2054 - Overestimation of Contralateral Hilar Lymph Node Metastasis in Non-Metastatic Non-Small Cell Lung Cancer and Its Predictive Model: HAM

Z. Hou¹, B. Dong², Z. Lin¹, Y. Zhang³, X. Liu³, C. Wu¹, Q. Xu¹, K. Chen⁴, Q. Li¹, X. Lin⁵, and M. Chen²; ¹Department of Radiation Oncology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Sun Yat-sen University, Guangzhou, China, ²Department of Radiation Oncology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Sun Yat-sen University, Guangzhou, Guangdong, China, ³Department of Radiation Oncology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Guangzhou, China, ⁴Department of Population Health Sciences, Duke University School of Medicine, Durham, NC, ⁵Department of Nuclear Medicine, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Sun Yat-sen University, Guangzhou, China

Purpose/Objective(s): Metastasis of non-metastatic non-small cell lung cancer (NMNSCLC) to contralateral hilar lymph nodes (CHLN) eliminates the opportunity for radical surgery or definitive radiotherapy, underscoring the critical importance of accurate CHLN metastasis assessment in clinical decision-making. This study aims to analyze whether CHLN metastasis in NMNSCLC is commonly overestimated in clinical practice and to establish a predictive model for enhanced precision. Materials/

Methods: We conducted a retrospective analysis of 571 pathologically confirmed NMNSCLC patients. Lasso regression was used to select predictive factors, and a multivariate binary logistic regression model named HAM was constructed. Internal validation was performed using ten-fold cross-validation.
Results: The CHLN metastasis rate for the NMNSCLC patients in this study was 6.0% (34/571). The positive predictive value (PPV) and sensitivity for CT diagnosis were 55.8% and 70.6%, respectively, and for PET-CT they were 42.4% and 73.5%, respectively. The five optimal predictive factors (age, emphysema or bullae, central-type lung cancer, CHLN metastasis diagnosed by CT and PET-CT) were used to develop the HAM model. The Area under curve (AUC) values for CT, PET-CT, and HAM model in diagnosing CHLN metastasis were 0.84, 0.84, and 0.97, respectively (CT vs. PET-CT, p=0.99; CT vs. HAM, p<0.001; PET-CT vs. HAM, p<0.001). The F1 scores for CT and PET-CT were 0.62 and 0.54, respectively, while the maximum F1 score of our model was 0.71, with corresponding PPV and sensitivity of 69.4%, and 73.5%, respectively. The Brier score of our model was 0.027, and the AUC value from internal ten-fold cross-validation was 0.94, with a Brier score of 0.034.
Conclusion: CHLN metastasis is rare in NMNSCLC patients. Diagnostic assessments using CT and PET-CT, particularly PET-CT, overestimate CHLN metastasis. The HAM model accurate determines CHLN metastasis in NMNSCLC, thereby improving clinical decision-making.