多模态功能磁共振成像联合Ki-67指数评估头颈部横纹肌肉瘤病理分型

doi:10.3969/j.issn.1674-8115.2026.09.010

摘要/Abstract

摘要：

目的·基于由弥散加权成像（diffusion-weighted imaging，DWI）获得的表观扩散系数（apparent diffusion coefficient，ADC）及由动态增强磁共振成像（dynamic contrast-enhanced magnetic resonance imaging，DCE-MRI）获得的时间-信号强度曲线（time-intensity curve，TIC）类型，探讨多模态功能MRI联合细胞增殖标志物Ki-67在头颈部横纹肌肉瘤（rhabdomyosarcoma，RMS）术前病理分型评估中的临床价值。方法·回顾性纳入2015年1月至2026年3月在上海交通大学医学院附属第九人民医院就诊并行手术切除的头颈部RMS患者。收集患者的人口学信息、临床资料及影像学资料。依据术后病理分型将患者分为胚胎型RMS（n=21）、腺泡型RMS（n=14）与梭形细胞/硬化型RMS（n=15）共3组。比较3组间多模态功能MRI参数（ADC值与TIC类型）、年龄、性别、Ki-67指数等差异，并分析ADC值与Ki-67指数的相关性。采用组内相关系数与Cohen's κ系数评估多模态功能MRI参数在观察者内和观察者间的一致性。采用多元Logistic回归模型分析多模态功能MRI参数与Ki-67指数对3种不同病理分型的预测作用，并绘制混淆矩阵评估模型性能。结果·ADC值与TIC类型的观察者间与观察者内一致性均较高。3种不同病理分型RMS在ADC值（P<0.001）、Ki-67指数（P<0.001）及TIC类型（P=0.031）方面差异具有统计学意义。Spearman相关分析显示ADC值与Ki-67指数呈负相关（r=-0.362，P=0.010）。多元Logistic回归显示，以胚胎型RMS为参照，ADC值与腺泡型RMS和梭形细胞/硬化型RMS独立相关，Ki-67指数与梭形细胞/硬化型RMS独立相关。模型在胚胎型、腺泡型与梭形细胞/硬化型3种不同病理分型RMS中的预测准确率分别为66.7%、57.1%、93.3%，总体预测准确率为72.0%。结论·多模态功能MRI联合Ki-67可有效预测头颈部RMS的病理分型，对术前风险分层与预后评估具有重要临床价值。

关键词: 横纹肌肉瘤, 头颈部, 多模态功能磁共振成像, 病理分型

Abstract:

Objective ·To investigate the clinical value of multimodal functional MRI combined with the cell proliferation marker Ki‑67 in the preoperative pathological classification of head and neck rhabdomyosarcoma (RMS), based on the apparent diffusion coefficient (ADC) derived from diffusion‑weighted imaging (DWI) and time‑intensity curve (TIC) types obtained from dynamic contrast‑enhanced magnetic resonance imaging (DCE‑MRI). Methods ·Patients diagnosed with head and neck RMS who underwent surgical resection at Shanghai Ninth People′s Hospital, Shanghai Jiao Tong University School of Medicine between January 2015 and March 2026 were retrospectively enrolled. Demographic, clinical, and imaging data of all patients were collected. According to postoperative pathological types, patients were divided into three groups: embryonal RMS (n=21), alveolar RMS (n=14), and spindle cell/sclerosing RMS (n=15). Intergroup differences in multimodal functional MRI parameters (ADC values and TIC types), age, sex, and Ki‑67 index were analyzed, and the correlation between ADC values and Ki‑67 index was analyzed. Intraclass correlation coefficient (ICC) and Cohen′s κ coefficient were calculated to assess the inter‑ and intra-reader consistency of multimodal functional MRI parameters. Multinomial Logistic regression was performed to evaluate the predictive value of ADC values, TIC types, and Ki-67 index for the three pathological types, and a confusion matrix was used to assess model performance. Results ·Both ADC values and TIC types exhibited good inter‑reader and intra‑reader consistency. Statistically significant differences were observed among the three RMS pathological types in ADC values (P<0.001), Ki‑67 index (P<0.001), and TIC types (P=0.031). Spearman correlation analysis revealed a significant negative correlation between ADC value and Ki‑67 index (r=-0.362, P=0.010). Multinomial Logistic regression demonstrated that, using the embryonal RMS as the reference, the ADC value was independently associated with both the alveolar RMS and the spindle cell/sclerosing RMS, while the Ki-67 index was independently associated with the spindle cell/sclerosing RMS. The predictive accuracy of the model for embryonal, alveolar, and spindle cell/sclerosing RMS was 66.7%, 57.1%, and 93.3%, respectively, with an overall predictive accuracy of 72.0%. Conclusion ·The combination of multimodal functional MRI and Ki‑67 index enables effective prediction of pathological types of head and neck RMS, and has crucial clinical value for preoperative risk stratification and prognostic assessment.

Key words: rhabdomyosarcoma (RMS), head and neck, multimodal functional MRI, pathological classification

中图分类号:

R445.2

肖华, 张紫旻, 姜梦达, 郭久红. 多模态功能磁共振成像联合Ki-67指数评估头颈部横纹肌肉瘤病理分型[J]. 上海交通大学学报（医学版）, 2026, 46(9): 1252-1259.

Xiao Hua, Zhang Zimin, Jiang Mengda, Guo Jiuhong. Multimodal functional magnetic resonance imaging combined with Ki-67 index for assessing pathological typing of head and neck rhabdomyosarcoma[J]. Journal of Shanghai Jiao Tong University (Medical Science), 2026,(9): 1252-1259.

图/表 7

表1 不同病理分型RMS患者的人口学信息及临床资料

Tab 1 Demographic and clinical characteristics of RMS patients with different pathological types

Pathological type	E-RMS (n=21)	A-RMS (n=14)	SS-RMS (n=15)
Age/year	13.3±8.7	21.8±12.8	17.4±15.2
Sex/n(%)
Female	9 (42.9)	4 (28.6)	2 (13.3)
Male	12 (57.1)	10 (71.4)	13 (86.7)
Location/n(%)
Masticator space	7 (33.3)	1 (7.1)	8 (53.3)
Orbit	9 (42.9)	2 (14.3)	2 (13.3)
Nasal cavity & paranasal sinuses	0 (0)	8 (57.1)	0 (0)
Nasal region	0 (0)	0 (0)	3 (20.0)
Masseter muscle	0 (0)	1 (7.1)	1 (6.7)
Buccal region	2 (9.5)	0 (0)	0 (0)
Lip	0 (0)	2 (14.3)	0 (0)
Parotid gland	1 (4.8)	0 (0)	0 (0)
Mandible	0 (0)	0 (0)	1 (6.7)
Palate	1 (4.8)	0 (0)	0 (0)
Temporal bone	1 (4.8)	0 (0)	0 (0)

图1 头颈部横纹肌肉瘤的功能MRI表现及相应病理图像Note： A‒C. E-RMS in a 6-year-old male (right masticator space). A. type Ⅱ TIC. B. ADC value (0.780×10-3 mm2·s-1). C. Pathological image (H-E staining, ×200), showing poorly differentiated stellate and small round cells with round or oval, hyperchromatic nuclei and easily observable mitotic figures. D‒F. A-RMS in a 17-year-old male (right maxillary sinus). D. type Ⅱ TIC. E. ADC value (0.750×10-3 mm2·s-1). F. Pathological image (H-E staining, ×200), showing tumor cells arranged in sheets or nests, with central tumor cells losing adhesion and shedding into the alveolar lumen due to degeneration, forming characteristic alveolar-like structures. G‒I. SS-RMS in a 28-year-old male (left masticator space). G. type Ⅱ TIC. H. ADC value (1.110×10-3 mm2·s-1). I. Pathological image (H-E staining, ×200), showing elongated spindle cells arranged in fascicles or storiform patterns, with elongated oval nuclei, deeply eosinophilic cytoplasm, mild nuclear atypia, and visible mitotic figures.

Fig 1 Functional MRI features of head and neck rhabdomyosarcoma and the corresponding pathological images

表2 多模态功能MRI参数的观察者内及观察者间一致性分析

Tab 2 Inter- and intra-reader agreement of multimodal functional MRI parameters in RMS

Parameter	ICC (95%CI)/κ (95%CI)
Parameter	Inter-reader	Intra-reader
ADC value	0.918 (0.866‒0.951)	0.902 (0.847‒0.941)
TIC type	0.876 (0.812‒0.940)	0.935 (0.891‒0.964)

表3 3种病理分型RMS的多模态功能MRI参数与Ki-67指数比较

Tab 3 Comparison of multimodal functional MRI parameters and Ki-67 index among three pathological types of RMS

Parameter	E-RMS (n=21)	A-RMS (n=14)	SS-RMS (n=15)	H/χ² value	P value
ADC value/(×10^-3 mm²·s^-1)	0.890 (0.770, 1.060)	0.745 (0.658, 0.845)	1.010 (1.010, 1.300)	26.688	<0.001
TIC type/n(%)				6.845	0.031
Ⅰ	5 (23.8)	5 (35.7)	10 (66.7)
Ⅱ	16 (76.2)	9 (64.3)	5 (33.3)
Ki-67 index/%	40.000 (30.000, 60.000)	40.000 (37.500, 50.000)	20.000 (15.000, 30.000)	17.019	<0.001

图2 ADC值与Ki-67指数相关性（A）及多模态功能MRI参数联合Ki-67指数预测RMS不同病理分型模型的效能矩阵（B）

Fig 2 Correlation between ADC values and Ki-67 index (A), and performance matrix of the model combining multimodal functional MRI parameters with Ki-67 index for predicting different pathological types of RMS (B)

表4 基于多模态功能MRI参数联合Ki-67指数的RMS病理分型的多元Logistic回归分析

Tab 4 Multinomial Logistic regression analysis of RMS pathological subtypes based on multimodal functional MRI parameters combined with Ki-67 index

Parameter	B	OR (95%CI)	P value
A-RMS^①
ADC value/(×10^-3 mm²·s^-1)	-8.841	0 (0‒0.154)	0.013
TIC type	0.164	1.178 (0.216‒6.441)	0.850
Ki-67 index/%	-0.026	0.975 (0.925‒1.027)	0.331
SS-RMS^①
ADC value/(×10^-3 mm²·s^-1)	7.634	2 067.165 (6.461‒661 414.539)	0.009
TIC type	0.886	2.425 (0.242‒24.344)	0.452
Ki-67 index/%	-0.079	0.924 (0.862‒0.991)	0.027

表5 多模态功能MRI参数联合Ki-67指数预测RMS不同病理分型的准确率

Tab 5 Accuracy of the model combining multimodal functional MRI parameters with Ki-67 index in predicting different pathological types of RMS

Ground truth	Predicted value					Accuracy/%
Ground truth	E-RMS		A-RMS	SS-RMS		Accuracy/%
Total	‒	‒			‒	72.0
E-RMS	14		4	3		66.7
A-RMS	6		8	0		57.1
SS-RMS	1		0	14		93.3

参考文献 27

[1]	中国抗癌协会小儿肿瘤专业委员会, 中华医学会儿科学分会血液学组, 中华医学会小儿外科学分会肿瘤组. 中国儿童及青少年横纹肌肉瘤诊疗建议(CCCG－RMS－2016)[J]. 中华儿科杂志, 2017, 55(10): 724-728.
	Chinese Children′s Cancer Group, Chinese Anti-Cancer Association, the Subspecialty Group of Hematology, the Society of Pediatrics, Chinese Medical Association, the Subspecialty Group of Oncology, the Society of Pediatric Surgery, Chinese Medical Association. Recommendations for the diagnosis and treatment of rhabdomyosarcoma in Chinese children and adolescents (CCCG-RMS-2016)[J]. Chinese Journal of Pediatrics, 2017, 55(10): 724-728.
[2]	Darwish C, Shim T, Sparks A D, et al. Pediatric head and neck rhabdomyosarcoma: an analysis of treatment and survival in the United States (1975‒2016)[J]. Int J Pediatr Otorhinolaryngol, 2020, 139: 110403.
[3]	Sbaraglia M, Bellan E, Dei Tos A P. The 2020 WHO classification of soft tissue tumours: news and perspectives[J]. Pathologica, 2021, 113(2): 70-84.
[4]	中华医学会病理学分会儿科病理学组, 中国抗癌协会小儿肿瘤专业委员会病理学组, 福棠儿童医学发展研究中心病理专业委员会. 儿童横纹肌肉瘤病理诊断规范化专家共识[J]. 中华病理学杂志, 2021, 50(10): 1110-1115.
	Expert Group of Pediatrics of Chinese Society of Pathology, Expert Group of Pathology of Chinese Children Cancer Group, Expert Group of Pathology Committee of Futang Research Center of Pediatric Development. Consensus on the diagnosis of pediatric rhabdomyosarcoma[J]. Chinese Journal of Pathology, 2021, 50(10): 1110-1115.
[5]	Gennaro N, Marrari A, Renne S L, et al. Multimodality imaging of adult rhabdomyosarcoma: the added value of hybrid imaging[J]. Br J Radiol, 2020, 93(1112): 20200250.
[6]	Agaram N P. Evolving classification of rhabdomyosarcoma[J]. Histopathology, 2022, 80(1): 98-108.
[7]	Wang X Y, Song L Y, Chong V, et al. Multiparametric MRI findings of sinonasal rhabdomyosarcoma in adults with comparison to carcinoma[J]. J Magn Reson Imaging, 2017, 45(4): 998-1004.
[8]	Inarejos Clemente E J, Navallas M, Barber Martínez de la Torre I, et al. MRI of rhabdomyosarcoma and other soft-tissue sarcomas in children[J]. Radiographics, 2020, 40(3): 791-814.
[9]	Parlak Ş, Gümeler E, Bulut E. Pediatric head and neck rhabdomyosarcoma; the role of MRI in differentiation from other common pediatric malignant head and neck tumors[J]. Turk J Pediatr, 2022, 64(3): 519-530.
[10]	Saboo S S, Krajewski K M, Zukotynski K, et al. Imaging features of primary and secondary adult rhabdomyosarcoma[J]. Am J Roentgenol, 2012, 199(6): W694-W703.
[11]	Chhabra A, Ashikyan O, Slepicka C, et al. Conventional MR and diffusion-weighted imaging of musculoskeletal soft tissue malignancy: correlation with histologic grading[J]. Eur Radiol, 2019, 29(8): 4485-4494.
[12]	Yuan J H, Xie D S, Fang S B, et al. Qualitative and quantitative MRI analysis of alveolar soft part sarcoma: correlation with histological grade and Ki-67 expression[J]. Insights Imaging, 2024, 15(1): 142.
[13]	Yuan Y, Zeng D W, Liu Y J, et al. DWI and IVIM are predictors of Ki67 proliferation index: direct comparison of MRI images and pathological slices in a murine model of rhabdomyosarcoma[J]. Eur Radiol, 2020, 30(3): 1334-1341.
[14]	Shomal Zadeh F, Pooyan A, Alipour E, et al. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in differentiation of soft tissue sarcoma from benign lesions: a systematic review of literature[J]. Skeletal Radiol, 2024, 53(7): 1343-1357.
[15]	Vallières M, Serban M, Benzyane I, et al. Investigating the role of functional imaging in the management of soft-tissue sarcomas of the extremities[J]. Phys Imaging Radiat Oncol, 2018, 6: 53-60.
[16]	Kataoka M, Honda M, Sagawa H, et al. Ultrafast dynamic contrast-enhanced MRI of the breast: from theory to practice[J]. J Magn Reson Imaging, 2024, 60(2): 401-416.
[17]	Del Grande F, Subhawong T, Weber K, et al. Detection of soft-tissue sarcoma recurrence: added value of functional MR imaging techniques at 3.0 T[J]. Radiology, 2014, 271(2): 499-511.
[18]	Lv C L, Xue X L, Huang M G, et al. The dynamic contrast enhanced-magnetic resonance imaging and diffusion-weighted imaging features of alveolar soft part sarcoma[J]. Quant Imaging Med Surg, 2023, 13(10): 7269-7280.
[19]	Lee J H, Yoon Y C, Seo S W, et al. Soft tissue sarcoma: DWI and DCE-MRI parameters correlate with Ki-67 labeling index[J]. Eur Radiol, 2020, 30(2): 914-924.
[20]	Yang B T, Wang Y Z, Dong J Y, et al. MRI study of solitary fibrous tumor in the orbit[J]. Am J Roentgenol, 2012, 199(4): W506-W511.
[21]	Soffer S, Amitai M M, Shimon O, et al. Rhabdomyosarcoma disease spread evaluation on CT scans: association with primary tumor size and Ki-67 proliferation marker[J]. Clin Imag, 2019, 56: 41-46.
[22]	McCarville M B. What MRI can tell us about neurogenic tumors and rhabdomyosarcoma[J]. Pediatr Radiol, 2016, 46(6): 881-890.
[23]	van Ewijk R, Chatziantoniou C, Adams M, et al. Quantitative diffusion-weighted MRI response assessment in rhabdomyosarcoma: an international retrospective study on behalf of the European paediatric Soft tissue sarcoma Study Group Imaging Committee[J]. Pediatr Radiol, 2023, 53(12): 2539-2551.
[24]	Liu J H, Qi M, Huang W H, et al. The magnetic resonance characteristics of sinonasal rhabdomyosarcoma in adults: analysis of 27 cases and comparison with pathological subtypes[J]. BMC Med Imaging, 2023, 23(1): 98.
[25]	Zeng J J, Liu L, Li J Y, et al. MRI features of different types of sinonasan rhabdomyosarcomas: a series of eleven cases[J]. Dentomaxillofac Radiol, 2021, 50(8): 20210030.
[26]	Yao R, Cheng A L, Zhang Z W, et al. Correlation between apparent diffusion coefficient and the Ki-67 proliferation index in grading pediatric glioma[J]. J Comput Assist Tomogr, 2023, 47(2): 322-328.
[27]	Sun M T, Cheng J L, Zhang Y, et al. Application of DWIBS in malignant lymphoma: correlation between ADC values and Ki-67 index[J]. Eur Radiol, 2018, 28(4): 1701-1708.