The San Diego Supercomputer Center (SDSC) and Moores Cancer Center have worked with CureMatch to develop a machine learning model to diagnose bladder cancer and identify what stage its at.
Called multi-layer perceptron or MLP, the approach is expected to provide a non-invasive diagnostic option and analyse a patient’s metabolites and their chemical descriptors to detect the cancer.
The research team noted that the new technology could accurately classify bladder cancer stages in a patient. It is hoped that this model would help the current invasive and costly diagnosis process.
San Diego Supercomputer Center neurosciences department research professor Igor Tsigelny said: “From my point of view, it can be very easy for patients [to] just give a sample of urine and our ML system can produce a ‘red flag’ analysis result telling them to go immediately to an oncologist for testing.
“Our machine learning model uses metabolites and corresponding genes to determine if a patient has bladder cancer and if so, at what stage.”
The team used multiple computational tools to examine the pathways associated with different bladder cancer stages that are useful in diagnostics and the monitoring of disease progression.
Urine metabolites data from patients at various stages of the cancer was then used to train the MLP software, which reviews the chemical descriptor of the metabolites related to each stage and generates artificial intelligence (AI) models of these profiles.
The American Cancer Society data shows that out of more than 81,000 US people diagnosed with bladder cancer last year, more than 17,000 died from the disease.
It is expected that the new machine learning model would help reduce this number.
