RAPID AI-ENABLED PICO SCOPING TO SUPPORT EUROPEAN JOINT CLINICAL ASSESSMENTS: A 4TH-LINE METASTATIC COLORECTAL CANCER CASE STUDY
Author(s)
Turgay Ayer, PhD1, Sumeyye Samur, PhD1, Ismail F. Yildirim, MSc1, Mine Tekman, PhD1, Jag Chhatwal, PhD2;
1Value Analytics Labs, Boston, MA, USA, 2Massachusetts General Hospital/Harvard Medical School, Boston, MA, USA
1Value Analytics Labs, Boston, MA, USA, 2Massachusetts General Hospital/Harvard Medical School, Boston, MA, USA
OBJECTIVES: To evaluate whether an AI-enabled, multi-agent architecture can systematically extract, harmonize, and consolidate PICO requirements across European HTA bodies, and to assess its utility for efficient, transparent PICO scoping in preparation for Joint Clinical Assessment, using fourth-line metastatic colorectal cancer as a case study.
METHODS: An agentic AI architecture using over 20 specialized sub-agents specialized for HEOR, ValueGen.AI, was applied to a focused case study in 4th-line metastatic colorectal cancer (mCRC). The platform automatically retrieved and synthesized publicly available HTA guidance, appraisal documents, and pivotal trial evidence from major European HTA bodies, including HAS (France), G-BA/IQWiG (Germany), AIFA (Italy), AEMPS (Spain), and others. Explicit and implicit PICO elements were extracted, including population definitions, interventions, comparators, outcomes, and subgroup specifications, and organized into a consolidated JCA-aligned PICO framework.
RESULTS: Across HTA bodies, the 4th-line mCRC population was consistently defined as adults with metastatic disease who progressed after fluoropyrimidine-, oxaliplatin-, and irinotecan-based chemotherapy, typically with ECOG performance status 0-1. Interventions commonly assessed included regorafenib, trifluridine/tipiracil (with or without bevacizumab), and fruquintinib. Best supportive care was universally accepted as a comparator, while acceptance of active comparators varied by country. Overall survival was consistently specified as the primary outcome, with progression-free survival, health-related quality of life (EORTC QLQ-C30; EQ-5D-5L), and adverse events included as secondary outcomes. Several HTA bodies additionally specified subgroup analyses based on prior treatment exposure and molecular characteristics.
CONCLUSIONS: ValueGen.AI enabled rapid consolidation of national PICO specifications into a single, structured, JCA-aligned framework, supporting cross-country scoping while preserving flexibility for national decision-making and limiting additive PICO expansion. The scoping process, which traditionally requires weeks of expert review, was completed within hours with full traceability to source HTA documents.
METHODS: An agentic AI architecture using over 20 specialized sub-agents specialized for HEOR, ValueGen.AI, was applied to a focused case study in 4th-line metastatic colorectal cancer (mCRC). The platform automatically retrieved and synthesized publicly available HTA guidance, appraisal documents, and pivotal trial evidence from major European HTA bodies, including HAS (France), G-BA/IQWiG (Germany), AIFA (Italy), AEMPS (Spain), and others. Explicit and implicit PICO elements were extracted, including population definitions, interventions, comparators, outcomes, and subgroup specifications, and organized into a consolidated JCA-aligned PICO framework.
RESULTS: Across HTA bodies, the 4th-line mCRC population was consistently defined as adults with metastatic disease who progressed after fluoropyrimidine-, oxaliplatin-, and irinotecan-based chemotherapy, typically with ECOG performance status 0-1. Interventions commonly assessed included regorafenib, trifluridine/tipiracil (with or without bevacizumab), and fruquintinib. Best supportive care was universally accepted as a comparator, while acceptance of active comparators varied by country. Overall survival was consistently specified as the primary outcome, with progression-free survival, health-related quality of life (EORTC QLQ-C30; EQ-5D-5L), and adverse events included as secondary outcomes. Several HTA bodies additionally specified subgroup analyses based on prior treatment exposure and molecular characteristics.
CONCLUSIONS: ValueGen.AI enabled rapid consolidation of national PICO specifications into a single, structured, JCA-aligned framework, supporting cross-country scoping while preserving flexibility for national decision-making and limiting additive PICO expansion. The scoping process, which traditionally requires weeks of expert review, was completed within hours with full traceability to source HTA documents.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
HTA47
Topic
Health Technology Assessment
Topic Subcategory
Decision & Deliberative Processes
Disease
SDC: Oncology