Abstract：
OBJECTIVE: To evaluate the accuracy of preoperative contrast-enhanced magnetic resonance imaging (MRI) in the assessment of radiological depth of invasion (rDOI) and bone invasion in patients with oral cavity cancer, and the prognostic value of preoperative rDOI.
METHODS: This retrospective study included patients with surgically resected oral cavity cancer and preoperative MRI acquired within four weeks before surgery. Two readers evaluated the MRI to assess the superficial and deep bone invasion, preoperative T stage, and measured the rDOI. The rDOI was compared to the histopathological DOI (pDOI), used as reference standard. Prognostic value of preoperative features for the disease-specific survival was evaluated using the Kaplan-Meier curve and multivariable Cox proportional hazards analysis.
RESULTS: The final population included 80 patients (50 males, mean age 67.7 ± 13.6 years). There was a strong statistically significant correlation between the rDOI (median 10 mm) and the pDOI (median 9 mm) (ρ: 0.978, p < 0.001). The agreement between MRI and histopathological T stage was excellent (k = 0.93, 95% CI 0.86, 0.99). The sensitivity and specificity of preoperative MRI were 93.3% and 98.8% for deep bone invasion, while they were 75.0% and 95.8% for superficial bone invasion, respectively. The rDOI > 10 mm was associated with poorer disease-specific survival (log-rank p = 0.016). The rDOI remained the only independent preoperative predictor associated with poorer disease-specific survival at multivariable analysis (hazard ratio 5.5; 95% CI 1.14, 26.58; p = 0.033).
CONCLUSIONS: Preoperative MRI is accurate for the assessment of DOI and bone invasion. The rDOI is an independent preoperative predictor of disease-specific survival in patients with oral cavity cancer.
METHODS: This retrospective study included patients with surgically resected oral cavity cancer and preoperative MRI acquired within four weeks before surgery. Two readers evaluated the MRI to assess the superficial and deep bone invasion, preoperative T stage, and measured the rDOI. The rDOI was compared to the histopathological DOI (pDOI), used as reference standard. Prognostic value of preoperative features for the disease-specific survival was evaluated using the Kaplan-Meier curve and multivariable Cox proportional hazards analysis.
RESULTS: The final population included 80 patients (50 males, mean age 67.7 ± 13.6 years). There was a strong statistically significant correlation between the rDOI (median 10 mm) and the pDOI (median 9 mm) (ρ: 0.978, p < 0.001). The agreement between MRI and histopathological T stage was excellent (k = 0.93, 95% CI 0.86, 0.99). The sensitivity and specificity of preoperative MRI were 93.3% and 98.8% for deep bone invasion, while they were 75.0% and 95.8% for superficial bone invasion, respectively. The rDOI > 10 mm was associated with poorer disease-specific survival (log-rank p = 0.016). The rDOI remained the only independent preoperative predictor associated with poorer disease-specific survival at multivariable analysis (hazard ratio 5.5; 95% CI 1.14, 26.58; p = 0.033).
CONCLUSIONS: Preoperative MRI is accurate for the assessment of DOI and bone invasion. The rDOI is an independent preoperative predictor of disease-specific survival in patients with oral cavity cancer.
摘要：
目的：评估术前对比增强磁共振成像（MRI）评估口腔癌患者放射浸润深度（rDOI）和骨浸润的准确性，术前rDOI的预后价值。
方法：这项回顾性研究包括手术切除的口腔癌患者和手术前四周内获得的术前MRI。两名读者评估了MRI以评估浅表和深层骨侵犯，术前T期,并测量了RDOI。将rDOI与组织病理学DOI（pDOI）进行比较，用作参考标准。使用Kaplan-Meier曲线和多变量Cox比例风险分析评估术前特征对疾病特异性生存的预后价值。
结果：最终人群包括80名患者（50名男性，平均年龄67.7±13.6岁)。rDOI(中值10mm)和pDOI(中值9mm)之间存在很强的统计学显著相关性(ρ:0.978,p<0.001)。MRI与组织病理学T分期之间的一致性非常好（k=0.93，95％CI0.86，0.99）。术前MRI对深部骨侵犯的敏感性和特异性分别为93.3%和98.8%,而浅表骨侵犯分别为75.0%和95.8%，分别。rDOI>10mm与较差的疾病特异性存活率相关(log-rankp=0.016)。在多变量分析中，rDOI仍然是唯一与疾病特异性生存率较差相关的独立术前预测因子(风险比5.5;95%CI1.14,26.58;p=0.033)。
结论：术前MRI可准确评估DOI和骨侵犯。rDOI是口腔癌患者疾病特异性生存的独立术前预测因子。
方法：这项回顾性研究包括手术切除的口腔癌患者和手术前四周内获得的术前MRI。两名读者评估了MRI以评估浅表和深层骨侵犯，术前T期,并测量了RDOI。将rDOI与组织病理学DOI（pDOI）进行比较，用作参考标准。使用Kaplan-Meier曲线和多变量Cox比例风险分析评估术前特征对疾病特异性生存的预后价值。
结果：最终人群包括80名患者（50名男性，平均年龄67.7±13.6岁)。rDOI(中值10mm)和pDOI(中值9mm)之间存在很强的统计学显著相关性(ρ:0.978,p<0.001)。MRI与组织病理学T分期之间的一致性非常好（k=0.93，95％CI0.86，0.99）。术前MRI对深部骨侵犯的敏感性和特异性分别为93.3%和98.8%,而浅表骨侵犯分别为75.0%和95.8%，分别。rDOI>10mm与较差的疾病特异性存活率相关(log-rankp=0.016)。在多变量分析中，rDOI仍然是唯一与疾病特异性生存率较差相关的独立术前预测因子(风险比5.5;95%CI1.14,26.58;p=0.033)。
结论：术前MRI可准确评估DOI和骨侵犯。rDOI是口腔癌患者疾病特异性生存的独立术前预测因子。
