Financial Disclosures 3D Mammography: The Latest Developments in the Breast Imaging Arena I have no financial disclosures Dr. Katharine Lampen-Sachar Breast and Body Radiologist Radiology Associates of South Florida Learning Objectives Highlight the basics of the technology behind breast tomosynthesis (DBT) and the C view Understand the indications and potential uses of DBT Understand the potential benefits of DBT as well as some of criticisms Discuss future directions, including tomosynthesis-guided biopsies and localizations Background Screening mammography saves lives. Potential to reduce breast cancer related deaths by 30% (TabarL et al) Frequently and highly visible disputed topic, in the literature and in the newspapers. As digital mammography generates a image of the 3D breast, superimposition of normal breast tissue may hide masses, distortion and other features of malignancy (Roth R et al; Park JM et al). Criticisms of screening digital mammography include (Bleyer and Welch): High false positive rate Limited sensitivity Potential for overdiagnosis 1
From The New York Times. Background Long standing quest for a modality for screening that will: Lower the false positives Increase the true positives that are worth treating Improve sensitivity Hologic Selenia Dimensions Generating the images Take a group of low dose exposures from different angles (projection images) 2
15 degrees Tube Arc Generating the images Drawing shows how DBT images are obtained. Multiple low-dose x-ray projection images are acquired in an arc and used to reconstruct a 3D image of the breast Published in: "Digital Breast Tomosynthesis: Lessons Learned from Early Clinical Implementation" Take a group of low dose exposures from different angles (projection images) Projection images reconstructed through software into a series of high resolution slices (approximately 1 mm in thickness) Number of slices determined by the breast thickness Additional slices generated due to paddle tilt RadioGraphics Vol. 34, No. 4: E89-E102 RSNA, 2014 Issue of Superimposition Issue of Superimposition Constant struggle with conventional digital mammography Park et al.: When screening mammograms demonstrate a questionable finding, the results of follow up diagnostic mammography and US, MR or biopsy ultimately determine whether the finding is significant. The process causes anxiety for patients and incurs additional healthcare costs for findings that are frequently proved benign. 3D tomosynthesisimages are intended to minimize the effect of overlapping breast tissue and to enhance lesion conspicuity and lesion localization (Skaaneet al.; Michellet al. ). Additionally, DBT can help to prevent recall for pseudolesions, which are due to superimposition from overlapping fibroglandular tissue. 3
A few examples.. Example 2. Cancer Case 2 3D Breast Tomosynthesis Slice 16 4
10/13/2014 Cancer Case 2 Screening Combo 12/01/11 Cancer Case 2 RCC 3D Breast Tomosynthesis Slice 28 RMLO 3D Breast Tomosynthesis Slice 28 3D Breast Tomosynthesis Slice 16 Cancer Case 6 Example 3. LCC 5
10/13/2014 Cancer Case 6 LCC 3D SIP Case 1 Example 4. RCC RCC 3D 6
Image Interpretation Invasive Lobular Carcinoma Each generated image depicts the objects (breast tissue, mass, microcalcifications) that are in focus in that particular plane. Objects that are above or below the plane will appear blurry to varying degrees (depending on their distance from the plane imaged). Microcalcificationswill rapidly become blurry and out of focus as the reader scrolls through the stack of images away from plane of the microcalcifications Roth R et al. RadioGraphics 2014 C-view C View Standard 2 D Synthesized image from the 3D tomosynthesis data Goal is for the C-view to replace the conventional digital mammographic image Currently,our plan is to still obtain conventional digital images along with the C-view and tomosynthesisimages: the combo mode TomosynthesisSlices Software Algorithm Generated Images 7
C-View Idea is that and 3D images complement one another in mammographic interpretation Helpful for comparison with prior 2 D conventional digital images C-view intended to preserve the important details from the tomosynthesis images What about the radiation?? Current radiation from DBT combined with the standard bilateral 2 view conventional mammography: Results in 2 x the current digital mammography dose Approximately 2.65 mgy-below the limits defined by the FDA (< 3 mgy) Dose will be cut in half if we abandon conventional for the C-view The Oslo Study. Comparison of Digital Mammography Alone and Mammography Plus Tomosynthesis in a Population Based Screening Program Skaane P et al. Radiology 2013 Prospective trial of over 12,000 examinations 27% increase in cancer detection rate utilizing DBT plus mammography (8.0 per 1000) vsmammography alone (6.1 per 1000) 15.1% decrease in the false positive rate for the DBT plus mammography arm PPV for recalled patients with cancers was similar for the two groups 40% increase in detection of invasive cancers for the DBT plus mammography arm (25 additional cancers) Mean interpretation time was 91 seconds for DBT plus mammography versus 45 seconds for mammography alone. Breast Cancer Screening Using Tomosynthesisin Combination with Digital Mammography JAMA 2014 8
Friedewald SM, et al. Objective: To determine if mammography in combination with tomosynthesisis associated with better performance of screening programs in the USA Retrospective analysis of 13 academic and nonacademic centers metrics of screening performance Examined 2 periods of time for each center: Period 1: digital screening mammography 1 year prior to implementation of tomosynthesis Period 2: Digital mammography + DBT from start dates of implementation of tomosynthesis Over 280,000 digital mammograms, over 170,000 digital mammograms + DBT Friedewald SM, et al. Overall decrease in recall rate by -16 per 1000 screens Decreased model adjusted biopsy rate for women screened with combination vs mammo alone (16.1 vs 19.3 per 1000) Overall increased cancer detection rate of 1.2% combined mammo+ DBT compared to mammoalone (Cancer detection rate for combination group: 5.4 per 1000) Increase in detection of invasive cancers utilizing the combination technique (1.2%) with no change in DCIS detection rate (1.4 per 1000). PPV for recall for mammo+ DBT: 6.4% (vs4.3% for mammo alone) PPV for biopsy for mammo+ DBT: 29.2% (vs24.2% for mammo alone) Friedewald SM, et al. Authors concluded that the addition of DBT to conventional digital mammography is associated with a decreased recall rate and an increase in cancer detection rate. Speculate that possibly the relative yield for each recall will increase Infer that increased cancer detection rate may optimize patient outcomes from mammographic screening Which patients will benefit most from tomosynthesis? Fatty or dense breast tissue or both groups? 9
Clearly, mammographically dense breast tissue will benefit. Who will benefit from tomosynthesis? WaldherrC., et al.: Examined 144 patients Sensitivity and negative predictive value of DBT was superior to digital mammography for women with fatty breasts and women with dense breasts Haas BM et al. : Utilization of DBT with conventional mammography decreased recall rates for all patients and all breast densities Statistically significant decrease in recall rate for scattered, heterogeneously dense, and dense breasts (P <0.05). Statistically significant decrease in recall rate for women < 40 yrs, 40-49 yrs, 50-59 yrs, and 60-69 yrs. Statistically significant differences for all subgroups EXCEPT those with fatty breast density and those patients > 70yrs. Greatest reductions in recall rates for patients with dense breasts and those < 50 yrs(less than 20 examinations needed to prevent one recall) Scattered Fibroglandular Breast Tissue LCC LCC 3D Applications of Tomosynthesis: Screening Mammography Rafferty EA et al: Improved diagnostic accuracy compared to digital mammo Decreased recall rate (6%- 67%) Improved sensitivity greatest for invasive cancers Roth R et al: Reduction in callback rate from 10.4% to 8.8% Odds Ratio: 1.24 the probability that a patient would be recalled from screening decreased by 24% with implementation of tomosynthesis + Trend toward increased cancer detection rate Rose SL et al: Decreased recall rate: 8.7% 5.5% (37% decrease) Increased cancer detection rate: 4.0 per 1000 screens 5.4 per 1000 screens (35% increase) Decrease in biopsy rates 15.2 13.5 per 1000 screens ( 11% decrease) 10
Applications of Tomosynthesis: Diagnostic Mammography Michel MJ et al.: The Area Under Curve (AUC) values for ROC demonstrated statistically significant (P= 0.0001) improvement in diagnostic accuracy in the arm utilizing tomosynthesiswith digital mammo(auc = 0.9671) versus in the arms without tomosynthesis(auc= 0.8949 and 0.7882). Effect significantly greater for soft tissue lesions than for microcalcifications Improved detection of architectural distortion and lesion margins Roth et al. : Several of their DBT-only cancers dissipated on standard spot compression views. Triangulation: Improved ability to determine the true 3D location of a target within the breast helpful with target for US May decrease the number of additional diagnostic views needed to triangulate a lesion may decrease overall radiation dose? Helpful with localizing lesions to the skin (Roth et al). Improved ability to estimate the extent of disease within the breast. Improved lesion characterization Improved lesion characterization Patient presenting for diagnostic mammography with a palpable abnormality Conventional CC mammogram demonstrating heterogeneous fibroglandular tissue without mammographic abnormality DBT image demonstrates a oval shaped mass with a possible capsule and containing both fat and dense breast tissue Published in: "Digital Breast Tomosynthesis in the Analysis of Fat-containing Lesions" RadioGraphics Vol. 34, No. 2: 343-358 RSNA, 2014 Published in: "Digital Breast Tomosynthesis in the Analysis of Fat-containing Lesions" RadioGraphics Vol. 34, No. 2: 343-358 RSNA, 2014 11
Improved lesion characterization Magnified CC DBT image clearly demonstrates a pseudocapsule surrounded by fat: Classic for a Hamartoma. Invasive Lobular Carcinoma: Multifocality better delineated with the DBT images Published in: "Digital Breast Tomosynthesis in the Analysis of Fat-containing Lesions" RadioGraphics Vol. 34, No. 2: 343-358 RSNA, 2014 Benefits of DBT Increased cancer detection rate Decreased recall rate Increased PPV for recall from DBT and increased PPV for biopsy based on DBT findings From the literature, these benefits hold true both in academic and community settings. Criticisms/ Disadvantages/ Questions Disadvantage: no magnification images possible at this time currently limits our diagnostic evaluation of microcalcifications Variable data regarding detection and characterization of microcalcifications in the literature Large coarse calcifications create artifacts on multiple slices Not intended currently for large breasted women: image may be clipped. False positives (Roth et al): Benign masses that would have been concealed by superimposition (lymph nodes, cysts) are detected more easily may prompt recall and further work up (Roth et al). Architectural distortion from benign lesions (radial scars) is much more evident. False negatives (Roth et al): Not everything is detectable by DBT! If a cancer does not have classic mass-like features or architectural distortion (ILC), it may not be visualized by DBT. Illustrates importance of additional evaluation with US and MRI. 12
Criticisms/ Disadvantages/ Questions Future Directions Time to interpret the images As with most new imaging modalities, there is a learning curve (Roth et al): Increased recall rate initially Recall of benign masses (LN, cysts) and architectural distortion (radial scars) HUGE files for storage: almost 10x greater than a standard 4 view mammogram Lack of long term follow up from all of the studies (Pisano E and YaffeM): limited ability to accurately determine false negative results rates, interval cancer rate and overdiagnosis Need for a randomized controlled trial like the DMIST trial (demonstrated superiority of digital mammography to film screen mammography) Tomosynthesis guided biopsy and localization Magnification views? CAD for DBT CPT code? Tomosynthesis guided biopsy Conclusions A lot of new data supporting the usage of DBT in both screening and diagnostic settings There remain several concerns, disadvantages that will hopefully be elucidated with further studies DBT is here at Baptist! 13
References BleyerA., Welch HG. Effect of Three Decades of Screening Mammography on Breast-Cancer Incidence. NEJM. 2012; 367: 1998-2005. FriedewaldSM, Rafferty EA, Rose SL, et al. Breast Cancer Screening Using Tomosynthesisin Combination With Digital Mammography. JAMA. 2014;311(24):2499-2507. Haas BM, KairaV, Geisel J, et al. Comparison of TomosynthesisPlus Digital Mammography and Digital Mammography Alone for Breast Cancer Screening. Radiology 2013; 269(3): 694-700. MichellMJ, IqbalA, WasanRK, et al. A comparison of the accuracy of film-screen mammography, full field digital mammography, and digital breast tomosynthesis. ClinRadiol2012; 67 (10): 976-981. Park JM et al. Breast Tomosynthesis: Present considerations and Future Applications. RadioGraphics2007; 27: S231-S240. Pisano E and YaffeM. Breast Cancer Screening: Should TomosynthesisReplace Digital Mammography? JAMA; 2014; 311(24): 2488-2489. Rafferty EA., Park JM., PhilpottsLE, et al. Assessing radiologist performance using combined digital mammography and breast tomosynthesiscompared with digital mammography alone: results of a multicenter, multireadertrial. Radiology 2013; 266 (1): 104-113. Rose SL, Tidwell AL, BujnochLI, et al. Implementation of Breast Tomosynthesisin a Routine Screening Practice: An Observational Study. AJR. 2013; 200(6): 1401-8. Roth R, MaidmentA, WeinstenS., et al. Digital Breast Tomosynthesis: Lessons Learned from Early Clinical Implementation. RadioGraphics2014; 34: E89- E102. SkaaneP., BandosAI, GuillenR, et al. Comparison of digital mammography alone and digital mammography plus tomosynthesisin a population-based screening program. Radiology 2013; 267 (1): 47-56. TabarL., VitakB., Chen TH, et al. Swedish Two County Trial: impact of mammographic screening on breast cancer mortality during 3 decades. Radiology 2011; 260 (3): 658-663. WaldherrC., CernyR., AltermattH., et al. Value of One-View Breast TomosynthesisVersus Two View Mammography in Diagnostic Workup of Women With Clinical Signs and Symptoms and in Women Recalled From Screening. AJR 2013; 200: 226-231. 14