AI predicts severity of periodontal disease post-treatment


LAHORE, PAKISTAN: Because treating periodontal disease requires a nuanced approach, methods to predict the most likely treatment requirements can help dental professionals tailor treatment plans to individual patients. In a new study, researchers in Pakistan evaluated whether using a home-grown machine learning model, a form of artificial intelligence (AI), could help predict the severity of the most complex course of periodontal disease after treatment. Although using limited data, they found that artificial intelligence can impact the treatment of periodontal disease, suggesting a greater role for machine learning in patient care.

The application of artificial intelligence in healthcare has received considerable research in recent years, and its potential to improve diagnostic and treatment outcomes has been demonstrated. For example, it has been successfully used to diagnose periodontal disease using panoramic X-rays. However, there are few studies on its use in predicting the course and outcome of periodontal disease.

After generating a comprehensive data set of 1,000 patients, focusing on variables such as age, smoking status and disease severity before and after treatment, the researchers used a linear regression machine learning model for predictive analysis. Artificial patients ranged in age from 20-80 years, with a median age of 45 years. Half were smokers and about half had received periodontal treatment. Periodontal disease severity is scaled from 0 (healthy) to 10 (severe), with observations after treatment showing a general decrease in disease severity.

Correlation analysis found no significant relationship between smoking habits, age, and disease severity before and after treatment. The correlation between age and treatment outcome was weak, with a surprising lack of significant relationship between smoking and post-treatment disease severity, and a positive correlation between pre- and post-treatment disease severity. Notably, patients who had severe disease before treatment also tended to develop severe disease after treatment, suggesting that more severe cases may be more difficult to treat effectively. The model illuminates subtle interactions between demographic and disease variables, but has limited predictive success, in part because the study assumes that the treatments given are generally effective and does not take into account, for example, the skills of the clinician.

Clinically, the findings underscore the need for personalized care that takes into account the nuances of individual patients. From an artificial intelligence perspective, the study highlights challenges in healthcare prediction and highlights the potential and need for continuous improvements in artificial intelligence. The study authors suggest that future research should address limitations of the study, such as artificially generated data, and potentially incorporate advanced artificial intelligence methods as we get closer to AI-driven predictive healthcare. They also suggested that future research could explore other methods of training models, such as gradient boosting or neural networks, for comparison.

The study, titled “The Role of Artificial Intelligence in Periodontology,” was published on May 27, 2023 in pakistan journal of medicine and health sciences.

Label:




Source link

Leave a Reply

Your email address will not be published. Required fields are marked *