Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea.
Department of Pathology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea.
Eur J Radiol. 2022 Sep;154:110396. doi: 10.1016/j.ejrad.2022.110396. Epub 2022 Jun 11.
To elucidate whether a high level of microsatellite instability (MSI-high) in colon cancer influences the CT assessment of regional lymph node (rLN) morphology and diagnostic performance on predicting pathological node-negative (pN0) patients.
We retrospectively reviewed 507 patients with cecal/proximal ascending colon cancer (age, 63.0 ± 11.6 years; MSI-stable, n = 398; MSI-high, n = 109) who underwent right hemicolectomy between July 1, 2009, and December 31, 2018. Preoperative CT images were assessed for number of rLNs, long/short diameter of the largest rLN, and CT LN grade (CTN0, low probability of metastasis; CTN1, borderline; CTN2, high probability). Sensitivity, specificity, positive predictive value and negative predictive value for predicting pN0 was calculated. Multivariable logistic regression analysis was performed. Statistical significance was defined as P < 0.05.
A study population of 507 patients (mean age ± standard deviation, 63.0 ± 11.6; 264 women) were evaluated. In patients with rLN metastasis, the rLN long axis (pN1: P = 0.013, pN2: P = 0.039) and short axis (pN1: P = 0.001, pN2: P = 0.009) were both longer in MSI-high tumors compared with MSI-stable tumors. High specificity for predicting pN0 was only achieved in MSI-high tumors [sensitivity = 58.3% (n = 137/235), specificity = 71.2% (n = 116/163); sensitivity = 38.4% (n = 33/86), specificity = 91.3% (n = 21/23)]. Multivariable logistic regression indicated MSI-high (P < 0.001, odds ratio = 3.701), smaller LN long axis (P = 0.023, odds ratio = 0.905), and lower CT LN grade (CTN0: P = 0.009, odds ratio = 2.987; CTN1: P = 0.014, odds ratio = 2.195) as significant parameters in predicting pN0.
MSI-high colon cancer is associated with larger rLNs and high specificity for predicting pN0 on CT assessment.
阐明结肠癌中高水平微卫星不稳定性（MSI-high）是否会影响 CT 评估区域淋巴结（rLN）形态以及预测病理阴性淋巴结（pN0）患者的诊断性能。
我们回顾性分析了 2009 年 7 月 1 日至 2018 年 12 月 31 日期间接受右半结肠切除术的 507 例盲肠/升结肠近端癌患者（年龄，63.0±11.6 岁；MSI-稳定，n=398；MSI-high，n=109）。对术前 CT 图像进行 rLN 数量、最大 rLN 长/短径以及 CTLN 分级（CTN0，转移概率低；CTN1，边界；CTN2，转移概率高）评估。计算预测 pN0 的敏感性、特异性、阳性预测值和阴性预测值。进行多变量逻辑回归分析。定义统计学意义为 P<0.05。
对 507 例患者（平均年龄±标准差，63.0±11.6；264 例女性）进行了评估。在有 rLN 转移的患者中，MSI-high 肿瘤的 rLN 长轴（pN1：P=0.013，pN2：P=0.039）和短轴（pN1：P=0.001，pN2：P=0.009）均长于 MSI-稳定肿瘤。MSI-high 肿瘤对预测 pN0 具有较高的特异性[敏感性=58.3%（n=137/235），特异性=71.2%（n=116/163）；敏感性=38.4%（n=33/86），特异性=91.3%（n=21/23）]。多变量逻辑回归表明 MSI-high（P<0.001，优势比=3.701）、较小的 LN 长轴（P=0.023，优势比=0.905）和较低的 CTLN 分级（CTN0：P=0.009，优势比=2.987；CTN1：P=0.014，优势比=2.195）是预测 pN0 的显著参数。
MSI-high 结肠癌与 rLN 较大且 CT 评估预测 pN0 的特异性较高有关。
