A Concept for Mining Transitive Sequential Patterns from Pancreatic Cancer Patient Journeys

Area of Science:

• Oncology
• Medical Informatics
• Data Science

Background:

• Pancreatic cancer is aggressive with a poor prognosis, requiring optimized treatment strategies.
• The sequence of clinical trial procedures, like chemo-radiotherapy, is crucial but potentially incomplete.
• Other temporal sequences in patient medical histories may influence treatment response and outcomes.

Purpose of the Study:

• To explore the utility of extracting transitive sequential patterns from patient medical trajectories.
• To identify temporal characteristics in complex diseases like pancreatic cancer.
• To demonstrate how discovered sequential patterns can aid pancreatic cancer research and patient care.

Main Methods:

• Utilizing data mining techniques to extract transitive sequential patterns from electronic health records.
• Analyzing patient medical trajectories to identify significant temporal sequences.
• Applying pattern discovery to pancreatic cancer patient data.

Main Results:

• Demonstrated a method for discovering complex temporal patterns in patient data.
• Highlighted the potential for these patterns to reveal previously unrecognized therapeutic sequences.
• Showcased the application of this methodology in the context of pancreatic cancer.

Conclusions:

• Transitive sequential pattern extraction offers a novel approach to understanding disease progression.
• Identifying temporal characteristics can lead to more effective pancreatic cancer treatment strategies.
• This methodology holds promise for advancing personalized medicine and improving patient care in oncology.