AI Improving Dose Decisions and Patient Outcomes in Oncology- with Shefali Kakar of Novartis

2026-05-26 · Source: The AI in Business Podcast · Field: Health & Wellbeing — Pharmaceuticals & Biotechnology, Clinical Care & Medical Practice, Health & Medical Research · Depth: Intermediate, long

Summary

Shefali Kakar, Global Head of PK Sciences and Oncology at Novartis, discusses how AI-driven modeling and deeper data integration are transforming oncology drug development. These approaches provide earlier clarity on dose decisions, safety signals, and patient variability, addressing limitations of traditional single-study methods. Kakar highlights the use of longitudinal analysis, exposure-response modeling, and covariate evaluation to reduce unnecessary sub-studies, tailor dosing for diverse patient groups, and strengthen cross-functional decision-making. The discussion also touches on "in silico" drug discovery and clinical trials, which leverage mass data to model chemical changes and predict drug success probabilities, potentially refining capital allocation and accelerating drug development by enabling faster failure identification.

Key takeaway

For Directors of AI/ML or R&D executives evaluating early-stage drug investments, integrating AI-driven modeling and comprehensive data access is crucial. You should prioritize building robust data lakes and applying advanced analytics like exposure-response modeling to gain earlier confidence in dose decisions and safety profiles. This approach enables faster identification of non-viable candidates, reducing patient exposure to ineffective drugs and optimizing capital allocation in complex oncology programs.

Key insights

AI-driven data integration refines oncology dose decisions, safety, and patient variability, accelerating drug development.

Principles

• Integrate data across development phases for precise decisions.
• Utilize "in silico" modeling for early drug discovery and trials.
• Learn from past failures to inform future drug development.

Method

Employ longitudinal analysis, exposure-response modeling, and covariate evaluation to reduce sub-studies, tailor dosing, and strengthen cross-functional decision-making.

In practice

• Implement data lakes for comprehensive data access and formatting.
• Refine probability of success models for capital allocation.
• Identify drug failures earlier (e.g., Phase 1 vs. Phase 3).

Topics

• AI in Oncology
• Drug Development
• Dose Optimization
• Clinical Trials
• Pharmacokinetics
• Data Integration
• Capital Allocation

Best for: Director of AI/ML, Consultant, Executive

Related on AIssential

• 🔬 Training Transformers to solve 95% failure rate of Cancer Trials — Ron Alfa & Daniel Bear, Noetik — Improving cancer drug success requires advanced AI models for precise patient selection based on multimodal biological data.
• AI is reengineering drug discovery by speeding up testing and scanning petabytes of data for connections between diseases — AI significantly enhances drug discovery by analyzing vast data to predict targets, efficacy, and disease interrelationships, improving success rates.
• From Data Integration to Decision Intelligence: The Next Transformation in Clinical and Commercial… — AI-powered decision intelligence transforms fragmented healthcare data into predictive insights for operational workflows.
• End-user validation of BRIGHT with custom-developed graphical user interface applied to cervical cancer brachytherapy — BRIGHT with its new GUI offers an excellent, user-friendly semi-automated planning tool for cervical cancer brachytherapy.
• Immunotherapy drug target identification using machine learning and patient-derived tumour explant validation — MIDAS, a multimodal GNN, effectively identifies novel immuno-oncology targets by integrating diverse biological data.
• AI identifies early risk patterns for skin cancer — AI models can predict melanoma risk with high accuracy using routine healthcare data, enabling targeted screening.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The AI in Business Podcast.