Li Xu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Li Xu has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Fangrong Yan
Department of Biostatistics, China Pharmaceutical University, Nanjing, China
1 shared publication - Fengnan Niu
Department of Pathology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University, Nanjing, China
1 shared publication - Haitao Guan
Department of Endocrinology, Tongren Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, P.R. China
1 shared publication Jian Wang
Department of Radiology, Zhujiang Hospital, Southern Medical University, Haizhu District, Guangzhou, Guangdong, P.R. China
1 shared publication- Jun Yang
Department of Pathology, Nanjing Drum Tower Hospital, the Affiliated Hospital of Nanjing University Medical School, 321 Zhongshan Road, Nanjing 210008, China
1 shared publication - Junkai Zhu
State Key Laboratory of Natural Medicines, Research Center of Biostatistics and Computational Pharmacy, China Pharmaceutical University, Nanjing, P.R. China
1 shared publication
Latest Publications
- Open Access
- 3 November 2023
- Published by Wolters Kluwer Health in Medicine
- Vol. 102 (44) , e35830
Development and validation of radiomics nomograms for preoperative prediction of characteristics in non-small cell lung cancer and circulating tumor cells
To develop and validate 3 radiomics nomograms for preoperative prediction of pathological and progression diagnosis in non-small cell lung cancer (NSCLC) as well as circulating tumor cells (CTCs). A total of 224 and 134 patients diagnosed with NSCLC were respectively gathered in 2018 and 2019 in this study. There were totally 1197 radiomics features that were extracted and quantified from the images produced by computed tomography. Then we selected the radiomics features with predictive value by least absolute shrinkage and selection operator and combined them into radiomics signature. Logistic regression models were built using radiomics signature as the only predictor, which were then converted to nomograms for individualized predictions. Finally, the performance of the nomograms was assessed on both cohorts. Additionally, immunohistochemical correlation analysis was also performed. As for discrimination, the area under the curve of pathological diagnosis nomogram and progression diagnosis nomogram in NSCLC were both higher than 90% in the training cohort and higher than 80% in the validation cohort. The performance of the CTC-diagnosis nomogram was somehow unexpected where the area under the curve were range from 60% to 70% in both cohorts. As for calibration, nonsignificant statistics (P > .05) yielded by Hosmer–Lemeshow tests suggested no departure between model prediction and perfect fit. Additionally, decision curve analyses demonstrated the clinically usefulness of the nomograms. We developed radiomics-based nomograms for pathological, progression and CTC diagnosis prediction in NSCLC respectively. Nomograms for pathological and progression diagnosis were demonstrated well-performed to facilitate the individualized preoperative prediction, while the nomogram for CTC-diagnosis prediction needed improvement.
- 3 November 2023
- Published by Wolters Kluwer Health in Medicine
- Vol. 102 (44) , e35830
Open AccessDevelopment and validation of radiomics nomograms for preoperative prediction of characteristics in non-small cell lung cancer and circulating tumor cells
To develop and validate 3 radiomics nomograms for preoperative prediction of pathological and progression diagnosis in non-small cell lung cancer (NSCLC) as well as circulating tumor cells (CTCs). A total of 224 and 134 patients diagnosed with NSCLC were respectively gathered in 2018 and 2019 in this study. There were totally 1197 radiomics features that were extracted and quantified from the images produced by computed tomography. Then we selected the radiomics features with predictive value by least absolute shrinkage and selection operator and combined them into radiomics signature. Logistic regression models were built using radiomics signature as the only predictor, which were then converted to nomograms for individualized predictions. Finally, the performance of the nomograms was assessed on both cohorts. Additionally, immunohistochemical correlation analysis was also performed. As for discrimination, the area under the curve of pathological diagnosis nomogram and progression diagnosis nomogram in NSCLC were both higher than 90% in the training cohort and higher than 80% in the validation cohort. The performance of the CTC-diagnosis nomogram was somehow unexpected where the area under the curve were range from 60% to 70% in both cohorts. As for calibration, nonsignificant statistics (P > .05) yielded by Hosmer–Lemeshow tests suggested no departure between model prediction and perfect fit. Additionally, decision curve analyses demonstrated the clinically usefulness of the nomograms. We developed radiomics-based nomograms for pathological, progression and CTC diagnosis prediction in NSCLC respectively. Nomograms for pathological and progression diagnosis were demonstrated well-performed to facilitate the individualized preoperative prediction, while the nomogram for CTC-diagnosis prediction needed improvement.