以像素值評估婦女乳房之乳腺密度

Evaluation of Breast Density with Pixel Assessment

趙真(Chen Chao)

指導教授 ： 鄭凱元

中臺科技大學/健康科學院/醫學影像暨放射科學系暨研究所/碩士(2012年)

https://doi.org/10.6822/CTUST.2012.00050

摘要

近年來國內外文獻均顯示乳房X 光攝影是乳癌最主要的篩檢工具。目前國內在進行乳房攝影影像之品質評估作業，皆利用50%脂肪-50%乳腺的假體進行測試及校正。本研究目的在於分析國內婦女乳腺-脂肪比例，經由研究結果進而選擇適合之假體進行儀器品管測試，以確保在不同的乳腺-脂肪比例下，儀器仍能維持最佳之影像品質及劑量輸出之穩定性。研究利用模擬人類乳房組成，內含不同平均腺體組織的乳房組織當量假體（tissue-equivalent phantom），以數位乳房X 光機進行攝影，分析假體中不同乳腺比例區域之像素平均值。以回溯的方式，將467 位婦女未經處理的乳房X 光影像（raw image），經由Matlab 程式推估乳腺-脂肪比例。結果顯示，467 位不同年齡層受檢者之乳腺-脂肪推估比例為：40-49 歲30.13%、
50-59 歲14.82%、60-69 歲8.88%。3 位放射師對臨床影像之判定等級，與經由像素值推估之乳腺-脂肪比例等級，p＜0.005，證明兩組判定等級的結果相近，且具相關性。研究結果顯示國內婦女乳腺-脂肪之比例並非50％-50％組成，如能選擇適合之假體進行儀器品管測試，才能調整適合之攝影參數並降低乳腺劑量，讓乳房攝影品管作業更臻完善。

關鍵字

像素值 ； 乳房攝影 ； 乳腺密度

並列摘要

Recent studies have indicated that mammography is still the major screening tool for breast cancer. The aim of this study was to analyze the glandular/adipose composition ratio in the breast of women in Taiwan to further select the suitable phantom for the quality assurance of mammography, and also ensure the best image quality and stable of the mammography equipment when analyzing various breast densities. In this study, the authors selected the Tissue-Equivalent Phantom with different glandular/adipose composition that mimics the human breast tissue and examined it through the digital mammography that took the image and analyzed it through pixel values in area with different degrees of glandular and adipose ratio compositions were evaluated through the Matlab system. The results showed the glandular compositions of the 467 tests women of different ages were: 30.13 % for 40-49 years old, 14.82 % for 50-59 years old and 8.88 % for 60-69 years old. The results also showed that the level of estimated glandular and adipose ratio determined by the three radiological technologists were similar, the Kendall’s coefficient of concordance test (p<0.005), with significant correlation. According to this result, the ratio of glandular and adipose in Taiwanese women was not composed of 50% to 50% ratio. If an appropriate phantom can be selected for mammography to lower the glandular dosage, the process and results of mammography can also be improved.

並列關鍵字

pixel value ； breast density ； mammography

參考文獻

[21] 2007國民健康局年報，行政院衛生署國民健康局 97 年9 月出版，第68頁。http://www.bhp.doh.gov.tw/

[1] 中華民國98年癌症登記報告，行政院衛生署國民健康局編印， 101年2月出版。http://www.bhp.doh.gov.tw/

[3] Norman F. Boyd , Lisa J. Martin , Johanna M. Rommens, Andrew D. Paterson, Salomon Minkin, Yaffe J. Martin, Jennifer Stone, and John L. Hopper, “Mammographic Density: A Heritable Risk Factor for Breast Cancer,” Methods of Molecular Biology, 2009; 472(2): 343-360.

[2] Thomas R. Nelsona, Laura I. Cervino, John M. Boone, and Karen K. Lindfors, “Classification of breast computed tomography data,” Medical Physics, 2008; 35(3): 1078-1086.

[5] Norman F Boyd, Johanna M Rommens, Kelly Vogt, Vivian Lee, John L Hopper, and Andrew D Paterson, “Mammographic breast density as an intermediate phenotype for breast cancer,” Lancet Oncol, 2005; 6: 798–808.