This observational study aims to develop and validate a novel, AI-assisted electrochemical sensor platform for saliva-based diagnostics in periodontitis. Periodontitis is a chronic inflammatory disease affecting the gums and supporting tissues of the teeth. Despite its high global prevalence, early diagnosis remains challenging because the disease often progresses silently until irreversible damage has occurred.
Saliva offers a promising, non-invasive diagnostic medium that reflects both oral and systemic health. However, its biological complexity and variability have limited its clinical use. This project addresses these challenges by combining advanced electrochemical sensing with artificial intelligence (AI) and synthetic data generation to improve diagnostic precision and reliability.
The study involves the collection of saliva samples from adult participants with diagnosed periodontitis and from healthy controls. The samples will be analyzed using a modular sensor platform equipped with multiple electrodes that detect electrochemical signals from a wide range of salivary biomarkers. The sensor data will then be processed using machine learning models trained on both real and synthetic data to classify disease states.
The main goals are to:
Evaluate the performance of the electrochemical sensor array for saliva analysis.
Develop and validate AI-based algorithms for detecting and differentiating between healthy and diseased samples.
Generate feasibility data supporting future clinical implementation of saliva-based diagnostics for periodontitis.
This interdisciplinary project combines expertise in clinical dentistry, biomedical engineering, and computer science. It is conducted in collaboration between Linköping University and Malmö University, with patient sampling carried out at an affiliated dental clinic.
The study is expected to result in a working proof-of-concept device that enables real-time, non-invasive detection of periodontitis at the point of care. By enabling earlier diagnosis and more personalized treatment, this technology may transform periodontal care and serve as a foundation for future saliva-based diagnostics targeting other oral and systemic diseases.
None
None
Adult participants (≥18 years) will be recruited from an affiliated dental clinic. The study population consists of two groups: (1) patients clinically diagnosed with periodontitis according to the 2018 AAP/EFP classification, and (2) periodontally healthy adults with no signs of periodontal disease. All participants must be able to provide informed consent and a saliva sample according to study protocol.
Case-control
Prospective
Not listed
Sensitivity and Specificity of the AI-Assisted Electrochemical Saliva Sensor for Detecting Periodontitis
This outcome assesses the diagnostic accuracy of the AI-assisted electrochemical sensor platform by calculating sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and the area under the receiver operating characteristic curve (AUC). The platform analyzes electrochemical signal patterns in saliva samples, and an AI-based classification model predicts whether participants have periodontitis. Clinical periodontal status (periodontitis vs. periodontal health) will be established using a full-mouth clinical examination according to the 2017 World Workshop classification criteria. Diagnostic accuracy metrics from the sensor platform will be compared against this clinical gold standard.
Within-Run Repeatability of Electrochemical Saliva Sensor Signals (Coefficient of Variation)
This outcome will assess the within-run repeatability of electrochemical signal measurements generated by the sensor array when analyzing saliva samples. Repeatability will be quantified using the coefficient of variation (CV%) calculated from multiple consecutive measurements of the same saliva sample under standardized laboratory conditions.
Correlation Between Sensor Outputs and Biochemical Reference Analyses
To evaluate the correlation between electrochemical sensor responses and conventional biochemical or molecular reference methods (e.g., LC-MS, H-NMR) for validation of analytical accuracy.
System Usability Scale (SUS) Score for Clinical Use of the AI-Assisted Saliva Sensor Platform
Feasibility of clinical implementation will be evaluated using the System Usability Scale (SUS), a validated 10-item questionnaire that provides a global usability score from 0 to 100. The SUS score will assess perceived usability, learnability, and integration into dental workflow among clinical staff using the AI-assisted saliva sensor platform. Higher scores indicate better usability and feasibility for routine clinical implementation. The SUS score will be summarized using mean and standard deviation.
Click on the lead sponsor to view all of their trials.
Ostergotland County Council, Sweden
Lead sponsor
Malmö University
Collaborator
Linkoeping University
Collaborator
This trial is not recruiting at the moment. You can still explore other options:
