-
骨肉瘤是青少年最常见的原发性恶性骨肿瘤,其致死和致残率高[1]。目前新辅助化疗、手术加术后化疗使骨肉瘤患者的5年生存率提高到了70%~80%[2]。新辅助化疗后骨肉瘤的肿瘤组织坏死率被认为是评价预后的“金标准”,而预后的好坏是确定是否可以行局部切除保肢治疗的依据。然而,这个“金标准”只能在术后获得,无法用于判断保肢手术进行与否。目前,用于术前评价新辅助化疗疗效的方法主要有临床评价和传统影像学评价,前者依据肿瘤大小变化、疼痛评分等化疗前后的症状或体征进行评价;后者主要包括对肿瘤边界、大小等形态学变化进行评价。这些方法虽然有一定的参考价值,但缺乏量化、标准化的指标,不能准确反映化疗疗效[3]。影像学评价方法具有很大的价值,其包括形态学评估及功能学评估,例如动态对比增强(DCE)-MRI、18F-FDG PET/CT等[4-6],但这些检查目前缺乏特异性高的标准化参数。MRI弥散加权成像(diffusion weighted imaging, DWI)功能显像方法简单、方便、无创,近年来的很多研究结果显示,表观扩散系数(apparent diffusion coefficient,ADC)在反映骨肉瘤等恶性肿瘤新辅助化疗疗效中具有重要价值[7-10]。但是由于肿瘤的异质性很大,其结果并不稳定,无法满足精准医学的要求。影像组学是将影像数据定量分析与机器学习相结合,筛选出具有诊断价值的影像组学特征,建立诊断模型,为临床诊疗提供更准确的信息[11]。本研究探讨基于MRI DWI的影像组学模型对骨肉瘤新辅助化疗疗效评估的价值。
-
41例骨肉瘤患者中,疗效好组10例、疗效差组31例。由表1可知,2组患者的ADC0的差异无统计学意义(t=1.14,P>0.05);疗效好组的ADC1、ADC2高于疗效差组,且差异均有统计学意义(t=−2.92,P<0.05;Z=−2.61,P<0.01)。ROC曲线分析结果显示,当ADC1≥1.34×10−3 mm2/s时,其对骨肉瘤新辅助化疗疗效评估的灵敏度为80%,特异度为81%,AUC为0.797(95%CI:0.629~0.965);当ADC2≥1.51×10−3 mm2/s时,其对骨肉瘤新辅助化疗疗效评估的灵敏度为90%,特异度为71%,AUC为0.777(95%CI:0.588~0.967)(图2)。
组别 ADC0
( )$\bar x\pm s $ ADC1
( )$\bar x\pm s $ ADC2
[M(Q1,Q3)或 ]$\bar x\pm s $ 疗效好组(n=10) 0.95±0.05 1.44±0.10 1.68(1.55, 1.85) 疗效差组(n=31) 1.05±0.05 1.10±0.06 1.33±0.06 检验值 1.14a −2.92a −2.61b P值 0.260 0.006 0.009 注:a表示t值;b表示Z值。ADC为表观扩散系数;ADC0为新辅助化疗前的ADC;ADC1为化疗一期结束后5 d内的ADC;ADC2为完成整个化疗后的ADC 表 1 新辅助化疗后不同疗效的2组骨肉瘤患者的ADC比 较(×10−3 mm2/s)
Table 1. Comparison of apparent diffusion coefficient values between two groups of osteosarcoma patients with different therapeutic effects after neoadjuvant chemotherapy (×10−3 mm2/s)
图 2 ADC预测骨肉瘤患者新辅助化疗疗效的受试者工作特征曲线
Figure 2. Receiver operating characteristic curves of apparent diffusion coefficient values in prediction of the efficacy of neoadjuvant chemotherapy for osteosarcoma
从化疗一期结束后的DWI和ADC图像中共提取出1409个影像组学特征,按6∶4的比例随机分为训练集和测试集[24(疗效好:6,疗效坏:18)∶17(疗效好:4,疗效坏:13)],将训练集数据扩充为70(疗效好:20,疗效坏:50),经筛选降维后获得5个最优影像组学特征,分别为Interquartile Range、Skewness、Uniformity、Median、Maximum,均为纹理特征,如图3所示,确定了LASSO法的最优α参数为−log(α)=2.32。以最优α参数对应的5个特征及其系数分别为:original_firstorder_Interquartile Range(系数为−0.289894)、original_firstorder_Skewness(系数为−0.409912)、original_firstorder_Uniformity(系数为0.015519)、original_firstorder_Median(系数为−0.373091)、original_firstorder_Maximum(系数为0.811615)。
图 3 影像组学特征的均方误差随α参数变化趋势图(A)及影像组学降维图(B)
Figure 3. Mean square error of radiomics features change trend according to α parameters (A) and dimension reduction diagram of radiomics features (B)
使用此影像组学模型预测骨肉瘤患者新辅助化疗疗效,结果显示,训练集中,影像组学模型预测骨肉瘤患者新辅助化疗疗效的ROC曲线的AUC为0.881(95%CI:0.811~0.942) ,灵敏度为90%,特异度为74%,测试集中AUC为0.769(95%CI:0.515~0.933),灵敏度为75%,特异度为69%(图4)。
基于MRI DWI的影像组学模型对骨肉瘤新辅助化疗疗效的评估价值
Evaluation of the efficacy of neoadjuvant chemotherapy in osteosarcoma based on MRI DWI radiomics model
-
摘要:
目的 探讨基于MRI弥散加权成像(DWI)的影像组学模型对骨肉瘤新辅助化疗疗效的评估价值。 方法 回顾性分析河北医科大学第三医院2015年6月至2017年11月经术后组织病理学检查结果证实,且在接受新辅助化疗前、后均行MRI检查的41例骨肉瘤患者[男性26例、女性15例,年龄(22.0±11.0)岁,范围11~49岁]的病历及影像资料。根据术后组织病理学检查结果,将肿瘤组织坏死率≥90%者纳入疗效好组,<90%者纳入疗效差组。分别于新辅助化疗前、化疗一期结束后5 d内和完成整个化疗后测量所有患者的表观扩散系数(ADC,分别记为ADC0、ADC1、ADC2,比较疗效好组和疗效差组ADC间的差异。在化疗一期结束后的DWI(b=1000 s/mm2)和ADC图像上手动勾画病灶的感兴趣区,提取影像组学特征,用随机分组法将数据按6∶4的比例分为训练集和测试集,采用SMOTE算法对训练集上的数据进行扩充,采用方差阈值、SelectKBest、最小绝对收缩和选择算子(LASSO)法进行影像组学特征筛选,采用逻辑回归分类器构建出影像组学模型。采用独立样本t检验或Wilcoxon秩和检验进行2组间比较。采用受试者工作特征(ROC)曲线评估传统影像学(ADC)及影像组学模型对骨肉瘤新辅助化疗疗效的预测效能。 结果 疗效好组10例、疗效差组31例。2组患者ADC0的差异无统计学意义[(0.95±0.05)×10−3 mm2/s对(1.05±0.05)×10−3 mm2/s,t=1.14,P>0.05)];疗效好组的ADC1、ADC2高于疗效差组,且差异均有统计学意义[(1.44±0.10)×10−3 mm2/s对(1.10±0.06)×10−3 mm2/s,t=−2.92,P<0.05;1.68(1.55,1.85)×10−3 mm2/s对(1.33±0.06)×10−3 mm2/s,Z=−2.61,P<0.01]。ROC曲线分析结果显示,当ADC1≥1.34×10−3 mm2/s时,其对骨肉瘤新辅助化疗疗效评估的灵敏度为80%,特异度为81%,曲线下面积(AUC)为0.797(95%CI:0.629~0.965);当ADC2≥1.51×10−3 mm2/s时,其对骨肉瘤新辅助化疗疗效评估的灵敏度为90%,特异度为71%,AUC为0.777(95%CI:0.588~0.967)。从化疗一期结束后的DWI和ADC图像中共提取出1409个影像组学特征,按6∶4的比例随机分为训练集和测试集[24(疗效好:6,疗效坏:18)∶17(疗效好:4,疗效坏:13)],将训练集数据扩充为70(疗效好:20,疗效坏:50),经影像组学特征筛选后,最终得到5个最优影像组学特征,包括Interquartile Range、Skewness、Uniformity、Median、Maximum。采用逻辑回归分类器构建影像组学模型,训练集中该模型预测骨肉瘤新辅助化疗疗效的ROC曲线的AUC为0.881(95%CI:0.811~0.942),灵敏度为90%,特异度为74%;测试集中AUC为0.769(95%CI:0.515~0.933),灵敏度为75%,特异度为69%。 结论 基于MRI DWI的影像组学模型在评估骨肉瘤新辅助化疗疗效中的效能优于传统影像学(ADC),在临床应用中潜力较大。 Abstract:Objective To investigate the value of MRI DWI-based radiomics models for evaluating the treatment response in osteosarcoma after neoadjuvant chemoherapy. Methods A retrospective analysis was conducted on the medical records and imaging data of 41 patients with osteosarcoma (26 males and 15 females; aged (22.0±11.0) years, range of 11–49 years) who underwent MRI examinations before and after receiving neoadjuvant chemotherapy, and confirmed by postoperative histopathological examinations at the Third Hospital of Hebei Medical University from June 2015 to November 2017. In accordance with the postoperative histopathological examination results, patients with a tumor tissue necrosis rate of ≥90% were included in the good-efficacy group, and those with a necrosis rate of <90% were included in the poor-efficacy group. The apparent diffusion coefficient (ADC, denoted as ADC0, ADC1, and ADC2) were measured in all patients before neoadjuvant chemotherapy, within 5 days after the end of the first stage of chemotherapy, and after the completion of the entire chemotherapy. The differences in ADC were compared between the two groups. The region of interest of the lesion was manually delineated on DWI (b=1 000 s/mm2) and ADC images after the end of the first stage of chemotherapy, and the radiomics features were extracted. Data were divided into training set and validation set by using random grouping at 6∶4. The SMOTE algorithm was used to expand the data on the training set. The variance threshold, SelectKBest, and least absolute shrinkage and selection operator (LASSO) algorithm were used to screen the radiomics features. A radiomics model was constructed using a logistic regression classifier. Independent sample t-test or Wilcoxon rank sum test was used to compare the two groups. Receiver operating characteristic (ROC) curves were used to evaluate the predictive efficacy of traditional imaging (ADC) and radiomics models on the efficacy of neoadjuvant chemotherapy for osteosarcoma. Results A total of 10 and 31 cases were included in the good-efficacy and poor-efficacy groups, respectively. No statistically significant difference was found in the ADC0 value between the two groups ((0.95±0.05)×10−3 mm2/s vs. (1.05±0.05)×10−3 mm2/s, t=1.14, P>0.05)). The values of ADC1 and ADC2 in the good-efficacy group were higher than those in the poor-efficacy group, with statistical significance ((1.44±0.10)×10−3 mm2/s vs. (1.10±0.06)×10−3 mm2/s, t=−2.92, P<0.05; 1.68 (1.55, 1.85)×10−3 mm2/s vs. (1.33±0.06)×10−3 mm2/s, Z=−2.61, P<0.01). ROC curve analysis showed that when ADC1 ≥1.34×10−3 mm2/s, the sensitivity for evaluating the efficacy of neoadjuvant chemotherapy in osteosarcoma was 80%, the specificity was 81%, and the area under the curve (AUC) was 0.797 (95%CI: 0.629–0.965). When ADC2 ≥1.51×10−3 mm2/s, the sensitivity for evaluating the efficacy of neoadjuvant chemotherapy in osteosarcoma was 90%, the specificity was 71%, and the AUC was 0.777 (95%CI: 0.588–0.967). A total of 1 409 radiomics features were extracted from the DWI and ADC images after the end of the first stage of chemotherapy. They were randomly divided into training set and validation set at a ratio of 6∶4 (24 (good efficacy: 6, poor efficacy: 18)∶17 (good efficacy: 4, poor efficacy: 13)). The training set data were expanded to 70 (good efficacy: 20, poor efficacy: 50). After the radiomics features were screened, five optimal radiomics features were ultimately obtained, including InterquartileRange, Skewness, Uniformity, Median, and Maximum. Logistic regression classifier was used to construct a radiomics model. The ROC curves showed that in the training set, the AUC of the model for predicting the efficacy of neoadjuvant chemotherapy in osteosarcoma was 0.881 (95%CI: 0.811–0.942), with sensitivity of 90% and specificity of 74%. Meanwhile, in the validation set, the AUC was 0.769 (95%CI: 0.515–0.933), with sensitivity of 75% and specificity of 69%. Conclusion The radiomics model based on MRI DWI outperforms the traditional imaging (ADC) in evaluating the efficacy of neoadjuvant chemotherapy for osteosarcoma, showing great potential in clinical applications. -
表 1 新辅助化疗后不同疗效的2组骨肉瘤患者的ADC比 较(×10−3 mm2/s)
Table 1. Comparison of apparent diffusion coefficient values between two groups of osteosarcoma patients with different therapeutic effects after neoadjuvant chemotherapy (×10−3 mm2/s)
组别 ADC0
( )$\bar x\pm s $ ADC1
( )$\bar x\pm s $ ADC2
[M(Q1,Q3)或 ]$\bar x\pm s $ 疗效好组(n=10) 0.95±0.05 1.44±0.10 1.68(1.55, 1.85) 疗效差组(n=31) 1.05±0.05 1.10±0.06 1.33±0.06 检验值 1.14a −2.92a −2.61b P值 0.260 0.006 0.009 注:a表示t值;b表示Z值。ADC为表观扩散系数;ADC0为新辅助化疗前的ADC;ADC1为化疗一期结束后5 d内的ADC;ADC2为完成整个化疗后的ADC -
[1] Sadykova LR, Ntekim AI, Muyangwa-Semenova M, et al. Epidemiology and risk factors of osteosarcoma[J]. Cancer Invest, 2020, 38(5): 259−269. DOI: 10.1080/07357907.2020.1768401. [2] Jundt G. Updates to the WHO classification of bone tumours[J]. Pathologe, 2018, 39(2): 107−116. DOI: 10.1007/s00292-017-0396-4. [3] Manfredi R, Maresca G, Smaniotto D, et al. Cervical cancer response to neoadjuvant therapy: MR imaging assessment[J]. Radiology, 1998, 209: 819−824. DOI: 10.1148/radiology.209.3.9844681. [4] Lim HJ, Ong CAJ, Tan JWS, et al. Utility of positron emission tomography/computed tomography (PET/CT) imaging in the evaluation of sarcomas: a systematic review[J]. Crit Rev Oncol Hematol, 2019, 143: 1−13. DOI: 10.1016/j.critrevonc.2019.07.002. [5] Hao YW, An R, Xue YS, et al. Prognostic value of tumoral and peritumoral magnetic resonance parameters in osteosarcoma patients for monitoring chemotherapy response[J]. Eur Radiol, 2021, 31(5): 3518−3529. DOI: 10.1007/s00330-020-07338-y. [6] Oh C, Bishop MW, Cho SY, et al. 18F-FDG PET/CT in the management of osteosarcoma[J]. J Nucl Med, 2023, 64(6): 842−851. DOI: 10.2967/jnumed.123.265592. [7] Wang JF, Sun ML, Liu DW, et al. Correlation between apparent diffusion coefficient and histopathology subtypes of osteosarcoma after neoadjuvant chemotherapy[J]. Acta Radiol, 2017, 58(8): 971−976. DOI: 10.1177/0284185116678276. [8] Yu H, Gao L, Shi RQ, et al. Monitoring early responses to neoadjuvant chemotherapy and the factors affecting neoadjuvant chemotherapy responses in primary osteosarcoma[J]. Quant Imaging Med Surg, 2023, 13(6): 3716−3725. DOI: 10.21037/qims-22-1095. [9] Degnan AJ, Chung CY, Shah AJ. Quantitative diffusion-weighted magnetic resonance imaging assessment of chemotherapy treatment response of pediatric osteosarcoma and Ewing sarcoma malignant bone tumors[J]. Clin Imaging, 2017, 47: 9−13. DOI: 10.1016/j.clinimag.2017.08.003. [10] Habre C, Dabadie A, Loundou AD, et al. Diffusion-weighted imaging in differentiating mid-course responders to chemotherapy for long-bone osteosarcoma compared to the histologic response: an update[J]. Pediatr Radiol, 2021, 51(9): 1714−1723. DOI: 10.1007/s00247-021-05037-4. [11] 王芳, 夏雨薇, 柴象飞, 等. 影像组学分析流程及临床应用的研究进展[J]. 中华解剖与临床杂志, 2021, 26(2): 236−241. DOI: 10.3760/cma.j.cn101202-20200701-00200.
Wang F, Xia YW, Chai XF, et al. Analysis process and clinical application of radiomics[J]. Chin J Anat Clin, 2021, 26(2): 236−241. DOI: 10.3760/cma.j.cn101202-20200701-00200.[12] He LJ, Yang HN, Huang JS. The tumor immune microenvironment and immune-related signature predict the chemotherapy response in patients with osteosarcoma[J/OL]. BMC Cancer, 2021, 21(1): 581[2022-11-22]. https://bmccancer.biomedcentral.com/articles/10.1186/s12885-021-08328-z. DOI: 10.1186/s12885-021-08328-z. [13] Prabowo Y, Setiawan I, Kamal AF, et al. Correlation between prognostic factors and the histopathological response to neoadjuvant chemotherapy in osteosarcoma: a retrospective study[J]. Int J Surg Oncol, 2021, 2021: 8843325. DOI: 10.1155/2021/8843325. [14] Ritter J, Bielack SS. Osteosarcoma[J]. Ann Oncol, 2010, 21(Suppl 7): Svii320−vii325. DOI: 10.1093/annonc/mdq276. [15] Kerwin WS, O'Brien KD, Ferguson MS, et al. Inflammation in carotid atherosclerotic plaque: a dynamic contrast-enhanced MR imaging study[J]. Radiology, 2006, 241(2): 459−468. DOI: 10.1148/radiol.2412051336. [16] Laux CJ, Berzaczy G, Weber M, et al. Tumour response of osteosarcoma to neoadjuvant chemotherapy evaluated by magnetic resonance imaging as prognostic factor for outcome[J]. Int Orthop, 2015, 39(1): 97−104. DOI: 10.1007/s00264-014-2606-5. [17] Choi SH, Jung SC, Kim KW, et al. Perfusion MRI as the predictive/prognostic and pharmacodynamic biomarkers in recurrent malignant glioma treated with bevacizumab: a systematic review and a time-to-event meta-analysis[J]. J Neurooncol, 2016, 128(2): 185−194. DOI: 10.1007/s11060-016-2102-4. [18] Zeitoun R, Shokry AM, Khaleel SA, et al. Osteosarcoma subtypes: magnetic resonance and quantitative diffusion weighted imaging criteria[J]. J Egypt Natl Canc Inst, 2018, 30(1): 39−44. DOI: 10.1016/j.jnci.2018.01.006. [19] Liu HH, Zhang CY, Wang LJ, et al. MRI radiomics analysis for predicting preoperative synchronous distant metastasis in patients with rectal cancer[J]. Eur Radiol, 2019, 29(8): 4418−4426. DOI: 10.1007/s00330-018-5802-7. [20] Kumar V, Gu YH, Basu S, et al. Radiomics: the process and the challenges[J]. Magn Reson Imaging, 2012, 30(9): 1234−1248. DOI: 10.1016/j.mri.2012.06.010. [21] Kayal EB, Kandasamy D, Khare K, et al. Texture analysis for chemotherapy response evaluation in osteosarcoma using MR imaging[J]. NMR Biomed, 2021, 34(2): e4426. DOI: 10.1002/nbm.4426. [22] 苗文杰, 杨光杰, 聂佩, 等. 高分辨率CT影像组学联合传统影像学征象预测肺腺癌微血管浸润的价值[J]. 国际放射医学核医学杂志, 2020, 44(9): 541−547. DOI: 10.3760/cma.j.cn121381-201909006-00075.
Miao WJ, Yang GJ, Nie P, et al. Value of HRCT radiomics combined with traditional imaging features in predicting microvascular invasion of lung adenocarcinoma[J]. Int J Radiat Med Nucl Med, 2020, 44(9): 541−547. DOI: 10.3760/cma.j.cn121381-201909006-00075.[23] Lee SK, Jee WH, Jung CK, et al. Prediction of poor responders to neoadjuvant chemotherapy in patients with osteosarcoma: additive value of diffusion-weighted MRI including volumetric analysis to standard MRI at 3T[J/OL]. PLoS One, 2020, 15(3): e0229983[2022-11-22]. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0229983. DOI: 10.1371/journal.pone.0229983.