Chang Song,1,2,* Chun-Yan Zhao,1,2,* Shu-Lin Song,3,* Yan-Rong Lin,1 Chao-Yan Xu,1 Hang-Biao Qiang,1 Ren-Hao Liu,2 Qi Li,4 Qing-Dong Zhu1 

1Department of Tuberculosis, The Fourth People’s Hospital of Nanning, Nanning, People’s Republic of China; 2Clinical Medical School, Guangxi Medical University, Nanning, People’s Republic of China; 3Department of Radiology, The Fourth People’s Hospital of Nanning, Nanning, People’s Republic of China; 4Department of Tuberculosis, Beijing Chest Hospital, Capital Medical University, Beijing, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Qing-Dong Zhu, Department of Tuberculosis, The Fourth People’s Hospital of Nanning, No. 1, Lane 2, Changgang Road, Xingning District, Nanning, Guangxi Zhuang Autonomous Region, 530021, People’s Republic of China, Email zhuqingdong2003@163.com Qi Li, Department of Tuberculosis, Beijing Chest Hospital, Capital Medical University, No. 97, Machang, Tongzhou District, Beijing, 101100, People’s Republic of China, Email liqilisa@126.com

Purpose: This study aimed to develop and validate a preoperative CT-based radiomics nomogram model incorporating intratumoral and peritumoral features to accurately differentiate lung adenocarcinoma (LUAD) from pulmonary tuberculosis (PTB) nodules in HIV/AIDS patients.
Patients and Methods: This retrospective study analyzed clinical and CT imaging data from 187 hIV/AIDS patients (84 with LUAD and 103 with PTB) treated at the Fourth People’s Hospital of Nanning. Patients were randomly divided into training and validation cohorts in a 7:3 ratio. Radiomics features were extracted from both the intratumoral region and a 2 mm peritumoral region, then combined with clinical factors (eg, fever, C-reactive protein levels, and cardiac disease) to develop multiple predictive models, including clinical model, intra model, peri 2mm model, fusion model, and combined model (which integrates clinical and fusion models). Diagnostic performance was evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and other metrics.
Results: The combined model achieved the highest AUC in both the training (0.978) and validation cohorts (0.969) cohorts, significantly outperforming the other models while mitigating the overfitting observed in the clinical model. Hosmer-Lemeshow (HL) tests, Integrated Discrimination Improvement (IDI), Net Reclassification Index (NRI), and decision curve analysis (DCA) confirmed its superior performance.
Conclusion: The CT-based radiomics nomogram model, intratumoral and peritumoral radiomics features, enables accurate differentiation between LUAD and PTB in HIV/AIDS patients, providing a non-invasive tool for preoperative diagnosis.

Keywords: HIV/AIDS, lung adenocarcinoma, tuberculosis pulmonary nodules, radiomics, differential diagnosis