A Herndon teen who graduated from Thomas Jefferson High School for Science and Technology this past June is one of 20 students from around the U.S. to be named 2018 Davidson Fellows, the Davidson Institute for Talent Development announced on Aug. 16.
Kavya Kopparapu, 18, earned a $50,000 Davidson Fellows Scholarship for cancer research that resulted in the development of a new, cost-effective method for detecting glioblastoma, an especially aggressive form of cancer that occurs in the brain and spinal cord.
“Kavya’s project is a major first step to personalized, targeted treatment for patients with cancer,” says a Davidson Institute for Talent Development profile for Kopparapu. “It identifies a significant disconnect in the treatment of patients in the current medical system. As a society, we’ve made leaps and bounds towards creating therapies targeted for a cancer’s specific genetic mutations, but have not significantly sped up the process for obtaining this tumor information.”
Kopparapu has been fascinated by science for as long as she can remember.
An aspiring doctor as a child, she grew up building robots and playing with Legos, while watching shows like Mythbusters and reading the magazine Scientific American with her younger brother.
Attending Thomas Jefferson only added fuel to her passion all things science-related.
After taking an Advanced Placement computer science course as a freshman, Kopparapu knew she wanted to explore the intersection of science and technology, especially in terms of how modern tech can be applied to address medical challenges.
Kopparapu’s interest in studying glioblastoma in particular took root last summer when she read articles about Beau Biden, the son of former Vice President Joe Biden, and Sen. John McCain (R-Ariz.), both of whom were diagnosed with the cancer and eventually died from it.
Glioblastoma is an especially malignant cancer, quick to spread and difficult to treat since the tumors affect different kinds of cells and have finger-like tentacles that make them hard to remove, according to the American Brain Tumor Association.
Glioblastoma represents about 15.4 percent of all primary brain tumors, increasing in frequency with age. The median survival for patients, meaning the point when the same number of patients improve and worsen, is 14.6 months from diagnosis.
Given that sobering prognosis, better ways to treat glioblastoma are no doubt needed, but Kopparapu says that, as she learned more about the disease, she became especially concerned by the challenges that doctors face when diagnosing tumors and determining a treatment plan.
Recent medical advances have given professionals the ability to develop personalized treatments that target the unique characteristics of an individual person’s disease, giving rise to a field known as precision medicine.
“This is such an awesome technology and to be able to use it would really impact the number of people that we could put into remission with cancer,” Kopparapu said. “The problem is, in order to get this information about a tumor, it takes three to four weeks of genome sequencing technology, which is very expensive and only available at the top-tier cancer centers. So, it’s not accessible to all patients, and it takes a long time.”
Using artificial intelligence skills that she has honed both through classes at Thomas Jefferson and on her own, Kopparapu came up with a possible solution.
With GlioVision, the Herndon resident developed an algorithm that can analyze a scanned image of a biopsy to predict different characteristics of the tumor, essentially doing the same work as standard gene-sequencing technology without the lag time and exorbitant costs.
This is not the first time that Kopparapu has worked with deep learning, where a computer is trained to recognize patterns in data that can be used to recognize speech, detect objects, or make predictions.
The teenager previously created a smartphone-based diagnostic system for diabetic retinopathy, a loss of vision caused by diabetes. That system could determine whether a patient had a retinal disorder based on an image of the person’s retina.
As a result, Kopparapu says the challenges she faced in developing GlioVision stemmed less from the technology itself and more from the data collection process.
She reached out to cancer researchers around the country seeking data and eventually acquired diagnosis-level images along with demographic information on about 350 glioblastoma patients going back to 2005.
Since the outcome of the diagnosis was already known for these patients, Kopparapu could work backwards from there to train her algorithm.
Now, she has partnered with the Georgetown University Medical Center on a clinical trial where she obtains information on new glioblastoma patients that come in, allowing her to compare the GlioVision results with those found by the medical center.
“So far, it’s been [an] 100 percent match between my algorithm and the traditional method for these new patients,” Kopparapu said.
The Davidson Institute for Talent Development awards $50,000, $25,000, and $10,000 scholarships to people 18 and under who have completed a college graduate-level project in science, technology, engineering, mathematics, literature, music, or philosophy that has the potential to benefit society.
Kopparapu was in California doing a summer internship for Apple’s machine learning division when she learned that she had been named a Davidson Fellow, though she had to read the email four or five times and call her parents before believing it was real.
“I would always read about the kids that won the award and read about their projects, and it was always really, really cool work they were doing that would impact so many people in the future,” Kopparapu said. “It’s kind of crazy to be among those kids and be a part of something so special.”
The $50,000 scholarship will go toward Kopparapu’s college tuition when she attends Harvard University this fall.
She plans to study computer science and biology with the goal of combining them for computational biology. Harvard offers that field of study through its school of public health but only as master and doctoral programs.
In the meantime, Kopparapu intends to continue working on GlioVision, which she has already patented, and she is also organizing an artificial intelligence summit for high school students that will be held at the Steven F. Udvar-Hazy Air and Space Museum in Chantilly.
Scheduled for Oct. 6 and 7, the summit is the work of the Girls Computing League, a nonprofit that Kopparapu founded because she wanted to encourage other female students to pursue computer science.
She still serves as the organization’s CEO, and its leadership team consists of other current and former Thomas Jefferson students.
Kopparapu credits her teachers at Thomas Jefferson for giving her a strong foundation in science and technology, and for supporting her as she takes on ambitious projects like GlioVision.
“They had a really deep understanding of the technology and what I was trying to accomplish,” Kopparapu said. “Going to TJ was invaluable for all the scientific pursuits that I’ve been able to do in my time in high school.”