Accuracy of Ai-Driven Detection of Marginal Bone Loss Around Dental Implants Through Intra-Oral Radiographic Analysis- A Systematic Review

Abstract

Purpose: Over the last decade, AI has witnessed considerable innovation to aid digital dentistry. AI has been evolving as a vital component in providing faster and effective healthcare as an auxiliary tool. Considering the work in implantology, research is going on regarding the application of AI for diagnosis and treatment planning of periodontal bone loss and dental implants. Aim: To assess the accuracy and compare the usage of AI for detection of peri-implant bone loss using radiographs with that of dental clinicians. Methods: The information was acquired by looking for articles in well-known search engines; which were released between January 2014 till December 2024, following the PICO format. Most of the included studies have reported testing precision, sensitivity and specificity and peri-implant bone loss as the outcomes of AI models. QUADAS-2 was the Risk of Bias analysis tool in the included studies. Results: AI system performs about as well as human clinicians in detecting marginal bone loss. The level of agreement between AI and the experts is moderate to substantial with kappa scores of 0.547 for general bone loss and 0.568 for bone loss around implants. Conclusion: The review concluded that the AI model can be utilized and can act as a promising diagnostic tool to measure the radiographic peri-implant bone loss and AI performs comparably or shows moderate agreement with that of the expert implantologists.