- Email: [email protected]
Performance of Photon-Counting Breast Computed Tomography, Digital Mammography, and Digital Breast Tomosynthesis in Evaluating Breast Specimens
ARTICLE IN PRESS
Original Investigation
Performance of Photon-Counting Breast Computed Tomography, Digital Mammography, and Digital Breast Tomosynthesis in Evaluating Breast Specimens Ann-Christin Rößler, Willi Kalender, Daniel Kolditz, Christian Steiding, Veikko Ruth, Caroline Preuss, Sandra Christina Peter, Barbara Brehm, Matthias Hammon, Rüdiger Schulz-Wendtland, Evelyn Wenkel Rationale and Objectives: This study compared a novel photon-counting breast computed tomography (pcBCT) system with digital mammography (DM) and digital breast tomosynthesis (DBT) systems. For this reason, surgical specimens were examined with all three techniques and rated by three observers. Materials and Methods: A total of 30 surgical specimens were investigated with DM, DBT, and pcBCT; the associated images were shown to three experienced radiologists. Findings (22 microcalcifications and 23 mass lesions) were recorded and compared to the results of the pathological examination. Sensitivity and specificity for detection of microcalcifications and lesions were calculated and displayed using receiver operating characteristic curves. Results: Sensitivity for microcalcifications was 82% for DM, 70% for DBT, and 85% for pcBCT. Specificity for microcalcifications was 71% for DM, 75% for DBT, and 83% for pcBCT. Sensitivity for lesions was 45% for DM, 62% for DBT, and 65% for pcBCT. Specificity for lesions was 76% for DM, 62% for DBT, and 76% for pcBCT. Conclusions: pcBCT showed a comparable or superior performance compared to the clinically approved DM and DBT systems. Mass lesion detectability can be increased further by the use of contrast media. Key Words: Breast computed tomography; digital mammography; digital breast tomosynthesis; photon-counting detector; specimen. © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
B
reast cancer is the most frequent solid malignant tumor among women in industrial nations. In 2012, breast cancer had an incidence of 464,000 cases in Europe and was the leading cause of cancer death in women (1). Early detection is essential to reduce the mortality rate. Each millimeter of tumor diameter is associated with a percent higher chance of death (2). For this reason, screening programs have been established in most European countries (3). Acad Radiol 2016; ■:■■–■■ From the Institute of Medical Physics, University of Erlangen-Nuremberg, Henkestrasse 91, Erlangen 91052 (A.-C.R., W.K., D.K., C.S., V.R.); AB-CT GmbH, Erlangen (W.K., D.K., C.S.); Department of Radiology, University Hospital Erlangen, Erlangen (C.S., S.C.P., B.B., M.H., R.S.-W., E.W.); Department of Gynecology and Obstetrics, University Hospital of Erlangen, Erlangen, Germany (C.P.). Received August 25, 2016; revised September 21, 2016; accepted September 25, 2016. Address correspondence to: A.-C.R. e-mail: [email protected] © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.acra.2016.09.017
Digital mammography (DM) is the workhorse of breast imaging but weakens its effectiveness in dense breast tissue due to superposition of tissue structures. Mammographic sensitivity in lesion detection for fatty breasts rises up to 98% but drops down to 48%–30% in very dense breast tissue (4,5). On the other hand, studies reported up to a fivefold increased breast cancer risk in women with dense breast tissue (6). Sensitivity of DM has improved significantly with the additional use of digital breast tomosynthesis (DBT), which has a slightly higher radiation dose compared to conventional mammography (7). The sensitivity of DBT alone was 43% higher than mammography in clinical trials (8). Recall rates in screening programs could be reduced if DBT was conducted in addition to mammography (8–11). Unfortunately, problems occurred in the detection of calcifications and sensitivity was higher for DM than for DBT in some studies (12). The use of dedicated breast computed tomography (BCT) for the detection and diagnosis of lesions is a novel approach in breast imaging. Several groups have developed and tested such systems in recent years (13). 1
Academic Radiology, Vol ■, No ■■, 2016
RÖßLER ET AL
Results show a comparable performance of BCT to DM in the detection of lesions especially when contrast media are applied (14,15). Detection of calcifications, however, appears to be slightly worse using BCT systems that utilize flat panel detectors, when compared to DM systems (16). In addition, BCT systems produce images with poor spatial resolution that cannot resolve structures smaller than 290 μm; this is considered to be a reason for the lower detection rate of microcalcifications (17). Patient doses for such systems range between 6 and 16 mGy, depending on breast size (18). In the present study, a novel photon-counting breast computed tomography (pcBCT), using a photon-counting cadmium-telluride detector, with a three-dimensional imaging technique is used. This technique achieves a high spatial resolution and is to detect structures down to 100 μm at a dose below 5 mGy, which gives the ability to delineate microcalcifications clearly (19). The purpose of the study presented here is to compare detection statistics between images produced from pcBCT, DM, and DBT. MATERIALS AND METHODS In total, 30 surgical specimens were evaluated for the present study from November to December 2015. Specimens of women who had a Breast Imaging Reporting and Data System (BI-RADS (20)) 4, 5, or 6 (one case) lesions were included. Fourteen women underwent lumpectomy and 16 women underwent mastectomy. The mean age of the examined patients was 58.6 years (range: 41–79 years). Ethical approval was confirmed and informed consent was obtained from all patients. All specimens were investigated directly after surgery by DM, DBT, and pcBCT before pathological examination. DM and DBT examinations were performed on two standard clinical systems made by two different manufacturers (Siemens Mammomat, Siemens Healthcare GmbH, Erlangen, Germany; and Hologic Selenia Dimensions, Hologic, Bedford, MA). Specimens were investigated with standard clinical settings at a tube voltage of 26 kV for DM and 27 kV for DBT. Spatial resolution of the systems was better than 100 μm for DM, DBT (only in-plane), and pcBCT (all directions, isotropic resolution). The examination systems for DM and DBT were chosen randomly for each specimen. pcBCT scans were performed with a tube voltage of 60 kV at an experimental scanner equipped with a photon-counting high-resolution cadmiumtelluride detector. The scan was made in spiral mode. A filtered backprojection was used for image reconstruction. Average glandular dose was kept below 5 mGy (21). The evaluation of images was performed by three radiologists with 12, 5, and 3 years of breast imaging experience, respectively. The software ImpactView (AB-CT GmbH, Erlangen, Germany) was used for image viewing on a dedicated workstation. Observers were allowed to change window settings and to slice through the data sets independently. For pcBCT exams, all views (transversal, coronal, and sagittal) were recorded and used. Images were grouped according to imaging modalities and shown in a random order to the observers 2
without knowledge of the results of the pathological examination or the other imaging modalities. First, DM then DBT and pcBCT images were shown to the observers in different random orders. The time it took the physicians to evaluate the images was recorded. DM scans were used for determining breast density according to American College of Radiology BI-RADS density categories as follows: (a) almost entirely fatty, (b) areas of fibroglandular tissue, (c) heterogeneously dense, and (d) extremely dense (20). In each specimen, two different kinds of findings were interpreted for all modalities: microcalcifications and mass lesions. Microcalcifications and mass lesions were reported as being present or not by the three observers. Each observer stated the certainty of his or her answer on a scale from 1 (not certain) to 5 (completely certain). The most experienced observer classified the findings analogously to the BI-RADS descriptors for the diagnostic DM images. The results between the observers were averaged. Sensitivity and specificity were calculated for each modality and compared to each other and to the results of the pathological examination. Differences in sensitivity and specificity between lumpectomy and mastectomy specimens were calculated. Receiver operating characteristic curves were constructed with the help of certainty values for DM, DBT, and pcBCT and were classified into microcalcifications and lesions. Areas under the curves (AUCs) were calculated. RESULTS Pathology revealed 16 invasive carcinomas with an average size of 20.3 mm (range: 2–84 mm). Seven of these carcinomas were associated with microcalcifications. Twelve carcinomas were additionally associated with ductal carcinoma in situ. Seven pure ductal carcinomas in situ were found pathologically from which six were associated with microcalcifications. Additionally, one case of atypical ductal hyperplasia without microcalcifications and five cases of fibrocystic changes (three associated with microcalcifications) were detected. In one of the specimens, a BI-RADS 6 case, none of the previously mentioned carcinomas were found after the patient underwent neoadjuvant chemotherapy. An overview of the pathological findings separated into lumpectomies and mastectomies is shown in Table 1. Twenty-two of the specimens contained radiologically detected microcalcifications. In six of the specimens, microcalcifications were not explicitly described in the pathological report but were clearly visible in all three imaging techniques and were defined as true positive. Breast tissue was classified as rather dense in mammography with a density of b–c. Lumpectomy specimens were rated denser (averaged c) than mastectomy specimens (averaged b). A characterization of findings was made by observer 1 in DM images according to the BI-RADS classification standards and is shown in Figure 1. If the finding was not visible
Academic Radiology, Vol ■, No ■■, 2016
IMAGING TECHNIQUES FOR EVALUATING BREAST SPECIMENS
TABLE 1. Overview of Pathological Findings in Lumpectomy and Mastectomy Specimens Lumpectomy
Mastectomy
Without Calcifications
With Calcifications
Without Calcifications
With Calcifications
0 2 0 1 2
1 3 2 0 3
3 4 1 0 3
0 3 4 0 0
IC without DCIS IC with DCIS Pure DCIS ADH Fibrocystic changes
ADH, atypical ductal hyperplasia; DCIS, ductal carcinoma in situ; IC, invasive carcinoma.
Figure 1.
Characterization of microcalcifications (n = 22) and mass lesions (n = 23).
in DM, characterization was done with DBT and pcBCT images. For lesion specification only, a few BI-RADS descriptors were chosen because radiological specimens were evaluated. The reading of pcBCT images took more time than the DM and DBT images. Times varied between observers. The mean reading times of the observers are given in Table 2. The sensitivity for microcalcifications was 82% for DM, 70% for DBT, and 85% for pcBCT. The specificity for microcalcifications was 71% for DM, 75% for DBT, and 83%
TABLE 2. Average Reading Time for DM, DBT, and pcBCT Reading Time (s)
Reader 1 Reader 2 Reader 3
DM
DBT
pcBCT
77 61 66
122 54 63
131 83 119
DBT, digital breast tomosynthesis; DM, digital mammography; pcBCT, photon-counting breast computed tomography.
3
RÖßLER ET AL
Academic Radiology, Vol ■, No ■■, 2016
Figure 2. Image example for soft-tissue delineation in dense breasts using digital mammography (a), digital breast tomosynthesis (b), and pcBCT (c). Notice that in pcBCT, there is less superimposition by other tissue layers. pcBCT, photoncounting breast computed tomography.
Figure 3. Image example for calcification detection using DM (a), DBT (b), and pcBCT (c). Calcification with a size of about 1.5 mm was clearly visible in pcBCT (arrow) but not in DM and DBT due to superposition with soft-tissue structures. DBT, digital breast tomosynthesis; DM, digital mammography; pcBCT, photon-counting breast computed tomography.
for pcBCT. The sensitivity for lesions was 45% for DM, 62% for DBT, and 65% for pcBCT. The specificity for lesions was 76% for DM, 62% for DBT, and 76% for pcBCT. An example of soft-tissue delineation in the three imaging modalities is shown in Figure 2. An example of microcalcification detection is shown in Figure 3. Based on the results, the study determined that sensitivities and specificities differed between lumpectomies and mastectomies. For microcalcifications, DM and DBT showed a higher sensitivity for the less dense mastectomy specimens. pcBCT microcalcification sensitivity was better for lumpectomy specimens and equal for mastectomy specimens when compared to DM and DBT. For lesions, DBT showed the highest sensitivity for lumpectomy specimens and pcBCT for mastectomies. Specificities for both were almost equal. Detailed results are shown in Figure 4. Receiver operating characteristic curves for all modalities and specimens are shown in Figure 5. AUC values for microcalcifications were 0.733 for DM, 0.672 for DBT, and 0.812 for pcBCT. AUC values for lesions were 0.420 for DM, 0.493 for DBT, and 0.583 for pcBCT. DISCUSSION The present study compared the performance of three different breast imaging modalities for the detection of microcalcifications and mass lesions in breast specimens. DM 4
and DBT systems were in daily clinical use; the investigated pcBCT system was a prototype system working with a novel photon-counting detector. Although readers were not used to BCT images, pcBCT showed a comparable or better performance than DM and DBT. It was expected that a minor experience in reading BCT might impede the interpretation of the computed tomography scans like it was in the case when DM was introduced. In the OSLO I trial, for example, DM was equal or inferior to conventional mammography. One argument was that this might have been caused by the lack of experience with digital mammograms. In the OSLO II trial, when the examiners where much more familiar with the digital system, DM was superior to conventional screen film mammography (22,23). Observers in the present study were most experienced in DM and DBT but not in computed tomography. Nevertheless, lesion detection for pcBCT was as high as that for DBT. In the patient studies to follow, contrast media will be applied, which offers the advantage of increasing lesion detectability significantly as shown in former studies by Boone et al. (15) and O’Connell et al. (24). pcBCT outperformed DM and DBT in the detection of microcalcifications. Eighty-five percent of the microcalcifications were visible at a specificity of 83%. One explanation for this may be that BCT images are free of superimpositions and have an isotropic-like resolution so that there are advantages, especially in dense breast tissue. This was already shown in studies
Academic Radiology, Vol ■, No ■■, 2016
IMAGING TECHNIQUES FOR EVALUATING BREAST SPECIMENS
Figure 4. Sensitivity and specificity of DM, digital breast tomosynthesis, and photoncounting breast computed tomography for detecting calcifications (a) and lesions (b) for all specimens as well as for lumpectomies and mastectomies (separately shown). BCT breast computed tomography; BT, breast tomosynthesis; DM, digital mammography.
Figure 5. ROC curves for detectability of calcifications (a) and lesions (b) using digital mammography, digital breast tomosynthesis, and pcBCT. ROC, receiver operating characteristic; pcBCT, photon-counting breast computed tomography.
5
RÖßLER ET AL
Academic Radiology, Vol ■, No ■■, 2016
Figure 6. Microcalcification details found in images produced by digital mammography (a), digital breast tomosynthesis (b), and pcBCT (c). A marking clip and many small microcalcifications are visible. For pcBCT, a maximum intensity projection image over the whole volume is shown to make it comparable to the other imaging techniques in this image. pcBCT, photon-counting breast computed tomography.
with other BCT systems, for example, O’Connell et al.’s study (24). In those systems, however, microcalcifications are only visible if they are bigger than 250 μm. An example of the visibility of a calcification cluster in the present study is shown in Figure 6. Further proofs for pcBCT’s ability to detect small calcifications is given in Kalender et al.’s study (21). In conclusion, sensitivities and specificities differed between lumpectomy and mastectomy specimens. The reasons for this result could be due to the higher-rated density of lumpectomy specimens and their smaller size, making it more difficult to distinguish between normal and pathological breast structures. Sensitivity for lesion detection in pcBCT was higher than that in DM and DBT for the large mastectomy because there is no superimposition of structures in pcBCT. Superposition is not crucial in smaller specimens. There are some limitations to the present study. The study was performed on specimens and the main focus was on detecting and not classifying the radiological pathology according to BI-RADS descriptors. Microcalcifications and mass lesions were characterized by only one reader. The authors feel that those aspects should be addressed in a clinical study with patients and not with radiological specimens. No contrast medium was used in the present study due to the investigation of specimens. When contrast medium is applied, sensitivity is expected to be much higher for mass lesions as it was shown in former studies, for example, for contrast-enhanced spectral mammography compared to DM (25). The positive effects of contrast media application were 6
investigated for other BCT systems and are expected for the pcBCT system too (14). In conclusion, pcBCT showed reliable performance in the detection of microcalcifications and mass lesions in surgical breast specimens. pcBCT outperformed DM and DBT especially in specimens with dense breast tissue. Further clinical studies are needed to validate these findings. REFERENCES 1. Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J, et al. Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012. Eur J Cancer 2013; 49:1374–1403. 2. Michaelson JS, Silverstein M, Sgroi D, et al. The effect of tumor size and lymph node status on breast carcinoma lethality. Cancer 2003; 98:2133– 2143. 3. Giordano L, von Karsa L, Tomatis M, et al. Mammographic screening programmes in Europe: organization, coverage and participation. J Med Screen 2012; 19(suppl 1):72–82. 4. Mandelson M, Oestreicher N, Porter P, et al. Breast density as a predictor of mammographic detection: comparison of interval- and screendetected cancers. J Natl Cancer Inst 2000; 92:1081–1087. 5. Kolb TM, Lichy J, Newhouse JH. Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations. Radiology 2002; 225:165–175. 6. McCormack VA, dos Santos Silva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 2006; 15:1159–1169. 7. van Engen R, Bosmans H, Bouwman R, et al. Protocol for the quality control of the physical and technical aspects of digital breast tomosynthesis systems. EUREF; Draft Version 0.10. 2013. 8. Lang K, Andersson I, Rosso A, et al. Performance of one-view breast tomosynthesis as a stand-alone breast cancer screening modality: results
Academic Radiology, Vol ■, No ■■, 2016
9.
10.
11.
12.
13. 14.
15. 16.
17.
IMAGING TECHNIQUES FOR EVALUATING BREAST SPECIMENS
from the Malmo Breast Tomosynthesis Screening Trial, a populationbased study. Eur Radiol 2016; 26:184–190. Haas BM, Kalra V, Geisel J, et al. Comparison of tomosynthesis plus digital mammography and digital mammography alone for breast cancer screening. Radiology 2013; 269:694–700. Rafferty E, Park J, Philpotts L, et al. Assessing radiologist performance using combined digital mammography and breast tomosynthesis compared with digital mammography alone: results of a multicenter, multireader trial. Radiology 2013; 266:104–113. Skaane P, Bandos AI, Gullien R, et al. Prospective trial comparing fullfield digital mammography (FFDM) versus combined FFDM and tomosynthesis in a population-based screening programme using independent double reading with arbitration. Eur Radiol 2013; 23:2061–2071. Spangler ML, Zuley ML, Sumkin JH, et al. Detection and classification of calcifications on digital breast tomosynthesis and 2D digital mammography: a comparison. AJR Am J Roentgenol 2011; 196:320–324. Sarno A, Mettivier G, Russo P. Dedicated breast computed tomography: basic aspects. Med Phys 2015; 42:2786–2804. He N, Wu Y-P, Kong Y, et al. The utility of breast cone-beam computed tomography, ultrasound, and digital mammography for detecting malignant breast tumors: a prospective study with 212 patients. Eur J Radiol 2016; 85:392–403. Prionas ND, Lindfors KK, Ray S, et al. Contrast-enhanced dedicated breast CT: initial clinical experience. Radiology 2010; 256:714–723. O’Connell A, Conover DL, Zhang Y, et al. Cone-beam CT for breast imaging: radiation dose, breast coverage, and image quality. AJR Am J Roentgenol 2010; 195:496–509. Koning Corporation. User’s manual Koning Breast CT. 2015.
18. O’Connell AM, Kawakyu-O’Connor D. Dedicated Cone-beam breast computed tomography and diagnostic mammography: comparison of radiation dose, patient comfort, and qualitative review of imaging findings in BIRADS 4 and 5 lesions. J Clin Imaging Sci 2012; 2:7. 19. Kalender WA, Beister M, Boone JM, et al. High-resolution spiral CT of the breast at very low dose: concept and feasibility considerations. Eur Radiol 2012; 22:1–8. 20. Sickles E, D’Orsi CJ, Bassett LW. ACR BI-RADS® mammography. In: ACR BI-RADS® atlas, breast imaging reporting and data system. Reston, VA: American College of Radiology, 2013. 21. Kalender WA, Kolditz D, Steiding C, et al. Technical feasibility proof for high-resolution low-dose photon-counting CT of the breast. Eur Radiol 2016; 1–6. doi:10.1007/s00330-016-4459-3. 22. Skaane P, Hofvind S, Skjennald A. Randomized trial of screen-film versus full-field digital mammography with soft-copy reading in populationbased screening program: follow-up and final results of Oslo II study. Radiology 2007; 244:708–717. 23. Hofvind S, Skaane P, Elmore JG, et al. Mammographic performance in a population-based screening program: before, during, and after the transition from screen-film to full-field digital mammography. Radiology 2014; 272:52–62. 24. O’Connell AM, Karellas A, Vedantham S. The potential role of dedicated 3D breast CT as a diagnostic tool: review and early clinical examples. Breast J 2014; 20:592–605. 25. Fallenberg EM, Dromain C, Diekmann F, et al. Contrast-enhanced spectral mammography: DOES mammography provide additional clinical benefits or can some radiation exposure be avoided? Breast Cancer Res Treat 2014; 146:371–381.
7
Original Investigation
Performance of Photon-Counting Breast Computed Tomography, Digital Mammography, and Digital Breast Tomosynthesis in Evaluating Breast Specimens Ann-Christin Rößler, Willi Kalender, Daniel Kolditz, Christian Steiding, Veikko Ruth, Caroline Preuss, Sandra Christina Peter, Barbara Brehm, Matthias Hammon, Rüdiger Schulz-Wendtland, Evelyn Wenkel Rationale and Objectives: This study compared a novel photon-counting breast computed tomography (pcBCT) system with digital mammography (DM) and digital breast tomosynthesis (DBT) systems. For this reason, surgical specimens were examined with all three techniques and rated by three observers. Materials and Methods: A total of 30 surgical specimens were investigated with DM, DBT, and pcBCT; the associated images were shown to three experienced radiologists. Findings (22 microcalcifications and 23 mass lesions) were recorded and compared to the results of the pathological examination. Sensitivity and specificity for detection of microcalcifications and lesions were calculated and displayed using receiver operating characteristic curves. Results: Sensitivity for microcalcifications was 82% for DM, 70% for DBT, and 85% for pcBCT. Specificity for microcalcifications was 71% for DM, 75% for DBT, and 83% for pcBCT. Sensitivity for lesions was 45% for DM, 62% for DBT, and 65% for pcBCT. Specificity for lesions was 76% for DM, 62% for DBT, and 76% for pcBCT. Conclusions: pcBCT showed a comparable or superior performance compared to the clinically approved DM and DBT systems. Mass lesion detectability can be increased further by the use of contrast media. Key Words: Breast computed tomography; digital mammography; digital breast tomosynthesis; photon-counting detector; specimen. © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
B
reast cancer is the most frequent solid malignant tumor among women in industrial nations. In 2012, breast cancer had an incidence of 464,000 cases in Europe and was the leading cause of cancer death in women (1). Early detection is essential to reduce the mortality rate. Each millimeter of tumor diameter is associated with a percent higher chance of death (2). For this reason, screening programs have been established in most European countries (3). Acad Radiol 2016; ■:■■–■■ From the Institute of Medical Physics, University of Erlangen-Nuremberg, Henkestrasse 91, Erlangen 91052 (A.-C.R., W.K., D.K., C.S., V.R.); AB-CT GmbH, Erlangen (W.K., D.K., C.S.); Department of Radiology, University Hospital Erlangen, Erlangen (C.S., S.C.P., B.B., M.H., R.S.-W., E.W.); Department of Gynecology and Obstetrics, University Hospital of Erlangen, Erlangen, Germany (C.P.). Received August 25, 2016; revised September 21, 2016; accepted September 25, 2016. Address correspondence to: A.-C.R. e-mail: [email protected] © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.acra.2016.09.017
Digital mammography (DM) is the workhorse of breast imaging but weakens its effectiveness in dense breast tissue due to superposition of tissue structures. Mammographic sensitivity in lesion detection for fatty breasts rises up to 98% but drops down to 48%–30% in very dense breast tissue (4,5). On the other hand, studies reported up to a fivefold increased breast cancer risk in women with dense breast tissue (6). Sensitivity of DM has improved significantly with the additional use of digital breast tomosynthesis (DBT), which has a slightly higher radiation dose compared to conventional mammography (7). The sensitivity of DBT alone was 43% higher than mammography in clinical trials (8). Recall rates in screening programs could be reduced if DBT was conducted in addition to mammography (8–11). Unfortunately, problems occurred in the detection of calcifications and sensitivity was higher for DM than for DBT in some studies (12). The use of dedicated breast computed tomography (BCT) for the detection and diagnosis of lesions is a novel approach in breast imaging. Several groups have developed and tested such systems in recent years (13). 1
Academic Radiology, Vol ■, No ■■, 2016
RÖßLER ET AL
Results show a comparable performance of BCT to DM in the detection of lesions especially when contrast media are applied (14,15). Detection of calcifications, however, appears to be slightly worse using BCT systems that utilize flat panel detectors, when compared to DM systems (16). In addition, BCT systems produce images with poor spatial resolution that cannot resolve structures smaller than 290 μm; this is considered to be a reason for the lower detection rate of microcalcifications (17). Patient doses for such systems range between 6 and 16 mGy, depending on breast size (18). In the present study, a novel photon-counting breast computed tomography (pcBCT), using a photon-counting cadmium-telluride detector, with a three-dimensional imaging technique is used. This technique achieves a high spatial resolution and is to detect structures down to 100 μm at a dose below 5 mGy, which gives the ability to delineate microcalcifications clearly (19). The purpose of the study presented here is to compare detection statistics between images produced from pcBCT, DM, and DBT. MATERIALS AND METHODS In total, 30 surgical specimens were evaluated for the present study from November to December 2015. Specimens of women who had a Breast Imaging Reporting and Data System (BI-RADS (20)) 4, 5, or 6 (one case) lesions were included. Fourteen women underwent lumpectomy and 16 women underwent mastectomy. The mean age of the examined patients was 58.6 years (range: 41–79 years). Ethical approval was confirmed and informed consent was obtained from all patients. All specimens were investigated directly after surgery by DM, DBT, and pcBCT before pathological examination. DM and DBT examinations were performed on two standard clinical systems made by two different manufacturers (Siemens Mammomat, Siemens Healthcare GmbH, Erlangen, Germany; and Hologic Selenia Dimensions, Hologic, Bedford, MA). Specimens were investigated with standard clinical settings at a tube voltage of 26 kV for DM and 27 kV for DBT. Spatial resolution of the systems was better than 100 μm for DM, DBT (only in-plane), and pcBCT (all directions, isotropic resolution). The examination systems for DM and DBT were chosen randomly for each specimen. pcBCT scans were performed with a tube voltage of 60 kV at an experimental scanner equipped with a photon-counting high-resolution cadmiumtelluride detector. The scan was made in spiral mode. A filtered backprojection was used for image reconstruction. Average glandular dose was kept below 5 mGy (21). The evaluation of images was performed by three radiologists with 12, 5, and 3 years of breast imaging experience, respectively. The software ImpactView (AB-CT GmbH, Erlangen, Germany) was used for image viewing on a dedicated workstation. Observers were allowed to change window settings and to slice through the data sets independently. For pcBCT exams, all views (transversal, coronal, and sagittal) were recorded and used. Images were grouped according to imaging modalities and shown in a random order to the observers 2
without knowledge of the results of the pathological examination or the other imaging modalities. First, DM then DBT and pcBCT images were shown to the observers in different random orders. The time it took the physicians to evaluate the images was recorded. DM scans were used for determining breast density according to American College of Radiology BI-RADS density categories as follows: (a) almost entirely fatty, (b) areas of fibroglandular tissue, (c) heterogeneously dense, and (d) extremely dense (20). In each specimen, two different kinds of findings were interpreted for all modalities: microcalcifications and mass lesions. Microcalcifications and mass lesions were reported as being present or not by the three observers. Each observer stated the certainty of his or her answer on a scale from 1 (not certain) to 5 (completely certain). The most experienced observer classified the findings analogously to the BI-RADS descriptors for the diagnostic DM images. The results between the observers were averaged. Sensitivity and specificity were calculated for each modality and compared to each other and to the results of the pathological examination. Differences in sensitivity and specificity between lumpectomy and mastectomy specimens were calculated. Receiver operating characteristic curves were constructed with the help of certainty values for DM, DBT, and pcBCT and were classified into microcalcifications and lesions. Areas under the curves (AUCs) were calculated. RESULTS Pathology revealed 16 invasive carcinomas with an average size of 20.3 mm (range: 2–84 mm). Seven of these carcinomas were associated with microcalcifications. Twelve carcinomas were additionally associated with ductal carcinoma in situ. Seven pure ductal carcinomas in situ were found pathologically from which six were associated with microcalcifications. Additionally, one case of atypical ductal hyperplasia without microcalcifications and five cases of fibrocystic changes (three associated with microcalcifications) were detected. In one of the specimens, a BI-RADS 6 case, none of the previously mentioned carcinomas were found after the patient underwent neoadjuvant chemotherapy. An overview of the pathological findings separated into lumpectomies and mastectomies is shown in Table 1. Twenty-two of the specimens contained radiologically detected microcalcifications. In six of the specimens, microcalcifications were not explicitly described in the pathological report but were clearly visible in all three imaging techniques and were defined as true positive. Breast tissue was classified as rather dense in mammography with a density of b–c. Lumpectomy specimens were rated denser (averaged c) than mastectomy specimens (averaged b). A characterization of findings was made by observer 1 in DM images according to the BI-RADS classification standards and is shown in Figure 1. If the finding was not visible
Academic Radiology, Vol ■, No ■■, 2016
IMAGING TECHNIQUES FOR EVALUATING BREAST SPECIMENS
TABLE 1. Overview of Pathological Findings in Lumpectomy and Mastectomy Specimens Lumpectomy
Mastectomy
Without Calcifications
With Calcifications
Without Calcifications
With Calcifications
0 2 0 1 2
1 3 2 0 3
3 4 1 0 3
0 3 4 0 0
IC without DCIS IC with DCIS Pure DCIS ADH Fibrocystic changes
ADH, atypical ductal hyperplasia; DCIS, ductal carcinoma in situ; IC, invasive carcinoma.
Figure 1.
Characterization of microcalcifications (n = 22) and mass lesions (n = 23).
in DM, characterization was done with DBT and pcBCT images. For lesion specification only, a few BI-RADS descriptors were chosen because radiological specimens were evaluated. The reading of pcBCT images took more time than the DM and DBT images. Times varied between observers. The mean reading times of the observers are given in Table 2. The sensitivity for microcalcifications was 82% for DM, 70% for DBT, and 85% for pcBCT. The specificity for microcalcifications was 71% for DM, 75% for DBT, and 83%
TABLE 2. Average Reading Time for DM, DBT, and pcBCT Reading Time (s)
Reader 1 Reader 2 Reader 3
DM
DBT
pcBCT
77 61 66
122 54 63
131 83 119
DBT, digital breast tomosynthesis; DM, digital mammography; pcBCT, photon-counting breast computed tomography.
3
RÖßLER ET AL
Academic Radiology, Vol ■, No ■■, 2016
Figure 2. Image example for soft-tissue delineation in dense breasts using digital mammography (a), digital breast tomosynthesis (b), and pcBCT (c). Notice that in pcBCT, there is less superimposition by other tissue layers. pcBCT, photoncounting breast computed tomography.
Figure 3. Image example for calcification detection using DM (a), DBT (b), and pcBCT (c). Calcification with a size of about 1.5 mm was clearly visible in pcBCT (arrow) but not in DM and DBT due to superposition with soft-tissue structures. DBT, digital breast tomosynthesis; DM, digital mammography; pcBCT, photon-counting breast computed tomography.
for pcBCT. The sensitivity for lesions was 45% for DM, 62% for DBT, and 65% for pcBCT. The specificity for lesions was 76% for DM, 62% for DBT, and 76% for pcBCT. An example of soft-tissue delineation in the three imaging modalities is shown in Figure 2. An example of microcalcification detection is shown in Figure 3. Based on the results, the study determined that sensitivities and specificities differed between lumpectomies and mastectomies. For microcalcifications, DM and DBT showed a higher sensitivity for the less dense mastectomy specimens. pcBCT microcalcification sensitivity was better for lumpectomy specimens and equal for mastectomy specimens when compared to DM and DBT. For lesions, DBT showed the highest sensitivity for lumpectomy specimens and pcBCT for mastectomies. Specificities for both were almost equal. Detailed results are shown in Figure 4. Receiver operating characteristic curves for all modalities and specimens are shown in Figure 5. AUC values for microcalcifications were 0.733 for DM, 0.672 for DBT, and 0.812 for pcBCT. AUC values for lesions were 0.420 for DM, 0.493 for DBT, and 0.583 for pcBCT. DISCUSSION The present study compared the performance of three different breast imaging modalities for the detection of microcalcifications and mass lesions in breast specimens. DM 4
and DBT systems were in daily clinical use; the investigated pcBCT system was a prototype system working with a novel photon-counting detector. Although readers were not used to BCT images, pcBCT showed a comparable or better performance than DM and DBT. It was expected that a minor experience in reading BCT might impede the interpretation of the computed tomography scans like it was in the case when DM was introduced. In the OSLO I trial, for example, DM was equal or inferior to conventional mammography. One argument was that this might have been caused by the lack of experience with digital mammograms. In the OSLO II trial, when the examiners where much more familiar with the digital system, DM was superior to conventional screen film mammography (22,23). Observers in the present study were most experienced in DM and DBT but not in computed tomography. Nevertheless, lesion detection for pcBCT was as high as that for DBT. In the patient studies to follow, contrast media will be applied, which offers the advantage of increasing lesion detectability significantly as shown in former studies by Boone et al. (15) and O’Connell et al. (24). pcBCT outperformed DM and DBT in the detection of microcalcifications. Eighty-five percent of the microcalcifications were visible at a specificity of 83%. One explanation for this may be that BCT images are free of superimpositions and have an isotropic-like resolution so that there are advantages, especially in dense breast tissue. This was already shown in studies
Academic Radiology, Vol ■, No ■■, 2016
IMAGING TECHNIQUES FOR EVALUATING BREAST SPECIMENS
Figure 4. Sensitivity and specificity of DM, digital breast tomosynthesis, and photoncounting breast computed tomography for detecting calcifications (a) and lesions (b) for all specimens as well as for lumpectomies and mastectomies (separately shown). BCT breast computed tomography; BT, breast tomosynthesis; DM, digital mammography.
Figure 5. ROC curves for detectability of calcifications (a) and lesions (b) using digital mammography, digital breast tomosynthesis, and pcBCT. ROC, receiver operating characteristic; pcBCT, photon-counting breast computed tomography.
5
RÖßLER ET AL
Academic Radiology, Vol ■, No ■■, 2016
Figure 6. Microcalcification details found in images produced by digital mammography (a), digital breast tomosynthesis (b), and pcBCT (c). A marking clip and many small microcalcifications are visible. For pcBCT, a maximum intensity projection image over the whole volume is shown to make it comparable to the other imaging techniques in this image. pcBCT, photon-counting breast computed tomography.
with other BCT systems, for example, O’Connell et al.’s study (24). In those systems, however, microcalcifications are only visible if they are bigger than 250 μm. An example of the visibility of a calcification cluster in the present study is shown in Figure 6. Further proofs for pcBCT’s ability to detect small calcifications is given in Kalender et al.’s study (21). In conclusion, sensitivities and specificities differed between lumpectomy and mastectomy specimens. The reasons for this result could be due to the higher-rated density of lumpectomy specimens and their smaller size, making it more difficult to distinguish between normal and pathological breast structures. Sensitivity for lesion detection in pcBCT was higher than that in DM and DBT for the large mastectomy because there is no superimposition of structures in pcBCT. Superposition is not crucial in smaller specimens. There are some limitations to the present study. The study was performed on specimens and the main focus was on detecting and not classifying the radiological pathology according to BI-RADS descriptors. Microcalcifications and mass lesions were characterized by only one reader. The authors feel that those aspects should be addressed in a clinical study with patients and not with radiological specimens. No contrast medium was used in the present study due to the investigation of specimens. When contrast medium is applied, sensitivity is expected to be much higher for mass lesions as it was shown in former studies, for example, for contrast-enhanced spectral mammography compared to DM (25). The positive effects of contrast media application were 6
investigated for other BCT systems and are expected for the pcBCT system too (14). In conclusion, pcBCT showed reliable performance in the detection of microcalcifications and mass lesions in surgical breast specimens. pcBCT outperformed DM and DBT especially in specimens with dense breast tissue. Further clinical studies are needed to validate these findings. REFERENCES 1. Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J, et al. Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012. Eur J Cancer 2013; 49:1374–1403. 2. Michaelson JS, Silverstein M, Sgroi D, et al. The effect of tumor size and lymph node status on breast carcinoma lethality. Cancer 2003; 98:2133– 2143. 3. Giordano L, von Karsa L, Tomatis M, et al. Mammographic screening programmes in Europe: organization, coverage and participation. J Med Screen 2012; 19(suppl 1):72–82. 4. Mandelson M, Oestreicher N, Porter P, et al. Breast density as a predictor of mammographic detection: comparison of interval- and screendetected cancers. J Natl Cancer Inst 2000; 92:1081–1087. 5. Kolb TM, Lichy J, Newhouse JH. Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations. Radiology 2002; 225:165–175. 6. McCormack VA, dos Santos Silva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 2006; 15:1159–1169. 7. van Engen R, Bosmans H, Bouwman R, et al. Protocol for the quality control of the physical and technical aspects of digital breast tomosynthesis systems. EUREF; Draft Version 0.10. 2013. 8. Lang K, Andersson I, Rosso A, et al. Performance of one-view breast tomosynthesis as a stand-alone breast cancer screening modality: results
Academic Radiology, Vol ■, No ■■, 2016
9.
10.
11.
12.
13. 14.
15. 16.
17.
IMAGING TECHNIQUES FOR EVALUATING BREAST SPECIMENS
from the Malmo Breast Tomosynthesis Screening Trial, a populationbased study. Eur Radiol 2016; 26:184–190. Haas BM, Kalra V, Geisel J, et al. Comparison of tomosynthesis plus digital mammography and digital mammography alone for breast cancer screening. Radiology 2013; 269:694–700. Rafferty E, Park J, Philpotts L, et al. Assessing radiologist performance using combined digital mammography and breast tomosynthesis compared with digital mammography alone: results of a multicenter, multireader trial. Radiology 2013; 266:104–113. Skaane P, Bandos AI, Gullien R, et al. Prospective trial comparing fullfield digital mammography (FFDM) versus combined FFDM and tomosynthesis in a population-based screening programme using independent double reading with arbitration. Eur Radiol 2013; 23:2061–2071. Spangler ML, Zuley ML, Sumkin JH, et al. Detection and classification of calcifications on digital breast tomosynthesis and 2D digital mammography: a comparison. AJR Am J Roentgenol 2011; 196:320–324. Sarno A, Mettivier G, Russo P. Dedicated breast computed tomography: basic aspects. Med Phys 2015; 42:2786–2804. He N, Wu Y-P, Kong Y, et al. The utility of breast cone-beam computed tomography, ultrasound, and digital mammography for detecting malignant breast tumors: a prospective study with 212 patients. Eur J Radiol 2016; 85:392–403. Prionas ND, Lindfors KK, Ray S, et al. Contrast-enhanced dedicated breast CT: initial clinical experience. Radiology 2010; 256:714–723. O’Connell A, Conover DL, Zhang Y, et al. Cone-beam CT for breast imaging: radiation dose, breast coverage, and image quality. AJR Am J Roentgenol 2010; 195:496–509. Koning Corporation. User’s manual Koning Breast CT. 2015.
18. O’Connell AM, Kawakyu-O’Connor D. Dedicated Cone-beam breast computed tomography and diagnostic mammography: comparison of radiation dose, patient comfort, and qualitative review of imaging findings in BIRADS 4 and 5 lesions. J Clin Imaging Sci 2012; 2:7. 19. Kalender WA, Beister M, Boone JM, et al. High-resolution spiral CT of the breast at very low dose: concept and feasibility considerations. Eur Radiol 2012; 22:1–8. 20. Sickles E, D’Orsi CJ, Bassett LW. ACR BI-RADS® mammography. In: ACR BI-RADS® atlas, breast imaging reporting and data system. Reston, VA: American College of Radiology, 2013. 21. Kalender WA, Kolditz D, Steiding C, et al. Technical feasibility proof for high-resolution low-dose photon-counting CT of the breast. Eur Radiol 2016; 1–6. doi:10.1007/s00330-016-4459-3. 22. Skaane P, Hofvind S, Skjennald A. Randomized trial of screen-film versus full-field digital mammography with soft-copy reading in populationbased screening program: follow-up and final results of Oslo II study. Radiology 2007; 244:708–717. 23. Hofvind S, Skaane P, Elmore JG, et al. Mammographic performance in a population-based screening program: before, during, and after the transition from screen-film to full-field digital mammography. Radiology 2014; 272:52–62. 24. O’Connell AM, Karellas A, Vedantham S. The potential role of dedicated 3D breast CT as a diagnostic tool: review and early clinical examples. Breast J 2014; 20:592–605. 25. Fallenberg EM, Dromain C, Diekmann F, et al. Contrast-enhanced spectral mammography: DOES mammography provide additional clinical benefits or can some radiation exposure be avoided? Breast Cancer Res Treat 2014; 146:371–381.
7