Abstract
目的
探讨能谱CT (dual energy CT, DECT)诊断非小细胞肺癌(non-small cell lung cancer, NSCLC)纵隔淋巴结转移的应用价值。
方法
选择2018年4月至2019年10月在北京大学第三医院接受胸部DECT检查且经术后病理诊断证实的NSCLC患者病例资料进行回顾性分析，共收集到病例57例，两名放射科医师共同分析患者术前CT图像，将轴位图像上所有短径(short-axis diameter, S)≥5 mm的纵隔淋巴结纳入本研究。测量淋巴结形态学参数长径(long-axis diameter, L)、S、短径与长径比值(ratio of short-axis diameter to long-axis diameter, S/L)以及能谱参数动脉期及静脉期碘浓度(iodine concentration, IC)、标准化碘浓度(normalized iodine concentration, NIC)、能谱曲线斜率及有效原子序数。比较转移与非转移淋巴结形态学指标及其能谱参数的差异，将有统计学差异的参数纳入Logistic回归方程筛选出有诊断价值的参数，并生成诊断淋巴结转移的联合变量，对淋巴结S、静脉期NIC及联合变量进行受试者工作特征(receiver operating characteristic, ROC)曲线分析。
结果
57例患者中，术后病理诊断证实转移淋巴结49枚，非转移淋巴结938枚。CT轴位上共检出S≥5 mm纵隔淋巴结163枚(转移淋巴结49枚，非转移淋巴结114枚)。转移淋巴结的S、L及S/L均显著大于非转移淋巴结(P＜0.05)，转移淋巴结的能谱参数均显著低于非转移性淋巴结(P＜0.05)。S是诊断淋巴结转移的最佳单一形态学指标，ROC曲线下面积(area under curve, AUC)为0.752，阈值8.5 mm，灵敏度67.4%，特异度73.7%，准确率71.8%。静脉期NIC为最佳单一能谱参数，AUC为0.861，阈值0.53，灵敏度95.9%，特异度70.2%，准确率77.9%。多因素分析显示S、静脉期NIC是转移淋巴结的独立预测因子。联合S、静脉期NIC诊断淋巴结转移的AUC为0.895，灵敏度79.6%，特异度87.7%，准确率85.3%，明显高于S (P＜0.001)、静脉期NIC (P=0.037)。
结论
DECT定量参数鉴别NSCLC患者纵隔淋巴结转移的价值优于形态学参数，联合S和静脉期NIC可提高术前诊断淋巴结转移的准确率。
Keywords: 非小细胞肺癌, 淋巴结转移, 体层摄影术, X射线计算机
Abstract
Objective
To validate the value of dual energy CT (DECT) in the differentiation of mediastinal metastatic lymph nodes from non-metastatic lymph nodes in non-small cell lung cancer (NSCLC).
Methods
In the study, 57 surgically confirmed NSCLC patients who underwent enhanced DECT scan within 2 weeks before operation were enrolled. Two radiologists analyzed the CT images before operation. All mediastinal lymph nodes with short diameter≥5 mm on axial images were included in this study. The morphological parameters [long-axis diameter (L), short-axis diameter (S) and S/L of lymph nodes] and the DECT parameters [iodine concentration (IC), normalized iodine concentration (NIC), slope of spectral hounsfield unit curve (λHU) and effective atomic number (Zeff) in arterial and venous phase] were measured. The differences of morphological parameters and DECT parameters between metastatic and non-metastatic lymph nodes were compared. The parameters with significant difference were analyzed by the Logistic regression model, then a new predictive variable was established. Receiver operator characteristic (ROC) analyses were performed for S, NIC in venous phase and the new predictive variable.
Results
In 57 patients, 49 metastatic lymph nodes and 938 non-metastatic lymph nodes were confirmed by surgical pathology. A total of 163 mediastinal lymph nodes (49 metastatic, 114 non-metastatic) with S≥5 mm were detected on axial CT images. The S, L and S/L of metastatic lymph nodes were significantly higher than those of non-metastatic lymph nodes (P < 0.05). The DECT parameters of metastatic lymph nodes were significantly lower than those of non-metastatic lymph nodes (P < 0.05). The best single morphological parameter for differentiation between metastatic and nonmetastatic lymph nodes was S (AUC, 0.752; threshold, 8.5 mm; sensitivity, 67.4%; specificity, 73.7%; accuracy, 71.8%). The best single DECT parameter for differentiation between metastatic and nonmetastatic lymph nodes was NIC in venous phase (AUC, 0.861; threshold, 0.53; sensitivity, 95.9%; specificity, 70.2%; accuracy, 77.9%). Multivariate analysis showed that S and NIC were independent predictors of lymph node metastasis. The AUC of combined S and NIC in the venous phase was 0.895(sensitivity, 79.6%; specificity, 87.7%; accuracy, 85.3%), which were significantly higher than that of S (P < 0.001) and NIC (P=0.037).
Conclusions
The ability of quantitative DECT parameters to distinguish mediastinal lymph node metastasis in NSCLC patients is better than that of morphological parameters. Combined S and NIC in venous phase can be used to improve preoperative diagnostic accuracy of metastatic lymph nodes.
Keywords: Non-small cell lung cancer, Lymph node metastasis, Tomography, X-ray computed
肺癌是全球癌症死亡的最主要原因之一，约占总癌症死亡的18.4%[1]。非小细胞肺癌(non-small cell lung cancer, NSCLC)占肺癌的80%~85%[2]。纵隔淋巴结转移与肺癌分期、处理原则及生存期密切相关[3]。常规CT通过形态学指标检出淋巴结转移，常用指标包括淋巴结短径(short-axis diameter, S)＞10 mm、短径与长径比值(ratio of short-axis diameter to long-axis diameter, S/L)增加等[4]。胸部CT诊断纵隔淋巴结转移的灵敏度及特异度均有限, 并且差异较大[5]，转移及非转移性淋巴结的鉴别仍然是临床困惑的问题。
能谱CT (dual-energy CT, DECT)成像通过单个X线球管瞬时高、低能切换达到双能量成像，可同时得到物质分离图像和40~140 kV间任意千伏的单能量图像，经过图像后处理可得到如碘(水)、钙(水)、钙(碘)等基物质对浓度、能谱曲线斜率(slope of spectral hounsfield unit curve，λHU)及有效原子序数(effective atomic number, Zeff)等定量参数[6]。文献报道能谱CT有助于提高区分恶性与良性肿瘤的诊断结果的准确率[7-9]，并区分转移性与非转移性淋巴结[10]。本研究旨在通过术前DECT与NSCLC术后病理诊断结果的对照研究，评价形态学参数与能谱参数判断纵隔淋巴结转移的价值，并通过Logistic回归分析术前诊断纵隔淋巴结转移的独立预测因素。
1. 资料与方法
1.1. 研究对象
选择2018年4月至2019年10月在北京大学第三医院接受胸部DECT检查且经术后病理诊断证实的NSCLC患者病例资料进行回顾性分析。病例纳入标准:(1)原发性NSCLC; (2)不合并其他恶性肿瘤; (3)手术前或穿刺前2周内行胸部DECT检查; (4)术前或穿刺前未经任何抗肿瘤治疗。排除标准:(1)临床资料不完整; (2)图像伪影重。最终57例患者纳入本研究，17例女性，40例男性，平均年龄(61.5±11.8)岁(32~79岁)。
1.2. 检查方法
使用GE Revolution CT (GE Healthcare，美国)行胸部增强扫描，患者取仰卧位，扫描范围从肺尖至肺底，采用单源高低压瞬切能谱成像(gemstone spectral imaging, GSI)模式，80/140 kV瞬时管电压切换，管电流200 mA，转速为0.5 s，螺距为0.992 :1，扫描层厚5 mm，层间距5 mm。使用高压注射器，经肘静脉注入非离子型对比剂碘海醇(每mL 300 mg I) 1.5 mg/kg及生理盐水20 mL，注射速率3 mL/s。注射对比剂后约25 s行动脉期扫描，约50 s行静脉期扫描。扫描后对原始数据进行重建(层厚1.25 mm，层间距1.25 mm)，重建数据传送至GE AW4.7工作站，进行能谱数据分析。
1.3. 图像分析及标准
由分别具有3年和5年工作经验的2名放射科医师在不知道术后病理诊断结果的前提下，共同分析所有患者的CT图像。对于淋巴结的入选，两名医师不一致的地方，由二人协商达成一致。所有入选淋巴结的形态学参数及能谱参数均由两人独立测量，两位观察者间一致性较好，不一致时以更有经验医师的测量结果进行分析。利用GE AW4.7后处理工作站能谱CT分析软件GSI viewer，在70 kV单能量轴位图像上选取淋巴结的最大层面，测量淋巴结长径(long-axis diameter, L)、S，并勾画感兴趣区域(region of interest，ROI)，测量淋巴结动脉期、静脉期淋巴结碘浓度(iodine concentration, IC)和同层主动脉IC，动脉期及静脉期Zeff，并获得40~140 kV间隔为10 kV的单能量CT值。所有S≥5 mm的淋巴结纳入本研究。测量过程中保存由系统自动分析生成的.csv数据文件，导入电脑中用Microsoft Office Excel软件进行数据分析。计算淋巴结S/L，标准化碘浓度(normalized iodine concentration，NIC)=淋巴结IC/同层面主动脉IC，λHU= (CT40kV-CT100kV)/60，CT40kV、CT100kV分别为40 kV、100 kV单能量下ROI的CT值。
1.4. 手术病理分析及分组
参照国际肺癌研究协会(International Association for the Study of Lung Cancer, IASLC)区域淋巴结分区系统，将区域淋巴结分为14个区。病理诊断结果中，如某个分区内无淋巴结转移，则该区CT上所有S≥5 mm淋巴结均视为非转移淋巴结; 某个分区内有1个或1个以上淋巴结转移，则该区内S最大或较大的淋巴结视为转移淋巴结，该区内其他S≥5 mm的淋巴结均被剔除。
1.5. 统计学分析
采用SPSS 25.0统计学软件进行统计学分析。计算组内相关系数(interclass correlation coefficient, ICC)，评价两名医师及同一医师不同时间测量结果的一致性，ICC≥0.81，一致性较好; ICC位于0.61~0.80，一致性中等; 0.41~0.60，一致性一般; 0.11~0.40，一致性差; 0.1以下无一致性。比较转移及非转移淋巴结各形态学指标及能谱参数，对数据进行正态分布及方差齐性检验，符合正态分布的计量资料组间比较采用独立样本t检验。差异有统计学意义的参数作为自变量，是否为转移淋巴结作为因变量，进行二分类Logistic回归分析，筛选术前诊断淋巴结转移的独立预测因素并生成新的预测变量。ROC曲线分析评价预测因子及新的预测变量诊断淋巴结转移的效能，ROC曲线分析与比较采用Medcalc version 18.11.6统计学软件，分析形态学参数、能谱参数以及新的预测变量诊断淋巴结转移的最佳阈值，计算灵敏度、特异度、阳性预测值、阴性预测值及准确率。P < 0.05认为差异具有统计学意义。
2. 结果
2.1. 一般情况
57例患者中，19例患者发现纵隔淋巴结转移，其中5例为N1期，14例为N2期，男性12例，女性7例，平均年龄(58.2±10.8)岁，肿瘤平均最大径(2.7±1.8) cm，腺癌13例，鳞癌5例，腺鳞癌1例。38例未发现淋巴结转移，其中男性28例，女性10例，平均年龄(64.6±13.7)岁，肿瘤平均最大径(2.3±1.1) cm，腺癌21例，鳞癌14例，腺鳞癌3例。
2.2. 术后病理诊断结果及CT结果
手术共切除987枚淋巴结，均为原发肿瘤同侧，其中转移淋巴结共31组49枚，非转移淋巴结183组938枚。CT轴位上共检出S≥5 mm纵隔淋巴结163枚，其中，转移淋巴结49枚(图 1)，非转移淋巴结114枚(图 2)，淋巴结分区情况见表 1。
1.
患者女性，63岁，右肺下叶腺癌，T2bN2M0, 纵隔4R区转移淋巴结
A 63-year-old female patient with right lower lobe adenocarcinoma, clinical stage was T2bN2M0, and the right fourth group of lymph nodes were metastatic
A, 70 kV monochromatic contrast-enhancement image in arterial phase shows the lymph node with a short-axis diameter of 7.32 mm, and the ROI with an area of 33.18 mm2; B, iodine based material-decomposition image in arterial phase shows that iodine concentration (IC) of the lymph node is 13.59×100 μg/cm3 and IC of the thoracic aorta in the same slice is 164.38×100 μg/cm3; C, effective atomic number CT image in arterial phase shows the effective atomic number (Zeff) of the lymph node is 8.44; D, spectral attenuation curve of lymph node in arterial phase, and the slope of attenuation curve (λHU) is 1.61; E, venous phase 70 kV monochromatic contrast-enhanced CT image; F, iodine based material-decomposition image in venous phase, and IC of the lymph node is 27.35×100 μg/cm3 and IC of the thoracic aorta in the same slice is 52.24×100 μg/cm3; G, effective atomic number CT image in venous phase, and Zeff of the lymph node is 9.12; H, spectral attenuation curve of lymph node in venous phase. λHU is 3.24.
2.
患者男性，71岁，右肺上叶腺癌，T1N0M0, 纵隔4R区非转移淋巴结
A 71-year-old male patient with right upper lobe adenocarcinoma, clinical stage was T1N0M0, and the right fourth group of lymph nodes were non-metastatic
A, 70 kV monochromatic contrast-enhancement image in arterial phase (AP) shows the lymph node with a short-axis diameter of 9.98 mm, and the ROI with an area of 71.03 mm2; B, iodine based material-decomposition image in arterial phase shows that iodine concentration (IC) of the lymph node is 21.20×100 μg/cm3 and IC of the thoracic aorta in the same slice is 135.24×100 μg/cm3; C, effective atomic number CT image in arterial phase shows the effective atomic number (Zeff) of the lymph node is 8.81; D, spectral attenuation curve of lymph node in arterial phase, and the slope of attenuation curve (λHU) is 2.51; E, venous phase (VP) 70 kV monochromatic contrast-enhanced CT image; F, iodine based material-decomposition image in venous phase, and IC of the lymph node is 28.81×100 μg/cm3 and IC of the thoracic aorta in the same slice is 37.49×100 μg/cm3; G, effective atomic number CT image in venous phase, and Zeff of the lymph node is 9.14; H, spectral attenuation curve of lymph node in venous phase. λHU is 3.35.
1.
转移与非转移淋巴结在纵隔各区的分布情况
Distribution of metastatic and non-metastatic lymph nodes in the regions of mediastinum
| Group | 2R | 2L | 3A | 3P | 4R | 4L | 5 | 6 | 7 | 8R | 8L | 9R | 9L | 10R | 10L |
| R, right; L, left; A, anterior; P, posterior. | |||||||||||||||
| Metastatic (n=49) | 2 | 1 | 2 | 2 | 4 | 1 | 4 | 2 | 15 | 1 | 1 | 2 | 3 | 5 | 4 |
| Non-metastatic (n=114) | 1 | 0 | 0 | 1 | 14 | 19 | 8 | 7 | 28 | 4 | 0 | 5 | 1 | 15 | 11 |
| Total | 3 | 1 | 2 | 3 | 18 | 20 | 12 | 9 | 43 | 5 | 1 | 7 | 4 | 20 | 15 |
2.3. 测量结果一致性评价
同一医师两次重复测量、两名医师分别测量S、L及能谱参数的ICC＞0.81，一致性较好(表 2)。
2.
形态学参数与能谱参数测量同一观察者、不同观察者一致性评价
Intra-observer and inter-observer agreement on morphological parameters and dual-energy parameters measurement
| Items | Doctor A | Doctor A and B | |||
| ICC | 95%CI | ICC | 95%CI | ||
| ICC, interclass correlation coefficient; L, long-axis diameter; S, short-axis diameter; IC, iodine concentration; NIC, normalized iodine concentration; λHU, slope of spectral hounsfield unit curve; Zeff, effective atomic number. | |||||
| S | 0.872 | 0.566－0.947 | 0.849 | 0.743－0.912 | |
| L | 0.891 | 0.643－0.954 | 0.932 | 0.873－0.963 | |
| Arterial phase | |||||
| IC | 0.930 | 0.303－0.980 | 0.897 | 0.818－0.942 | |
| NIC | 0.880 | 0.798－0.930 | 0.853 | 0.730－0.919 | |
| λHU | 0.891 | 0.690－0.952 | 0.839 | 0.614－0.923 | |
| Zeff | 0.949 | 0.723－0.982 | 0.920 | 0.864－0.954 | |
| Venous phase | |||||
| IC | 0.833 | 0.484－0.929 | 0.901 | 0.420－0.967 | |
| NIC | 0.891 | 0.690－0.952 | 0.839 | 0.614－0.923 | |
| λHU | 0.903 | 0.335－0.970 | 0.812 | 0.340－0.927 | |
| Zeff | 0.869 | 0.150－0.960 | 0.927 | 0.874－0.958 | |
2.4. 转移及非转移淋巴结的形态学指标及能谱参数比较
淋巴结形态学指标及各能谱参数均符合正态分布，转移淋巴结的S、L及S/L均明显大于非转移淋巴结(P＜0.05，表 3); 转移淋巴结的能谱参数均显著低于非转移性淋巴结(P＜0.05，表 4)。
3.
非小细胞肺癌纵隔转移与非转移淋巴结形态学特征比较(x±s)
Comparison of morphologic indexes between metastatic and non-metastatic mediastinal lymph nodes in non-small cell lung cancer (x±s)
| Group | L/mm | S/mm | S/L |
| L, long-axis diameter; S, short-axis diameter; S/L, ratio of short-axis diameter to long-axis diameter. | |||
| Metastatic (n=49) | 13.10±2.99 | 9.90±2.55 | 0.76±0.09 |
| Non-metastatic (n=114) | 11.38±2.91 | 7.72±1.96 | 0.69±0.08 |
| t | 3.432 | 5.899 | 5.075 |
| P | 0.001 | ＜0.001 | ＜0.001 |
4.
非小细胞肺癌纵隔转移与非转移淋巴结能谱CT参数比较(x±s)
Comparison of quantitative dual-energy CT parameters between metastatic and non-metastatic mediastinal lymph nodes in non-small cell lung cancer (x±s)
| Characteristic | Metastatic (n=49) | Non-metastatic (n=114) | t | P |
| IC, iodine concentration; NIC, normalized iodine concentration; λHU, slope of spectral hounsfield unit curve; Zeff, effective atomic number. | ||||
| Arterial phase | ||||
| IC/(100 μg/cm3) | 22.14±5.11 | 24.94±3.79 | －3.448 | ＜0.001 |
| NIC | 0.25±0.04 | 0.36±0.09 | －10.300 | ＜0.001 |
| λHU | 2.13±0.21 | 2.33±0.23 | －5.194 | ＜0.001 |
| Zeff | 8.07±0.46 | 8.34±0.51 | －3.213 | 0.002 |
| Venous phase | ||||
| IC/(100 μg/cm3) | 27.24±5.68 | 30.26±5.60 | －3.139 | 0.002 |
| NIC | 0.43±0.07 | 0.60±0.14 | －10.875 | ＜0.001 |
| λHU | 2.46±0.31 | 2.84±0.29 | －7.405 | ＜0.001 |
| Zeff | 8.25±0.37 | 8.55±0.45 | －4.052 | ＜0.001 |
2.5. 淋巴结的形态学指标及能谱参数判断淋巴结转移的二元Logistic回归分析
以形态学参数和能谱参数作为自变量，是否淋巴结转移作为因变量建立回归模型，经逐步拟合，S、静脉期NIC进入模型(P＜0.05)，回归方程为LogitP=－3.092－0.305S+13.243静脉期NIC，将S、静脉期NIC代入回归方程产生诊断淋巴结转移的联合变量LogitP。
2.6. 淋巴结的形态学指标及能谱参数的ROC曲线分析
对形态学参数、能谱参数及联合变量LogitP进行ROC分析。不同指标诊断NSCLC纵隔淋巴结转移的AUC、最佳阈值、特异度、灵敏度、阳性预测值、阴性预测值、准确率见表 5。S是最佳单一形态学指标，判断淋巴结转移的AUC为0.752，最佳阈值是8.50 mm时的准确率为71.8%，优于L (Z=4.419，P＜0.001)，与S/L差异无统计学意义(Z=0.336，P=0.736)。静脉期NIC为判断淋巴结转移的最佳单一能谱参数，AUC为0.861，最佳阈值是0.53时准确率为77.9%。能谱参数中，动脉期NIC、静脉期NIC在鉴别淋巴结转移方面优于S (动脉期NIC比S，Z=1.997，P=0.045;静脉期NIC比S，Z=2.328，P=0.019)。在最佳预测概率阈值0.56时，联合S、静脉期NIC诊断淋巴结转移的AUC为0.895，灵敏度79.6%，特异度87.7%，阳性预测值73.6%，阴性预测值90.9%，准确率85.3%，明显高于S (Z=4.186，P＜0.001)、静脉期NIC (Z=2.085，P=0.037，图 3)。
5.
形态学参数与能谱参数鉴别NSCLC纵隔淋巴结转移的ROC曲线分析结果
Results of ROC analysis of morphologic indexes and dual-energy CT parameters in differential diagnosis of mediastinal metastatic and non-metastatic lymph nodes in non-small cell lung cancer (NSCLC)
| Items | AUC | Sensitivity/% | Specivicity/% | Positive predictive value/% | Negative predictive value/% | Accuracy/% | |
| Data in parentheses are 95%CI. LogitP, combination of short-axis diameter and normalized iodine concentration in venous phase; L, long-axis diameter; S, short-axis diameter; IC, iodine concentration; NIC, normalized iodine concentration; λHU, slope of spectral hounsfield unit curve; Zeff, effective atomic number. | |||||||
| Parameter threshold | |||||||
| S | 8.50 | 0.752(0.672-0.832) | 67.4(52.5-80.1) | 73.7(64.6-81.5) | 52.4(43.3-61.3) | 84.0(77.6-88.8) | 71.8(64.2 -78.5) |
| L | 10.94 | 0.654(0.564-0.745) | 79.6(65.7-89.8) | 50.0(40.5-59.5) | 40.6(35.2-46.3) | 85.1(76.1-91.1) | 58.9(50.9 -66.5) |
| S/L | 0.76 | 0.734(0.644-0.825) | 57.1(42.2-71.2) | 84.2(76.2-90.4) | 60.9(48.8-71.7) | 82.1(76.6-86.4) | 76.1(68.8-82.4) |
| Arterial phase | |||||||
| IC | 22.23 | 0.660(0.561-0.759) | 61.2(46.2-74.8) | 70.2(60.9-78.4) | 46.9(38.1-55.8) | 80.8(74.4-85.9) | 67.5(59.7-74.6) |
| NIC | 0.31 | 0.850(0.793-0.908) | 91.8(80.4-97.7) | 69.3(60.0-77.6) | 56.3(49.1-63.2) | 95.2(88.5-98.1) | 76.1(68.8-82.4) |
| λHU | 2.11 | 0.754(0.674-0.834) | 57.1(42.2-71.2) | 84.2(76.2-90.4) | 60.9(48.8-71.7) | 82.1(76.6-86.4) | 76.1(68.8-82.4) |
| Zeff | 8.34 | 0.657(0.569-0.744) | 73.5(58.9-85.1) | 56.1(46.5-65.4) | 41.9(35.5-48.5) | 83.1(75.0-89.0) | 61.4(53.4-68.9) |
| Venous phase | |||||||
| IC | 25.56 | 0.643(0.544-0.742) | 44.9(30.7-59.8) | 83.3(75.2-89.7) | 53.7(40.9-66.0) | 77.9(73.0-82.1) | 71.8(64.2-78.5) |
| NIC | 0.53 | 0.861(0.806-0.915) | 95.9(86.0-99.5) | 70.2(60.9-78.4) | 58.0(50.9-64.8) | 97.6(91.1- 99.4) | 77.9(70.8-84.0) |
| λHU | 2.72 | 0.807(0.736-0.878) | 81.6(68.0-91.2) | 66.7(57.2-75.2) | 51.3(44.0-58.5) | 89.4(82.2-93.9) | 71.2(63.6-78.0) |
| Zeff | 8.74 | 0.701(0.624-0.779) | 89.8(77.8-96.6) | 52.6(43.1-62.1) | 44.9(39.6-50.3) | 92.3(83.7-96.6) | 63.8(55.9-71.2) |
| LogitP | 0.56 | 0.895(0.846-0.943) | 79.6(65.7-89.8) | 87.7(80.3-93.1) | 73.6(62.6-82.3) | 90.9(85.1- 94.6) | 85.3(78.9-90.3) |
3.
S、静脉期NIC与联合变量LogitP鉴别NSCLC纵隔转移与非转移淋巴结的ROC曲线
ROC curves of short-axis diameter (S), normalized iodine concentration (NIC) in venous phase and combined S and NIC in venous phase (LogitP) for differentiating metastatic and non-metastatic mediastinal lymph nodes in patients with NSCLC
T, threshold.
3. 讨论
临床工作中，主要以淋巴结大小作为判断其是否转移的指标，缺乏统一标准，且特异度、灵敏度较差。目前大多数研究认为纵隔正常淋巴结在轴位CT图像上的短径上限值为10 mm。Vansteenkiste等[5]报道，常规CT诊断纵隔淋巴结转移的敏感度和特异度分别为51% (95% CI: 47%~54%)和85% (95% CI: 84%~88%)，认为CT判断或排除纵隔淋巴结转移的能力有限。De Leyn等[11]对235例CT未发现纵隔淋巴结增大的NSCLC患者行纵隔镜检查，对右上和左上气管旁淋巴结、隆突前淋巴结、右下和左下气管旁淋巴结，以及隆突下淋巴结进行活检，结果在47例(20%)患者中发现了转移淋巴结。本研究纵隔转移淋巴结的S、L均明显大于非转移淋巴结，但二者鉴别淋巴结是否转移的灵敏度、特异度及准确率均较低。S的诊断能力优于L，以短径8.50 mm作为诊断纵隔淋巴结转移的阈值，特异度、灵敏度及准确率分别为67.4%、73.7%、71.8%，与文献的研究结果类似。无论诊断标准如何，正常大小的淋巴结可能包含微小转移，增大的淋巴结可能由炎性反应导致，这种内在局限性导致通过直径判断淋巴结转移的灵敏度、特异度均较低。
淋巴结的形状也有助于判断其性质，良性淋巴结通常呈卵圆形，转移淋巴结由于S/L比例增加，形态上较良性淋巴结更圆。Fukuya等[12]研究发现胃癌转移淋巴结的平均S/L为0.81±0.15，而非转移淋巴结平均S/L为0.57±0.15，二者差异具有统计学意义(P < 0.001)。Yoshimura等[13]在一项4 043例乳腺癌腋窝淋巴结的研究中，使用长径10 mm加上L/S比值为1.6，检测淋巴结转移的灵敏度为79%，特异度为93%。本研究以S/L＞0.76作为诊断纵隔淋巴结的阈值，曲线下面积为0.734，准确率76.1%，与S相比差异无统计学意义。
DECT在碘基物质分离图像上可获得IC、Zeff和λHU等反映组织碘含量的特征定量参数，可作为反映影响组织血容量和血管渗透性的病理生理变化的生物标记物。本研究对57例NSCLC患者进行了两期增强DECT，发现转移淋巴结的动脉期IC、静脉期IC、动脉期NIC、静脉期NIC均显著低于非转移淋巴结，与既往研究结果类似[14]。Rizzo等[15]应用DECT评估转移与非转移淋巴结碘含量的差异以及碘分布与病理结构的关系，发现转移淋巴结碘含量明显低于非转移淋巴结，并且转移淋巴结内部碘分布较非转移淋巴结均匀; 非转移淋巴结由于内部血供不均，其内的碘分布也表现与血供分布一致的从中心向外周递减的梯度，而转移淋巴结内，肿瘤将淋巴结门血管结构推向外周，使其强化程度趋于均匀，同时由于肿瘤血管畸变，仅由单层内皮构成，使其血供水平低于非转移淋巴结，与能谱CT测量结果一致。
能谱曲线反映物质的能量衰减特征，由组织本身的理化性质决定，能谱曲线形状类似或走行一致则反映具有同样的结构及病理类型。本研究发现转移淋巴结的λHU显著低于非转移淋巴结(P＜0.001)，与Yang等[16]在结直肠癌区域淋巴结转移的研究结果一致，但与Zhang等[17]在乳腺癌腋窝淋巴结转移、Yang等[18]在NSCLC中的研究结果有所不同，后两项研究都显示转移淋巴结的λHU显著高于非转移淋巴结。有研究报道[19], 增强后λHU受组织中对比剂含量的影响，高级别NSCLC由于肿瘤生长速度快，超出其血液供应，肿瘤内的微血管密度相对低，增强后组织内的对比剂浓度相对低，低千伏下的CT值较低，导致其λHU低。本研究转移淋巴结的λHU显著低于非转移淋巴结，考虑可能与转移淋巴结的血供水平较低有关，与转移淋巴结的IC及NIC均低于非转移淋巴结的结果一致，而与Zhang等[17]、Yang等[18]的研究结果不同，可能是由于不同类型肿瘤、不同分化程度肿瘤的血供不同导致，需要扩大样本量进一步研究。
Zeff是由原子序数引申而来的新概念，假设某元素对X射线的质量衰减与某化合物的质量衰减系数相同，该元素的原子序数就是该物质的Zeff。本研究转移淋巴结动脉期及静脉期的Zeff显著低于非转移淋巴结，与崔元龙等[20]研究结果一致。但叶亚君等[21]应用DECT鉴别纵隔淋巴结良恶性，结果发现良恶性淋巴结的Zeff值差异无统计学意义。笔者认为，增强后的Zeff与组织中对比剂含量有关，因此会受对比剂用量、增强扫描延迟时间、患者体质量及血流动力学状态等的影响，导致不同研究的结果不一致。有学者计算Zeff与同层面主动脉的Zeff比值，获得标准化Zeff(normalized Zeff，nZeff)，发现nZeff在转移与非转移淋巴结之间差异有统计学意义[16, 18]。
已有研究发现定量能谱参数如IC、NIC、λHU、Zeff、nZeff等有助于结直肠癌[11]、肺癌[13]、乳腺癌[12]、胃癌[22]、下咽癌[23]等患者术前识别转移淋巴结。本研究ROC曲线分析结果表明，能谱参数动脉期NIC和静脉期NIC在区分转移和非转移淋巴结方面优于S，静脉期NIC是鉴别转移淋巴结的最佳单一能谱参数，准确率为77.9%。多因素分析显示S、静脉期NIC是区分转移和非转移淋巴结的独立预测因子。联合S、静脉期NIC可以显著提高CT的诊断能力，优于单一形态学参数或能谱参数，可将淋巴结转移的诊断准确率提高至85.3%。
本研究存在一定的局限性，第一，本研究仅纳入了短径≥5 mm的淋巴结，不排除发生微小转移淋巴结的短径＜5 mm; 第二，将CT轴位图像上短径最大的1个或n (≥2)个淋巴结视为转移淋巴结，造成样本选择偏倚，然而实际操作中，几乎不可能将CT图像上的淋巴结与术后病理诊断结果进行逐一对照; 第三，本研究主观绘制感兴趣区域，ROI仅为淋巴结最大层面，不能反映整个淋巴结信息。
综上所述，能谱CT定量参数鉴别NSCLC患者纵隔淋巴结转移的能力优于形态学参数，联合S和静脉期NIC可提高术前诊断淋巴结转移的准确率。
Funding Statement
国家自然科学基金(81871326)
Supported by the National Natural Foundation of China (81871326)
References
- 1.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin. 2018;68(1):7–30. doi: 10.3322/caac.21442. [DOI] [PubMed] [Google Scholar]
- 2.Ferlay J, Colombet M, Soerjomataram I, et al. Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods. Int J Cancer. 2019;144(8):1941–1953. doi: 10.1002/ijc.31937. [DOI] [PubMed] [Google Scholar]
- 3.中华医学会, 中华医学会肿瘤学分会, 中华医学会杂志社 中华医学会肺癌临床诊疗指南(2018版) 中华肿瘤杂志. 2018;40(12):935–964. doi: 10.3760/cma.j.issn.0253-3766.2018.12.012. [DOI] [PubMed] [Google Scholar]
- 4.Torabi M, Aquino SL, Harisinghani MG. Current concepts in lymph node imaging. https://www.ncbi.nlm.nih.gov/pubmed/15347718. J Nucl Med. 2004;45(9):1509–1518. [PubMed] [Google Scholar]
- 5.Vansteenkiste J, Dooms C, De Leyn P. Early stage non-small cell lung cancer: challenges in staging and adjuvant treatment: evidence-based staging. Ann Oncol. 2010;21(Suppl 7):vii189–vii195. doi: 10.1093/annonc/mdq424. [DOI] [PubMed] [Google Scholar]
- 6.宁 先英, 李 浩, 杨 明, et al. CT能谱定量分析对肺腺癌与鳞癌的鉴别诊断价值. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=fsxsj201703007 放射学实践. 2017;32(3):237–241. [Google Scholar]
- 7.Lv P, Lin XZ, Li J, et al. Differentiation of small hepatic hemangioma from small hepatocellular carcinoma: recently introduced spectral CT method. https://pubs.rsna.org/doi/10.1148/radiol.11101425. Radiology. 2011;259(3):720–729. doi: 10.1148/radiol.11101425. [DOI] [PubMed] [Google Scholar]
- 8.Martin SS, Weidinger S, Czwikla R, et al. Iodine and fat quantification for differentiation of adrenal gland adenomas from metas-tases using third-generation dual-source dual-energy computed tomography. Invest Radiol. 2018;53(3):173–178. doi: 10.1097/RLI.0000000000000425. [DOI] [PubMed] [Google Scholar]
- 9.Muenzel D, Lo GC, Yu HS, et al. Material density iodine images in dual-energy CT: detection and characterization of hypervascular liver lesions compared to magnetic resonance imaging. https://www.sciencedirect.com/science/article/abs/pii/S0720048X17303558. Eur J Radiol. 2017;95(10):300–306. doi: 10.1016/j.ejrad.2017.08.035. [DOI] [PubMed] [Google Scholar]
- 10.Liu X, Ouyang D, Li H, et al. Papillary thyroid cancer: dual-energy spectral CT quantitative parameters for preoperative diagnosis of metastasis to the cervical lymph nodes. Radiology. 2015;275(1):167–176. doi: 10.1148/radiol.14140481. [DOI] [PubMed] [Google Scholar]
- 11.De Leyn P, Vansteenkiste J, Cuypers P, et al. Role of cervical mediastinoscopy in staging of non-small cell lung cancer without enlarged mediastinal lymph nodes on CT scan. Eur J Cardiothorac Surg. 1997;12(5):706–712. doi: 10.1016/S1010-7940(97)00253-4. [DOI] [PubMed] [Google Scholar]
- 12.Fukuya T, Honda H, Hayashi T, et al. Lymph-node metastases: efficacy for detection with helical CT in patients with gastric cancer. Radiology. 1995;197(3):705–711. doi: 10.1148/radiology.197.3.7480743. [DOI] [PubMed] [Google Scholar]
- 13.Yoshimura G, Sakurai T, Oura S, et al. Evaluation of axillary lymph node status in breast cancer with MRI. Breast Cancer. 1999;6(3):249–258. doi: 10.1007/BF02967179. [DOI] [PubMed] [Google Scholar]
- 14.Li X, Meng X, Ye Z. Iodine quantification to characterize primary lesions, metastatic and non-metastatic lymph nodes in lung cancers by dual energy computed tomography: an initial experience. Eur J Radiol. 2016;85(6):1219–1223. doi: 10.1016/j.ejrad.2016.03.030. [DOI] [PubMed] [Google Scholar]
- 15.Rizzo S, Radice D, Femia M, et al. Metastatic and non-metastatic lymph nodes: quantification and different distribution of iodine uptake assessed by dual-energy CT. Eur Radiol. 2018;28(2):760–769. doi: 10.1007/s00330-017-5015-5. [DOI] [PubMed] [Google Scholar]
- 16.Yang Z, Zhang X, Fang M, et al. Preoperative diagnosis of regional lymph node metastasis of colorectal cancer with quantitative parameters from dual-energy CT. AJR Am J Roentgenol. 2019;213(6):1–9. doi: 10.2214/AJR.18.20843. [DOI] [PubMed] [Google Scholar]
- 17.Zhang X, Zheng C, Yang Z, et al. Axillary sentinel lymph nodes in breast cancer: quantitative evaluation at dual-energy CT. Radiology. 2018;289(2):337–346. doi: 10.1148/radiol.2018180544. [DOI] [PubMed] [Google Scholar]
- 18.Yang F, Dong J, Wang X, et al. Non-small cell lung cancer: spectral computed tomography quantitative parameters for preoperative diagnosis of metastatic lymph nodes. https://www.sciencedirect.com/science/article/pii/S0720048X1730030X. Eur J Radiol. 2017;89(4):129–135. doi: 10.1016/j.ejrad.2017.01.026. [DOI] [PubMed] [Google Scholar]
- 19.Lin LY, Zhang Y, Suo ST, et al. Correlation between dual-energy spectral CT imaging parameters and pathological grades of non-small cell lung cancer. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=6b3b5acce5079fb3ef58b70042756f90. Clin Radiol. 2018;73(4):412. doi: 10.1016/j.crad.2017.11.004. [DOI] [PubMed] [Google Scholar]
- 20.崔 元龙, 许 毛荣, 文 智. 能谱CT定量参数对非小细胞肺癌纵隔淋巴结转移中的应用价值. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=lcfsxzz201905016 临床放射学杂志. 2019;38(5):825–829. [Google Scholar]
- 21.叶 亚君, 莫 淑琼. CT能谱成像鉴别诊断纵隔淋巴结良恶性的临床意义. http://www.cnki.com.cn/Article/CJFDTotal-XDYY201808042.htm 现代医用影像学. 2018;27(8):2702–2703. [Google Scholar]
- 22.Pan Z, Pang L, Ding B, et al. Gastric cancer staging with dual energy spectral CT imaging. PLoS One. 2013;8(2):e53651. doi: 10.1371/journal.pone.0053651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Yang L, Luo D, Li L, et al. Differentiation of malignant cervical lymphadenopathy by dual-energy CT: a preliminary analysis. https://pubs.rsna.org/doi/10.1148/radiol.2451061804. Sci Rep. 2016;6(1):31020. doi: 10.1038/srep31020. [DOI] [PMC free article] [PubMed] [Google Scholar]