Peer-Reviewed Research
Published research papers describing AI model performance and real-world workflow impact of GenzAI Labs' AI-powered medical imaging platforms.
Research vs clinical performance — The metrics reported on this page are extracted from peer-reviewed research publications and describe AI model performance on retrospective research datasets. They are not clinical performance claims, regulatory certifications, or guarantees of accuracy in clinical use. Genz AI Labs products are clinical decision-support tools intended for use by qualified healthcare professionals and are not a substitute for clinical judgement, diagnosis, or treatment.
Integration of AI Stroke Analysis into Radiology Workflow: Real-World Experience with Stroke Insightz for MRI from Indian Perspective
DOI: 10.36106/ijsr | ISSN: 2277-8179
A PRISMA-guided systematic review of the key workflow limitations hindering timely stroke diagnosis in India, synthesising global and Indian evidence on AI-enabled stroke imaging. The paper presents an embedded real-world case illustration of the Stroke Insightz MRI pipeline deployed at a tertiary Indian centre, demonstrating scan-to-AI completion times of 6–10 minutes, immediate AI-to-notification, and a >60% reduction in MRI post-processing workload.
Experiences of Stroke Insightz AI for Stroke Analysis into MR Imaging Workflow: A Global Perspective
Paper ID: BJMHS450537
A PRISMA-2020 systematic synthesis of AI-powered MRI stroke analysis platforms evaluating diagnostic performance, workflow integration and clinical impact. The review confirms that MRI-based AI achieves pooled sensitivity and specificity of ~93% for acute ischaemic lesion detection, and positions Stroke Insightz as part of a new generation of clinical decision support tools standardising MRI stroke analysis and optimising turnaround time.
Validation of an AI-Based Tool for Detecting Radiographic Findings Suggestive of Tuberculosis: A Pilot Study
Paper ID: BJMHS450536
A pilot validation study of CXR Insightz (now TB Insightz) on a retrospective dataset of 170 chest X-rays (88 TB-positive, 82 normal). The AI tool achieved an overall accuracy of 93.53%, sensitivity of 100%, specificity of 86.59%, F1-score of 94.12%, ROC-AUC of 0.9329, and strong agreement with radiologist findings (Cohen's Kappa = 0.8698) — supporting its utility as a screening aid in large-scale TB workflows.
Request Full Papers
Contact us for full-text copies of our published research or to collaborate on clinical validation studies.