評估數位式乳房斷層攝影之平均乳腺劑量及影像品質：可壓迫假體之研究

Evaluation of average glandular dose and image quality in digital breast tomosynthesis: A compressible phantom study

陳瑜慧(Yu-Hui Chen)

指導教授 ： 董尚倫

中山醫學大學/醫學科技學院/生物醫學科學學系碩士班/碩士(2016年)

摘要

前言
近年來評估數位式乳房斷層攝影(digital breast tomosynthesis, DBT)、全景式乳房攝影(full-field digital mammography, FFDM)與重組合成乳房影像(synthesized mammogram, SM)之平均乳腺劑量(average glandular dose, AGD)和影像品質是一項熱門的臨床議題，臨床上是以硬質假體評估平均乳腺劑量與影像品質，本研究目的是應用可壓迫假體評估數位式乳房斷層攝影之平均乳腺劑量及影像品質。
材料與方法
本研究以軟質地Bolus材質設計可壓迫假體進行實驗，使用Mammomat Novation DR (SIEMENS)與Selenia Dimensions (Hologic)進行攝影。首先在厚度評估方面，以Bolus假體搭配放射治療劑量評估軟體(Doselab)建立數位影像為基礎的厚度評估方法(imaged-based breast thickness determination, IBBTD) method，用以修正系統顯示厚度(system indicated thickness, SIT)。其次進行劑量評估，將Bolus堆疊成不同厚度(20~80 mm)，以COMOBO模式進行左右乳房頭腳向(craniocaudal view, CC view)與內外斜位向(mediolateral oblique view, MLO view)攝影，再依歐洲規範計算one-view (MLO) DBT與two-view (CC與MLO) FFDM的AGD。接著進行影像品質評估，將不同大小之Testing objecting (0.127, 0.28與0.5 mm)包埋在不同厚度之Bolus(30, 60與90 mm)之上、中、下層，攝影後得到FFDM、SM與DBT之影像，由9位乳房攝影從業人員依據5分量表(5-steps-scale)進行閱片評分，以分析其總分、敏感性與準確度，以比較FFDM、SM與DBT之影像品質。
結果
在厚度評估方面，IBBTD法最大的厚度評估誤差為-2.4 mm (22.6-25)，小於品保測試規範中之厚度容許誤差±5 mm。在劑量評估方面，one-view DBT的AGD都比FFDM為低；舉例而言，在Bolus假體厚度為60 mm時，one-view DBT的AGD為1.49 mGy而two-view FFDM的AGD為2.75 mGy。在影像品質評估方面，DBT的上層與下層之影像評分沒有顯著差異(p>0.93)。在相同影像種類時，針對不同Testing object size而言，總分、敏感性與準確度都會隨著Testing object大小增加而上升；同時隨著假體厚度增加而下降。在比較FFDM與SM之影像品質方面，FFDM的總分 (70~162)比SM (58~148)高；針對相同大小Testing object而言，FFDM與SM之評分差異會隨著假體厚度增加而加大。
結論
Bolus為片狀假體，非常適合用來發展乳房攝影用之可壓迫假體，結果顯示IBBTD方法的誤差是可以接受的並可用來評估二維壓迫厚度之分布。在劑量評估方面，使用one-view (MLO)DBT攝影時可以降低平均乳腺劑量。在影像方面， FFDM之影像品質優於SM，同時Bolus假體厚度會影響閱片者之評分結果。上述假體實驗之結果將有助於日後進一步探討臨床上使用SM影像來降低DBT檢查中需外加FFDM攝影之需求，以達成抑低病患輻射劑量之策略。

關鍵字

數位式乳房斷層攝影 ； Bolus假體 ； 可壓迫假體 ； 平均乳腺劑量 ； 影像品質

並列摘要

Purpose
In recent years, the assessment of average glandular dose (AGD) and image quality for digital breast tomosynthesis (DBT), full-field digital mammography (FFDM), and synthesized mammogram (SM) is an emerging clinical issue. At present, the AGD and image quality of mammography are evaluated with rigid construction phantoms. The purpose of this study is the assessment of AGD and image quality of DBT using a compressible phantom.
Materials and Methods
In this study, Bolus soft slabs were used to design the compressible phantom. Two mammography systems, Novation DR (Siemens) and Selenia Dimensions (Hologic), were used. For thickness assessment, the imaged-based breast thickness determination (IBBTD) method was developed by using Bolus phantom and dose calculation program (Doselab) in radiation therapy. The system indicated thickness (SIT) of mammography unit was corrected. For dose assessment, difference thicknesses of Bolus phantom (20~80 mm) were imaging with craniocaudal (CC) and mediolateral oblique (MLO) view using COMBO mode. AGD values of one-view (MLO) DBT and two-view (CC and MLO) were calculated using the European protocol. To evaluate image quality, several testing objects (0.127, 0.28 and 0.5 mm) were embedded in bolus phantoms (30, 60 and 90 mm) at difference layers. The phantoms were imaged and DBT, FFDM, and SM images were acquired. A total of 9 readers participated in the assessment of image quality using the 5-steps-scale image criteria. The scores, sensitivity, and accuracy of DBT, FFDM, and SM images were analyzed and compared
Results
In thickness assessment, the maximum error produced from the IBBTD method was -2.4 mm (22.6-25) which is less than the tolerable thickness error ±5 mm in the quality control testing. For dose assessment, AGD of one-view (MLO) DBT is lower than two-view (CC and MLO) FFDM. For a 60 mm bolus phantom, the AGD values of one-view (MLO) DBT and two-view (CC+MLO) FFDM were 1.49 and 2.75 mGy, respectively. For image quality assessment, the total scores of upper and lower layers shows insignificant (p>0.93). For the same type of image, the total scores, sensitivity and accuracy increased with increasing size of testing object and decreased with increasing thickness of bolus phantom. For comparison of image quality, the total scores (70~162) of FFDM images was higher than the total scores (58~148) of SM images. For the same size of testing object, the difference on total scores between FFDM and SM images were increased with the increasing thickness of Bolus phantoms.
Conclusion
Bolus phantom is a slab structure which is suitable for developing compressible phantom in mammography. Results from this study show that the thickness error produced from IBBTD method is acceptable and 2D compressed thickness distribution can be acquired in mammography. For dose assessment, AGD reduction can be achieved by applying one-view (MLO) DBT alone in clinical practice. For image assessment, the image quality of FFDM is better than SM and reader interpretation is affect by the thickness of bolus phantom. The results mentioned above from phantom study may be helpful for the achievement of the dose reduction strategy with SM image by decreasing the need for additional FFDM during DBT-based imaging.

並列關鍵字

digital breast tomosynthesis ； Bolus phantom ； compressible phantom ； average glandular dose ； image quality

參考文獻

(引用時間: 2016/4/4)

4. Skaane P, Young K and Skjennald A. Population-based Mammography Screening: Comparison of Screen-Film and Full-Field Digital Mammography with Soft-Copy Reading—Oslo I Study 1. Radiology 2003; 229: 877-884.

5. Waldherr C, Cerny P, Altermatt HJ, Berclaz G, Ciriolo M, Buser K and Sonnenschein MJ. Value of one-view breast tomosynthesis versus two-view mammography in diagnostic workup of women with clinical signs and symptoms and in women recalled from screening. American Journal of Roentgenology 2013; 200: 226-231.

6. Wallis MG, Moa E, Zanca F, Leifland K and Danielsson M. Two-view and single-view tomosynthesis versus full-field digital mammography: high-resolution X-ray imaging observer study. Radiology 2012; 262: 788-796.

7. Gennaro G, Toledano A, Di Maggio C, Baldan E, Bezzon E, La Grassa M and Muzzio PC. Digital breast tomosynthesis versus digital mammography: a clinical performance study. European radiology 2010; 20: 1545-1553.