Analysis Pancreatic Cancer Risk with Manifold Fitting¶
Overview¶
Pancreatic cancer ranks among the top ten deadliest cancers worldwide, with a five-year survival rate of less than 10%. Due to the lack of noticeable symptoms in its early stages, most cases are diagnosed at advanced stages, resulting in poor treatment outcomes. Early risk screening is thus critical for effective pancreatic cancer prevention and control.
This study utilizes the UK Biobank database, which includes data from approximately 210,000 participants,and variables as demographic information, physical measurements, self-reported lifestyle factors, metabolic biomarkers, comorbidity diagnoses, and more, to identify significant features linked to pancreatic cancer risk. We aim to develop an interpretable, accurate, and scalable predictive model by integrating manifold fitting with deep learning techniques to capture the nonlinearities, heterogeneity, and complex interactions in high-dimensional data. Building on this, a novel hybrid model that combines the strengths of decision trees and deep neural networks is proposed.
This research highlights the potential to improve early diagnosis, enable personalized preventive strategies by targeting modifiable risk factors, and advance understanding of pancreatic cancer etiology, providing valuable data support for future studies and enabling researchers to gain deeper insights into the etiology of pancreatic cancer.
This study utilizes the UK Biobank database, which includes data from approximately 210,000 participants,and variables as demographic information, physical measurements, self-reported lifestyle factors, metabolic biomarkers, comorbidity diagnoses, and more, to identify significant features linked to pancreatic cancer risk. We aim to develop an interpretable, accurate, and scalable predictive model by integrating manifold fitting with deep learning techniques to capture the nonlinearities, heterogeneity, and complex interactions in high-dimensional data. Building on this, a novel hybrid model that combines the strengths of decision trees and deep neural networks is proposed.
This research highlights the potential to improve early diagnosis, enable personalized preventive strategies by targeting modifiable risk factors, and advance understanding of pancreatic cancer etiology, providing valuable data support for future studies and enabling researchers to gain deeper insights into the etiology of pancreatic cancer.