Predicting programmed death-ligand 1 expression level in non-small cell lung cancer using a combination of peritumoral and intratumoral radiomic features on computed tomography

Prediction of programmed death-1 expression status in non-small cell lung cancer based on intratumoural and peritumoral computed tomography (CT) radiomics nomogram

AIMS

This study aimed to predict the expression of programmed death-1 (PD-1) in non-small cell lung cancer (NSCLC) using intratumoral and peritumoral computed tomography (CT) radiomics nomogram.

MATERIALS AND METHODS

Two hundred patients pathologically diagnosed with NSCLC from two hospitals were retrospectively analyzed. Of these, 159 NSCLC patients from our hospital were randomly divided into a training cohort (n=96) and an internal validation cohort (n=63) at a ratio of 6:4, while 41 NSCLC patients from another medical institution served as the external validation cohort. The radiomic features of the gross tumor volume (GTV) and peritumoral volume (PTV) were extracted from the CT images. Optimal radiomics features were selected using least absolute shrinkage and selection operator regression analysis. Finally, a CT radiomics nomogram of clinically independent predictors combined with the best rad-score was constructed.

RESULTS

Compared with the 'GTV' and 'PTV' radiomics models, the combined 'GTV + PTV' radiomics model showed better predictive performance, and its area under the curve (AUC) values in the training, internal validation, and external validation cohorts were 0.90 (95% confidence interval [CI]: 0.83-0.97), 0.85 (95% CI: 0.74-0.96) and 0.78 (95% CI: 0.63-0.92). The nomogram constructed by the rad-score of the 'GTV + PTV' radiomics model combined with clinical independent predictors (prealbumin and monocyte) had the best performance, with AUC values in each cohort being 0.92 (95% CI: 0.85-0.98), 0.88 (95% CI: 0.78-0.97), and 0.80 (95% CI: 0.66-0.94), respectively.

CONCLUSION

The intratumoral and peritumoral CT radiomics nomogram may facilitate individualized prediction of PD-1 expression status in patients with NSCLC.

Ct-based intratumoral and peritumoral radiomics for predicting prognosis in osteosarcoma: A multicenter study

PURPOSE

To evaluate the performance of CT-based intratumoral, peritumoral and combined radiomics signatures in predicting prognosis in patients with osteosarcoma.

METHODS

The data of 202 patients (training cohort:102, testing cohort:100) with osteosarcoma admitted to the two hospitals from August 2008 to February 2022 were retrospectively analyzed. Progression free survival (PFS) and overall survival (OS) were used as the end points. The radiomics features were extracted from CT images, three radiomics signatures（RSintratumoral, RSperitumoral, RScombined）were constructed based on intratumoral, peritumoral and combined radiomics features, respectively, and the radiomics score (Rad-score) were calculated. Kaplan-Meier survival analysis was used to evaluate the relationship between the Rad-score with PFS and OS, the Harrell's concordance index (C-index) was used to evaluate the predictive performance of the radiomics signatures.

RESULTS

Finally, 8, 6, and 21 features were selected for the establishment of RSintratumoral, RSperitumoral, and RScombined, respectively. Kaplan-Meier survival analysis confirmed that the Rad-scores of the three RSs were significantly correlated with the PFS and OS of patients with osteosarcoma. Among the three radiomics signatures, RScombined had better predictive performance, the C-index of PSF prediction was 0.833 in the training cohort and 0.814 in the testing cohort, the C-index of OS prediction was 0.796 in the training cohort and 0.764 in the testing cohort.

CONCLUSIONS

CT-based intratumoral, peritumoral and combined radiomics signatures can predict the prognosis of patients with osteosarcoma, which may assist in individualized treatment and improving the prognosis of osteosarcoma patients.

Comparison of the radiomics-based predictive models using machine learning and nomogram for epidermal growth factor receptor mutation status and subtypes in lung adenocarcinoma

The purpose of this study is to develop the predictive models for epidermal growth factor receptor (EGFR) mutation status and subtypes [exon 21-point mutation (L858R) and exon 19 deletion mutation (19Del)] and evaluate their clinical usefulness. Total 172 patients with lung adenocarcinoma were retrospectively analyzed. The analysis of variance and the least absolute shrinkage were used for feature selection from plain computed tomography images. Then, radiomic score (rad-score) was calculated for the training and test cohorts. Two machine learning (ML) models with 5-fold were applied to construct the predictive models with rad-score, clinical features, and the combination of rad-score and clinical features. The nomogram was developed using rad-score and clinical features. The prediction performance was evaluated by the area under the receiver operating characteristic curve (AUC). Finally, decision curve analysis (DCA) was performed using the best ML and nomogram models. In the test cohorts, the AUC of the best ML and the nomogram model were 0.73 (95% confidence interval, 0.59-0.87) and 0.79 (0.65-0.92) in the EGFR mutation groups, 0.83 (0.67-0.99) and 0.85 (0.72-0.97) in the L858R mutation groups, as well as 0.77 (0.58-0.97) and 0.77 (0.60-0.95) in the 19Del groups. The DCA showed that the nomogram models have comparable results with ML models. We constructed two predictive models for EGFR mutation status and subtypes. The nomogram models had comparable results to the ML models. Because the superiority of the performance of ML and nomogram models varied depending on the prediction groups, appropriate model selection is necessary.