Skip to main content
AAE21 Main banner
Q&A

(PR26) Validation of Pre-operative Cyst Screening Tool Using Cone-beam Computed Tomography

    Presenter(s)

  • Michael A. Eskander, DDS PhD

    Graduate Endodontics Resident
    University of Texas at San Antonio
    UT Health San Antonio
    San Antonio, Texas, United States

    • Co-Author(s)

  • Vanessa Chrepa, D.D.S., M.S.

    Assistant Professor
    University of Texas at San Antonio
    san antonio, Texas, United States

    • Submitter(s)

  • Michael A. Eskander, DDS PhD

    Graduate Endodontics Resident
    University of Texas at San Antonio
    UT Health San Antonio
    San Antonio, Texas, United States

Abstract:

Introduction: Periapical granulomas and cysts are the most common lesions associated with periradicular pathology. True cysts are less likely to heal with non-surgical root canal therapy compared to endodontic microsurgery, but differentiation can only be accomplished with histopathological examination after surgery. We previously developed and published a pre-operative cyst screening tool using cone beam computed tomography (CBCT) volumetric measurements combined with radiographic assessments of periapical lesions.
 

Objective: To assess the validity of a previously established cyst versus granuloma screening tool.
 

Methods: A new set of preoperative CBCT scans from patients who underwent apicoectomies was evaluated by a blinded observer for root displacement and lesion volume using the Amira software. A decision tree model was applied to predict whether the lesion was a granuloma or cyst based on volume and root displacement. Briefly, if volume was >247mm3 or if volume was < 247mm3 and root displacement was present, the predicted diagnosis was a cyst.  Receiver operating characteristic curves were constructed in order to determine the predictive validity of our cyst screening tool.
 

Results: Preliminary model prediction data analysis of 23 cysts and 77 granulomas (n=100) showed 65% sensitivity, 96.1% specificity, 89% accuracy, 83.3% positive predictive value (PPV), and 90.2% negative predictive value (NPV). Notably, of the 18 lesions that the tool predicted as cysts, 15 were truly cysts per histopathological diagnosis.
 


Conclusions: The preliminary validation of our cyst screening tool showed excellent specificity, accuracy and PPV and has the potential of a useful clinical tool for endodontics.