Machine Learning Algorithms Offer Alternative to Imaging in NSCLC Staging
A recent study published in Journal of Thoracic Disease demonstrated the effectiveness of artificial intelligence (AI) in predicting lymph node metastasis in patients with non-small cell lung cancer (NSCLC). The model used only routine preoperative data, without the need for contrast-enhanced computer tomography (CT) or positron emission tomography (PET).
The study involved a retrospective analysis of 988 patients with NSCLC who underwent pulmonary resection and mediastinal lymph node dissection between January 2011 and October 2022. The AI model relied on accessible clinical data, including age, sex, smoking history, tumor characteristics, and blood tests, such as tumor markers.
Six machine learning algorithms—Support Vector Classification (SVC), k-nearest neighbor (k-NN), logistic regression (LR), random forest (RF), gradient boosting (GB), and multilayer perceptron (MLP)—were trained and validated to predict lymph node metastasis. The gradient boosting model achieved the highest predictive performance, with an accuracy of 80.0%, a specificity of 95.6%, and an area under the curve (AUC) of 0.75. These results indicate that AI models can offer similar predictive capability to current standard imaging methods, such as PET and contrast-enhanced CT, which have reported diagnostic sensitivities and specificities ranging from 59% to 90%.
Permutation feature importance (PFI) analysis revealed that serum carcinoembryonic antigen (CEA) and cytokeratin fragment 19 (CYFRA) levels, maximum solid tumor diameter, age, and consolidation-to-tumor ratio were the most influential variables in predicting lymph node metastasis. These factors align with established clinical predictors of metastasis in lung cancer. Additionally, the model identified tumors located in certain lung segments, particularly segments 6 and 3, as having a higher risk of lymph node involvement.
The implications of this study are significant for surgical planning in early-stage NSCLC. Lymph node metastasis is a critical factor influencing staging, prognosis, and treatment strategy, with higher rates of recurrence linked to lymph node involvement. Despite improvements in imaging technology, routine use of PET and contrast-enhanced CT for early-stage NSCLC remains debated due to concerns regarding cost and radiation exposure. The AI model described in this study provides a potential solution by utilizing noninvasive, low-cost diagnostic data to guide clinical decision-making.
“We demonstrated that an AI model without access to information from PET or contrast-enhanced CT could predict preoperative lymph node status with relatively high performance,” the researchers concluded. “Surgeons can assess the lymph node status by confirming the presence AI model score and PET or contrast-enhanced CT features, while also considering the range of lymph node metastasis in anatomical lung resection.”
Reference
Yoshimura R, Endo Y, Akashi T, et al. Diagnostic artificial intelligence model predicts lymph node status in non-small cell lung cancer using simplified examination. J Thorac Dis. 2024;16(11):7320-7328. doi:10.21037/jtd-24-1067
