Deep Learning for the Radiographic Detection of Apical Lesions


      • Apical lesions (ALs) were detected on panoramic radiographs using neural networks.
      • A 7-layer deep network was trained on a limited amount of image segments.
      • The network showed high specificity and moderate sensitivity.
      • The overall discriminatory ability of the network was satisfactory.
      • Applying neural networks may assist dentists' in reliably and accurately detecting ALs.



      We applied deep convolutional neural networks (CNNs) to detect apical lesions (ALs) on panoramic dental radiographs.


      Based on a synthesized data set of 2001 tooth segments from panoramic radiographs, a custom-made 7-layer deep neural network, parameterized by a total number of 4,299,651 weights, was trained and validated via 10 times repeated group shuffling. Hyperparameters were tuned using a grid search. Our reference test was the majority vote of 6 independent examiners who detected ALs on an ordinal scale (0, no AL; 1, widened periodontal ligament, uncertain AL; 2, clearly detectable lesion, certain AL). Metrics were the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and positive/negative predictive values. Subgroup analysis for tooth types was performed, and different margins of agreement of the reference test were applied (base case: 2; sensitivity analysis: 6).


      The mean (standard deviation) tooth level prevalence of both uncertain and certain ALs was 0.16 (0.03) in the base case. The AUC of the CNN was 0.85 (0.04). Sensitivity and specificity were 0.65 (0.12) and 0.87 (0.04,) respectively. The resulting positive predictive value was 0.49 (0.10), and the negative predictive value was 0.93 (0.03). In molars, sensitivity was significantly higher than in other tooth types, whereas specificity was lower. When only certain ALs were assessed, the AUC was 0.89 (0.04). Increasing the margin of agreement to 6 significantly increased the AUC to 0.95 (0.02), mainly because the sensitivity increased to 0.74 (0.19).


      A moderately deep CNN trained on a limited amount of image data showed satisfying discriminatory ability to detect ALs on panoramic radiographs.

      Key Words

      To read this article in full you will need to make a payment


        • Segura-Egea J.J.
        • Martin-Gonzalez J.
        • Castellanos-Cosano L.
        Endodontic medicine: connections between apical periodontitis and systemic diseases.
        Int Endod J. 2015; 48: 933-951
        • Huumonen S.
        • Suominen A.L.
        • Vehkalahti M.M.
        Prevalence of apical periodontitis in root filled teeth: findings from a nationwide survey in Finland.
        Int Endod J. 2017; 50: 229-236
        • Connert T.
        • Truckenmuller M.
        • ElAyouti A.
        • et al.
        Changes in periapical status, quality of root fillings and estimated endodontic treatment need in a similar urban German population 20 years later.
        Clin Oral Investig. 2019; 23: 1373-1382
        • Kanagasingam S.
        • Hussaini H.M.
        • Soo I.
        • et al.
        Accuracy of single and parallax film and digital periapical radiographs in diagnosing apical periodontitis-a cadaver study.
        Int Endod J. 2017; 50: 427-436
        • Leonardi Dutra K.
        • Haas L.
        • Porporatti A.L.
        • et al.
        Diagnostic accuracy of cone-beam computed tomography and conventional radiography on apical periodontitis: a systematic review and meta-analysis.
        J Endod. 2016; 42: 356-364
        • Grandreviewresearch
        Dental X-Ray Market Analysis By Product (Analog, Digital), By Type (Intraoral - Bitewing, Periapical, Occlusal; Extraoral - Panoramic, Cone Beam Computed Tomography), By Application (Medical, Cosmetic, Forensic), And Segment Forecasts, 2018 - 2024.
        2016 (Available at: Accessed May 18, 2019)
      1. KZBV. KZBV Jahrbuch 2017. Available at: Accessed May 29, 2019.

        • Nardi C.
        • Calistri L.
        • Grazzini G.
        • et al.
        Is panoramic radiography an accurate imaging technique for the detection of endodontically treated asymptomatic apical periodontitis?.
        J Endod. 2018; 44: 1500-1508
        • Nardi C.
        • Calistri L.
        • Pradella S.
        • et al.
        Accuracy of orthopantomography for apical periodontitis without endodontic treatment.
        J Endod. 2017; 43: 1640-1646
        • Ahlqwist M.
        • Halling A.
        • Hollender L.
        Rotational panoramic radiography in epidemiological studies of dental health. Comparison between panoramic radiographs and intraoral full mouth surveys.
        Swed Dent J. 1986; 10: 73-84
        • Parker J.M.
        • Mol A.
        • Rivera E.M.
        • Tawil P.Z.
        Cone-beam computed tomography uses in clinical endodontics: observer variability in detecting periapical lesions.
        J Endod. 2017; 43: 184-187
        • Lin P.L.
        • Huang P.Y.
        • Huang P.W.
        Automatic methods for alveolar bone loss degree measurement in periodontitis periapical radiographs.
        Comp Methods Programs Biomed. 2017; 148: 1-11
        • Becker A.S.
        • Marcon M.
        • Ghafoor S.
        • et al.
        Deep learning in mammography: diagnostic accuracy of a multipurpose image analysis software in the detection of breast cancer.
        Invest Radiol. 2017; 52: 434-440
        • Esteva A.
        • Kuprel B.
        • Novoa R.A.
        • et al.
        Dermatologist-level classification of skin cancer with deep neural networks.
        Nature. 2017; 542: 115-118
        • Gulshan V.
        • Peng L.
        • Coram M.
        • et al.
        Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs.
        JAMA. 2016; 316: 2402-2410
        • Lee J.H.
        • Kim D.H.
        • Jeong S.N.
        • Choi S.H.
        Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm.
        J Dent. 2018; 77: 106-111
        • Lee J.H.
        • Kim D.H.
        • Jeong S.N.
        • Choi S.H.
        Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm.
        J Periodontal Implant Sci. 2018; 48: 114-123
        • Goodfellow I.
        • Bengio Y.
        • Courville A.
        Deep Learning.
        MIT Press, Cambridge, MA2016
        • Mazurowski M.
        • Buda M.
        • Saha A.
        • Bashir M.
        Deep learning in radiology: an overview of the concepts and a survey of the state of the art.
        J Magn Reson Imaging. 2019; 49: 939-954.
        • Litjens G.
        • Kooi T.
        • Bejnordi B.E.
        • et al.
        A survey on deep learning in medical image analysis.
        Med Image Anal. 2017; 42: 60-88
        • Fleiss J.
        Measuring nominal scale agreement among many raters.
        Psychol Bull. 1971; 76: 378-382
        • Landis J.
        • Koch G.
        The measurement of observer agreement for categorical data.
        Biometrics. 1977; 33: 159-174
        • LeCun Y.
        • Bengio Y.
        • Hinton G.
        Deep learning.
        Nature. 2015; 521: 436-444
      2. Ioffe S, Szegedy C. Batch normalization: accelerating deep network training by reducing internal covariate shift. arXiv 2015: arXiv:1502.03167v3 [cs.LG], 2015.

        • Claesen M.
        • De Moor B.
        Hyperparameter search in machine learning.
        (Available at:) (Accessed May 18, 2019)
        • Hussain Z.
        • Gimenez F.
        • Yi D.
        • Rubin D.
        Differential data augmentation techniques for medical imaging classification tasks.
        AMIA Annu Symp Proc. 2018; 2017 (Annual Symposium proceedings): 979-984
        • Buda M.
        • Maki A.
        • Mazurowski M.
        A systematic study of the class imbalance problem in convolutional neural network.
        Neural Netw. 2018; 106: 249-259
        • Bossuyt P.M.
        • Reitsma J.B.
        • Bruns D.E.
        • et al.
        STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies.
        BMJ. 2015; 351: h5527
        • Walsh T.
        Fuzzy gold standards: Approaches to handling an imperfect reference standard.
        J Dent. 2018; 74: S47-S49
        • Masood M.
        • Masood Y.
        • Newton J.T.
        The clustering effects of surfaces within the tooth and teeth within individuals.
        J Dent Res. 2015; 94: 281-288

      Supplemental Appendix S1 References

        • Fleiss J.L.
        Measuring nominal scale agreement among many raters.
        Psychol Bull. 1971; 76: 378-382
        • Gwet K.L.
        Computing inter-rater reliability and its variance in the presence of high agreement.
        Br J Math Stat Psychol. 2008; 61: 29-48
        • Scott W.A.
        Reliability of content analysis: the case of nominal scale coding.
        Public Opin Q. 1955; 19: 321
        • Sokolova M.
        • Lapalme G.
        A systematic analysis of performance measures for classification tasks.
        Inf Process Manag. 2009; 45: 427-437