21P - Improving early cancer screening efficacy by adjusting tumor burden spectrum bias of liquid biopsy biomarkers

Background

Liquid biopsy biomarkers are important for early cancer screening. However, the available samples are usually from patients who have been clinically diagnosed, resulting in relatively higher tumor burdens compared to the point-of-care population for cancer screening. Due to this tumor burden spectrum bias, those biomarkers may significantly lose their screening efficacy in early cancer detection.

Methods

We provided a statistical model of biomarker distribution along with tumor burden based on the Tumor-derived Biomarker Circulation Model (TBCM). According to the observed sample with high tumor burden, we estimated and thereafter optimized the diagnostic efficacy of biomarker panel under given low tumor burden distribution. We test our method via simulated data based on literature of lung cancer and renal cancer. The efficacies of biomarker panels were measured using the area under the receiver operating curve (AUROC), sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Tumor burden spectrum bias was estimated comparing the efficacy measures between the observed samples and the target screening population. Additionally, we compared the efficacy measures of biomarker panels obtained from our method against those from traditional methods.

Results

For lung cancer, the traditional biomarker panel had AUC 0.881 among observed samples, which decreased into AUC 0.675 among target screening population. Our improved biomarker panel achieved AUC 0.752 among target screening population. For renal cancer, the traditional biomarker panel had AUC 0.937 among observed samples, which decreased into AUC 0.581 among target screening population. Our improved biomarker panel achieved AUC 0.711 among target screening population.

Conclusions

Tumor burden spectrum bias is a significant factor contributing to the failure of liquid biopsy biomarkers in early cancer screening. Our method can correctly adjust this bias and substantially improve the efficacy of biomarkers in the target screening population.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

The author.

Funding

National Natural Science Foundation of China.

Disclosure

The author has declared no conflicts of interest.