AI technologies have been increasingly deployed to understand the most effective, and least risky, treatment for various conditions.  The latest such tool comes from a team at the National University of Singapore.  In a recent study, they describe how they can predict the survival rate and treatment outcomes of early-stage lunch cancer patients.

The tool uses 29 extracellular matrix (ECM) genes that have been identified due to their abnormal expression in lung cancer patients compared to those with healthy lung tissues.  ECM is important because it defines the space that surrounds cells, whilst providing structural and biochemical support to those cells.  Studies have suggested there is a clear link between tissue stiffness and the risk of cancer because the cells in the tumor produce a fibrous protein that eventually form an ECM for the cancer cells to attach to.

Predicting lung cancer

“The traditional way of targeting cancer has been a ‘one size fits all’ approach for patients. Yet, although two persons may have the same type of cancer, how the disease manifests and progresses is unique to each individual,” the researchers say.  “In our research, we look at non-small cell lung cancer, which is the most common type of lung cancer. Our novel tool successfully identified early-stage patients who derived survival benefit from adjuvant chemotherapy. This is a very exciting development as we have taken a big step forward in enabling treatments to be customized for cancer patients to improve survival rates.”

They believe that their work is another important step towards personalized medicine, with the right treatment given to the right person, in the right dose and at the right time.

“When precision medicine meets big data, its potential is even greater. With the increase of global joint efforts in sharing large-scale data, we were able to explore the genomic data across multiple cancer types through various databases,” they say.

By identifying 29 specific ECM components that could be used as biomarkers for lung cancer diagnosis, the team made the first step towards the creation of a gene panel for clinical application.  The panel was tested on over 2,000 early-stage lung cancer patients, with the results validating the tool’s ability to predict survival outcomes and chemotherapy success rates.

“Our study demonstrates how we can harness and transform unprecedented amount of genomic data into a useful decision-making tool that can be implemented in routine clinical practice. We are excited about the potential of applying our novel bioinformatics approach into the emerging area of liquid biopsy, which serves as an alternative to invasive and painful tissue biopsy,” the researchers explain.

The team hope to further their work and explore whether the gene panel could also be useful in predicting the survival rate and treatment outcomes in 11 other cancer types.  It’s another interesting example of how data can be useful in tackling cancer.