New data on common mutations associated with prostate cancer have been instrumental in creating a new computer model used to predict the growth of prostate cancer, according to an international team of scientists.
Using Predictive Analysis to Help Differentiate Cancer Diagnosis
Diagnosing different types of prostate cancer can be a difficult task. There’s a lot riding on an accurate diagnosis, as identifying an aggressive cancer correctly creates opportunities to treat that cancer much more effectively. If it’s misdiagnosed as less aggressive, these opportunities slip away; likewise, if a less aggressive cancer is characterized incorrectly, a patient may end up undergoing treatments unnecessarily.
The challenge of differentiating cancer diagnoses more accurately has been taken up by researchers from both Denmark, from the University of Copenhagen, and Germany, from Heidelberg’s European Molecular Biology Laboratory. Their adopted approach, according to a research paper published in the journal Cell, has been to use computer modeling in order to predict prostate cancer growth by feeding cell mutation patient data into a simulation.
How the New Approach Works
Researchers took data from close to 300 men suffering from early-onset prostate cancer. The process included genetic sequencing each patient’s cancer to identify every mutation present within the tumor. Then, based on this voluminous amount of information, the computer model can be used to check a new patient’s sequenced data against the existing database to make predictions on how this new cancer will react based on its own mutations.
The computer model helps to determine what types of changes a new patient’s cancer are most likely to undergo in the future, based on what other patients with the same mutations experienced themselves. This allows doctors to predict how the cancer will change and grow, and this, in turn, makes it easier to anticipate what specific types of treatment a patient may need in the future.
Researchers say that while having the full cancer genome from 300 patients is obviously a good start, the computer model will become more capable of accurate predictions as more data is added. The end goal is to have several thousand patients’ worth of mutation data to pull from, which will be added over the next several years as this computer model is implemented in cancer clinics around the world; there’s already one such clinic, located in Germany, where the model is up and running.
