# Incorporating the inflammation-related parameters enhances the performance of the nomogram for predicting local control in lung cancer patients treated with stereotactic body radiation therapy

## TL;DR

Adding inflammation-related factors improves a prediction model for lung cancer patients undergoing radiation therapy.

## Contribution

The study shows that including inflammation-related parameters enhances nomogram performance for predicting local control in lung cancer patients.

## Key findings

- Inflammation-related parameters improved model performance in AIC, C-index, and AUC metrics.
- The ACPBLN model showed higher net benefit across threshold probabilities compared to the ACPB model.
- Lymphocyte and neutrocyte counts were significant predictors of local control.

## Abstract

The study aims to investigate whether including the inflammation-related parameters would enhance the accuracy of a nomogram for local control (LC) prediction in lung cancer patients undergoing stereotactic body radiation therapy (SBRT).

158 primary or metastatic lung cancer patients treated with SBRT were retrospectively analyzed. The clinical, dosimetric and inflammation-related parameters were collected for the Cox regression analysis. The ACPB model was constructed by employing the clinical and dosimetric factors. And the ACPBLN model was established by adding the inflammation-related factors to the ACPB model. The two models were compared in terms of ROC, Akaike Information Criterion (AIC), C-index, time-dependent AUC, continuous net reclassification index (NRI), integrated discrimination improvement (IDI), calibration plots and decision curve analysis (DCA).

Multivariate Cox regression analysis revealed that six prognostic factors were independently associated with LC, including age, clinical stage, planning target volume (PTV) volume, BED of the prescribed dose (BEDPD), the lymphocyte count and neutrocyte count. The ACPBLN model performed better in AIC, bootstrap-corrected C-index, time-dependent AUC, NRI and IDI than the ACPB model. The calibration plots showed good consistency between the probabilities and observed values in the two models. The DCA curves showed that the ACPBLN nomogram had higher overall net benefit than the ACPB model across a majority of threshold probabilities.

The inflammation-related parameters were associated with LC for lung cancer patients treated with SBRT. The inclusion of the inflammation-related parameters improved the predictive performance of the nomogram for LC prediction.