Elsevier

European Journal of Radiology

Volume 85, Issue 12, December 2016, Pages 2161-2168
European Journal of Radiology

Diagnostic performance of digital breast tomosynthesis with a wide scan angle compared to full-field digital mammography for the detection and characterization of microcalcifications

https://doi.org/10.1016/j.ejrad.2016.10.004Get rights and content

Highlights

  • Wide scan-angle DBT alone shows a high detection rate for microcalcifications.

  • DBT and FFDM can characterize microcalcifications at a comparable level.

  • Characterization is influenced by reader and by lesion type (benign vs malignant).

  • DBT might be used as a stand-alone technique for the assessment of microcalcifications.

Abstract

Objectives

To assess the diagnostic performance of digital breast tomosynthesis (DBT), with a wide scan-angle, compared to full-field digital mammography (FFDM), for the detection and characterization of microcalcifications.

Methods

IRB approval was obtained for this retrospective study. We selected 150 FFDM and DBT (50 benign and 50 malignant histologically verified microcalcifications, 50 cases classified as BI-RADS 1). Four radiologists evaluated, in separate sessions and blinded to patients’ history and histology, the presence of microcalcifications. Cases with microcalcifications were assessed for visibility, characteristics, and grade of suspicion using BI-RADS categories. Detection rate and diagnostic performance were calculated. Visibility, lesions’ characteristics and reading time were analysed.

Results

Detection rate and visibility were good for both FFDM and DBT, without intra-reader differences (P = 0.510). Inter-reader differences were detected (P < 0.018). Only two lesions were not detected by any reader on either FFDM or DBT. Diagnostic performance with DBT was as good as that of FFDM, but a significant inter-reader difference was found (P = 0.041). High inter-reader variability in the use of the descriptors was found. Reading time for DBT was almost twice that for FFDM (44 and 25 s, respectively).

Conclusion

Wide scan-angle DBT enabled the detection and characterization of microcalcifications with no significant differences from FFDM. Inter-reader variability was seen.

Introduction

Digital breast tomosynthesis (DBT) is an increasingly used technique in breast cancer screening and assessment [1], [2], [3]. DBT is able to overcome several limitations of mammography, in particular, tissue superimposition due to the acquisition of multiple x-ray projections and the reconstruction of pseudo-tomographic images. Various commercial systems are available at the present time, with different acquisition and reconstruction parameters [4].

Regardless of technical parameters, the added value of DBT in the evaluation of masses, asymmetries, and architectural distortions has been confirmed [5]. However, this does not apply to the assessment of microcalcifications [6]. Microcalcifications are a common finding in the breast that can be caused by benign changes or may represent an early sign of malignant disease [7]. A careful characterization of their morphology and distribution is essential to stratify the risk of malignancy and guide clinical management decisions, such as the need for further diagnostic work-up or standard follow-up [8]. Only a few studies have assessed the value of DBT in the detection and characterization of microcalcifications [9], [10], [11], [12]. Kopans et al. [9] reported a very good visibility and good image quality for microcalcifications in a side-by-side evaluation of mammography and DBT. Spangler et al. [10] concluded that digital mammography maintains a higher sensitivity and specificity for microcalcifications, compared to DBT. Tagliafico et al. [12] showed that DBT can miss malignant clusters of microcalcifications that can be easily detected with mammography. These preliminary results opened the discussion about whether DBT is suitable for the study of microcalcifications, a relevant issue that must be considered if DBT is to be used as a primary screening modality [6], [12].

It should be noted that the majority of studies that analysed microcalcifications used acquisition parameters characterized by relatively narrow scan angles, and always evaluated DBT in association with 2D imaging [10], [12], while data about the evaluation of microcalcifications using DBT with a wide scan-angle alone is still scarce. The scan-angle is the total angular range covered by the projections acquired during the examination. Scan-angle is one of the main acquisition parameters that affect the image quality of DBT, along with the number of projections and their distribution [13], [14]. Scan-angle is highly variable in different devices, ranging from 15° (narrow angle) to 50° (wide angle). The optimal combination of the different acquisition parameters is still a topic of intense discussion [13].

The aim of this study was to assess the performance of DBT with a wide scan-angle for the detection and characterization of microcalcifications, and to compare it with full-field digital mammography (FFDM). Inter-reader variability was taken into consideration. Reading time was also measured.

Section snippets

Patient selection

Eligible subjects for this IRB-approved, retrospective study were patients who had undergone digital breast tomosynthesis (DBT) as a screening or diagnostic examination at our institution between January 2010 and August 2012. Overall, 761 patients were examined.

Inclusion criteria used to generate our study cohort were: (a) availability of images from at least one breast with two views in FFDM and DBT examinations; (b) histopathological verification of microcalcifications; and (c) at least two

Detection rate and visibility

The detection rate for microcalcifications with FFDM and DBT is shown in Table 2.

Differences between detection rates were found for readers (P = 0.001) but not for modalities (P = 0.510). Also when benign and malignant microcalcifications were considered separately (Table 2) there was a significant inter-reader difference in the percentage of detection rates (P = 0.003), but there was no significant difference between the modalities (P > 0.496).

In two cases, microcalcifications were not detected with

Discussion

Our single-centre experience shows that DBT with a wide scan-angle alone performs as well as FFDM for the detection and characterization of the microcalcifications usually encountered in the clinical setting. The variability in the assessment of microcalcifications is influenced by the reader but not by the modality under evaluation.

The appropriate detection and characterization of microcalcifications plays a central role in the diagnosis of breast lesions. Preliminary results, obtained from

Conclusions

In conclusion, DBT with a wide scan-angle enables the detection and characterization of microcalcifications at a level comparable to that of FFDM. Therefore, DBT alone can be considered appropriate for the evaluation of microcalcifications. Detection and characterization were not influenced by the modality used, while, for both modalities, there was a significant inter-reader variability.

Conflicts of interest

No conflicts of interest.

Acknowledgments

The study was supported by a grant from Siemens Healthcare, (Erlangen, Germany) and Novomed, (Vienna, Austria) and a seeding grant from the “Privatstiftung für Brustgesundheit” (Vienna, Austria). We would like to thank Mary McAllister for the language revision of this text.

References (34)

  • D. Kopans et al.

    Calcifications in the breast and digital breast tomosynthesis

    Breast J.

    (2011)
  • M.L. Spangler et al.

    Detection and classification of calcifications on digital breast tomosynthesis and 2D digital mammography: a comparison

    AJR Am. J. Roentgenol.

    (2011)
  • M.G. Wallis et al.

    Two-view and single-view tomosynthesis versus full-field digital mammography: high-resolution X-ray imaging observer study

    Radiology

    (2012)
  • A. Tagliafico et al.

    Characterisation of microcalcification clusters on 2D digital mammography (FFDM) and digital breast tomosynthesis (DBT): does DBT underestimate microcalcification clusters? Results of a multicentre study

    Eur. Radiol.

    (2015)
  • I. Sechopoulos

    A review of breast tomosynthesis. Part I. The image acquisition process

    Med. Phys.

    (2013)
  • I. Sechopoulos et al.

    Optimization of the acquisition geometry in digital tomosynthesis of the breast

    Med. Phys.

    (2009)
  • N. Perry et al.

    European Guidelines for Quality Assurance in Breast Cancer Screening and Diagnosis

    (2006)
  • Cited by (40)

    • Breast cancer: introduction

      2022, Targeted Nanomedicine for Breast Cancer Therapy
    • Effect of Dose Level on Radiologists’ Detection of Microcalcifications in Digital Breast Tomosynthesis: An Observer Study with Breast Phantoms

      2022, Academic Radiology
      Citation Excerpt :

      The improvement was attributed mostly to increased detection of masses and architectural distortion. Studies of detectability of microcalcifications in DBT reported conflicting results (6,9–13), which may be caused by differences in the DBT system design and other factors that affected image quality (14–16). A number of approaches has been studied to improve the detection of microcalcifications in DBT, including the use of DM in combination with DBT, the development of machine learning, or artificial intelligent techniques, to assist in detection (17–22), the enhancement of the visibility of microcalcifications by improving reconstruction and image processing methods (23–30), and the synthesis of a DM-like image to replace the directly acquired DM.

    • Diagnostic Performance of Digital Breast Tomosynthesis for Breast Suspicious Calcifications From Various Populations: A Comparison With Full-field Digital Mammography

      2019, Computational and Structural Biotechnology Journal
      Citation Excerpt :

      Similarly, they evaluated the DBT images randomly while blinded to the FFDM images and the patient's clinical information. To minimize the learning and memory bias, they evaluated the DBT images and assigned the BI-RADS category with a one-month interval at least [24,29]. When the assigned BI-RADS categories within the same imaging modality were inconsistent among the radiologists, a consensus was reached through discussion [15].

    View all citing articles on Scopus
    1

    Both authors equally contributed to the paper.

    View full text