Nutritional, inflammatory and functional biomarkers in lung cancer : identifying patients at risk of adverse outcomes through two retrospective cohort studies

Abstract

Background

Lung cancer is the commonest cause of cancer death worldwide. A range of biomarkers are associated with adverse outcomes in lung cancer, but these have not been assimilated into routine clinical practice. The aim of the two studies was to identify predictive variables within existing healthcare data for adverse outcomes following lung cancer treatment, with a view to informing optimal treatment selection for future patients.

Methods

Two retrospective cohort studies of lung cancer patients in South East Scotland were undertaken using demographic and clinical data from healthcare records. A range of explanatory variables were explored using descriptive statistics, logistic regression and survival analysis for treatment-related outcomes. These included overall survival (OS), early mortality and treatment completion.

Results

194 patients were included the chemoradiotherapy study, median OS 19 months. Low skeletal muscle attenuation (MA), (odds ratio [OR] 1.61 [95% CI 1.16, 2.23, p=0.004) independently predicted reduced OS. Independent predictors of death within 90 days of treatment completion were Eastern Cooperative Oncology Group Performance Status ≥2 (OR 3.97 [1.20, 13.08], p=0.024) and body mass index (BMI) ≤20 (OR 3.91 [1.24, 12.38], p=0.020).

397 patients were included in the palliative chemotherapy study, median OS 6.9 months. Independent predictors of reduced OS were: neutrophil-to-lymphocyte ratio ≥4, albumin <35, MA and low skeletal muscle mass. Patients who did not receive guideline-recommended treatment (GRT) had a median OS of 3.3 months. Independent predictors of non-GRT receipt

were: non-small cell lung cancer, BMI ≤20, neutrophil count ≥7, lymphocyte count <1 and MA

<31.55.

Discussion

A range of routinely available biomarkers can identify patients with lung cancer at increased risk of adverse outcomes. Optimal treatment selection for each patient could be improved by routine utilising these biomarkers. Biomarkers may also be useful to identify patients for integrated supportive care during their cancer treatment. Further research is needed.