Medalyze: Lightweight Medical Report Summarization Application Using FLAN-T5-Large

TL;DR

Medalyze tackles the challenge of accessible medical text by delivering three specialized Flan-T5-Large models for medical report summarization, health-issues extraction from conversations, and identifying central questions. It enables web and mobile usage with a YugabyteDB-backed data layer and a Flask-based API for lightweight, offline-ready deployment that preserves privacy. Experimental comparisons against GPT-4 across multiple medical-text tasks show strong semantic retention and competitive performance, while maintaining a significantly lighter footprint. The work provides a practical, privacy-preserving solution to improve information accessibility in healthcare and support non-specialist users in understanding complex medical content.

Abstract

Understanding medical texts presents significant challenges due to complex terminology and context-specific language. This paper introduces Medalyze, an AI-powered application designed to enhance the comprehension of medical texts using three specialized FLAN-T5-Large models. These models are fine-tuned for (1) summarizing medical reports, (2) extracting health issues from patient-doctor conversations, and (3) identifying the key question in a passage. Medalyze is deployed across a web and mobile platform with real-time inference, leveraging scalable API and YugabyteDB. Experimental evaluations demonstrate the system's superior summarization performance over GPT-4 in domain-specific tasks, based on metrics like BLEU, ROUGE-L, BERTScore, and SpaCy Similarity. Medalyze provides a practical, privacy-preserving, and lightweight solution for improving information accessibility in healthcare.

Figures (9)

• Figure 1: System Architecture of Medalyze
• Figure 2: Diagram of Data Source Preparation for Model Training
• Figure 3: Workflow of Model Training and Fine-Tuning
• Figure 4: Sequence Diagram for Text Summarization Process
• Figure 5: User History Section Interface of Medalyze