Wen Qingyun, Yue Yong, Shang Jin, Lu Xiaomei, Gao Lu, Hou Yang
Department of Radiology, Shengjing Hospital of China Medical University, Shenyang, China.
CT Clinical Science, Philips Healthcare, Shenyang, China.
Quant Imaging Med Surg. 2021 Feb;11(2):521-532. doi: 10.21037/qims-20-2.
Differentiating between malignant solitary pulmonary nodules (SPNs) and other lung diseases remains a substantial challenge. The latest generation of dual-energy computed tomography (CT), which realizes dual-energy technology at the detector level, has clinical potential for distinguishing lung cancer from other benign SPNs. This study aimed to evaluate the performance of dual-layer spectral detector CT (SDCT) for the differentiation of SPNs.
Spectral images of 135 SPNs confirmed by pathology were retrospectively analyzed in both the arterial phase (AP) and the venous phase (VP). Patients were classified into two groups [the malignant group (n=93) and the benign group (n=42)], with the malignant group further divided into small cell lung cancer (SCLC, n=30) and non-small cell lung cancer (NSCLC, n=63) subtypes. The slope of the spectral Hounsfield Unit (HU) curve (λ), normalized iodine concentration (NIC), CT values of 40 keV monochromatic images (CT), and normalized arterial enhancement fraction (NAEF) in contrast-enhanced images were calculated and compared between the benign and malignant groups, as well as between the SCLC and NSCLC subgroups. ROC curve analysis was performed to assess the diagnostic performance of the above parameters. Seventy cases were randomly selected and independently measured by two radiologists, and intraclass correlation coefficient (ICC) and Bland-Altman analyses were performed to calculate the reliability of the measurements.
Except for NAEF (P=0.23), the values of the parameters were higher in the malignant group than in the benign group (all P<0.05). NIC, λ, and CT performed better in the VP (NIC, λ, and CT) (P<0.001), with an area under the ROC curve (AUC) of 0.93, 0.89, and 0.89 respectively. With respective cutoffs of 0.31, 1.83, and 141.00 HU, the accuracy of NIC, λ, and CT was 91.11%, 85.19%, and 88.15%, respectively. In the subgroup differentiating NSCLC and SCLC, the diagnostic performances of NIC (AUC =0.89) were greater than other parameters. NIC had an accuracy of 86.02% when the cutoff was 0.14. ICC and Bland-Altman analyses indicated that the measurement of SDCT has great reproducibility.
Quantitative measures from SDCT can help to differentiate benign from malignant SPNs and may help with the further subclassification of malignant cancer into SCLC and NSCLC.
鉴别恶性孤立性肺结节（SPN）与其他肺部疾病仍然是一项重大挑战。最新一代的双能计算机断层扫描（CT），即探测器层面实现双能技术，在区分肺癌与其他良性SPN方面具有临床潜力。本研究旨在评估双层光谱探测器CT（SDCT）对SPN的鉴别性能。
回顾性分析135个经病理证实的SPN在动脉期（AP）和静脉期（VP）的光谱图像。患者分为两组[恶性组（n = 93）和良性组（n = 42）]，恶性组进一步分为小细胞肺癌（SCLC，n = 30）和非小细胞肺癌（NSCLC，n = 63）亚型。计算并比较良性组和恶性组以及SCLC和NSCLC亚组之间光谱Hounsfield单位（HU）曲线的斜率（λ）、归一化碘浓度（NIC）、40 keV单色图像的CT值（CT）以及对比增强图像中的归一化动脉增强分数（NAEF）。进行ROC曲线分析以评估上述参数的诊断性能。随机选择70例病例，由两名放射科医生独立测量，并进行组内相关系数（ICC）和Bland-Altman分析以计算测量的可靠性。
除NAEF（P = 0.23）外，恶性组参数值均高于良性组（所有P < 0.05）。NIC、λ和CT在VP中表现更好（NIC、λ和CT）（P < 0.001），ROC曲线下面积（AUC）分别为0.93、0.89和0.89。NIC、λ和CT的截断值分别为0.31、1.83和141.00 HU时，准确率分别为91.11%、85.19%和88.15%。在区分NSCLC和SCLC的亚组中，NIC的诊断性能（AUC = 0.89）优于其他参数。截断值为0.14时，NIC的准确率为86.02%。ICC和Bland-Altman分析表明SDCT测量具有很高的可重复性。
SDCT的定量测量有助于区分良性和恶性SPN，并可能有助于将恶性肿瘤进一步细分为SCLC和NSCLC。
