The clinical application of multimodal ultrasound features in predicting Luminal subtype invasive ductal carcinoma
-
摘要:
目的 探讨基于多模态超声特征预测Luminal亚型浸润性导管癌（IDC）的临床价值。 方法 选取内蒙古自治区人民医院2021年6月~2023年12月病理证实为Luminal型的IDC患者85例，其中Luminal A（LA）型38例，Luminal B（LB）型47例。比较两组病灶常规超声、弹性成像及超声造影特征的差异；将差异有统计学意义的特征进行单、多因素逻辑回归分析，并建立逻辑回归模型；绘制ROC曲线分析其预测LB型IDC的诊断效能。 结果 与LA组相比，LB组病灶最大直径、富血供型占比、病灶内部最大杨氏模量（Emax）、瘤周平均杨氏模量（Emean shell-2.0）、最大杨氏模量（Emax shell-2.0）、造影灌注缺损占比、峰值强度（PI）、AUC、平均渡越时间均较大；始增时间（AT）、瘤周高回声晕占比较小，差异均有统计学意义（P＜0.05）。多因素回归分析显示：AT、瘤周高回声晕为预测LB型的独立保护因素，PI、灌注缺损、Emax shell-2.0为预测LB型的独立危险因素。构建预测LB型IDC的逻辑回归模型：Logit(P)=-9.868-2.004×瘤周高回声晕+2.896×灌注缺损-0.399×AT+0.379×PI+0.030×Emax shell-2.0。ROC曲线分析显示，逻辑回归模型诊断效能最高，曲线下面积为0.945，敏感度、特异度分别为89.4%、89.5%。 结论 不同Luminal亚型IDC的多模态超声特征具有一定的差异性，基于多模态超声特征所建立的逻辑回归模型在预测Luminal亚型中具有应用价值，能够为各亚型IDC的个体化诊疗提供重要参考依据。 Abstract:Objective To investigate the clinical value of predicting Luminal subtype invasive ductal carcinoma (IDC) based on multimodal ultrasound features. Methods Eighty-five patients with Luminal IDC diagnosed by pathology from June 2021 to December 2023 in Inner Mongolia Autonomous Region People's Hospital were selected, including 38 cases in the Luminal A (LA) group and 47 cases in the Luminal B (LB) group. Differences in conventional ultrasound, elastography and contrast-enhanced ultrasound characteristics of the lesions in the two groups were compared; characteristic parameters with statistically significant differences were subjected to single and multifactorial logistic regression analyses with the LB type, and a logistic regression model was established; and the diagnostic efficacy in predicting the LB type of IDC was analyzed by plotting the ROC curves. Results Compared with the LA group, the LB group had a larger maximum diameter of the lesion, percentage of blood-rich type, maximum Young's modulus inside the lesion (Emax), average peritumour Young's modulus (Emean shell-2.0), maximum Young's modulus (Emax shell-2.0), percentage of contrast perfusion defect, peak intensity (PI), area under the curve (AUC), and mean transit time; arrival time (AT) and percentage of peritumoural hyperechoic halo were smaller, and all the differences were statistically significant (P<0.05). Multifactorial regression analysis showed that AT and peritumoural hyperechoic halo were independent protective factors for the prediction of LB, and PI, perfusion defect and Emax shell-2.0 were independent risk factors for the prediction of LB. A logistic regression model was constructed to predict LB-type IDC: Logit (P)=-9.868-2.004×peritumoural hyperechoic halo +2.896×perfusion defect-0.399×AT+0.379×PI +0.030×Emax shell-2.0. ROC curve analysis showed that the logistic regression model had the highest diagnostic efficacy, with an area under the curve of 0.945 with a sensitivity and specificity of 89.4% and 89.5%, respectively. Conclusion The multimodal ultrasound characteristics of different IDC subtypes have certain differences, and the logistic regression model based on multimodal ultrasound characteristics has value in predicting Luminal subtypes, which can provide an important reference for individualised diagnosis and treatment of IDC of different subtypes. -
Key words:
- multimodal ultrasound /
- invasive ductal carcinoma /
- molecular staging
-
图 1 LA组与LB组常规超声图像
Figure 1. Conventional ultrasound images of group LA and group LB. A, B: Conventional ultrasound images of a lesion of type LA group showed: the maximum diameter of the lesion was 23 mm, the border was clear, the morphology was irregular, the internal echogenicity was not homogeneous, microcalcified foci could be seen, and the periphery could be seen as a hypoechoic halo; Adler's classification of grade 1, RI: 0.75; C, D: Conventional ultrasound images of a lesion of type LB group showed: the maximum diameter of the lesion was 26 mm, the border was unclear, the morphology was irregular, the internal echogenicity was not homogeneous, microcalcified foci could be seen, and a hyperechoic halo could be seen in the periphery; Adler's classification was grade 2-3, RI: 0.78.
图 2 LA组与LB组SWE图像
Figure 2. SWE images of group LA and group LB. A: SWE image of a lesion in type LA group showed: a "hard ring" sign around the lesion, with an internal Emax 78.53 kPa, Emax shell-2.0 119.13 kPa, Emean 50.27 kPa; B: SWE image of a lesion in type LB group showed: a "hard ring" sign around the lesion, with an internal Emax 111.61 kPa, Emax shell-2.0 173.15 kPa, Emean shell-2.0 52.37 kPa. SWE: Shear wave elastography.
图 3 LA组与LB组超声造影图像
Figure 3. Contrast-enhanced ultrasound images of group LA and group LB. A, B: Contrast-enhanced ultrasound images of a lesion in type LA group showed: no area of perfusion defect within the lesion; quantitative analysis of TIC: AT 12.27 s, PI 23.18 dB, AUC 1888.18 dBs, MTT 82.67 s; C, D: Contrast-enhanced ultrasound images of a lesion in type LB group showed: area of perfusion defect was visible inside the lesion; TIC quantitative analysis: AT 10.13 s, PI 28.99 dB, AUC 2579.15 dBs, MTT 105.67s.
图 4 瘤周高回声晕、血流分级、灌注缺损、Emax shell-2.0、AT、PI、AUC及联合模型预测Luminal B型IDC的ROC曲线图
Figure 4. ROC curves of peritumoural hyperechoic halo, blood flow grade, perfusion defect, Emaxshell-2.0, AT, PI, AUC, and combine model predicting IDC in Luminal B.
表 1 LA、LB组间常规超声参数比较
Table 1. Comparison of conventional ultrasound parameters between Luminal A and B groups
Variable Total (n=85) Luminal A (n=38) Luminal B (n=47) χ²/t P Maximum diameter (mm, Mean±SD) 25.88±7.13 23.95±7.43 27.45±6.54 -2.24 0.023 RI (Mean±SD) 0.77±0.08 0.77±0.06 0.76±0.09 0.56 0.580 H/W 2.72 0.099 ＜1 43(50.59) 23(60.53) 20(42.55) ≥1 42(49.41) 15(39.47) 27(57.45) Microcalcified foci [n(%)] 2.61 0.106 No 28(32.94) 16(42.11) 12(25.53) Yes 57(67.06) 22(57.89) 35(74.47) Peritumoural hyperechoic halo [n(%)] 14.33 ＜0.001 No 44(51.76) 11(28.95) 33(70.21) Yes 41(48.24) 27(71.05) 14(29.79) Posterioracoustic features [n(%)] 2.32 0.128 No Posterior acoustic features 37(43.53) 20(52.63) 17(36.17) Posterior acoustic attenuation 48(56.47) 18(47.37) 30(63.83) Blood flow grade [n(%)] 15.88 ＜0.001 Lack of blood supply 40(47.06) 27(71.05) 13(27.66) Rich blood supply 45(52.94) 11(28.95) 34(72.34) RI: Resistance index; H/W: Height/Width. 
下载: 导出CSV
表 2 LA、LB组间剪切波弹性参数比较
Table 2. Comparison of shear wave elastography parameters between Luminal A and B groups (Mean±SD)
Variable Total (n=85) Luminal A(n=38) Luminal B(n=47) t P Emean 35.46±5.91 34.68±6.12 36.09±5.71 -1.09 0.277 Emax 118.12±30.85 106.42±24.36 127.57±32.50 -3.43 ＜0.001 Emean shell-2.0 44.26±11.07 41.26±10.68 46.68±10.90 -2.30 0.024 Emax shell-2.0 136.58±30.47 122.63±27.36 147.86±28.33 -4.15 ＜0.001 Emean: Mean Young's modulus inside the lesion; Emax: Maximum Young's modulus inside the lesion; Emean shell-2.0: Mean Young's modulus 2 mm around tumour; Emax shell-2.0: Maximum Young's modulus 2 mm around tumour.
下载: 导出CSV
表 3 LA、LB组间超声造影参数比较
Table 3. Comparison of ultrasonographic parameters between Luminal A and B groups
Variable Total (n=85) Luminal A (n=38) Luminal B (n=47) χ²/t P AT (Mean±SD) 11.42±2.51 12.56±2.52 10.50±2.11 4.10 ＜0.001 TTP (Mean±SD) 25.42±4.38 25.15±3.49 25.65±5.01 -0.53 0.600 PI (Mean±SD) 27.23±3.91 24.83±3.44 29.16±3.14 -6.05 ＜0.001 AS (Mean±SD) 0.86±0.41 0.82±0.28 0.90±0.50 -0.93 0.356 DT/2 (Mean±SD) 98.13±18.83 95.34±14.69 100.38±21.51 -1.23 0.222 DS (Mean±SD) 0.18±0.10 0.17±0.07 0.20±0.12 -1.07 0.288 AUC (Mean±SD) 2106.04±393.47 1913.87±242.72 2261.41±424.44 -4.74 ＜0.001 MTT (Mean±SD) 92.32±20.04 87.31±16.70 96.37±21.71 -2.11 0.038 Enhancement method [n(%)] 3.45 0.063 Equal enhancement 12(14.12) 38(21.05) 4(8.51) High enhancement 73(85.88) 30(78.95) 43(91.49) Extent of lesion [n(%)] 1.76 0.185 Non-expansion 13(15.29) 8(21.05) 5(10.64) Expanded 72(84.71) 30(78.95) 43(89.36) Perfusion defect [n(%)] 19.21 ＜0.001 No perfusion defect 47(55.29) 31(81.58) 16(34.04) Perfusion defect 38(44.71) 7(18.42) 41(65.96) AT: Arrival time; TTP: Time to peak; PI: Peak intensity; AS: Ascending slope; DT/2: Decending time/2; DS: Descending slope; AUC: Area under curve; MTT: Mean transit time.
下载: 导出CSV
表 4 Logistic回归分析
Table 4. Logistic regression analysis
Variable Β S.E. Wals P OR(95% CI) Hyperechoic halo -2.004 0.815 6.051 0.014 0.135（0.027-0.665） Perfusion defect 2.896 0.867 11.167 0.001 18.095（3.311-98.892） AT -0.399 0.159 6.283 0.012 0.671（0.491-0.917） PI 0.379 0.122 9.630 0.002 1.460（1.150-1.855） Emax shell-2.0 0.030 0.013 5.362 0.021 1.031（1.005-1.057） Constant -9.868 4.009 6.060 0.014 ＜0.001
下载: 导出CSV
表 5 ROC曲线分析
Table 5. Analysis of ROC curves
Mode OR（95% CI） Truncation value Jordon index Sensitivity（%） Specificity（%） Combine 0.945（0.879-0.993） 0.58 0.788 89.4 89.5 Peritumoural hyperechoic halo 0.706（0.593-0.819） No hyperechoic halo 0.430 70.2 71.1 Blood flow grade 0.717（0.605-0.829） Rich blood supply 0.434 72.3 71.1 Perfusion defect 0.738（0.630-0.846） No perfusion defect 0.475 66.0 81.6 Emax shell-2.0 0.737（0.629-0.846） 124.48 kPa 0.483 85.1 63.2 AT 0.759（0.653-0.866） 11.90 s 0.498 78.7 71.1 PI 0.837（0.743-0.930） 26.66 dB 0.646 83.0 81.6 AUC 0.812（0.712-0.912） 2083.95 dBs 0.656 78.7 86.8
下载: 导出CSV
-
[1] Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA Cancer J Clin, 2021, 71(3): 209-49. doi: 10.3322/caac.21660 [2] 中华医学会外科学分会乳腺外科学组. 中国浸润性乳腺癌诊治临床实践指南(2022版)[J]. 中国实用外科杂志, 2022, 42(2): 121-7. [3] 安蕾, 李丽霞, 孙芳, 等. 浸润性乳腺癌的多模态影像学特征与分子分型的相关性[J]. 滨州医学院学报, 2022, 45(1): 51-7. [4] 孙静, 马千清, 沈春云, 等. 多模态超声结合免疫组化参数构建列线图预测乳腺癌分子分型相关预后[J]. 中国超声医学杂志, 2024, 40(6): 632-6. [5] 王楠, 庞坤坤, 张飞雪, 等. 乳腺浸润性导管癌5种分子亚型临床及超声特征[J]. 中国现代普通外科进展, 2024, 27(3): 204-8. [6] 赵玉哲, 吴天琦, 赵怡然, 等. 老年乳腺癌超声影像特征与分子分型的相关性研究[J]. 实用老年医学, 2024, 38(2): 166-70. [7] 廖春旭, 张慧娟, 张丽娟, 等. T1期乳腺癌多模态超声影像特征与分子分型相关性研究[J]. 医学影像学杂志, 2024, 34(2): 52-6. [8] 阚光娟, 傅晓凤, 杨晶, 等. 年轻乳腺癌不同分子亚型的超声多模态特征[J]. 医学影像学杂志, 2023, 33(12): 2220-3. [9] Adler DD, Carson PL, Rubin JM, et al. Doppler ultrasound color flow imaging in the study of breast cancer: preliminary findings[J]. Ultrasound Med Biol, 1990, 16(6): 553-9. doi: 10.1016/0301-5629(90)90020-D [10] Zhang XM. Molecular classification of breast cancer: relevance and challenges[J]. Arch Pathol Lab Med, 2023, 147(1): 46-51. doi: 10.5858/arpa.2022-0070-RA [11] Jackson HW, Fischer JR, Zanotelli VRT, et al. The single-cell pathology landscape of breast cancer[J]. Nature, 2020, 578(7796): 615-20. doi: 10.1038/s41586-019-1876-x [12] Zhao F, Cai CJ, Liu MH, et al. Identification of the lymph node metastasis-related automated breast volume scanning features for predicting axillary lymph node tumor burden of invasive breast cancer via a clinical prediction model[J]. Front Endocrinol, 2022, 13: 881761. doi: 10.3389/fendo.2022.881761 [13] Ades F, Zardavas D, Bozovic-Spasojevic I, et al. Luminal B breast cancer: molecular characterization, clinical management, and future perspectives[J]. J Clin Oncol, 2014, 32(25): 2794-803. doi: 10.1200/JCO.2013.54.1870 [14] 中国抗癌协会肿瘤标志专业委员会乳腺癌标志物协作组. 基于靶标指导乳腺癌精准治疗标志物临床应用专家共识(2022版)[J]. 中国癌症防治杂志, 2022, 14(4): 346-62. [15] Tamaki K, Sasano H, Ishida T, et al. The correlation between ultrasonographic findings and pathologic features in breast disorders[J]. Jpn J Clin Oncol, 2010, 40(10): 905-12. doi: 10.1093/jjco/hyq070 [16] 毛羡仪, 梁伟翔, 蒋殿虎, 等. 乳腺癌高回声晕超声特征与ER、PR表达水平的相关性分析[J]. 医学影像学杂志, 2020, 30(6): 1001-4, 1013. [17] Chung WS, Chen SC, Ko TM, et al. An integrative clinical model for the prediction of pathological complete response in patients with operable stage Ⅱ and stage Ⅲ triple-negative breast cancer receiving neoadjuvant chemotherapy[J]. Cancers, 2022, 14(17): 4170. doi: 10.3390/cancers14174170 [18] 崔亚云. 声触诊弹性成像技术诊断乳腺癌及其分子分型的价值研究[D]. 合肥: 安徽医科大学, 2023. [19] Skerl K, Vinnicombe S, Thomson K, et al. Anisotropy of solid breast lesions in 2D shear wave elastography is an indicator of malignancy[J]. Acad Radiol, 2016, 23(1): 53-61. doi: 10.1016/j.acra.2015.09.016 [20] 贾美红, 李海康, 薛继平, 等. 乳腺癌剪切波弹性模量最大值与其免疫组化相关性研究[J]. 中华超声影像学杂志, 2018, 27(6): 510-4. doi: 10.3760/cma.j.issn.1004-4477.2018.06.011 [21] Yoo J, Seo BK, Park EK, et al. Tumor stiffness measured by shear wave elastography correlates with tumor hypoxia as well as histologic biomarkers in breast cancer[J]. Cancer Imaging, 2020, 20(1): 85. doi: 10.1186/s40644-020-00362-7 [22] Wang M, Feng HL, Liu YQ, et al. Angiogenesis research in mouse mammary cancer based on contrast-enhanced ultrasonography: exploratory study[J]. Acad Radiol, 2018, 25(7): 889-97. doi: 10.1016/j.acra.2017.12.004 [23] Wen BJ, Kong WT, Zhang YD, et al. Association between contrast-enhanced ultrasound characteristics and molecular subtypes of breast cancer[J]. J Ultrasound Med, 2022, 41(8): 2019-31. doi: 10.1002/jum.15886 [24] 唐恩红, 杨少玲, 王凤翎, 等. 分子亚型乳腺癌的超声特征对照研究[J]. 中国超声医学杂志, 2022, 38(1): 22-5. doi: 10.3969/j.issn.1002-0101.2022.01.008 [25] Liang XY, Li ZY, Zhang L, et al. Application of contrast-enhanced ultrasound in the differential diagnosis of different molecular subtypes of breast cancer[J]. Ultrason Imaging, 2020, 42(6): 261-70. doi: 10.1177/0161734620959780 [26] 余小琴, 郭亮, 王俊芳, 等. 小乳腺癌超声征象与其分子亚型的相关性研究[J]. 医学影像学杂志, 2023, 33(12): 2224-7. [27] 王玉敏, 梁丹艳, 王芳, 等. 基于超声造影参数的Logistic回归模型预测浸润性乳腺癌Ki-67表达水平的临床价值[J]. 临床超声医学杂志, 2023, 25(11): 905-9. doi: 10.3969/j.issn.1008-6978.2023.11.013 [28] Shin I. HER2 signaling in breast cancer[J]. Adv Exp Med Biol, 2021, 1187: 53-79. [29] Wang CY, Wei W, Santiago L, et al. Can imaging kinetic parameters of dynamic contrast-enhanced magnetic resonance imaging be valuable in predicting clinicopathological prognostic factors of invasive breast cancer?[J]. Acta Radiol, 2018, 59(7): 813-21. doi: 10.1177/0284185117740746 [30] 康佳, 吴桐, 张蕾, 等. 不同分子分型乳腺癌的多模态超声特征和临床病理对照研究[J]. 中华超声影像学杂志, 2020, 29(4): 330-6. doi: 10.3760/cma.j.cn131148-20190926-00591 [31] 苏荃利, 李玲玲, 邓芸霞, 等. 超声造影定性特征和VueBox定量参数评估乳腺癌分子分型的应用价值[J]. 中华超声影像学杂志, 2023, 32 (5): 399-405. doi: 10.3760/cma.j.cn131148-20221107-00753 -
点击查看大图
计量
- 文章访问数: 50
- HTML全文浏览量: 26
- PDF下载量: 9
- 被引次数: 0
