Model Uses Blood Proteins to Predict Lung Cancer Risk

Imagine if you could use a simple blood sample to predict lung cancer—one of the deadliest cancers in the world. This type of diagnostic could be a game changer for early lung cancer detection. And while the work is still underway, the preliminary results look promising, thanks to a new algorithm that’s helping unravel information from proteomics data.

Using proteomics data from cancer-free people and those who went on to develop cancer, researchers at the International Agency for Research on Cancer are creating a protein-based prediction tool. The researchers’ tool will help decide who to screen for lung cancer using imaging (i.e., low dose computed tomography).

You can read about this extensive research effort, which is drawing together NCI researchers, along with scientists from 25 institutes from around the globe, in two recent reports.

Codes for the lung cancer proteomics study are available on GitHub.

The authors say they’re encouraged by this protein-based prediction model’s early results. The next step will be to further refine and validate the model with additional data from the Lung Cancer Cohort Consortium (LC3) and, eventually, to translate these findings to a clinical setting.

As noted by Dr. Danielle Mercatante Carrick, of NCI’s Division of Cancer Control & Population Sciences’ Genomic Epidemiology Branch, if this type of blood-based risk assessment tool shows widespread success, it could significantly improve early detection of lung cancer. She notes, “From a public health perspective, the benefits of this type of screening could help offset the costs and efforts involved in routine use of a blood-based tool such as this.”

Vote below about this page’s helpfulness.