Introducing a Dental Caries Marking Software and Evaluate Radiologists’ Disagreement in Caries Detection Using this Software

Document Type : Original Article(s)

Authors
H. Baseri 1
R. Rafeh 2
F. Soltani Tafreshi 2
M. Houshyar 3
L. Khojastepour 4
1 School of Dentistry, Shiraz University of Medical Sciences, Iran.
2 Department of Computer Engineering, Faculty of Engineering, Arak University, Arak, Iran
3 Department of Oral and Maxillofacial Radiology, Arak University of Medical Sciences, Arak, Iran
4 Department of Oral and Maxillofacial Radiology, School of Dentistry, Shiraz University of Medical Sciences, Shiraz, Iran
Abstract
Statement of Problem: Bitewing radiograph is the main reference for diagnosis of proximal and recurrent caries. There is no software for gathering and analyzing radiologists’ opinion in the field of caries detection on digital bitewing radiograph (DBR).Objectives: The main aim of this study is to introduce the first windows application that could be used for marking caries on DBR. This software is called Dental Caries Marking Software (DCMS). The second aim is to create the first DBR caries dataset to be used for future software development projects in the field of automatic caries detection; also gathering and documenting the disagreements and critiques regarding DCMS.Materials and Methods: DCMS has been designed and implemented by the researchers of this study. This software is divided into two parts. The first part is DCMS writer that is used for gathering the user’s opinion and The second part is DCMS analyzer that is used for reading and analyzing the user’s opinion file.  Eight radiologists with different experiences used DCMS for marking dental caries on 200 DBR, 50 of which were repeated twice for assessing the accuracy of each  radiologist.Results: A total of 3833 points were marked by 8 users on 150 non repetitive DBR. Only 35 points were marked similarly by 8 users; in other words, 8 users totally agreed with 4% of the caries points. According to 50 repetitive DBR, the maximum accuracy of users was 69% and the minimum was 50%.Conclusions: There is significant debate over the diagnosis of caries on DBR; therefore, for unifying the radiologist’s opinions, the need for intelligent caries detection software is apparent. DCMS is useful software for gathering caries data. Moreover, the use of conventional display monitor has negative impacts on accurate diagnosis of caries on DBR.