Zhonghua Yi Xue Za Zhi. 2025 Dec 2;105(44):4056-4064. doi: 10.3760/cma.j.cn112137-20250619-01495.
ABSTRACT
Objective: To construct and validate a prognostic model for patients with advanced lung squamous cell carcinoma (LUSC) receiving chemotherapy combined with immunotherapy based on quantitative CT features. Methods: A total of 96 patients with advanced LUSC who received chemotherapy combined with immunotherapy at Union Hospital, Tongji Medical College, Huazhong University of Science and Technology from November 2020 to October 2023 were retrospectively included. They were randomly divided into a training set (n=72) and an internal validation set (n=24) at a ratio of 3∶1 (random sequence was generated by R Studio 4.4.2 software). Additionally, patients with advanced LUSC who visited Yichang Central People’s Hospital from October 2020 to June 2024 were enrolled as the external validation set (n=58) according to the same criteria. Pretreatment chest CT images were obtained from patients, and quantitative CT features were extracted. Based on the training set data, LASSO regression and univariate and multivariate Cox regression models were used to screen the independent variables, and the model was constructed with the progression-free survival (PFS, progress=1, no progression=0) as the dependent variable. Taking the progression at 180 d as the outcome variable and the quantitative CT score as the evaluation index, the receiver operating characteristic (ROC) curve was drawn. The calibration curve and decision curve were used to evaluate the model efficacy. Based on the optimal cut-off value determined by the ROC curve, patients were divided into high-risk and low-risk groups, and the log-rank test was used to compare the survival differences between the two groups. Results: A total of 154 patients with LUSC were included, including 147 males and 7 females, with an age of (64.7±8.1) years; there were no statistically significant differences in demographic characteristics and TNM stage among the training set, internal validation set, and external validation set (all P>0.05). Four quantitative CT features were associated with PFS, including the percentage of low attenuation area with CT value < -910 HU in the whole lung (LAA%-910_lung) (HR=0.013, 95%CI: 0.002-0.313), the area percentage of low attenuation area with CT value < -950 HU in the right lower lung (LAA%-950_right_lower) (HR=0.011, 95%CI: 0.001-0.012), the minimum wall thickness of grade 2 airwall (minThicknessOfAirwall_2) (HR=0.117, 95%CI: 0.029-0.463) and the mean diameter of grade 1 airway (meanDiameterOfAirway_1) (HR=0.767, 95%CI: 0.687-0.857), which were used to construct the quantitative CT (QCT) score: QCTscore=-4.346×(LAA%-910_lung)-4.513×(LAA%-950_right_lower)-2.14×(minThicknessOfAirwall_2)-0.265×(meanDiameterOfAirway_1). The optimal cutoff value for the score was -9.45. The areas under the ROC curve (AUC) for predicting 180 d survival was 0.843 (95%CI: 0.773-0.952) in the training set, 0.778 (95%CI: 0.527-0.999) in the internal test set, and 0.762 (95%CI: 0.615-0.921) in the external test set. The Hosmer-Lemeshow goodness-fit test results showed that the model had a good fitting effect (training set: χ2=8.058, P=0.428; internal validation set: χ2=12.883, P=0.116; internal validation set: χ2=3.141, P=0.925). The decision curve shows that when the risk threshold of the training set is 0 to 79%, that of the internal validation set is 0 to 81%, and that of the independent validation set is 0 to 82%, the prediction model has a clinical net benefit. In the three data sets, the 180 d PFS rate of the high-risk group (QCT score ≥ -9.45) was lower than that of the low-risk group (QCT score < -9.45) (training set: 5.1% vs 66.7%; internal validation set: 180 d progression-free survival rate 12.5% vs 75%; external validation set: 180 d progression-free survival rate 68.6% vs 100.0%, all P<0.05). Conclusion: Based on the QCT image features, a prognostic prediction model for chemotherapy combined with immunotherapy in LUSC was constructed, providing an effective means for predicting the response of chemotherapy combined with immunotherapy in LUSC patients.
PMID:41320660 | DOI:10.3760/cma.j.cn112137-20250619-01495
