THE ROLE OF ARTIFICIAL INTELLIGENCE (AI) LEARNING ALGORITHMS IN ENHANCING QUANTITATIVE IMAGE ANALYSIS MEASUREMENTS
Saraaz Khalil, Bradford Teaching Hospitals NHS Foundation Trust
Background:
The study suggests future research should refine protocols and departmental policies to maximize AI’s use in ultrasound.
Methodology:
This study employed a comparative analysis approach, evaluating the performance of AI and ML algorithms used in case studies and trials against traditional quantitative image measurement analysis approaches. The dataset comprised of 100 cases which were all anonymised to comply with ethical standards. The AI / ML algorithms were pre-developed and trained using a subset of images to deep learn what was expected with organ volume and tissue characterisation being extracted using both traditional and ML algorithms. The performance of the methods were assessed based on accuracy, consistency and processing time.
Results:
AI and ML algorithms significantly improved the accuracy of quantitative image analysis measurements, reducing operator-induced variability and decreasing processing time by over 30%. The study suggests future research should refine protocols and departmental policies to maximize AI's use in ultrasound.
Conclusion:
AI and ML algorithms show promise for enhancing quantitative measurements in ultrasound. However, further research is needed to assess the impact and risks associated with AI in ultrasound practice to understand its broader applicability.