AI- and LLM-Driven Discovery to Predictive Medicine of Curcumin and Demethoxycurcumin in Oral Squamous Cell Carcinoma Using Swalife Research Platforms

Abstract

Background: Oral Squamous Cell Carcinoma (OSCC) is a major global health burden, particularly in developing
countries like India, with strong associations to tobacco use, chronic inflammation, and metabolic dysregulation.
Conventional therapies often face limitations such as toxicity, recurrence, and resistance. Curcumin and its analog
demethoxycurcumin, derived from Curcuma longa, have demonstrated multi-target anticancer potential.
However, systematic integration of AI-driven approaches for mechanistic discovery to predictive medicine
remains underexplored.
Objective: To develop an AI- and large language model (LLM)-driven integrated research pipeline using Swalife
platforms to evaluate the mechanistic, preclinical, clinical, and predictive potential of curcumin and
demethoxycurcumin in OSCC.
Methods: A multi-stage workflow was employed combining literature mining (PubMed, Google Scholar,
ScienceDirect), AI-driven prompt-based analysis (Perplexity AI), and Swalife research tools. Six major models
were developed: Target & Mechanism, Lead Identification & Optimization, In vitro Design, In vivo Design,
Clinical & Pharmacovigilance (PV), and Market & IPR. Data were structured into standardized datasets and

subjected to cross-validation and consistency checks to improve reliability. This study was conducted as an AI-
assisted, in silico analysis based on systematic literature mining and biomedical databases. No new in vitro, in

vivo, or clinical experiments were performed. Data were synthesized using AI tools and validated through cross-
checking with published studies.

Results: Curcumin has been reported to demonstrate multi-target modulation across key OSCC pathways
including NF-κB, PI3K/Akt, STAT3, and Wnt/β-catenin.
Conclusion: This study presents an AI-driven integrated framework for mapping curcumin and
demethoxycurcumin from molecular mechanisms to predictive medicine in OSCC. This workflow demonstrates
a scalable and reproducible model for drug discovery and translational research using natural compounds.