Machine learning prediction of the case-fatality of COVID-19 and risk factors for adverse outcomes in patients with non-small cell lung cancer

Abstract

Background

Since the emergence of coronavirus disease 2019 (COVID-19) across the globe, patients with cancer have been found to have an increased risk of infection with COVID-19 and are highly likely to experience a severe disease course. This study analyzed the clinical outcomes of COVID-19 in patients with non-small cell lung cancer (NSCLC) and identified the risk factors for adverse outcomes.

Methods

The study included patients diagnosed with COVID-19 between January 2020 and April 2022 at the Samsung Medical Center in Seoul, Republic of Korea, who have a history of NSCLC. The case-fatality rate and risk factors for COVID-19 were analyzed using a machine-learning prediction method. Additionally, the study investigated the effect of COVID-19 on the systemic treatment of patients with advanced-stage NSCLC.

Results

Overall, 1,127 patients were included in this study, with 10.3% of the patients being older than 75 years; of these patients, 51.8% were ex- or current smokers. Among the 584 patients cured after surgery, 91 had stable disease after concurrent chemo-radiotherapy, and 452 had recurrent or metastatic NSCLC. Among 452 patients with recurrent or metastatic NSCLC, 387 received systemic treatment in a palliative setting during COVID-19. Of these, 188 received targeted therapy, 111 received cytotoxic chemotherapy, 63 received immunotherapy +/− chemotherapy, and 26 received other agents. Among them, 94.6% of patients continued systemic treatment after the COVID-19 infection. Only one patient discontinued treatment because of complications of the COVID-19 infection, and 18 patients changed their systemic treatment because of disease progression. The case fatality rates were 0.86% for patients with early-stage NSCLC, 4.4% for patients with locally advanced NSCLC, and 9.96% for patients with advanced NSCLC. The factors associated with fatalities included palliative chemotherapy, type of palliative chemotherapy, age (≥75 years), diabetes, smoking history, history of lung radiotherapy, hypertension, sex, and chronic obstructive pulmonary disease (COPD). The predictive model using logistic regression and eXtreme Gradient Boosting (XGB) performed well [area under the curve (AUC) for logistic regression =0.84 and AUC for XGB =0.84].

Conclusions

The case fatality rate in patients with NSCLC was 4.8%, while most patients with advanced NSCLC continued to receive systemic treatment. However, patients with risk factors require careful management of COVID-19 complications.