2012年北美放射学年会乳腺影像学研究进展

赵莉芸，周纯武，李静

2012年北美放射学年会(RSNA)共收录乳腺影像学摘要约339篇，其中大会发言133篇，科学展板82篇，教育展览124篇。从成像技术方面，涉及MRI的文章约占30%，其次分别为超声成像、数字X线摄影(digital mammography,DM)、数字合成乳腺X线体层成像(digital breast tomosynthesis,DBT)和乳腺分子影像(molecular breast imaging,MBI)。从研究内容方面，主要集中在乳腺成像和诊断新技术的应用、乳腺癌筛查、新辅助化疗(neoadjuvant chemotherapy,NAC)疗效评价和定位活检等，笔者将从以上四个方面分别进行阐述，以期对今后乳腺影像学研究提供参考。

1 乳腺成像和诊断新技术

1.1 分子影像

以X线摄影和超声为代表的形态学成像是乳腺传统的检查方法，磁共振扩散、灌注及波谱等功能成像也正逐渐应用于临床；而MBI可从细胞、分子甚至基因水平反映活体的生理和病理过程，目前已成为乳腺影像学研究的新热点。

2002年，Brem等[1]率先将乳腺专用伽马成像(breast-specific gamma imaging,BSGI)应用于临床并取得成功。BSGI具有无痛、易于被患者接受、检查时间短、比较经济等特点，因而适合于乳腺筛查。Coffey等[2]对364例X线表现正常的乳腺癌高危人群行BSGI，认为BSGI是常规乳腺筛查的有益补充，可使检出率提高2.5%，尤其对于致密型乳腺。Velez等[3]研究显示将BSGI引入传统乳腺筛查中，可提高乳腺癌和高危病变检出的敏感度，但特异度却无明显改变。除此之外，Rapelyea等[4]研究显示，BSGI上表现为弥漫性放射性摄取增高的乳腺浸润性导管癌，经保乳手术后阳性切缘的发生率较高，往往需要再次手术，因此术前BSGI可提示这类患者要扩大手术切除范围。PET领域的研究主要集中在乳腺癌治疗效果评价和预后预测方面。Koo等[5]回顾性分析了548例乳腺癌患者的PET-CT图像，发现不同分子亚型的乳腺癌，其18F-FDG摄取也不同，三阴性乳腺癌FDG摄取最高(SUVmax 9.55±5.90)，而Luminal A型者最低(SUVmax 4.62±3.42)。说明PET-CT可从分子水平反映疾病的生物学特性。Kiyoto等[6]分析了87例乳腺癌(90个病灶)患者NAC前后PET-CT上SUVmax，结果显示ΔSUVmax>80.8%往往提示为完全病理缓解者(pathological complete response,pCR)，且可获得相对延长的无病生存(disease-free survival，DFS)。Nakajima等[7]对49例经过乳腺切除和术后放疗的患者行术前PET-CT检查，认为治疗前的SUV-max是这类患者预后的独立预测指标。

光学分子成像仍处于发展的初期，是分子影像领域新的研究方向。Ahn等[8]的研究认为，与单独超声检查相比，扩散光学成像(optical diffusion imaging)联合超声可明显提高乳腺病变诊断的准确度和特异度，减少假阳性的产生。Choi等[9]对240例乳腺癌患者行术前扩散光学成像和PET-CT检查，结果显示扩散光学成像参数氧饱和度(SO2)与SUVmax存在相关性，SO2高者，SUVmax也相应较高，NAC后pCR率也较高。Sklair-Levy等[10]采用一种新型红外线 (RUTH) 成像技术获得了乳腺超分辨率3D血管图像，结果显示乳腺浸润性癌的血管密度评分高于正常乳腺，而大多数患者血管图像中显示的异常增多血管是人眼无法识别的；因此，认为这种快速、非侵入性的成像方式有望检出早期乳腺癌。

1.2 MRI

扩散加权成像(diffusion weighted imaging，DWI)在乳腺MRI的应用已较为广泛，DWI与动态增强MRI联合可提高乳腺癌诊断的准确度、特异度和阳性预测值[11-13]；DWI亦可鉴别乳腺浸润性癌与导管原位癌[14]。Pediconi等[15]研究显示DWI联合T2 IDEAL序列评价乳腺癌NAC疗效的敏感度和特异度等同于MRI动态对比增强，而前者不需注射对比剂。Schiani等[16]研究认为ADC值与乳腺癌预后因子(ER、PR、HER2和Ki-67)间存在相关性，Benveniste等[17]的研究结果则不然。

扩散张量成像(diffusion tensor imaging，DTI)已广泛应用于中枢神经系统，而在乳腺的应用鲜有报道。由于乳腺导管的平均直径(约60 μm)要比脑白质纤维束(1～2 μm)大得多，因此需延长扩散时间来充分实现各向异性。Cho等[18]行6个方向扩散敏感梯度的乳腺DTI，结果显示不同扩散时间获得的各向异性指数FA和扩散特征值λ3有显著性差异，因而证实延长扩散时间可增加乳腺组织对各向异性的敏感度，有望鉴别正常乳腺组织与恶性肿物及获得乳腺导管结构的扩散张量图。Penn等[19]对36例患者行DTI检查，发现ADC值联合DMB(diffusion morphological blooming)有助于鉴别乳腺良恶性肿瘤，ROC曲线下面积达0.90(单独应用ADC值为0.77)；而与对比增强MRI相比，DTI不需注射对比剂、费用更低。

1.3 超声剪切波弹性成像(shear wave elastography，SWE)

SWE定量参数(如弹性评分、弹性指数等)的测定具有可重复性，常规超声检查联合SWE可提高乳腺肿瘤诊断的准确度和特异度，而不降低敏感度[20-24]。Evans等[25]分析了266例乳腺癌患者术前SWE，有17例表现为良性，其中5例为导管原位癌(5/11)，12例为浸润性癌(12/255)，可见导管原位癌在SWE上更容易被误诊为良性病变，而SWE上表现为良性征象的浸润性癌在组织学上多为Ⅰ级且易被X线检出；而这些SWE误诊为良性的病例在常规超声检查中不易被误诊，因此二者联合可提高病变诊断的准确度。Chang等[26]研究了乳腺浸润性癌SWE上弹性指数与肿瘤大小、受体状态、组织学类型和分级之间的相关性，结果显示Luminal A型乳腺癌的平均弹性指数最低，肿瘤体积较大、组织学分化较差和三阴性乳腺癌的平均弹性指数则较高。

2 乳腺筛查

X线摄影和超声检查是最基本的乳腺影像学筛查手段，MRI在乳腺筛查，特别是乳腺癌高危人群和致密型乳腺者筛查中的价值也逐渐被肯定。2007年美国癌症协会的筛查指南建议乳腺癌高危人群需行每年一次MRI检查作为乳腺X线筛查的补充。No等[27]研究显示，许多患者和医师由此夸大了MRI在乳腺筛查中的作用，从而使这些患者摒弃了每年一次的X线筛查。MRI是乳腺癌高危人群筛查的有益补充，而不可替代常规X线摄影。Schrading等[28]评估了乳腺MRI在一般人群筛查中的价值，结果显示对于X线和超声无法发现的乳腺癌，MRI的额外检出率为5.4‰，因此认为MRI用于一般人群的乳腺筛查亦是非常必要的。另外两项研究结果显示，MRI可作为常规X线的有益补充而用于乳腺癌病史患者的筛查，其检出率分别为4.4%[29]和2.1%[30]。Kaiser等[31]研究认为，与单独行乳腺X线筛查相比，MRI检查的费用尽管较高，但从长远考虑，则是一种经济的筛查策略。Johnson等[32]研究认为乳腺癌高危人群MRI筛查的召回率与读片者的经验关系不大，而与X线筛查的召回率相关，因此其客观性相对较强。以往考虑到女性体内激素水平随月经周期而出现波动，乳腺背景的强化程度亦随激素水平的波动而发生变化，因此推荐MRI筛查应安排在月经周期的早期(即受检者体内激素水平最低时)进行。Dontchos等[33]回顾性分析了243名未绝经妇女的319项MRI检查，认为MRI的诊断结果不受月经周期影响，因而妇女可选择月经周期的任何时段行乳腺MRI检查。

数字合成体层成像技术用于乳腺X线检查的时间并不长，DBT可呈现三维空间的影像信息，从而消除常规DM检查中组织重叠造成的观察障碍。Miravete等[34]回顾了5718名乳腺筛查者的普查数据，结果显示与传统的DM筛查比较，引入DBT后的联合筛查的检出率可由原来的1.50%提高到2.33%，且对于不同密度的乳腺腺体的检出率无明显差别。Rose等[35]、Hass等[36]、Destounis等[37]的研究结果显示DM联合DBT筛查可减少召回率，尤其是对于致密型乳腺和<50岁的患者。尽管DBT是一种很有前景的筛查方式，但也存在一定局限性，如工作时间延长、射线剂量高于常规DM等。Kalra等[38]的研究结果显示DM联合DBT阅片时间约为传统DM的2倍，这将会影响临床工作的效率。Garrett等[39]利用先验图像压缩感知算法(prior image constrained compressed sensing,PICCS)降低DBT检查约60%的剂量，而其诊断效果与常规剂量DBT相当，且PICCS可提高低对比病灶的信噪比而不影响其重要细节的显示。

3 乳腺癌NAC疗效评价和预测

NAC前后MRI参数(血流动力学参数、ADC值等)的改变评价乳腺癌NAC疗效的价值已得到大量研究证实[40-42]。Arlinghaus等[43]对25例乳腺癌患者行化疗前和NAC 1周期后MRI检查，结果显示NAC 1周期后ADC值的改变可早期预测pCR，因此有望提早帮助临床医师决定治疗方案，避免不必要的化疗。Li等[44]对28例乳腺癌患者NAC早期疗效预测的研究亦显示，NAC 1周期后动态增强定量参数Ktrans、Kep和Vp的改变可早期预测NAC的疗效。然而，这些仅仅是小样本研究，NAC 1周期的MRI参数能否预测NAC最终疗效还需大宗病例和Meta分析的进一步证实。Telesca等[45]使用一种高弛豫率对比剂钆贝葡胺(Gd-BOPTA，莫迪司)行3.0 T增强 MRI评价乳腺癌NAC疗效，结果显示其测得的病灶最大径与病理标本的测量值呈正相关(r=0.746)，因此认为这种方法可准确评价乳腺癌NAC后残余病灶的大小。由于Gd-BOPTA的弛豫率约为Gd-DTPA的2倍，因此1/2剂量即可达到同样的强化效果，适用于灌注成像；且初步研究显示[46]，Gd-BOPTA还可用于肝肾功能不全者，从而拓展了其应用范围。另外，亦有学者报道了影响MRI评价乳腺癌NAC疗效准确性的一些因素。Ko等[47]的174例大样本研究显示，对于导管原位癌、分化程度较好、基线MRI表现为非肿块样强化或NAC后残余病灶较大者，肿瘤大小评价NAC疗效的准确性较低。Drisis等[48]对91例乳腺癌患者行动态增强MRI检查，发现定量参数评价和预测乳腺癌NAC疗效的能力受分子亚型影响，对于三阴性和HER2阳性乳腺癌，其评价效能较高，对luminal型则较低。

4 乳腺定位技术

导丝定位由于置入的导丝容易移位脱落，且手术阳性切缘率较高、患者易产生不适感及并发症，近年来受到质疑。Ananyhasivan等[49]采用放射性同位素引导隐匿性病灶定位(radio-guided occult lesion localization,ROLL)技术对乳腺肿瘤及前哨淋巴结进行术前定位，即在超声或X线引导下将放射性同位素锝标记硫胶体注入病灶内，术中用伽马探针对病灶进行精确定位，结果显示与导丝定位相比，ROLL是一种准确、简便、性价比高且容易被患者接受的定位方法。然而ROLL的手术阳性切缘率还需进一步评价。King等[50]采用放射性粒子125I植入定位(radioactive seed localization，RSL)技术对195名妇女行乳腺术前定位，结果显示单独采用RSL定位者手术阳性切缘率为10%，RSL和导丝联合定位者为14%，两者无统计学差异。因而，RSL是一种安全、有效、快速且并发症较少的定位方法。

纵观2012年RSNA乳腺影像学报道，MRI在乳腺疾病诊断和研究中具有广阔前景，分子成像等新技术的兴起将会开辟乳腺影像学研究的新局面；同时，寻找更好的、适合本国国情的早期发现和诊断乳腺癌的影像筛查手段，降低死亡率，仍是放射科医师今后继续努力的方向。

[References]

[1]Brem RF,Schoonjans JM,Kieper DA,et al.High-resolution scintimammography:a pilot study.J Nucl Med,2002,43(7):909-915.

[2]Coffey CM,Brem RF,Lieberman JH,et al.Breast-specific gamma imaging (BSGI) for the detection of mammographically occult breast cancer in high-risk women.Chicago:RSNA,2012:VSMI21-09.

[3]Velez MB,Kieper DA,Chang TS,et al.The role of molecular breast imaging in patient management in a community breast center.Chicago:RSNA,2012:VSMI21-10.

[4]Rapelyea JA,Vasnani RJ,Rechtman LR,et al.Predicting surgical treatment outcomes of invasive breast cancers with breast-specific gamma imaging (BSGI).Chicago:RSNA,2012:LL-BRS-MO3D.

[5]Koo HR,Moon WK,Cho N,et al.Relationship between quantitative FDG uptake on PET and immunohistochemical subtypes of breast cancers.Chicago:RSNA,2012:VSMI21-14.

[6]Kiyoto S,Sugawara Y,Hosokawa K,et al.FDG PET/CT can predict pathological complete response and clinical outcome after neoadjuvant chemotherapy in breast cancer patients.Chicago:RSNA,2012:SSA18-08.

[7]Nakajima N,Kataoka,M,Sugawara Y,et al.Prognostic value of pretreatment standardized uptake value (SUV) parameters in stage IIIII breast cancer treated with postmastectomy radiotherapy (PMRT).Chicago:RSNA,2012:MSRO32-03.

[8]Ahn HS,Yun BL,Kim SM,et al.Does adding optical diffusion imaging to us improve the differentiation between benign and malignant breast lesions? Observer performance study.Chicago:RSNA,2012:SSG01-03.

[9]Choi WJ,Kim HH,Kim H,et al.Optical diffusion breast imaging;correlation of parameters in optical imaging with maximum standardized uptake value in breast cancer patients.Chicago:RSNA,2012:VSMI21-02.

[10]Sklair-Levy M,Mayer A,Sella T,et al.Non-invasively detection of vascular map asymmetry in breast cancer using novel functional infrared imaging.Chicago:RSNA,2012:VSMI21-03.

[11]Pinker-Domenig K,Bickel H,Helbich TH,et al.BI-RADS®-based combination of CE-MRI and DWI obviates unnecessary breast biopsies.Chicago:RSNA,2012:VSBR51-09.

[12]Yoo,H,Shin HJ,Cha JH,et al.The diagnostic value of apparent diffusion coefficient and quantitative kinetic analysis for predicting malignancy in patients with newly diagnosed breast cancer.Chicago:RSNA,2012:LL-BRS-MO1A.

[13]Olson ML,Kurland B,Chai X.Detection and characterization of mammographically and clinically occult suspicious breast MRI lesions with diffusion weighted imaging at 3 T.Chicago:RSNA,2012:LL-BRS-TU5C.

[14]Bickel H,Pinker-Domenig K,Bogner W,et al.Diffusion-weighted imaging:a noninvasive biomarker for breast cancer invasiveness.Chicago:RSNA,2012:SSG01-07.

[15]Pediconi F,Vasselli F,Roselli,A,et al.Breast MRI at 3 T:role of unenhanced MRI (DWI combined with T2 IDEAL sequence) vs CEMRI in the assessment of response to neo-adjuvant chemotherapy.Chicago:RSNA,2012:VSBR51-08.

[16]Schiani E,Cristel G,Losio C,et al.Diffusion-weighted magnetic resonance imaging in breast cancer:a promising tool to predict tumor aggressiveness.Chicago:RSNA,2012:LL-BRS-MO2D.

[17]Benveniste AP,Dogan BE,Fox PS,et al.Is There a role for ADC values in predicting breast cancer prognostic factors? Chicago:RSNA,2012:LL-BRS-TU3C.

[18]Cho G,Moccaldi M,Moy L,et al.Observation of time-dependent DTI Anisotropy in breast fibroglandular tissue.Chicago:RSNA,2012:SSG01-09.

[19]Penn AI,Rockvile,Hruska CB,et al.Discrimination of benign from malignant breast lesions on DTI using novel feature obtained from ADC on lesion and ADC gradient in peri-tumor.Chicago:RSNA,2012:LL-BRS-SU5B.

[20]Au WF,Ghai S,Crystal P,et al.Quantitative shear wave elastography:performance in characterization of solid breast masses.Chicago:RSNA,2012:LL-BRS-TU1C.

[21]Lee SH,Moon WK,Cho N,et al.Two-view vs single-view shearwave elastography (SWE):a comparison of observer performance in differentiating benign from malignant solid breast masses.Chicago:RSNA,2012:LL-BRS-TU2D.

[22]Lee SH,Chang JM,Moon WK,et al.Added value of shear-wave elastography (SWE) color scoring system and quantitative elasticity values in evaluation of ultrasound (US) detected breast masses.Chicago:RSNA,2012:LL-BRS-WE6A.

[23]Gweon HM,Youk JH,Kim J,et al.Qualitative assessment in shearwave elastography for breast masses:comparative analysis of E value and color overlay pattern classifi cation.Chicago:RSNA,2012:LL-BRS-WE3C.

[24]Kim MY,Cho N,Yi A,et al.Complex cystic breast masses:role of sonoelastography in distinguishing benign from malignant lesions and in the biopsy decision.Chicago:RSNA,2012:VSBR41-17.

[25]Evans A，Whelehan P,Vinnicombe SJ,et al.What are the pathological characteristics of breast cancers misclassifi ed as benign by ultrasound shear wave elastography? Chicago:RSNA,2012:VSBR41-15.

[26]Chang JM,Lee,SH,Moon WK,et al.Stiffness values measured by shear-wave elastography correlated with aggressive breast cancer phenotype.Chicago:RSNA,2012:SSK01-09.

[27]No CJ,Sanders L,Chin JW,et al.The impact of breast cancer screening with MRI on patient adherence to annual mammography screening schedules in a high risk population in a community setting.Chicago:RSNA,2012:LL-BRS-TU2C.

[28]Schrading S,Schild HH,Kuhl CK.MRI screening of women at average risk of breast cancer.Chicago:RSNA,2012:SSJ02-02.

[29]Wooten K,Weinstock C,Stokum C,et al.Screening magnetic resonance imaging (MRI) as an adjunct to mammography (MG)in asymptomatic breast cancer survivors.Chicago:RSNA,2012:SSM02-02.

[30]Tsang Mui Chung M,Lourenco AP,Mainiero MB,et al.Screening breast mri in women with a personal history of breast cancer.Chicago:RSNA,2012:LL-BRS-WE2B.

[31]Kaiser CG,Reich C,Wasser K,et al.Economic aspects of MR-mammography in dense breasts.Chicago:RSNA,2012:LL-BRS-MO3B.

[32]Johnson KS,Soo MS,Holbrook AI,et al.Are screening MRI recall rates related to temperament or experience? Chicago:RSNA,2012:LL-BRS-TU1A.

[33]Dontchos B,Demartini WB,Rahbar H,et al.Infl uence of menstrual cycle timing on screening breast MRI performance in pre-menopausal women.Chicago:RSNA,2012:SSK01-07.

[34]Miravete PM,Sainz M,Garcia-Lallana A,et al.The role of additional digital tomosynthesis combined with digital mammography.Chicago:RSNA,2012:SSE02-02.

[35]Rose SL,Ice MF,Nordmann AS,et al.A comparison of recall rates between full field digital mammography (FFDM) and full field digital mammography plus tomosynthesis in a community setting.Chicago:RSNA,2012:SSE02-01.

[36]Hass B,Kalra VB,Raghu M,et al.Performance of digital breast tomosynthesis compared to conventional digital mammography for breast cancer screening.Chicago:RSNA,2012:SSE02-04.

[37]Destounis SV,Murphy PF,Seifent PJ,et al.Experience with combination 2D/3D breast tomosynthesis vs FFDM in the screening environment.Chicago:RSNA,2012:LL-BRS-MO3C.

[38]Kalra VB,Haas B,Butler RS,et al.Time for tomosynthesis:screening combined digital breast tomosynthesis vs.full-fi eld digital mammography interpretation time in clinical practice:a prospective study of 300 cases.Chicago:RSNA,2012:VSBR41-11.

[39]Garrett JW,Tang J,Burnside ES,et al.Dose reduction in digital breast tomosynthesis with prior image constrained compressed sensing(PICCS).Chicago:RSNA,2012:SSE02-03.

[40]Pinker-Domening K,Bickel H,Bogner W,et al.Assessment of response to neoadjuvant chemotherapy with high resolution MRI at 3 T:correlation with pathological response.Chicago:RSNA,2012:SSM02-05.

[41]Bickel H,Pinker-Domening K,Bogner W,et al.Assessment of response to neoadjuvant chemotherapy with diffusion weighted imaging at 3 T.Chicago:RSNA,2012:LL-BRS-TU5D.

[42]Budreau D,Giger ML,Lan L,et al.Breast MRI-based feature analysis in predicting neoadjuvant therapy response.Chicago:RSNA,2012:SSM01-04.

[43]Arlinghaus LR,Abramson RG,Chakravathy AB,et al.DW-MRI for predicting the response of locally advanced breast cancer to neoadjuvant chemotherapy.Chicago:RSNA,2012:LL-BRS-TH5A.

[44]Li X,Arlinghaus LR,Abramson RG,et al.Towards optimization of DCE-MRI analysis for predicting the response of breast cancer to neoadjuvant chemotherapy.Chicago:RSNA,2012:LL-BRS-WE3B.

[45]Telesca M,Pediconi F,Luciani ML,et al.Accuracy of 3 T magnetic resonance imaging with a high-relaxivity contrast agent in assessing treatment response in patients with advanced breast cancer undergoing neoadjuvant chemotherapy.Chicago:RSNA,2012:LL-BRS-WE4B.

[46]Davies BE,Kirchin MA,Bensel K,et al.Pharmacokinetics and safety of gadobenate dimeglumine (Multihance) in subjects with impaired liver function.Invest Radiol,2002,37(5):299-308.

[47]Ko ES,Han B,Shin JH,et al.Analysis of factors infl uencing accuracy of MRI for predicting response after neoadjuvant chemotherapy in locally advanced breast cancer.Chicago:RSNA,2012:LL-BRS-WE5A.

[48]Drisis S,Stathopoulos K,Capelan M,et al.The impact of breast cancer subtypes on DCE-MRI as a surrogate marker for locally advanced breast cancer.Chicago:RSNA,2012:SSM01-06.

[49]Ananyhasivan R,Parameshwaran R,Nagaraj C,et al.Radio-guided occult lesion localization (ROLL) for non-palpable breast lesions:the indian experience.Chicago:RSNA,2012:LL-BRS-MO4C.

[50]King V,Sung JS,Thornton C,et al.Safety and effi cacy of radioactive seed localization with I-125 prior to lumpectomy and/or excisional biopsy.Chicago:RSNA,2012:SST01-02.