Clinical and Operational Validation of an Artificial Intelligence-Based Platform for Population-Level Multi-Cancer Screening Using Routine Laboratory Tests.

• Grant number: 25/10717-7
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: November 01, 2025
• End date: July 31, 2027
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Marco Aurelio Kohara
• Grantee: Marco Aurelio Kohara
• Principal investigators: Daniella Castro Araujo
• Associated research grant: 23/14898-0 - Ambispective clinical study to develop artificial intelligence models using routine blood biomarkers to support the early identification of breast cancer., AP.PIPE
• Associated scholarship(s): 25/22717-1 - Clinical and Operational Validation of an AI-Based Platform for Population-Wide Multi-Cancer Screening Using Routine Lab Tests: A Clinical Study of the Colorectal Cancer Case Group., BP.TT; 25/27008-9 - Clinical and operational validation of an Artificial Intelligence-based platform for population screening of multiple cancers: Refinement and External Validation of the AI model via prospective testing., BP.TT; 25/24195-2 - Model Refinement & External Validation (Retrospective Study) and Prospective Study Cross-sectional = Colorectal Cancer (CRC), BP.TT; 25/24275-6 - Model Refinement & External Validation (Retrospective Study) and Prospective Cross-Sectional Study - Lung Cancer, BP.TT

Abstract

Huna develops artificial intelligence (AI) algorithms for cancer screening, utilizing routine laboratory tests. Following the successful validation of breast cancer in projects funded by FAPESP - PIPE Phases 1 and 2 - this project aims to clinically and operationally validate two new predictive algorithms focused on screening for lung and colorectal cancers. The goal is to transform simple and widely available tests-such as complete blood count (CBC), C-reactive protein (CRP), ferritin, D-dimer, among others-into population risk stratification tools with the potential to impact care pathways and enable early detection of these tumors.The proposed methodology is based on an ambispective clinical study conducted at the Hospital de Amor, involving 1,200 participants divided into case groups (patients with confirmed diagnoses) and control groups (asymptomatic individuals with confirmation of no disease through gold-standard exams such as chest CT or colonoscopy). Blood biomarker panels will be collected based on scientific literature and prior internal research by Huna. The lab data will be used as input for supervised machine-learning algorithms that have been previously trained and refined by the team through retrospective analysis.At the end of the 21-month project, the goal is to demonstrate the accuracy and clinical applicability of the models, expanding the social impact of Huna's technology. The solution represents a cost-effective alternative for cancer screening in both public and private healthcare systems-especially in settings with underreporting and limited access to imaging exams. The successful execution of this project will position Huna as the first multi-cancer screening platform based exclusively on routine blood tests, strengthening its market presence in Brazil and setting the stage for international expansion. (AU)