Radiographic Identification of Separated Instruments Retained in the Apical Third of Root Canal–filled Teeth

Detection of the separated root canal instrument on panoramic radiograph: a comparison of LSTM and CNN deep learning methods

OBJECTIVES

A separated endodontic instrument is one of the challenging complications of root canal treatment. The purpose of this study was to compare two deep learning methods that are convolutional neural network (CNN) and long short-term memory (LSTM) to detect the separated endodontic instruments on dental radiographs.

METHODS

Panoramic radiographs from the hospital archive were retrospectively evaluated by two dentists. A total of 915 teeth, of which 417 are labeled as "separated instrument" and 498 are labeled as "healthy root canal treatment" were included. A total of six deep learning models, four of which are some varieties of CNN (Raw-CNN, Augmented-CNN, Gabor filtered-CNN, Gabor-filtered-augmented-CNN) and two of which are some varieties of LSTM model (Raw-LSTM, Augmented-LSTM) were trained based on several feature extraction methods with an applied or not applied an augmentation procedure. The diagnostic performances of the models were compared in terms of accuracy, sensitivity, specificity, positive- and negative-predictive value using 10-fold cross-validation. A McNemar's tests was employed to figure out if there is a statistically significant difference between performances of the models. Receiver operating characteristic (ROC) curves were developed to assess the quality of the performance of the most promising model (Gabor filtered-CNN model) by exploring different cut-off levels in the last decision layer of the model.

RESULTS

The Gabor filtered-CNN model showed the highest accuracy (84.37 ± 2.79), sensitivity (81.26 ± 4.79), positive-predictive value (84.16 ± 3.35) and negative-predictive value (84.62 ± 4.56 with a confidence interval of 80.6 ± 0.0076. McNemar's tests yielded that the performance of the Gabor filtered-CNN model significantly different from both LSTM models (p < 0.01).

CONCLUSIONS

Both CNN and LSTM models were achieved a high predictive performance on to distinguish separated endodontic instruments in radiographs. The Gabor filtered-CNN model without data augmentation gave the best predictive performance.

Influence of different viewing conditions on the detection of fractured endodontic instruments using periapical radiographs at 3 projection angles

OBJECTIVES

The objective of this study was to assess the influence of ambient lighting levels and horizontal viewing angle (HVA) on the radiographic detection of fractured endodontic instruments in root canals in extracted human mandibular molars.

STUDY DESIGN

Thirty-two root canals were selected. Endodontic instrument fracture was induced in 24 canals, and all canals were endodontially treated. Periapical radiographs were obtained at orthoradial, mesioradial, and distoradial projection angles. Five oral radiologists examined the radiographs for detection of fractured endodontic instruments under 3 ambient lighting levels (high, 450 lux; medium, 28 lux; and low, 2.5 lux) and 3 HVAs (90°, 67.5°, and 45°). Area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, and specificity were calculated and compared by 2-way analysis of variance using Tukey post hoc tests (α = .05).

RESULTS

AUROC data indicated that ambient lighting (P = .739) and HVA (P = .238) did not significantly influence the detection of fractured endodontic instruments. Overall mean diagnostic values were 0.745, 0.57, 0.46, and 0.90 for AUROC, accuracy, sensitivity, and specificity, respectively.

CONCLUSIONS

The viewing conditions assessed in this study did not influence the radiographic detection of fractured endodontic instruments in root canals.

Detection of Fractured Endodontic Instruments in Root Canals: Comparison between Different Digital Radiography Systems and Cone-beam Computed Tomography

INTRODUCTION

This study compared the detection of fractured instruments in root canals with and without filling by periapical radiographs from 3 digital systems and cone-beam computed tomographic (CBCT) images with different resolutions.

METHODS

Thirty-one human molars (80 canals) were used. Root canals were divided into the following groups: the control group, without fillings; the fracture group, without fillings and with fractured files; the fill group, filled; and the fill/fracture group, filled and with fractured files. Digital radiographs in ortho-, mesio-, and distoradial directions were performed in 2 semidirect systems (VistaScan [Dürr Dental, Beitigheim-Bissinger, Germany] and Express [Instrumentarium Imaging, Tuusula, Finland]) and a direct system (SnapShot [Instrumentarium Imaging]). CBCT images were acquired with 0.085-mm and 0.2-mm voxel sizes. All images were assessed and reassessed by 4 observers for the presence or absence of fractured files on a 5-point scale. The sensitivity, specificity, and accuracy were calculated.

RESULTS

In the absence of filling, accuracy values were high, and there were no statistical differences among the radiographic techniques, different digital systems, or the different CBCT voxels sizes. In the presence of filling, the accuracy of periapical radiographs was significantly higher than CBCT images. In general, SnapShot showed higher accuracy than VistaScan and Express.

CONCLUSIONS

Periapical radiographs in 1 incidence were accurate for the detection of fractured endodontic instruments inside the root canal in the absence or presence of filling, suggesting that this technique should be the first choice as well as the direct digital radiographic system. In the presence of filling, the decision to perform a CBCT examination must take into consideration its low accuracy.