTS Detection rates, including those for invasive and in situ cancers, were 6.1 per 1000 examinations for mammography alone and 8.0 per 1000 examinations for mammography plus tomosynthesis (27% increase, adjusted for reader; P = .001). False-positive rates before arbitration were 61.1 per 1000 examinations with mammography alone and 53.1 per 1000 examinations with mammography plus tomosynthesis (15% decrease, adjusted for reader; P < .001). After arbitration, positive predictive values for recalled patients with cancers verified later were comparable (29.1% and 28.5%, respectively, with mammography alone and mammography plus tomosynthesis; P = .72). Twenty-five additional invasive cancers were detected with mammography plus tomosynthesis (40% increase, adjusted for reader; P < .001). The mean interpretation time was 45 seconds for mammography alone and 91 seconds for mammography plus tomosynthesis (P < .001). CONCLUSION The use of mammography plus tomosynthesis in a screening environment resulted in a significantly higher cancer detection rate and enabled the detection of more invasive cancers. Clinical trial registration no. NCT01248546.

Diagnostic Accuracy and Recall Rates for Digital Mammography and Digital Mammography Combined With One-View and Two-View Tomosynthesis: Results of an Enriched Reader Study

OBJECTIVE The purpose of this study was to compare two methods of combining tomosynthesis with digital mammography by assessing diagnostic accuracy and recall rates for digital mammography alone and digital mammography combined with one-view tomosynthesis and two-view tomosynthesis. SUBJECTS AND METHODS Three hundred ten cases including biopsy-proven malignancies (51), biopsy-proven benign findings (47), recalled screening cases (138), and negative screening cases (74) were reviewed by 15 radiologists sequentially using digital mammography, adding one-view tomosynthesis, and then two-view tomosynthesis. Cases were assessed for recall and assigned a BI-RADS score and probability of malignancy for each imaging method. Diagnostic accuracy was assessed using receiver operating characteristic (ROC) analysis. Screening recall rates were compared using pooled logistical regression analysis. A p value of < 0.0167 was considered significant. RESULTS The area under the ROC curve (AUC) for digital mammography (DM), DM plus one-view tomosynthesis, and DM plus two-view tomosynthesis was 0.828, 0.864, and 0.895, respectively. Both one-view and two-view tomosynthesis plus DM were significantly better than DM alone (Δ AUCs 0.036 [p = 0.009] and 0.068 [p < 0.001]). Average noncancer recall rates for digital mammography, DM plus one-view tomosynthesis, and DM plus two-view tomosynthesis were 44.2%, 27.2%, and 24.0%, respectively. Combined with DM, one-view and two-view tomosynthesis both showed significantly lower noncancer recall rates than digital mammography alone (p < 0.001). Digital mammography with two-view tomosynthesis showed a significantly lower recall rate than digital mammography with one-view tomosynthesis (p < 0.001). Diagnostic accuracy for dense (Δ AUC, 0.091%; p < 0.001) and nondense (Δ AUC, 0.035%; p = 0.001) breasts improved with DM plus two-view tomosynthesis compared with digital mammography alone. Compared with digital mammography, diagnostic sensitivity for invasive cancers increased with the addition of both one-view (Δ12.0%, p < 0.001) and two-view (Δ21.7%, p < 0.001) tomosynthesis. CONCLUSION The addition of one-view tomosynthesis to conventional digital mammography improved diagnostic accuracy and reduced the recall rate; however, the addition of two-view tomosynthesis provided twice the performance gain in diagnostic accuracy while further reducing the recall rate.

Clinical Performance of Breast Tomosynthesis as a Function of Radiologist Experience Level

This study reports the performance of breast tomosynthesis (3D images) combined with digital mammography (2D images), compared to digital mammography alone, as a function of the experience of the radiologist. In this trial, twelve readers analyzed 316 image sets, giving BIRADS (and other) scores first for the digital mammograms, and subsequently for the combined datasets of tomosynthesis and digital mammograms. Clinical performance was measured using two metrics: area under the ROC curve (AUROC) and recall rate, and was analyzed as a function of the experience level. The study found that all radiologists AUROC improved when using 3D+2D compared to 2D, with no correlation with experience level, increasing by 0.077 ± 0.058 (mean ± 1 standard deviation) for the 5 least experienced and increasing by 0.078 ± 0.029 for the 5 most experienced. Similarly, the use of 3D+2D compared to 2D imaging showed a mean decrease in recall rate of 39.2% for the most experienced and a decrease of 39.6% for the least experienced, again with no correlation found with experience level. In summary, radiologists with a range of experience demonstrated improved performance using tomosynthesis in combination with digital mammography (3D+2D), as measured using recall rate reduction and AUROC metrics.

A new generation FFDM/tomosynthesis fusion system with selenium detector

A new generation of digital breast tomosynthesis system has been designed and is commercially available outside the US. The system has both a 2D mode and a 3D mode to do either conventional mammography or tomosynthesis. Uniquely, it also has a fusion mode that allows both 3D and 2D images to be acquired under the same breast compression, which results in co-registered images from the two modalities. The aim of this paper is to present a technical description on the design and performance of the new system, including system details such as filter options, doses, AEC operation, 2D and 3D images co-registration and display, and the selenium detector performance. We have carried out both physical and clinical studies to evaluate the system. In this paper the focus will be mainly on technical performance results.

Lesion visibility in low dose tomosynthesis

Visibility of lesions in mammography are significantly reduced by the presence of anatomical, or structure, noise. Breast tomosynthesis offers the possibility of reducing this noise. We have compared the detection of low contrast and microcalcification objects with tomosynthesis imaging as a function of dose to full field digital mammography (FFDM) performed at a standard screening dose. The measurements were performed with a variety of phantoms and complex backgrounds. The complex backgrounds greatly reduced object visibility using FFDM; much less so for the tomosynthesis images. In summary, visibility of low contrast objects using tomosynthesis was superior to visibility of these objects in FFDM, even when the tomosynthesis imaging was performed at 1/4 or less of a FFDM dose. Tomosynthesis also showed superior visibility to FFDM for 160-180 micron microcalcifications at 1/2 the FFDM dose.

Experimental and theoretical spectral optimization for digital mammography

The detection characteristics of digital x-ray and film-screen mammography systems are different and thus current film-screen techniques are not ideal for digital mammography. Therefore optimum technical parameters required for digital mammography are likely to be different compared with film-screen mammography. The goal of this study is to evaluate the optimum technical parameters for full-field digital mammography by experimental and computer simulation methods. A General Electric Full Field Digital Mammography (FFDM) prototype unit using Cesium Iodide (CsI) on an amorphous Silicon photodiode array was used for the experimental measurements. Using breast equivalent phantoms, images were acquired for a set of x-ray target-filters for a range of peak kilovoltage, varying breast composition and thickness, with and without an anti-scatter grid. The signal-to-noise ratio (SNR) and figure-of-merit (FOM) were determined for simulated calcification and mass targets, independently by the two methods. The results for noise, contrast, SNR and FOM were compared and agree within 5% and 6% respectively. Combined results are presented for the case of 50% glandular - 50% adipose tissue breast composition using the grid and for the calcification target. Based on the FOM approach, preliminary results suggest that a Rhodium target-filter combination will be beneficial for higher breast thickness and for denser breasts.

The dependence of tomosynthesis imaging performance on the number of scan projections

In general, the use of more projections results in fewer tomosynthesis reconstruction artifacts. However, under a fixed dose, an excess number of projections will make the detector noise more pronounced in each of the x-ray shots and thus degrade image quality. Even in the absence of detector noise the advantages of higher projection numbers eventually have diminishing returns, making more projections unnecessary. In this study, we explore the dependence of tomosynthesis imaging performance on the number of projections, while keeping other factors fixed. We take the contrast-to-noise ratio as the figure of merit to search for the range of optimal projection number. The study is carried out through both simulations and experiments, with phantoms consisting of micro-calcification and mass objects, and a cadaver breast. The goal of this paper is to describe our methodology in general, and use a prototype tomosynthesis system as an example. The knowledge learned will help the design of future generation clinical tomosynthesis systems.

Statistical properties of a utility measure of observer performance compared to area under the ROC curve

The receiver operating characteristic (ROC) curve has become a common tool for evaluating diagnostic imaging technologies, and the primary endpoint of such evaluations is the area under the curve (AUC), which integrates sensitivity over the entire false positive range. An alternative figure of merit for ROC studies is expected utility (EU), which focuses on the relevant region of the ROC curve as defined by disease prevalence and the relative utility of the task. However if this measure is to be used, it must also have desirable statistical properties keep the burden of observer performance studies as low as possible. Here, we evaluate effect size and variability for EU and AUC. We use two observer performance studies recently submitted to the FDA to compare the EU and AUC endpoints. The studies were conducted using the multi-reader multi-case methodology in which all readers score all cases in all modalities. ROC curves from the study were used to generate both the AUC and EU values for each reader and modality. The EU measure was computed assuming an iso-utility slope of 1.03. We find mean effect sizes, the reader averaged difference between modalities, to be roughly 2.0 times as big for EU as AUC. The standard deviation across readers is roughly 1.4 times as large, suggesting better statistical properties for the EU endpoint. In a simple power analysis of paired comparison across readers, the utility measure required 36% fewer readers on average to achieve 80% statistical power compared to AUC.

Comparative Statistical Properties of Expected Utility and Area Under the ROC Curve for Laboratory Studies of Observer Performance in Screening Mammography

RATIONALE AND OBJECTIVES Our objective is to determine whether expected utility (EU) and the area under the receiver operator characteristic (AUC) are consistent with one another as endpoints of observer performance studies in mammography. These two measures characterize receiver operator characteristic performance somewhat differently. We compare these two study endpoints at the level of individual reader effects, statistical inference, and components of variance across readers and cases. MATERIALS AND METHODS We reanalyze three previously published laboratory observer performance studies that investigate various x-ray breast imaging modalities using EU and AUC. The EU measure is based on recent estimates of relative utility for screening mammography. RESULTS The AUC and EU measures are correlated across readers for individual modalities (r = 0.93) and differences in modalities (r = 0.94 to 0.98). Statistical inference for modality effects based on multi-reader multi-case analysis is very similar, with significant results (P < .05) in exactly the same conditions. Power analyses show mixed results across studies, with a small increase in power on average for EU that corresponds to approximately a 7% reduction in the number of readers. Despite a large number of crossing receiver operator characteristic curves (59% of readers), modality effects only rarely have opposite signs for EU and AUC (6%). CONCLUSIONS We do not find any evidence of systematic differences between EU and AUC in screening mammography observer studies. Thus, when utility approaches are viable (i.e., an appropriate value of relative utility exists), practical effects such as statistical efficiency may be used to choose study endpoints.

Automatic patient motion detection in digital breast tomosynthesis

Patient motion is frequently a problem in mammography, especially when the x-ray exposure is long, resulting in image quality degradation. At present, patient motion can only be identified by inspecting the image subjectively after image acquisition. As digital breast tomosynthesis (DBT) takes longer time to complete the data acquisition than conventional mammography, there is more chance for patient motion to happen in DBT. Therefore it is important to understand the potential motion problem in DBT and incorporate a design to minimize it. In this paper we present an automatic method to detect patient motions in DBT. The method is developed based on an understanding that, features of breast should move along predictable trajectory in a time-series of projection measurements; deviations from it are linked to patient motion. Motion distance is estimated by analyzing skin lines and large calcifications (if exist) in all projection images and then a motion score is derived for a DBT scan. Effectiveness and robustness of this method will be demonstrated with clinical data, together with discussions on different motion patterns observed clinically. The impacts of this work could be far-reaching. It allows real-time detection and objective evaluation of patient motions, applicable to all breasts. Patient with severe motion can be re-scanned immediately before leaving the room. Data with moderate motions can go through additional targeted image processing to minimize motion artifacts. It also enables a powerful tool to evaluate and optimize different DBT designs to minimize the patient motion problem. Besides, this method can be extended to other imaging modalities, e.g. breast CT, to study patient motions.