Advances in Clinical Medicine
Vol. 14  No. 03 ( 2024 ), Article ID: 82416 , 6 pages
10.12677/ACM.2024.143692

乳腺X线摄影及其新技术结合人工智能在乳腺癌诊断中的应用

穆亚斯尔·艾尼,陈 跃,刘宏伟*

新疆医科大学第五附属医院影像中心,新疆 乌鲁木齐

收稿日期:2024年2月8日;录用日期:2024年3月2日;发布日期:2024年3月12日

摘要

乳腺癌是占我国女性发病率和死亡率首位的恶性肿瘤。早期发现、早期治疗不仅可以降低乳腺癌患者的死亡率,还可以提高患者五年生存率及生活质量。乳腺钼靶X线检查(mammography, MG)是目前唯一被证明可以降低乳腺癌患者死亡率的检查方法。其衍生出的数字乳腺断层摄影(digital breast tomosynthesis, DBT)、对比增强乳腺X线摄影(contrast-enhanced mammography, CEM)和对比增强数字化乳腺断层摄影(contrast-enhanced digital breast tomosynthesis, CEDBT)等新检查技术分别展现了独特的优势。人工智能(artificial intelligence, AI)作为当今科技发展的核心技术之一,在乳腺影像领域也取得了长足的进展。本文就人工智能结合乳腺X线检查及其新技术在乳腺癌诊断中的发展和应用现状进行分析及综述。

关键词

乳腺钼靶X线检查,人工智能,数字乳腺断层摄影,对比增强乳腺X线摄影,对比增强数字化乳腺断层摄影

Application of Mammography and Its Novel Techniques Combined with AI in the Diagnosis of Breast Cancer

Muyassar Gheni, Yue Chen, Hongwei Liu*

Imaging Center, The Fifth Affiliated Hospital of Xinjiang Medical University, Urumqi Xinjiang

Received: Feb. 8th, 2024; accepted: Mar. 2nd, 2024; published: Mar. 12th, 2024

ABSTRACT

Breast cancer is the malignant tumour that accounts for the first place in the morbidity and mortality among women in China. Early detection and treatment can reduce the mortality rate of breast cancer patients, in addition to improving their five-year survival rate and quality of life. Mammography (MG) is currently the only screening method proven to reduce the mortality rate of breast cancer patients. Its derivatives, such as digital breast tomosynthesis (DBT), contrast-enhanced mammography (CEM) and contrast-enhanced digital breast tomosynthesis (CEDBT) have shown unique advantages. As one of the core technologies in today’s scientific and technological development, artificial intelligence has also made great progress in the field of breast imaging. This article analyses and reviews the current development and application of AI combined with MG and its new techniques in breast cancer diagnosis.

Keywords:Mammography, Artificial Intelligence, Digital Breast Tomosynthesis, Contrast-Enhanced Mammography, Contrast-Enhanced Digital Breast Tomosynthesis

Copyright © 2024 by author(s) and Hans Publishers Inc.

This work is licensed under the Creative Commons Attribution International License (CC BY 4.0).

http://creativecommons.org/licenses/by/4.0/

1. 引言

世界卫生组织国际癌症研究机构(International Agency for Research on Cancer, IARC)发布的最新资料表明:2020年,乳腺癌已经超过肺癌成为全球最流行的恶性肿瘤,也是全球女性发病率和死亡率最高的恶性肿瘤。2020年乳腺癌在我国新发病数例达42万,是中国女性恶性肿瘤发病率之首,并且呈逐年增加的趋势 [1] 。虽然有研究表明女性发育迟缓、初产年龄偏高、哺乳时间短、口服避孕药及家族遗传等因素会增加乳腺癌的患病风险,但乳腺癌的发病原因还不明确 [2] ,所以对于乳腺癌的防范和治疗是以早期筛查和积极治疗为主。使用科学且有效的检测方法对适龄女性进行乳腺癌筛查,不仅可以有效降低乳腺癌的发病率和死亡率,还能延长此类患者的生存期,保障女性的健康和生命安全。相关研究表明对于乳腺癌早期的患者进行干预治疗,可以使患者5年生存率达到87% [3] 。本文就乳腺癌现有X线影像学诊断手段及新型的人工智能辅助诊断系统的价值进行阐述,旨在为有效辅助诊断医师发现早期乳腺癌提供参考,并为相关新辅助诊断系统提供进一步的探索思路。

2. 乳腺钼靶及其新技术与AI的结合

2.1. 乳腺钼靶

MG是目前唯一被证明可降低乳腺癌患者死亡率的方法, 也是我国抗癌协会乳腺癌诊疗指南推荐的首选检查方法 [4] [5] 。其使用X射线照射乳房、通过收集穿透乳房的残余X射线生成组织重叠的图像。根据不同区域的射线信息可确定该乳房区域内的正常组织和肿块、钙化、不对称致密影、结构扭曲及肿大的腋窝淋巴结等异常病变。乳腺疾病的检查和评估一般采用乳腺影像报告和数据系统(Breast Imaging Reporting and Data System, BI-RADS)。即使数字化乳腺摄影(digital mammography, DM)的应用进一步提高了图像的清晰度和对比度,重叠的纤维腺体组织对非钙化性病变的遮蔽仍对病变的诊断造成较大影响。乳腺X线检查新技术及基于影像组学的人工智能辅助诊断系统有望改进乳腺疾病的诊疗策略。

早在上个世纪60年代便已出现可用于乳腺癌影像诊断的计算机辅助诊断(Computer-Aided Diagnosis, CAD)系统 [6] 。智慧医疗飞速发展的今天,多种基于人工智能的辅助诊断系统在病灶的检出、分割、良恶性分类、BI-RADS分级等方面的有效性已在临床试验中被证实 [7] [8] 。Mao等 [9] 在2021年的一项多中心研究中用四种机器学习算法以0.92的AUC证明了在乳腺X线摄影中应用影像组学模型可以预测乳腺癌的风险。Kim等 [10] 开展的一项纳入了170,230例乳腺X射线检查结果的国际研究验证了基于大规模数据开发的人工智能模型在肿块探测、不对称致密影和扭曲结构识别方面表现出更好的诊断性能,也让放射科医生的工作表现有了明显改善。Lauritzen等 [11] 用人工智能对研究所涉及影像资料进行0~10评分,代表恶性肿瘤风险。正常的检查(评分 < 5)将被排除在放射科医生的阅读范围之外,可疑的检查(评分 > 召回阈值)将被召回。这使得放射科医生的工作量减少了62.6%。

2.2. 数字乳腺断层摄影

DBT于2011年获得美国食品药品监督管理局应用许可,通常与DM相结合应用于乳腺疾病的筛查和诊断。乳腺DBT成像是球管沿一定角度旋转进行的三维扫描过程,再经过计算机后处理重建出一系列高分辨率的断层图像。重建后的薄层图像可以以独立或连续的形式显示。DBT可以显著减少甚至消除由于腺体组织重叠所带来的影响,因而在乳腺中的肿块、结构扭曲、非对称性致密等非钙化性病变的检出具有比DM更高的敏感性和特异性 [12] [13] 。目前关于DBT的共识主要集中于该技术明显提高了无症状人群中的早期、微小浸润性乳腺癌的检出率 [14] [15] ,降低了乳腺检查结果的假阳性率和召回率 [16] 。

虽然DBT可能会提高癌症检出率,但一些研究表明,DBT检测到的大多数癌症恶性程度不高,这引起了人们对过度诊断的担忧 [17] 。目前可用的数据 [18] 显示,与DM相比,接受DBT筛查的妇女的间期癌症没有显著减少。然而,最近一项大型的单一机构研究 [19] 表明,与传统的DM相比,在第一轮和次级筛查中,DBT都检测到了预后不佳的侵袭性癌症。尽管有大量已公布的临床数据和广泛的临床使用,但与DM相比,DBT是否也能改善乳腺癌患者的短期和长期预后仍然不确定。研究人员提出,这可能是由于DBT检测到的主要是预后较理想的早期Luminal A型乳腺癌,不太可能影响患者的长期健康 [20] 。

由于同时行DBT和DM检查会增加辐射剂量,研究人员在DBT的基础上开发了合成二维乳腺X线摄影(synthesized two-dimensional mammograms, SM)。SM算法设置增强了钙化、肿块边缘棘状突起及结构扭曲等病变的显示 [12] 。由于SM由DBT重建而成,SM联合DBT的辐射剂量即为DBT的辐射剂量,也无需额外的图像采集时间。国外一项双中心回顾性研究 [21] 显示,DBT + SM取代DBT + DM后,乳腺癌的筛查效能仍保持在基准线之内。最近的一项研究 [22] 涉及的辅助诊断系统根据软组织密度的可疑程度生成CAD增强的合成图像,以此避免读者来回翻阅DBT图像。这一应用在灵敏度、特异度和召回率没有受到影响的情况下将读图时间减少了23%。

DBT算法的持续进步,如基于深度学习的重建,将提供具有更少图像噪音和更少伪影的高空间分辨率(合成)图像。国外一项研究 [23] 显示采用人工智能增强的6 mm层厚的切片代替1 mm层厚的重建方案可在不降低判读准确性的同时大大缩短DBT判读时间。DBT包含的图像是DM的100多倍以上,而且恶性特征通常只能在几个层面中可见。这种大的数据量和小的、微妙的发现的组合导致深度学习模型设计过程中对感兴趣区域(ROI)的分割、特征提取和建模过程更复杂。因此打破技术壁垒,优化DBT的性能是普及DBT在临床应用的必经之路。

2.3. 对比增强乳腺X线摄影

CEM在DM和DBT提供的解剖和形态信息的基础上增加了功能信息。肿瘤血管生成在原发性乳腺癌中的临床重要性是众所周知的。研究 [24] 表明,肿瘤内微血管生成是一个独立的预后指标,它与较高的转移发生率相关。CEM利用乳腺肿瘤细胞增殖快且分泌多种肿瘤血管生长因子刺激病灶产生新生血管,瘤体周围富集对比剂,真实地反映乳腺病灶血管网的分布及血流动力学特征,来提高乳腺疾病病灶的检出率及诊断准确率 [25] 。CEM检查需要双能量乳腺X线检查的基础上静脉注射碘化造影剂,获取标准的头尾位和斜侧位乳腺X线影像。每组影像由类似于DM的“低能量”图像和通过使用高于碘k边的千电子伏特来增强对比材料信号获得的“高能量”图像组成。低能量图像相当于用于检测密度和钙化的DM,取代了常规乳腺X线检查 [26] 。经后处理低能量和高能量图像可以去除背景信号,重新组合的仅碘图像有助于识别增强病变,在致密性腺体及隐匿性病灶中具有独特的优势 [27] 。

虽然CEM结合了DM (X射线剂量和对乳腺的挤压)和乳腺核磁共振成像(需要静脉注射造影剂)的缺点,但它可以克服DM中组织重叠的影响,并能够检测类似于MRI的肿瘤新生血管相功能信息,同时保持DM的高图像分辨力 [28] 。Sorin等 [29] 的研究显示,在611名腺体致密的中等乳腺癌风险妇女中,用CEM与DM相比,癌症的检测率提高了13.1‰。Sung等 [30] 报告了在904例患者的CEM检查中发现了6例未能在DM中发现的癌症,补充检出率为6.6‰。即使CEM的临床应用尚未广泛普及,研究人员已开始研发相关人工智能系统。国内一项研究 [31] 显示病灶形态学特征与不同分子亚型之间的关系,来判断乳腺肿瘤的生物学行为及预后,为临床提供精准治疗及术前间接评估预后提供参考。Marino等 [32] 通过将影像组学应用于CEM图像中,鉴别乳腺癌的侵袭性、激素受体状态及肿瘤分期,准确率从78.4%到100%。Petrillo等 [33] 的研究结果证实,从CEM图像中提取的影像组学纹理特征可以提供有关肿瘤性质和分级以及肿瘤的分子亚型高度相关的信息。这也证明人工智能算法与影像组学分析的概念相结合可以成功创建支持医生乳腺癌诊断决策的工具。

2.4. 对比增强乳腺X线摄影

CEDBT结合DBT三维成像的特征和CEM对病灶增强显示的特征,提供三维的对比增强影像。与CEM相比,CEDBT可以提供更准确的病变细节,如病灶形态、大小和位置。Huang等 [34] 的研究指出CEDBT的几项临床应用潜力:① 病灶特征的三维显示可能有利于评估新辅助化疗后的治疗反应,② 三维的对比增强和病灶的解剖信息为活检提供更好的指导,③ 与CEM + DBT相比,CEDBT的辐射计量有所减低,④ 与SM类似合成CEM可以由CEDBT数据中合成,并提供与CEM类似的病变对比增强信息。Chou等 [35] 开展的一项研究评估了CEM、CEDBT及DCE-MRI的AUC并表明通过将CEDBT添加到CEM,诊断准确性在统计学上没有显著提升。这也引起大家对CEDBT是否可以通过提供更好的病变特征和避免不必要的活检来合理化它的额外辐射剂量等问题的思考。国内外对于CEDBT的诊断性能及相关人工智能的研究非常少见。

3. 结语

乳腺癌的影像诊断是乳腺癌二级预防的重要手段。乳腺X线检查新技术在乳腺癌良恶性的判断、术前分期、新辅助化疗效果评估及复发转移风险预测等方面有广阔的发展前景。人工智能技术在大数据时代的优势是不容忽视的。虽然当前各项新技术的有效性和人工智能技术的可靠性均存在不少争议,但不得不确定的是随着技术的进步,人机结合是影像诊断领域的必然发展趋势。

文章引用

穆亚斯尔·艾尼,陈 跃,刘宏伟. 乳腺X线摄影及其新技术结合人工智能在乳腺癌诊断中的应用
Application of Mammography and Its Novel Techniques Combined with AI in the Diagnosis of Breast Cancer[J]. 临床医学进展, 2024, 14(03): 246-251. https://doi.org/10.12677/ACM.2024.143692

参考文献

  1. 1. Sung, H., Ferlay, J., Siegel, R.L., et al. (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA: A Cancer Journal for Clinicians, 71, 209-249. https://doi.org/10.3322/caac.21660

  2. 2. Kashyap, D., Pal, D., Sharma, R., et al. 2022 () Global Increase in Breast Cancer Incidence: Risk Factors and Preventive Measures. BioMed Research International, 2022, Article ID: 9605439. https://doi.org/10.1155/2022/9605439

  3. 3. Vanni, G., Pellicciaro, M., Materazzo, M., et al. (2020) Breast Cancer Diagnosis in Coronavirus­Era: Alert from Italy. Frontiers in Oncology, 10, Article 938. https://doi.org/10.3389/fonc.2020.00938

  4. 4. Telli Melinda, L., Gradishar William, J. and Ward John, H. (2019) NCCN Guidelines Updates: Breast Cancer. Journal of the National Comprehensive Cancer Network, 17, 552-555.

  5. 5. 国家卫生健康委员会医政医管局. 乳腺癌诊疗指南(2022年版) [J]. 中华肿瘤杂志, 2023, 45(10): 803-833.

  6. 6. Winsberg, F., Elkin, M., Josiah Macy, J., Bordaz, V. and Weymouth, W. (1967) Detection of Radio-graphic Abnormalities in Mammograms by Means of Optical Scanning and Computer Analysis. Radiology, 89, 211-215. https://doi.org/10.1148/89.2.211

  7. 7. (2023) Breast-Cancer Screening Gets a Boost from AI. Nature, 620, Article 471. https://doi.org/10.1038/d41586-023-02526-4

  8. 8. Freeman, K., Geppert, J., Stinton, C., et al. (2021) Use of Artifi-cial Intelligence for Image Analysis in Breast Cancer Screening Programmes: Systematic Review of Test Accuracy. BMJ, 374, n1872. https://doi.org/10.1136/bmj.n1872

  9. 9. Mao, N., Yin, P., Wang, Q., Liu, M., Dong, J., Zhang, X., Xie, H. and Hong, N. (2019) Added Value of Radiomics on Mammography for Breast Cancer Diagnosis: A Feasibility Study. Journal of the American College of Radiology, 16, 485-491. https://doi.org/10.1016/j.jacr.2018.09.041

  10. 10. Kim, H.E., Kim, H.H., Han, B.K., et al. (2020) Changes in Cancer Detection and False-Positive Recall in Mammography Us-ing Artificial Intelligence: A Retrospective, Multireader Study. The Lancet Digit Health, 2, E138-E148. https://doi.org/10.1016/S2589-7500(20)30003-0

  11. 11. Lauritzen, A.D., Rodríguez-Ruiz, A., Von Euler-Chelpin, M.C., Lynge, E., Vejborg, I., Nielsen, M., Karssemeijer, N. and Lillholm, M. (2022) An Artificial Intelligence-Based Mammography Screening Protocol for Breast Cancer: Outcome and Radiologist Workload. Radiology, 304, 41-49. https://doi.org/10.1148/radiol.210948

  12. 12. Bae, M.S. and Moon, W.K. (2018) Is Synthetic Mammography Com-parable to Digital Mammography for Detection of Microcalcifications in Screening? Radiology, 289, 639-640. https://doi.org/10.1148/radiol.2018181961

  13. 13. Hofvind, S., Holen, Å.S., Aase, H.S., et al. (2019) Two-View Digital Breast Tomosynthesis versus Digital Mammography in a Population-Based Breast Cancer Screening Programme (To-Be): A Randomised, Controlled Trial. The Lancet Oncology, 20, 795-805. https://doi.org/10.1016/S1470-2045(19)30161-5

  14. 14. Ikeda, D.M., Birdwell, R.L., O’Shaughnessy, K.F., Brenner, R.J. and, Sickles, E.A. (2003) Analysis of 172 Subtle Findings on Prior Normal Mammograms in Women with Breast Cancer Detected at Follow-Up Screening. Radiology, 226, 494-503. https://doi.org/10.1148/radiol.2262011634

  15. 15. Su, X.H., Lin, Q., Cui, Ch.X., et al. (2017) Non-Calcified Ductal Carcinoma in Situ of The Breast: Comparison of Diagnostic Accuracy of Digital Breast Tomosynthesis, Digital Mam-mography, and Ultrasonography. Breast Cancer, 24, 562-570. https://doi.org/10.1007/s12282-016-0739-7

  16. 16. Marinovich, M.L., Hunter, K.E., Macaskill, P. and Houssami, N. (2018) Breast Cancer Screening Using Tomosynthesis or Mammography: A Meta-Analysis of Cancer Detection and Recall. JNCI: Journal of the National Cancer Institute, 110, 942-949. https://doi.org/10.1093/jnci/djy121

  17. 17. Bahl, M., Mercaldo, S., Dang, P.A., McCarthy, A.M., Lowry, K.P. and Lehman, C.D. (2020) Breast Cancer Screening with Digital Breast Tomosynthesis: Are Initial Benefits Sustained? Radiology, 295, 529-539. https://doi.org/10.1148/radiol.2020191030

  18. 18. Kim, J.Y., Kang, H.J., Shin, J.K., et al. (2017) Biologic Profiles of Invasive Breast Cancers Detected Only with Digital Breast Tomosynthesis. American Journal of Roentgenology, 209, 1411-1418. https://doi.org/10.2214/AJR.17.18195

  19. 19. Lång, K. (2019) The Coming of Age of Breast Tomosyn-thesis in Screening. Radiology, 291, 31-33. https://doi.org/10.1148/radiol.2019190181

  20. 20. Houssami, N. (2022) Should Tomosynthesis Replace Mammogra-phy for Breast Cancer Screening? The Lancet Oncology, 23, 554-555. https://doi.org/10.1016/S1470-2045(22)00215-7

  21. 21. Herschorn, S.D, Zuckerman, S., Sprague, B., et al. (2020) Multicenter Evaluation of Breast Cancer Screening with Digital Breast Tomosynthesis in Combination with Synthetic versus Digital Mammography. Radiology, 297, 495-742. https://doi.org/10.1148/radiol.2020200240

  22. 22. Balleyguier, C., Arfi-Rouche, J., Levy, L., et al. (2017) Improving Digital Breast Tomosynthesis Reading Time: A Pilot Multi-Reader, Multi-Case Study Using Concurrent Comput-er-Aided Detection (CAD). European Journal of Radiology, 97, 83-89. https://doi.org/10.1016/j.ejrad.2017.10.014

  23. 23. Sauer, S.T., Christner, S.A., Kuhl, P.J., et al. (2023) Artifi-cial-Intelligence-Enhanced Synthetic Thick Slabs versus Standard Slices in Digital Breast Tomosynthesis. British Journal of Radiology, 96, Article ID: 20220967. https://doi.org/10.1259/bjr.20220967

  24. 24. Dey, M., Ayan, B., Yurieva, M., Unutmaz, D. and Ozbolat, I.T. (2021) Studying Tumor Angiogenesis and Cancer Invasion in a Three-Dimensional Vascularized Breast Cancer Mi-cro-Environment. Advanced Biology, 5, e2100090. https://doi.org/10.1002/adbi.202100090

  25. 25. Yuen, S., Monzawa, S., Gose, A., Yanai, S. and Yata, Y. (2022) Im-pact of Background Parenchymal Enhancement Levels on the Diagnosis of Contrast-Enhanced Digital Mammography in Evaluations of Breast Cancer: Comparison with Contrast-Enhanced Breast MRI. Breast Cancer, 29, 677-687. https://doi.org/10.1007/s12282-022-01345-1

  26. 26. Fallenberg, E.M., Dromain, C., Diekmann, F., et al. (2014) Con-trast-Enhanced Spectral Mammography: Does Mammography Provide Additional Clinical Benefits or Can Some Radia-tion Exposure Be Avoided? Breast Cancer Research and Treatment, 146, 371-381. https://doi.org/10.1007/s10549-014-3023-6

  27. 27. 文婵娟, 徐维敏, 曾辉, 等. 对比增强X线摄影对乳腺可疑病变的诊断价值[J]. 中华放射学杂志, 2019, 53(9): 737-741.

  28. 28. Dromain, C., Balleyguier, C., Muller, S., et al. (2006) Evaluation of Tumor Angiogenesis of Breast Carcinoma Using Contrast-Enhanced Digital Mammography. American Journal of Roentgenology, 187, W528-W537. https://doi.org/10.2214/AJR.05.1944

  29. 29. Sorin, V., Yagil, Y., Yosepovich, A., et al. (2018) Contrast-Enhanced Spectral Mammography in Women with Intermediate Breast Cancer Risk and Dense Breasts. American Journal of Roentgenology, 211, W267-W274. https://doi.org/10.2214/AJR.17.19355

  30. 30. Sung, J.S., Lebron, L., Keating, D., et al. (2019) Performance of Du-al-Energy Contrastenhanced Digital Mammography for Screening Women at Increased Risk of Breast Cancer. Radiology, 293, 81-88. https://doi.org/10.1148/radiol.2019182660

  31. 31. 李鸿恩, 谢汉民, 吕培锋, 等. 对比增强能谱乳腺X线摄影征象与不同分子亚型乳腺癌的相关性[J]. 现代肿瘤医学, 2023, 31(21): 4008-4014.

  32. 32. Marino, M.A., Leithner, D., Sung, J., Avendano, D., et al. (2020) Radiomics for Tumor Characterization in Breast Cancer Patients: A Feasibility Study Comparing Contrast-Enhanced Mammography and Magnetic Resonance Imaging. Diagnostics, 10, Article 492. https://doi.org/10.3390/diagnostics10070492

  33. 33. Petrillo, A., Fusco, R., Di Bernardo, E., et al. (2022) Prediction of Breast Cancer Histological Outcome by Radiomics and Artificial Intelligence Analysis in Contrast-Enhanced Mam-mography. Cancers, 14, Article 2132. https://doi.org/10.3390/cancers14092132

  34. 34. Huang, H., Scaduto, D.A., Liu, C., et al. (2019) Comparison of Contrast-Enhanced Digital Mammography and Contrast-Enhanced Digital Breast Tomosynthesis for Lesion Assessment. Journal of Medical Imaging, 6, Article ID: 031407. https://doi.org/10.1117/1.JMI.6.3.031407

  35. 35. Chou, C.P., Lewin, J.M., Chiang, C.L., et al. (2015) Clinical Evaluation of Contrast-Enhanced Digital Mammography and Contrast Enhanced Tomosynthesis—Comparison to Contrast-Enhanced Breast MRI. European Journal of Radiology, 84, 2501-2508. https://doi.org/10.1016/j.ejrad.2015.09.019

  36. NOTES

    *通讯作者。

期刊菜单