Li Hanting, Luo Qinyue, Zheng Yuting, Ding Chengyu, Yang Jinrong, Chen Leqing, Liu Xiaoqing, Guo Tingting, Fan Jun, Han Xiaoyu, Shi Heshui
Department of Radiology, Wuhan Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China.
Hubei Province Key Laboratory of Molecular Imaging, Wuhan, China.
Transl Lung Cancer Res. 2024 Oct 31;13(10):2617-2635. doi: 10.21037/tlcr-24-492. Epub 2024 Oct 28.
Different pathological stages of lung adenocarcinoma require different surgical strategies and have varying prognoses. Predicting their invasiveness is clinically important. This study aims to develop a nomogram to predict the invasiveness of lung adenocarcinoma manifesting as ground-glass nodules (GGNs) based on follow-up computed tomography (CT) imaging.
We retrospectively collected data of 623 GGNs from 601 patients who underwent two follow-up chest CT scans and were confirmed as lung adenocarcinoma by postoperative pathology between June 2017 and August 2023. These patients were randomly divided into training and testing sets in a 7:3 ratio. Eighty-seven GGNs from 86 patients who underwent surgery between September 2023 and April 2024 were prospectively collected as a validation set. The volume, mean density, solid component volume (SV), percentage of solid component (PSC), and mass of GGNs were evaluated using the InferRead CT Lung software. Patients were classified into Group A (atypical adenomatous hyperplasia, adenocarcinoma in situ, and minimally invasive adenocarcinoma) and Group B (invasive adenocarcinoma). Three predictive models were established: model 1 utilized clinical characteristics and morphological features on pre-surgical CT, model 2 incorporated clinical characteristics, morphological features and quantitative parameters on pre-surgical CT, and model 3 utilized all selected features on baseline and pre-surgical CT.
Model 3 achieved a satisfying area under the curves values of 0.911, 0.893, and 0.932 in the training, testing, and validation sets, respectively, demonstrating superior predictive performance than model1 (0.855, 0.858, and 0.816) and model2 (0.895, 0.891, and 0.903). A nomogram was constructed based on model 3. Calibration curves showed a good fit, and decision curve analysis showed that the nomogram was clinically useful.
The nomogram based on morphological features and quantitative parameters from follow-up CT images showed good discrimination and calibration abilities in predicting the invasiveness of lung adenocarcinoma manifesting as GGNs.
肺腺癌的不同病理阶段需要不同的手术策略，且预后各异。预测其侵袭性在临床上具有重要意义。本研究旨在基于随访计算机断层扫描（CT）影像开发一种列线图，以预测表现为磨玻璃结节（GGN）的肺腺癌的侵袭性。
我们回顾性收集了2017年6月至2023年8月期间601例接受两次胸部CT随访扫描且术后病理确诊为肺腺癌的患者的623个GGN数据。这些患者按7:3的比例随机分为训练集和测试集。前瞻性收集了2023年9月至2024年4月期间86例接受手术的患者的87个GGN作为验证集。使用InferRead CT Lung软件评估GGN的体积、平均密度、实性成分体积（SV）、实性成分百分比（PSC）和质量。患者分为A组（非典型腺瘤样增生、原位腺癌和微浸润腺癌）和B组（浸润性腺癌）。建立了三种预测模型：模型1利用术前CT的临床特征和形态学特征，模型2纳入术前CT的临床特征、形态学特征和定量参数，模型3利用基线和术前CT上所有选定的特征。
模型3在训练集、测试集和验证集中的曲线下面积值分别为0.911、0.893和0.932，显示出比模型1（0.855、0.858和0.816）和模型2（0.895、0.891和0.903）更好的预测性能。基于模型3构建了列线图。校准曲线显示拟合良好，决策曲线分析表明列线图在临床上有用。
基于随访CT图像的形态学特征和定量参数的列线图在预测表现为GGN的肺腺癌的侵袭性方面显示出良好的区分能力和校准能力。
