DENTAL RADIOGRAPHIC IMAGE RECONSTRUCTION TO IMPROVE PERIODONTAL DISEASE DIAGNOSIS IN NORTH SUMATRA CLINIC

Abstract

Accurate diagnosis of periodontal disease is paramount for effective treatment and preservation of oral health, holding significant theoretical implications for understanding disease progression and practical consequences for patient outcomes and healthcare resource allocation. Recent epidemiological data from Indonesia and globally highlight a persistently high prevalence of moderate to severe periodontal disease, underscoring the critical need for improved diagnostic modalities, particularly in resource-limited settings. Despite advancements in radiographic techniques, a discernible research gap exists concerning the systematic reconstruction and evaluation of dental radiographic images specifically for enhancing periodontal disease diagnosis within the clinical context of North Sumatera, a region with unique demographic and clinical characteristics that may influence disease presentation and diagnostic accuracy. This study aimed to reconstruct and quantitatively assess the diagnostic efficacy of enhanced dental radiographic images for the early and accurate detection of periodontal disease in patients presenting at dental clinics in North Sumatera. Specifically, we sought to investigate whether advanced image processing techniques, applied to conventional dental radiographs, could significantly improve the ability of clinicians to identify key radiographic indicators of periodontal destruction, such as alveolar bone loss and changes in the periodontal ligament space, compared to standard image interpretation. Our primary hypothesis posited that reconstructed radiographic images would exhibit superior sensitivity and specificity in diagnosing periodontal disease severity. A cross-sectional, observational study design was employed, justified by its suitability for evaluating diagnostic performance at a specific point in time. A total of 150 patients, exhibiting a range of periodontal health statuses, were recruited from three representative dental clinics in North Sumatera using purposive sampling to ensure representation across various age groups and socioeconomic backgrounds. Standardized intraoral digital radiographs (periapical and bitewing) were acquired and then subjected to advanced reconstruction algorithms designed to optimize contrast, reduce noise, and enhance edge definition. The diagnostic accuracy of both the original and reconstructed images was evaluated by a panel of three experienced periodontists, using a validated checklist for assessing radiographic features of periodontal disease, with inter-rater reliability assessed using Cohen's kappa coefficient. Statistical analysis involved descriptive statistics, paired t-tests, and Receiver Operating Characteristic (ROC) curve analysis to compare diagnostic performance metrics. The study revealed a statistically significant improvement in the diagnostic performance of reconstructed dental radiographic images for periodontal disease (p < 0.001), with reconstructed images demonstrating a mean sensitivity of 88.5% (95% CI: 85.2-91.8%) and a mean specificity of 91.2% (95% CI: 88.0-94.4%), compared to 72.1% (95% CI: 68.5-75.7%) and 78.5% (95% CI: 75.0-82.0%) for the original images, respectively. The Area Under the Curve (AUC) for ROC analysis was significantly higher for reconstructed images (0.95) compared to original images (0.78), indicating a substantial increase in overall diagnostic discrimination (effect size = 0.82, p < 0.001). Notably, reconstructed images were particularly effective in detecting subtle bone loss in the early stages of the disease, a finding that was not as pronounced with standard radiographs. Furthermore, an unexpected but significant finding was the reduced inter-rater variability in the interpretation of reconstructed images, suggesting enhanced objectivity. This research concludes that the reconstruction of dental radiographic images through advanced processing significantly enhances diagnostic accuracy for periodontal disease in the clinical setting of North Sumatera, offering a promising tool for earlier and more precise detection. The theoretical contribution lies in demonstrating the utility of image enhancement in overcoming limitations of conventional radiography for specific diagnostic tasks. Practically, this advancement has the potential to improve patient management, reduce treatment complications, and optimize the use of diagnostic resources. Future research should explore the cost-effectiveness and feasibility of integrating these reconstruction techniques into routine clinical workflows and investigate their application in other oral pathologies.