Machine Learning Algorithm Guiding Local Treatment Decisions to Reduce Pain for Lung Cancer Patients with Bone Metastases, a Prospective Cohort Study

Introduction As life expectancy increases for lung cancer patients with bone metastases, the need for personalized local treatment to reduce pain is expanding. Methods Patients were treated by a multidisciplinary team (MDT), and local treatment including surgery, percutaneous osteoplasty, or radiation. Visual analog scale (VAS) and quality of life (QoL) scores were analyzed. VAS at 12 weeks after treatment was the main outcome. We developed and tested machine learning models to predict which patients should receive local treatment. Model discrimination was evaluated by the area under curve (AUC), and the best model was used for prospective decision-making accuracy validation. Results Under the direction of MDT, 161 patients in the training set, 32 patients in the test set, and 36 patients in the validation set underwent local treatment. VAS in surgery, percutaneous osteoplasty, and radiation groups decreased significantly to 4.78 ± 1.28, 4.37 ± 1.36, and 5.39 ± 1.31 at 12 weeks, respectively ( p < 0.05), with no significant differences among the three datasets, and improved QoL was also observed ( p < 0.05). A decision tree (DT) model that included VAS, bone metastases character, Frankel classification, Mirels score, age, driver gene, aldehyde dehydrogenase 2, and enolase 1 expression had a best AUC in predicting whether patients would receive local treatment of 0.92 (95% CI 0.89–0.94) in the training set, 0.85 (95% CI 0.77–0.94) in the test set, and 0.88 (95% CI 0.81–0.96) in the validation set. Conclusion Local treatment provided significant pain relief and improved QoL. There were no significant differences in reducing pain and improving QoL among training, test, and validation sets. The DT model was best at determining whether patients should receive local treatment. Our machine learning model can help guide clinicians to make local treatment decisions to reduce pain. Trial registration Trial registration number ChiCRT-ROC-16009501.