JMIR Publications

ABSTRACT

Background:

Liquid biopsies based on blood samples have been widely accepted as a diagnostic and monitoring tool for cancers, but extremely high sensitivity is frequently needed due to the very low level of the specially selected DNA/RNA/protein biomarkers released into blood. On the other hand, routine blood indices tests are frequently ordered by physicians, easy-to-perform and cost effective. Meanwhile, artificial intelligence is broadly accepted for its capability of deciphering complex connections between multiple test data and diseases.

Objective:

To discover the association between lung cancer and routine blood indices and hence to help clinicians and patients to identify lung cancer based on routine blood indices.

Methods:

The machine learning method as RandomForest was adopted to build the identification model between routine blood indices and lung cancer. Ten-fold cross-validation test and external validation test were utilized to evaluate the reliability of the built model.

Results:

277 patients with 49 types of routine blood indices were included in this study, including 183 patients with lung cancer and other 94 patients without lung cancer. Astonishing correlation between the combination of 19 types of routine blood indices and lung cancer were found here. Lung cancer patients can be distinctly identified from other patients with sensitivity, specificity and total accuracy of 0.9630, 0.9497 and 0.9570 for the cross-validation validation test, respectively, especially from patients with tuberculosis, which usually has similar clinical symptoms with lung cancer. This identification method, called RBLC, which is promising to be of help as a useful tool for the identification of lung cancer based on routine blood indices for both clinicians and patients, is now online available at http://lishuyan.lzu.edu.cn/RBLC.

Conclusions:

Lung cancer can be identified based on the combination of 19 types of routine blood indices. It implies that artificial intelligence can find the remote relevance of a disease with the fundamental indices of blood, which may reduce the necessity of costly elaborate blood test techniques for this purpose. Correlation of the combination of multiple indices obtained from the routing blood test with other diseases may also be expected.