Monday, October 31, 2011

Computer program found useful in cancer diagnosis

An algorithm developed to recognize people at risk for developing cancer has proven itself to be useful for use by general practitioners. The program combines risk factors and symptoms that are provided by patient data, which results in a warning when this combination could lead to cancer in the future. Specifically, the tool is able to predict lung cancer and gastro-oesophagael cancer. By raising awareness to GP's, the tool could save about 5000 lives each year. That is what scientists conclude after conducting experiments with the program, which seems quite high for a tool that just raises attention to known risk factors and symptoms.

According to the study, 77 percent of the total number of patients in the UK that have been diagnosed with  lung cancer or gastro-oesophagael cancer were predicted by the computer program, highlighting how important future patient discovery by a computer algorithm can be. These patients represent 10 percent of the total number of people that were identified by the program as high risk, two years before the diagnosis was made. That means, two years after assessment by the program, one out of ten people identified as high risk will have cancer in the near future.

One of the advantages is that the tool can be incorporated with existing IT systems to analyse patient data. By using the tool on all known patient data, many more people with high risk on cancer could be discovered, potentially saving many lives. Because detecting cancer early is vital for chances of survival, identifying people with high risk and consequently monitoring them can prevent cancer deaths.

The results show that the developed computer program for GP's is a useful and seemingly important tool to use on patient data. It could also be a first step to the generation of more algorithms, that detect other forms of cancer, and possibly even other diseases. We keep discovering new risk factors and biomarkers that are indicators for disease, which makes it likely that more automated risk assessments will find their way to the clinic in the future. 

No comments:

Post a Comment