河北医科大学学报 ›› 2025, Vol. 46 ›› Issue (6): 649-655.doi: 10.3969/j.issn.1007-3205.2025.06.005

• • 上一篇    下一篇

MRI纹理分析+DWI预测鼻咽癌放化疗应答性的临床研究

  

  1. 江苏省连云港市第一人民医院,南京医科大学康达学院第一附属医院医学影像科,江苏 连云港 222000

  • 出版日期:2025-06-25 发布日期:2025-07-04
  • 作者简介:贡瑞敏(1998-),女,江苏东海人,江苏省连云港市第一人民医院初级技师,医学学士,从事头颈部影像诊断研究。

  • 基金资助:
    江苏省优势学科建设工程项目(YSHL2005-141)

Clinical study of MRI texture analysis + DWI in predicting the response to radiotherapy and chemotherapy in nasopharyngeal carcinoma

  1. Department of Medical Imaging, the First People′s Hospital of Lianyungang City/the First Affiliated 
    Hospital of Kangda College, Nanjing Medical University, Jiangsu Province, 
    Lianyungang 222000, China

  • Online:2025-06-25 Published:2025-07-04

摘要: 目的研究磁共振成像(magnetic resonance imaging,MRI)纹理分析+弥散加权成像(diffusion-weighted imaging,DWI)预测鼻咽癌放化疗应答性的临床价值。
方法选取2020年7月—2023年8月江苏省连云港市第一人民医院收治的行放化疗治疗的鼻咽癌患者102例,于放化疗前行MRI纹理分析+DWI扫描。根据放化疗应答性分为应答良好组(n=46)、不良组(n=56)。统计2组一般资料、MRI纹理分析、DWI扫描结果,分析MRI纹理分析、DWI扫描结果与鼻咽癌放化疗应答性相关性,分析MRI纹理分析、DWI扫描结果对鼻咽癌放化疗应答性的影响及预测价值。
结果不良组治疗前集群突出(群集显著性)[(816.40±220.36) vs. (492.13±163.58)]、Haralick相关性[(12 834.28±875.54) vs. (10 025.14±783.29)]、表观弥散系数(apparent diffusion coefficient,ADC)[(0.85±0.10)×10-3 mm2/s vs. (0.78±0.09)×10-3 mm2/s]高于良好组(P<0.05);群集显著性、Haralick相关性、ADC与鼻咽癌放化疗应答性呈负相关(r=-0.662、-0.594、-0.597,均P<0.05);偏回归分析显示,治疗前群集显著性、Haralick相关性、ADC均是鼻咽癌放化疗应答性的相关因素(P<0.05);群集显著性+Haralick相关性+ADC联合预测鼻咽癌放化疗应答不良的曲线下面积最大,为0.938,预测敏感度为82.14%,特异度为89.13%。
结论鼻咽癌患者MRI纹理参数、DWI参数与放化疗应答性密切相关,群集显著性、Haralick相关性、ADC三者联合预测鼻咽癌放化疗应答不良具有较高价值。


关键词: 鼻咽癌, MRI纹理, 放化疗应答性

Abstract: Objective To investigate the clinical value of magnetic resonance imaging (MRI) texture analysis plus diffusion-weighted imaging (DWI) in predicting the response to radiotherapy and chemotherapy in nasopharyngeal carcinoma (NPC). 
Methods A total of 102 patients with NPC treated by radiotherapy and chemotherapy in the First People′s Hospital of Lianyungang City from July 2020 to August 2023 were selected. MRI texture analysis and DWI scanning were performed before radiotherapy and chemotherapy. They were divided into good response group (n=46) and poor response group (n=56) according to the response to radiotherapy and chemotherapy. Baseline data, MRI texture analysis and DWI scanning results of the two groups were recorded to analyze the correlation of MRI texture analysis and DWI scanning results with the response to radiotherapy and chemotherapy for NPC, and to analyze the impact of MRI texture analysis and DWI scanning results on the response to radiotherapy and chemotherapy and predictive value for NPC. 
Results Before treatment, the poor response group exhibited higher values in cluster significance [(816.40±220.36) vs. (492.13±163.58)], Haralick correlation [(12 834.28±875.54) vs. (10 025.14±783.29)], and apparent diffusion coefficient (ADC) [(0.85±0.10)×10-3 mm2/s vs. (0.78±0.09)×10-3 mm2/s] compared with the good response group (P<0.05). Cluster significance, Haralick correlation and ADC were negatively correlated with the response to radiotherapy and chemotherapy (r=-0.662, -0.594, -0.597, all P<0.05). Partial regression analysis showed that cluster significance, Haralick correlation and ADC were correlated with the response to radiotherapy and chemotherapy before treatment (P<0.05). The area under the curve of cluster significance + Haralick correlation + ADC in combination for predicting poor response to radiotherapy and chemotherapy for NPCwas the largest (0.938), with sensitivity of 82.14% and specificity of 89.13%. 
Conclusion MRI texture parameters and DWI parameters are closely related to the response to radiotherapy and chemotherapy in patients with NPC. The combination of cluster significance, Haralick correlation and ADC has high value in predicting the poor response to radiotherapy and chemotherapy in NPC patients. 


Key words: nasopharyngeal carcinoma, MRI texture, response to chemoradiotherapy