RecuNet: A Novel, Low-Cost, & Automated Pipeline for the Spatiotemporal Prediction of Brain Tumor Recurrence

Overview

Gliomas affect 90,000 people annually in the United States and have 5-year survival rates as low as 7%, largely due to 5 -year recurrence rates of 52 -62%. Current MRI relies on contrast enhancement (CE) to visualize the glioma, which has several limitations: (1) tumor infiltration often extends far beyond CE margins; (2) After surgery, non -contrast enhancing tumor grows undetected & appears as “recurrence” in follow -up scans. Blood Oxygen Level -Dependent (BOLD) fMRI measures brain blood flow, which can be disrupted by the tumor microenvironment. Therefore, BOLD, combined with current imaging, could detect real-time tumor progression prior to radiologic “recurrence”, defined by CE. This project, RecuNet, aims to (1) establish a link between BOLD, non -contrast enhancing tumor, and recurrence, and use deep learning to (2) detect the non-contrast enhancing tumor and (3) spatiotemporally predict tumor recurrence. Both algorithms take standard (T1 + FLAIR) scans and BOLD fMRI as inputs. The detection algorithm uses a 3D-UNet Architecture with optimized loss functions and Attention Gated Networks (AGNs). The Prediction Model is a CNN that uses temporal-spatial convolutional layers and hybrid inputs to extract features in peritumoral regions. Preliminary detection res ults show an IoU of 94.1%, accurately detecting con -contrast enhancing tumor portions. The prediction algorithm has a 94.52% location accuracy and a mean average error of 5.3 days from recorded recurrence time, significantly outperforming current methods. RecuNet nearly eliminates recurrence risk by visualizing the entire, exact tumor and accurately predicting areas with a high risk of developing tumor growth, saving money, resources, and lives. Combating PFAS Contamination: A Preconcentration Method for Detection in Drinking Water Wanda Wu duPont Manual High, Louisville, KY Mentor, Dr. Xiao An Fu, University of Louisville Per- and polyfluoroalkyl substances (PFAS), nicknamed “forever chemicals,” are a group of harmful chemicals that are known to cause damage to the immune system and liver. Despite being found everywhere, including in drinking water, only 30 states have imp lemented PFAS regulations, only 11 of which include drinking water. Limited PFAS regulations is partly because PFAS usually exists in too low of concentrations to detect. Therefore, this research focuses on pre-concentrating PFAS to develop an on -site method for trace-level PFAS detection in drinking water. In this study, perfluorooctanoic-acid (PFOA) was pre-concentrated by flowing 8 mL of a 0.25 ng/L PFOA solution through a sorbent in a microfabricated -chip, followed by elution with 800 uL of phosphate-buffer solution (PBS). The solution was then run through a microfluidic sensor with two sandwiched gold electrodes, where a current was applied between the electrodes and the change in electrical impedance was used to quantify the PFAS present. A solution of 0.05ng/L PFOA was then flowed through the sorbents a t the highest -performing flow rate to simulate real -world contamination levels. Three sorbents (carboxen, carbopack, and tenax) were tested at two flow rates (0.25 and 0.5 mL/min), and it was hypothesized that carboxen at 0.25 mL/min would have the highest capture efficiency due to the largest surface area and longer contact time. However, the research found that tenax at 0.25 mL/min slightly surpassed carboxen, indicating the importance of pore size of the sorbent in PFAS adsorption.

Competition history

• JSHS 2025 Category not listed

Resources

• National JSHS abstract book ↗

Related projects

Closest projects by meaning, across every fair and year in the corpus.