AI Helps Forecast Brain Tumor Outcomes
A team of researchers, including scientists from Stanford Medicine, has created an artificial intelligence (AI) model capable of forecasting brain tumor outcomes by examining stained images of a type of brain cancer called glioblastoma. The researchers published a paper describing how the AI model could help physicians identify cellular markers for more aggressive tumors in patients before marking them for expedited follow-ups.
Brain cancers tend to have poor outcomes because their proximity to important brain tissue makes them incredibly hard to treat via conventional methods such as surgery and radiation without harming healthy tissue. The blood-brain barrier, which is a sort of filter that prevents dangerous material from entering the brain through the bloodstream, also prevents chemotherapy drugs from reaching brain tumors and reduces the effectiveness of brain cancer treatment.
Another reason why glioblastoma can be quite hard to treat is that the cellular makeup of glioblastoma tumors differs significantly from one person to another. This means that a treatment protocol that was effective for one person may not be as effective for another person, increasing the need for more personalized treatments.
Since patients often have an average life expectancy of only a year after receiving a glioblastoma diagnosis, physicians rarely have the time to develop individual treatment protocols for each patient before the disease becomes too advanced to treat.
The AI tool developed by the Stanford Medicine researchers can deal with the heterogeneity of glioblastoma tumors by analyzing the tumor cells’ genetic makeup to predict the tumor’s aggressiveness and the degree of cancer cells that will remain after surgery. Stanford Medicine postdoctoral scholar Yuanning Zheng, PhD, calls the model a “support system” for doctors.
At the moment, physicians identify brain tumors and create treatment plans by studying pictures of stained tumor tissue. The technique shows the location and shape of tumor cells but does not provide an in-depth look at the tumor cells, minimizing doctors’ ability to design effective treatment plans.
Olivier Gevaert, PhD, a biomedical informatics and data science associate professor, says that a recently developed imaging technique called spatial transcriptomics also reveals the genetic makeup of different types of cancer cells, allowing for more effective treatment and better outcomes. However, the technique is still too costly for the average patient.
Gevaert and his team developed the AI model using genetic data using more than 20 glioblastoma patients and spatial transcriptomics images. The research team has provided a proof-of-concept version of the AI model for researchers, but Zheng notes that it is still in the research phase and will need to be trained on more data before it can be released to physicians.
Once it is ready for release to physicians, the model could potentially allow doctors to identify cancer patients who are at risk of progressing at an accelerated rate, giving them the opportunity to deploy the appropriate treatment protocols early.
As it becomes easier to track or profile brain cancers, it is likely that central nervous system cancers could get more effective drugs from companies such as CNS Pharmaceuticals Inc. (NASDAQ: CNSP).
NOTE TO INVESTORS: The latest news and updates relating to CNS Pharmaceuticals Inc. (NASDAQ: CNSP) are available in the company’s newsroom at https://ibn.fm/CNSP
Please see full terms of use and disclaimers on the BioMedWire website applicable to all content provided by BMW, wherever published or re-published: http://BMW.fm/Disclaimer
A team of researchers, including scientists from Stanford Medicine, has created an artificial intelligence (AI) model capable of forecasting brain tumor outcomes by examining stained images of a type of brain cancer called glioblastoma. The researchers published a paper describing how the AI model could help physicians identify cellular markers for more aggressive tumors in patients before marking them for expedited follow-ups.
Brain cancers tend to have poor outcomes because their proximity to important brain tissue makes them incredibly hard to treat via conventional methods such as surgery and radiation without harming healthy tissue. The blood-brain barrier, which is a sort of filter that prevents dangerous material from entering the brain through the bloodstream, also prevents chemotherapy drugs from reaching brain tumors and reduces the effectiveness of brain cancer treatment.
Another reason why glioblastoma can be quite hard to treat is that the cellular makeup of glioblastoma tumors differs significantly from one person to another. This means that a treatment protocol that was effective for one person may not be as effective for another person, increasing the need for more personalized treatments.
Since patients often have an average life expectancy of only a year after receiving a glioblastoma diagnosis, physicians rarely have the time to develop individual treatment protocols for each patient before the disease becomes too advanced to treat.
The AI tool developed by the Stanford Medicine researchers can deal with the heterogeneity of glioblastoma tumors by analyzing the tumor cells’ genetic makeup to predict the tumor’s aggressiveness and the degree of cancer cells that will remain after surgery. Stanford Medicine postdoctoral scholar Yuanning Zheng, PhD, calls the model a “support system” for doctors.
At the moment, physicians identify brain tumors and create treatment plans by studying pictures of stained tumor tissue. The technique shows the location and shape of tumor cells but does not provide an in-depth look at the tumor cells, minimizing doctors’ ability to design effective treatment plans.
Olivier Gevaert, PhD, a biomedical informatics and data science associate professor, says that a recently developed imaging technique called spatial transcriptomics also reveals the genetic makeup of different types of cancer cells, allowing for more effective treatment and better outcomes. However, the technique is still too costly for the average patient.
Gevaert and his team developed the AI model using genetic data using more than 20 glioblastoma patients and spatial transcriptomics images. The research team has provided a proof-of-concept version of the AI model for researchers, but Zheng notes that it is still in the research phase and will need to be trained on more data before it can be released to physicians.
Once it is ready for release to physicians, the model could potentially allow doctors to identify cancer patients who are at risk of progressing at an accelerated rate, giving them the opportunity to deploy the appropriate treatment protocols early.
As it becomes easier to track or profile brain cancers, it is likely that central nervous system cancers could get more effective drugs from companies such as CNS Pharmaceuticals Inc. (NASDAQ: CNSP).
NOTE TO INVESTORS: The latest news and updates relating to CNS Pharmaceuticals Inc. (NASDAQ: CNSP) are available in the company’s newsroom at https://ibn.fm/CNSP
Please see full terms of use and disclaimers on the BioMedWire website applicable to all content provided by BMW, wherever published or re-published: http://BMW.fm/Disclaimer
