DYNAMIC BUDGET IMPACT OF ARTIFICIAL INTELLIGENCE (AI)-ASSISTED IMAGING INTERPRETATION FOR BRAIN METASTASES IN ADVANCED LUNG CANCER
Author(s)
Hsiao Ling Chen, BS, MS1, Ming-Yu Hong, MS2, Chen-Han Chueh, PhD3, Jia-Sheng Hong, MS.4, Chien-Yu Tseng, PharmD5, Yu-Te Wu, PhD4, Wan-Yuo Guo, PhD6, Shuu-Jiun Wang, PhD6, Yi-Wen Tsai, PhD7.
1Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taipei, Taiwan, 2Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taiwan, New Taipei City, Taiwan, 3University of California San Diego, San Diego, CA, USA, 4Institute of Biophotonics, National Yang Ming Chiao Tung University, Taipei, Taiwan, 5Department of Pharmaceutical Outcomes and Policy, College of Pharmacy, University of Florida, Gainesville, FL, USA, 6College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan, Taipei, Taiwan, 7National Yang Ming Chiao Tung University, Taipei, Taiwan.
1Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taipei, Taiwan, 2Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taiwan, New Taipei City, Taiwan, 3University of California San Diego, San Diego, CA, USA, 4Institute of Biophotonics, National Yang Ming Chiao Tung University, Taipei, Taiwan, 5Department of Pharmaceutical Outcomes and Policy, College of Pharmacy, University of Florida, Gainesville, FL, USA, 6College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan, Taipei, Taiwan, 7National Yang Ming Chiao Tung University, Taipei, Taiwan.
Presentation Documents
OBJECTIVES: Demand for AI-based software as a medical device (SaMD) is increasing as health systems seek to improve diagnostic accuracy and clinician efficiency. This study evaluates the budgetary and workforce impact of adopting the AI-assisted system DeepBT® for brain metastasis detection in advanced non-small cell lung cancer from the Taiwan National Health Insurance perspective.
METHODS: A Markov model was developed to estimate annual patient volumes and brain magnetic resonance imaging (MRI) utilization among patients with stage III and IV NSCLC. Health states were defined by cancer stage, brain metastasis status, treatment, and death, with annual transitions between states. Both incident and prevalent advanced NSCLC populations in Taiwan from 2027 to 2031 were modeled using projections from 2023 data. State-specific annual brain MRI utilization rates were derived from the National Health Insurance Research Database to estimate SaMD use and radiologist workload under standard and AI-assisted interpretation. The model assumed a per-scan SaMD cost of NTD 2,500 and improved radiologist efficiency, with reporting time reduced from 4.8 to 3.6 minutes per MRI. SaMD adoption was evaluated over a five-year budget impact horizon under two uptake scenarios. First- and fifth-year penetration rates were 9.4% and 53.8% for the conservative scenario, and 20.8% and 83.0% for the optimistic scenario, respectively.
RESULTS: Over five years, annual net budget impact increased from national Taiwan dollar (NTD) 2.79 to 16.94 million under the conservative scenario, with cumulative radiologist interpretation time savings of 1,248-8,014 minutes. Under the optimistic scenario, annual net budget impact ranged from NTD 6.17 to 26.14 million, with cumulative time savings of 2,764-11,739 minutes.
CONCLUSIONS: Adoption of SaMD for MRI interpretation was associated with increased budget impact and reduced radiologist interpretation time, with potential amplification under wider adoption across diverse clinical settings and health system.
METHODS: A Markov model was developed to estimate annual patient volumes and brain magnetic resonance imaging (MRI) utilization among patients with stage III and IV NSCLC. Health states were defined by cancer stage, brain metastasis status, treatment, and death, with annual transitions between states. Both incident and prevalent advanced NSCLC populations in Taiwan from 2027 to 2031 were modeled using projections from 2023 data. State-specific annual brain MRI utilization rates were derived from the National Health Insurance Research Database to estimate SaMD use and radiologist workload under standard and AI-assisted interpretation. The model assumed a per-scan SaMD cost of NTD 2,500 and improved radiologist efficiency, with reporting time reduced from 4.8 to 3.6 minutes per MRI. SaMD adoption was evaluated over a five-year budget impact horizon under two uptake scenarios. First- and fifth-year penetration rates were 9.4% and 53.8% for the conservative scenario, and 20.8% and 83.0% for the optimistic scenario, respectively.
RESULTS: Over five years, annual net budget impact increased from national Taiwan dollar (NTD) 2.79 to 16.94 million under the conservative scenario, with cumulative radiologist interpretation time savings of 1,248-8,014 minutes. Under the optimistic scenario, annual net budget impact ranged from NTD 6.17 to 26.14 million, with cumulative time savings of 2,764-11,739 minutes.
CONCLUSIONS: Adoption of SaMD for MRI interpretation was associated with increased budget impact and reduced radiologist interpretation time, with potential amplification under wider adoption across diverse clinical settings and health system.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
EE146
Topic
Economic Evaluation
Topic Subcategory
Budget Impact Analysis
Disease
SDC: Oncology