Clinical Predictor Tool

Limerick Digital Cancer Research Centre
Developed by: Meghana Kshirsagar and Gauri Vaidya

This project is a proof-of-concept of treatment personalization for patients with Head and Neck Squamous Cell Carcinoma (HNSCC) using structured outcomes from completed clinical trials with Large Language Models (LLMs).

It consists of two key components: one for generating patient profiles, and another for relevant treatment recommendations learned from trial outcome data.

Project Components

  1. Patient Profile Creation
    Generates a patient profile using a specified clinical trial ID.

  2. Treatment Recommendation
    Recommend treatments by analyzing trial outcomes in the context of the generated patient profile.

Usage Instructions

  1. For demonstration purposes, this tool uses the clinical trial ID NCT02268695. However, users are encouraged to explore additional completed trials by selecting different NCT IDs from clinicaltrials.gov.

  2. Open the Patient_Profile_Creation notebook and substitute the nct_id with any other and run all cells to generate a patient profile based on the selected clinical trial.

  3. Copy the generated patient profile and paste it into the Patient_Profile_Treatment_Recommendation notebook.

  4. Run the treatment recommendation notebook to receive a personalized recommendation based on trial outcome data.

Required Libraries

  • langchain
  • langchain_community
  • neo4j
  • transformers

Use the navigation bar to explore the notebooks.