Can diagnostic changes caused by cone beam computed tomography alter the clinical decision in impacted lower third molar treatment plan?

Objectives:

This study observed whether changes in diagnosis caused by analysis of three-dimensional images can lead to alterations in the treatment plans of impacted lower third molars (ILTMs).

Methods:

Sets of panoramic (PAN) – cone beam computed tomography (CBCT) of 218 patients were assessed for ILTM classification, contact with mandibular canal, contact and resorption of the lower second molar (LSM), intraoperative planning and post-operative expectations.

Results:

Percentage agreement and McNemar test compared PAN vs CBCT assessments. Logistic regression analyzed the dependency of change in surgical planning considering the changes in diagnostic features; descriptive statistics was used to observe the expectation of post-operative complications and paresthesia. Differences were found between PAN vs CBCT for classification of impaction and positioning, LSM relationship, choice for crown and root sectioning and expectation of post-operative complications (all with p < 0.001). Logistic regression indicated that the change in diagnosis caused by CBCT examination did not change the clinical decision to extract ILTM but altered the planning of intraoperative steps such as osteotomy, crown sectioning and relaxing incision. The expectation of post-operative complications decreased when professionals planned the ILTM removal using tri-dimensional images.

Conclusions:

We concluded that changes in the diagnosis after CBCT examination can lead to alterations in the treatment plan of impacted lower third molar.

Introduction

According to current guidelines, bi-dimensional (2D) examinations, such as panoramic radiography (PAN), should precede tri-dimensional (3D) exams, such as cone beam computed tomography (CBCT), in the process of dental diagnosis and treatment planning.^1–5 Considering impacted lower third molars (ILTMs), studies have shown that the use of 3D images can change the diagnostic classifications related to ILTM position, contact with the mandibular canal (MC), and resorption of adjacent lower second molar (LSM).^6–8 However, there is scarce information about if these diagnostic features can lead to alterations in clinical decision-making regarding treatment plan.

Neurosensory disturbance related to the inferior alveolar nerve (IAN) poses as the major concern in ILTM removal. Given that PAN is the first-choice imaging exam in the treatment of ILTM, seven radiographic signs have been suggested as potential indicators of the need for complementary 3D examinations.^9–11 These signs are root apex darkening, root deflection, root narrowing, veiled or bifid apexes, interruption of radiopaque canal lines, canal deviation, and mandibular canal narrowing.¹² These signs refer only to potential risk of nerve injury by observing the relationship between the ILTMs and the MC containing the IAN. However, other relevant diagnostic information should be taken into account to a complete treatment plan, such as the relationship of the ILTM with the adjacent lower second molar,⁶ the presence of retromolar canals or foramen,¹³ and lingual cortical perforations.¹⁴ Despite its advantages in the diagnosis and treatment plan, the indication of CBCT is still controversial due to the radiation dose, high cost, and lack of support in the literature.

The contribution of an image exam into the diagnosis and treatment process of a patient was discussed by Fryback and Thornbury in 1991, who proposed a six-level hierarchical model for studies evaluating the efficacy imaging exams: Level 1 concerns technical quality of the images; Level 2 addresses diagnostic accuracy, sensitivity, and specificity associated with interpretation of the images. Next, Level 3 focuses on whether the information produces change in the referring physician’s diagnostic thinking. Such a change is a logical prerequisite for Level 4 efficacy, which concerns the effect on the patient management plan. Level 5 efficacy studies measure (or compute) the effect of the information on patient outcomes. Finally, at Level 6, analyses examine societal costs and benefits of a diagnostic imaging technology.¹⁵

Therefore, the objective of this level 3 study was to evaluate the relationship between imaging diagnostic features and the treatment plan for ILTM, comparing the use of 2D and 3D images. The research hypothesis is that the diagnostic features observed on 2D or 3D images regarding ILTM position, relationship with MC, and adjacent LSM resorption do not alter clinical decisions on the surgical planning and post-operative expectation.

Methods and materials

Study design

This was a clinical, observational, transversal study that aimed to address the following questions:

1. Are there differences between PAN and CBCT regarding the diagnosis of ILTM imaging features, such as: classification of ILTM regarding impaction and position, contact of the ILTM with the MC and the LSM, and resorption of the LSM?

2. Would eventual differences between those imaging modalities alter the treatment plan or post-operative expectation?

The study design focuses on whether the image information produces change in the diagnostic thinking of specialists in Oral and Maxillofacial Surgery (study Level 3 from Fryback and Thornbury).¹⁵ In this study, the criterion was based on clinical decision-making by dental image.

The variable was the professionals’ responses to parameters related to ILTM diagnosis, treatment plan and post-operative expectation. The dependence of change in the diagnosis and treatment plan was the primary outcome.

Exam selection

Sets of CBCT and PAN obtained with a maximum interval of 3 months between exams, from patients presenting at least one ILTM, were retrieved from institutional diagnostic image databank. Sets presenting bone lesions in the region (from the lower second molar to the mandibular ramus) or poor image quality were excluded.

PANs were obtained on VATECH PaxX-400C (Vatech Global, Korea). CBCTs were acquired by i-CAT FLX (Imaging Sciences International, Hatfield, PA, USA) according to the following parameters: 0.25 mm voxel size, 120 kVp, and 36.12 mAs.

Diagnostic analysis, treatment planning, and prediction of post-operative complications

Two specialists in Oral and Maxillofacial Surgery were invited to evaluate the CBCT and PAN and, based on those, provide diagnostic analysis, treatment plan, and postoperative expectation, according to the parameters described in Figure 1. Both professionals had more than 10 years of experience in the field, having performed more than 100 ILTMs removals and with experience with the analysis of CBCT exams. The κ value for their evaluations were >.90 for all parameters analyzed.

Evaluation of the images was performed using a high-resolution monitor in a dimly lit room. The professionals initially received 218 anonymized and randomized PANs to be assessed using Radioimp software (Radio Memory, Belo Horizonte, Brazil). After 3 months, the same professionals received 218 CBCTs of the same cases, in randomized order and assessed them using Ondemand3D software (Cybermed Inc., Seoul, South Korea). Figure 2 illustrates one PAN/CBCT set.

Data analysis

Data were analyzed in SPSS software (v. 22, Chicago, IL). The percentual agreement was calculated for the variables studied (ILTM classification, surgical planning, and post-operative expectation) and compared by McNemar test. Logistic regression was used to assess the dependency of radiographic signs of ILMT and MC contact and those assessment on CBCT. Also, logistic regression analyzed the dependency of change in surgical planning considering the changes in diagnostic decision of ILMT impaction and positioning classifications, MC contact, and relation to the LSM. Finally, a descriptive analysis of the change in post-operative expectations was performed for when the imaging modalities were in agreement or disagreement concerning ILMT impaction and positioning classifications, IAN contact and LSM relation. A p-value below 0.05 was considered statistically significant.

Results

Considering 2D vs 3D images, differences were observed for classifications regarding impaction (Pell & Gregory’s)¹⁶ and positioning (Winter’s) (p < 0.001), relationship of ILTM with adjacent LSM (p < 0.001). Treatment planning involving crown sectioning and root sectioning (p = 0.006 and p < 0.001, respectively), and for expectation of post-operative complications (p < 0.001) also showed significant differences between imaging modalities (Table 1).

In none of the cases, the option to not remove the ILTM was selected by the professionals. Exodontia was planned for almost all cases, with exception of three cases in PAN and one case in CBCT where coronectomy was the treatment of choice in three cases with PAN and one case with CBCT, and those did not coincide. Prevalence of treatment plans and surgical procedures indicated with PAN and CBCT is further described in Table 2.

According to the logistic regression (Table 3), proximity to the MC observed in CBCT assessment showed dependency to radiographic sign of bifid apex (p < 0.05). The interruption of the MC cortical observed in CBCT was dependent on the interruption of the MC cortical and bifid apex observed on PAN (p < 0.05). Thinning of the MC on CBCT showed significant positive dependence with bifid apex and deviation of MC observed on PAN.

Table 4 shows that change on the diagnosis of ILTM impaction and positioning, MC contact, as well the relation to the LSM do not significantly alter the surgical plan regarding extraction or coronectomy. Changes on the Winter´s classification after CBCT assessment significantly influenced the option for relaxing incision. Additionally, there was a significant chance of changing the surgical plan considering osteotomy when CBCT altered the diagnosis of LSM contact area; the same was found for crown sectioning.

In Table 5, the expected post-operative complications according to ILTM impaction and position classifications in those cases that analyses on PAN and CBCT were in agreement or disagreement. When PAN and CBCT were not in agreement for both classifications, there was a major change in the expectation of post-operative complications to be absent, and in only one case, the CBCT evaluation changed the expectation of complications to be present. Consequently, it is possible to observe that when PAN and CBCT agree in ILTM classification the expectation of postoperative complications is greater for PAN and decreases in CBCT analyses. For paresthesia, when Winter´s classification differed between PAN and CBCT, 10% of the responses changed to a positive paresthesia expectation, while 8.1% changed to a negative expectation. Overall, numbers for paresthesia are similar in both situations.

Discussion

This study showed that the diagnostic classifications regarding ILTM position, relationship with the adjacent LSM and the MC may be different when a 2D or 3D exam is used, and that those differences potentially alter treatment plan and expectation of post-operative complications.

Although Pell and Gregory’s and Winter’s classifications are widely used, those are based on 2D imaging exams.^8,16,17 In our study, the data suggest that there is a significant difference between the perception of the ILTM position on 2D and 3D images. Such difference influenced the treatment plan proposed by the examiners, as seen from the results of the regression analysis, which showed that there is a higher chance of indication of a relaxing incision when 3D images altered the Winter’s classification from PAN. Indeed, Brasil et al (2019) showed that panoramic shows a tendency to underestimate the space for accommodation of the third molar compared with CBCT, justifying the changing in the intraoperative steps.⁸ About the expectation of post-operative complications crossed with the classification of ILTM position, it’s possible to note that in 18% of the cases the professionals changed their expectation from “post-operative complications” to “no post-operative complications” after CBCT examination; however, in only one case the professional changed the prediction from “no ” to “yes” after analyzing the CBCT, indicating that more often the analysis of 3D images will influence the professional to decrease the expectation of complications like pain, trismus and swelling. For paresthesia, the expected presence or absence showed similar values after CBCT examination.

One of the most considered aspects in ILTM treatment is its relationship with the MC due to the possibility of IAN injury.^9–11 According to Winstanley et al in 2018, the cortical interruption and root darkening seen on PAN, are the most predictive findings of ILTM contact with IAN.¹⁸ From the sample of this study, considering the seven radiographic signs analyzed on PAN, bifid apex showed a significant positive dependence with CBCT signs of contact with the MC, interruption of MC cortical and thinning of the MC. Importantly, it was observed a significant positive chance of true MC cortical interruption on 3D when it was suggested by 2D, indicating that the contact of ILTM and MC can be similarly observed on 2D and 3D.

The effect of a change in the diagnosis of contact of the ILTM with the MC from PAN to CBCT in the selection of required surgical procedures was assessed, and significant positive correlation was observed between MC deviation and osteotomy. Interestingly, the option of not removing the ILTM was not chosen in none of the cases. Coronectomy was chosen more from the analysis of PAN (3) than CBCT (1). Differently, in 2019 Matzen et al¹⁹ showed an increased number of coronectomies when a 3D exam was performed and reported a higher number of cases of paresthesia when only 2D exams were used. Regarding the contact of ILTM with the MC, although the selection of required surgical procedures did not alter significantly from PAN to CBCT analysis, the expectation of post-operative complications from 2D analysis tended to decrease after 3D. The inverse did not occur.¹⁹ The change in the expectation of paresthesia, which is the condition usually associated with most concerns of the professional, ranged from 7.7 to 9.7%, similar when analysis from 2D and 3D were in agreement or not. We hypothesize, thus, that maybe the 3D display of the surgical area can play a role in even minor aspects in the indication of surgical procedures, such as the required pressure to be applied on surgical movements. Indeed, in 2009 Ghaemina et al showed that CBCT was not more accurate at predicting IAN exposure during third molar removal, however, the authors affirm that did elucidate the 3D relationship of the third molar root to the mandibular canal dictates the surgical approach how to remove the third molar, so the IAN will not be subjected to pressure from movements of the roots or the careless use of burrs and elevators; what in our point of view, explain the decrease in the expectation of pain when the professionals visualized the ILTM on CBCT.²⁰ Besides, it is worth to mention that the professionals who worked on the treatment plan in this study were experienced Oral and Maxilofacial surgeons and therefore, maybe after analyzing the CBCT, based on their clinical experience, predicted less post-operative complications by deciding for more relaxing incisions, osteotomy or root sectioning, which would potentially decrease intraoperative time and, consequently, paresthesia and post-operative complications in a clinical scenario.

The presence of resorption of LSM caused by ILTM ranges from 20 to 40% of the cases in studies using 3D images.^21–23 In this study, the use of 2D or 3D images affected the detection of contact between those teeth and external resorption on the LSM, with a low agreement between the analysis from 2D and 3D images (28% of the cases, for resorption). In 2014, Oenning et al⁶ corroborated such results by detecting prevalence of LSM resorption up to four times higher on 3D examinations. The changes in the classifications related to the relationship between ILTM and LSM seems to have influenced the choice of required intraoperative procedures, such as osteotomy and crown sectioning, depending on the presence of contact. Herman et al in 2019 assessed impacted upper third molars regarding the presence of resorption on the adjacent second molar. They found that the analysis of CBCT often changed the treatment plan in comparison to PAN. In this study, regarding the post-operative, for LSM contact area and resorption, the expectation of post-operative complications also appears to decrease when the case is observed on 3D image.²⁴

One limitation of this study was the fact that the examiners that analyzed the images and provided the treatment plans are highly experienced professionals. Since this study deals with clinical decisions, the results of this study must be interpreted carefully in a broader scenario where different professionals may be at different stages in their clinical career; additionally, it is important to mention that the reference standard protocol for planning the treatment of ILTM involves clinical and image exams. Future studies, comparing senior and junior professionals in the context of ILTM treatment from the analysis of different imaging modalities could address this matter.

It is concluded that changes in diagnostic classifications from PAN to CBCT analysis did not change the clinical decision of recommending removal of the ILTM but altered the indication of intraoperative procedures such as osteotomy, crown sectioning and relaxing incision. The expectation of post-operative complications decreased when professionals planned the ILTM removal using 3D images.

Clinical relevance

CBCT is an image exam that can change the diagnosis in the impacted third molar, especially for detection of second molar resorption and contact with inferior alveolar nerve. In this study, we showed that these changes in the diagnosis can cause minor alterations in the treatment plan

REFERENCES