Abstract Introduction: Pancreatic cancer accounts for 3% of cancer cases and 7% of cancer deaths in the United States. It is currently in the third position with respect to cancer-related deaths, and it is expected to move to second place as of 2040. This is mainly due to its diagnosis in the later stage. If detected in the first or second stage surgery can be performed thus increasing the patient’s survival rate. The current methods of diagnosis cannot do early stage or asymptomatic detection. So, detecting cancer-specific biomarkers can aid in the prognosis. Serum protein-based biomarkers carbohydrate antigen CA242, CA19-9, and carcinoembryonic antigen CEA are specific to pancreatic cancer. Thus, in this preliminary study, we aim to develop a nanoengineered electrochemical biosensor combined with machine learning to detect pancreatic cancer-specific biomarkers CA242, CA19-9, and CEA. Methodology: Nanoengineering the biosensor: The gold working electrode of the gold- biosensor (DRP-250AT) was coated with graphene oxide nano colloidal solution and kept in UV overnight for the graphene oxide nanosheet to crosslink with the working electrode. Detection of biomarkers using an electrochemical biosensor: First, a Crosslinker (DSP) was added to ethanol cleaned biosensor to bind the analytes. Following that antibody specific to each biomarker were added to the biosensors. The addition of TBS superblock prevented non-specific binding. Then different concentrations of biomarkers were added, and the corresponding sensor was subjected to various electrochemical analyses such as electrochemical impedance spectroscopy and voltammetry. SEM-EDAX analysis: The antigen-antibody binding on the biosensor's surface was confirmed. Automation and machine learning: Python code was developed to automate the analysis process and machine learning models (support vector machine and neural network) were used to classify the concentrations into various risk levels. The sensitivity of the developed tool was high when compared with conventional methods like ELISA and confocal microscopy. Results: From the electrochemical studies we found that with increasing concentration of biomarkers, there was a trend in output parameters: capacitance, resistance, impedance and area under the curve. By analyzing this trend, the concentrations of biomarkers can be detected. Using this data as input the machine learning model (accuracy greater than 80%) classified the concentrations into risk levels (normal, low risk, and high risk). Significance: Nanoengineering with graphene oxide nanosheets enhances the sensitivity of electrochemical biosensors. Automation and machine learning models help in the analysis of large datasets produced by this sensor. Overall, this device aids in the prognosis of pancreatic cancer. Our preliminary results are encouraging but more thorough research is required to make this diagnostic tool a working application. Citation Format: Meenal Karunanidhi, Hemalatha Kanniyappan, Edith Zhan, Ruth Mathew, Yani Sun, Junyi Wu, Yan Yan, Gnanasekar Munirathinum, Mathew T. Mathew. PancreaAlert: Intelligent nanoengineered biosensor for pancreatic cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7289.
Read full abstract