![After a patient has received treatment for brain cancer, lesions often show up on MRI scans that resemble a recurring tumor. Researchers hope artificial intelligence tools could find that information quicker, without subjecting the patient to another invasive surgery. [create jobs 51 / Shutterstock]](https://npr.brightspotcdn.com/dims4/default/0f2fcc4/2147483647/strip/true/crop/3000x1679+0+0/resize/840x470!/format/webp/quality/90/?url=http%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2Flegacy%2Fuploads%2F2020%2F7%2F29%2Fshutterstock_695480482.jpg)
Case Western Reserve University researchers are investigating how artificial intelligence (AI) can improve cancer treatments.
The technology could possibly cut down on the number of surgeries needed, said lead researchers Pallavi Tiwari and Satish Viswanath.
Tiwari specializes in brain cancers, while Viswanath studies colorectal cancers. In both types of cancers, lesions or dead tissues often show up on MRI scans after a patient receives treatment, which can resemble recurring tumors, said Tiwari.
鈥淭he only definitive diagnosis comes from a surgical resection, which means that they go in and do a biopsy, or they go in and take the whole thing out 鈥� only to later discover that it鈥檚 a benign condition,鈥� she said.
The surgeries are invasive and often unnecessary, she said, and in some cases, they can significantly reduce a patient鈥檚 quality of life.
Tiwari and Viswanath are collaborating with imaging specialists, AI experts and local physicians to develop AI software that could help doctors quickly determine whether these masses are cancerous without having to operate on the patient.
鈥淚f we can use non-invasive, routinely required MRI scans to extract information, to mine information 鈥� then that can save a patient who has that benign condition the surgery that, to start with, was not needed,鈥� Tiwari said.
The tools would allow doctors to see measurements of the tumor, such as whether it has progressed or changed in size.
鈥淲hat we are trying to do is unlock all of the information that鈥檚 available by building these AI algorithms that can go in and extract information from the image that鈥檚 present 鈥� we just can鈥檛 see it,鈥� Viswanath said.
The tools are being developed for brain and colorectal cancer treatments for now, but the hope is that they can be used in other fields of study in the future, Viswanath said.
鈥淭he idea is, can we build them in such a way that anyone in the research community and then eventually, hopefully, the clinical community can run these tools and work them, and sort of disseminate them broadly to a big audience,鈥� he said.
The research is funded through a recent $1.15 million three-year grant from the National Cancer Institute.
