Accuracy of Artificial Intelligence Technology in Detecting Number of Root Canals in Human Mandibular First Molars Obturated and Indicated for Retreatment: Diagnostic Accuracy Experimental Study
Accuracy of Artificial Intelligence Technology in Detecting Number of Root Canals in Human Mandibular First Molars Obturated and Indicated for Retreatment: Diagnostic Accuracy Experimental Study
evaluate the accuracy of new AI technology for detecting root canals in mandibular first molars retreatment cases in comparison to dentist clinical access cavity and CBCT imaging.
evaluate the accuracy of new AI technology for detecting root canals in mandibular first molars retreatment cases in comparison to dentist clinical access cavity and CBCT imaging.
CBCT exmanation stage: In this stage, CBCT scanning was done using Soredex Cranex 3D Dental Imaging System, FINLAND, with the following parameters ((XS FOV dimensions 61 x 41 mm (HxD)) (XS FOV High resolution 90 kV / 4 - 12.5 mA / 6.1 s)).
The samples will be randomized using randomization software (Microsoft Office Excel, USA) and will be assigned randomly to 2 endodontists who are unaware of the findings of stage 2. After interpreting and segmenting the CBCT scans in DICOM Format using OnDemand software (USA), the number of canals identified will be recorded on a pre-established information guide.
The samples are coded based on the patient's file number, and the codes were undisclosed so that the CBCT examiners could not identify the samples. All images were interpreted from the axial section in the analysis of the tomographic sections, the number of canals are identified by the corresponding radiolucent orifices, regardless of their location along the root
Clinical Stage: This is a clinical stage where the thirty-five patients, as predetermined by power analysis, will be randomly distributed upon 6 Practitioners using randomization software (Microsoft Office Excel). Practitioners will then proceed in access formation under dental operating microscope, (Leica M320D using magnification 16X, using fully integrated 4K camera).
Access will be done using TR13 diamond stone (Mani, Japan) to remove caries and restorations.
Troughing will be done using ultrasonic tip (NSK E4 and E15D) power 3W.
Irrigation will be done using NAOCL (JK Dental Vision sodium hypochlorite, Egypt) with a concentration of 2.5%.
Gutta percha will be removed from the canal using M-pro rotary files:
At first orifice opener will be used to remove the coronal gutta percha then used the yellow file tapered 4% then confirm the working length by apex locator, after that using taper file 25 to remove the remaining gutta percha.
DG16 endodontic probe (Dentsply Sirona, Germany) will be used to locate canal orifices.
Upon confirmation by clinic PHD supervisors, the number of orifices found will be recorded on a pre-formed information guide, in one visit per patient. Access cavity will be aided by Leica M320D DOM
Artificial intelligence stage: The carrying out of this stage will be solely undertaken by the principal investigator. The CBCT images will be uploaded to convolutional neural network software (CNN) that uses a deep learning algorithm and CBCT segmentation. The software will then record the number of canals it found
The software utilized employs deep convolutional neural networks (CNNs) with a specific U-net inspired structure. The complete CBCT scan is uploaded onto the software, where all collected images are analyzed and each tooth in the 3D scan is precisely located and assessed. The software uses pattern recognition and statistical predictions to segment numerous slices of each tooth and determine the condition or pathosis present. This is achieved by analyzing previously fed photos that were used to train the software
Inclusion Criteria:
Exclusion Criteria: