The effect of imaging modality on the evaluation of the outcome of endodontic surgery

Abstract

Objectives:

The aim of this study was to assess the effect of the imaging modality on the evaluation of the outcome of modern surgical endodontic treatments, based on a systematic review of the literature.

Methods:

Strict inclusion criteria were adopted in order to identify studies that assessed the outcome of surgical endodontic treatments. Treatment success rates were pooled. The effect of the imaging modality used for the outcome assessment, and the methodological quality of the studies (based on the risk of bias (ROB)), were evaluated.

Results:

Nineteen articles were included. The success rates assessed by periapical (PA) radiography were significantly higher than when cases were evaluated by cone beam computed tomography (CBCT; 90 and 35% respectively). This difference was mainly due to a significant proportion of cases that were assessed by CBCT as uncertain healing (48%) compared to only 4% using PA. The success rates ranged between 86 and 92% in low ROB studies, and between 19–100% in high ROB studies.

Conclusions:

Outcome assessment based on CBCT may lead to significantly lower estimates of rate of success, and higher rates of uncertain healing, thus presenting a dilemma in the decision-making following surgical endodontic treatment. The success rates of studies with lower methodological quality are more variable than for high quality studies.

Introduction

Endodontic surgery is a viable and reliable treatment option for teeth with persistent radiographic and/or clinical signs and symptoms after failed orthograde endodontic treatment, ¹ where the aim is to create optimal conditions for healing through regeneration of the periapical tissues. ² For example, surgical endodontic treatment may be indicated for teeth with apical periodontitis when further non-surgical treatment is impractical or unlikely to improve previous results. ^3,4

Surgical endodontic treatment performed by the modern technique (mSET) is a precise procedure performed with the aid of magnification devices, such as the dental operative microscope (OM). ^5,6 These enable smaller osteotomies to be made while providing easier identification of root apices, and shallower resection angles with no or minimal bevels that preserve root length and cortical bone. ⁵ Furthermore, under high magnification and illumination, the resected root surface easily reveals any microfractures or isthmi and lateral canals. ⁵

Although mSET has been reported to achieve predictable clinical results, ⁷ the variability of study designs and treatment protocols in many recent studies makes it difficult to evaluate the influence of specific factors on the outcomes, or to establish a widely accepted protocol for mSET. Such differences in reported protocols include the variable use of magnification devices, such as the dental OM, ^5,6 endoscopes, ^8,9 and loupes. ^5,10 In addition, a variety of root-end filling materials ^11–13 have been used in an attempt to obtain a better seal and an improved apical tissue response. ¹⁴ These include mineral trioxide aggregate (MTA), ¹³ ethoxybenzoic acid (EBA) cement, ⁶ intermediate restorative material (IRM), ⁸ and bioceramic filling material. Furthermore, there has recently been a significant shift in the methods and criteria used to assess outcomes, from the traditional criteria that were based on periapical (PA) radiographic evaluation, to cone beam computed tomography (CBCT)-based evaluations. ¹⁴

Indeed, the integration of CBCT imaging in daily endodontic practice has now become very popular. Many studies, ^15–19 including several position statements, ²⁰ published in recent years, have advocated for the daily use of CBCT for a variety of endodontic purposes including the assessment of treatment outcomes. ²¹

Previous systematic reviews of the literature published in 2009 and in 2013, ^11,12 described mSET as a predictable method of therapy, with a successful outcome achieved in 89.0% of cases as judged by a follow-up of more than a year. ¹² Notably, the type of root-end filling material and magnification device employed could affect the outcome. ¹²

However, a number of more recent publications, ²² have replaced the traditional criteria to assess the outcome (i.e. based on PAs) with CBCT-based evaluations. Therefore, we predicted that an updated systematic review of the literature would reveal new and important information regarding the outcome of mSET, and identify factors, including the outcome assessment methods and criteria, ^11,12 that may influence the outcome.

The aim of this study was to assess the effect of the imaging modality on the evaluation of the outcome of mSET, based on a systematic review of the literature.

Methods and materials

The protocol for this review was registered in the PROSPERO database with the registration ID CRD42021292100.

Criteria for considering studies for this review

This systematic review includes clinical studies that reported the outcomes of mSET in patients with apical periodontitis in a tooth that had already been treated with a previous endodontic treatment.

The inclusion criteria were the following ^11,12 :

1. Studies that were randomized controlled trials (RCTs), controlled clinical trials (CCTs), or prospective case series (PCS).

2. mSET using magnification, root-end resection with minimal or no bevel, root-end cavity preparation with ultrasonic tips, and root-end filling.

3. Lesions located in the periapical area.

4. Studies with at least a 1-year follow-up.

5. Outcomes that were evaluated based on PA radiography ^23,24 (PA evaluation); OR based on CBCT radiography ^25,26 (CBCT evaluation).

The following exclusion criteria were applied:

1. Previous endodontic surgery (re-surgery cases)

2. Studies involving teeth with combined endodontic–periodontal defects or periodontal disease (periodontal pockets and/or mobility).

3. Cases in which guided tissue regeneration techniques were used

4. Studies that involved treatment for teeth with root fractures or root perforations.

Search methods for the identification of studies

All articles published in dental journals from 1966 to March 2020 were screened. The electronic databases searched included: MEDLINE using the PubMed search engine (http://www.ncbi.nlm.nih.gov/sites/pubmed), Embase (http://www.embase.com), Scopus (http://www.scopus.com), and Cochrane Central Register of Controlled Trials (http://www.cochrane.org). An initial search on MEDLINE used the following keywords: apicoectomy OR apicectomy OR periradicular surgery OR endodontic surgery OR apical surgery OR periapical surgery OR root-end surgery OR root-end resection AND healing with the application of the following limits: ‘‘humans’’ and ‘‘English language.’’ Additional searches were performed using the Embase, Scopus, and Cochrane databases using the same keywords and search limits. Related articles and the reference lists of the literature reviews that were retrieved by the PubMed search engine were manually checked for eligible articles. In addition, a manual search of issues from 1995 to March 2020 was made for the following journals: Australian Endodontic Journal, International Endodontic Journal, International Journal of Oral and Maxillofacial Surgery, Journal of Endodontics, Journal of Oral and Maxillofacial Surgery, Oral Surgery Oral Medicine Oral Pathology and Oral Radiology, and Endodontology. ^11,12

An additional search of the gray literature ²⁷ was made in order to identify additional relevant articles.

Data collection and analysis

Selection of studies

The articles were initially evaluated for relevance based on their titles and abstracts by three independent reviewers (IT, RS, and EK). Potentially eligible studies were subjected to a full-text evaluation, by obtaining the full text of the relevant study and assessing the relevance based on the inclusion and exclusion criteria described previously. If there was disagreement between the reviewers, the case was discussed until an agreement was reached. Articles identified as suitable were subjected to data extraction, assessment of the methodological quality, and data analysis.

Data extraction

The data were extracted by three independent reviewers (ER, RS, and EK). Cases for which there was disagreement were subjected to joint evaluation by the reviewers and discussions continued until an agreement was reached. The authors’ names and the date of publication were recorded for each study. In addition, factors thought to have the potential to influence the evaluation of the outcome were recorded. These included: patient’s age, gender, tooth location (divided into groups: mandibular anterior, maxillary anterior, mandibular posterior, and maxillary posterior teeth), presence of a post, magnification technique (i.e. OM, endoscope, or loupes), root-end filling material, lesion size (small if diameter <5 mm or large if diameter >5 mm), and type of radiographic assessment (PA vs CBCT).

The following methodological parameters were also noted: the method of generating the random sequence and maintaining the blind (for randomized studies), and the blinding of outcome assessments, completeness of outcome data, comparability of the study groups at entry, presence of clear selection criteria, recall rate, sample size, reasons for withdrawals, and length of follow-up period (for all studies). ^11,12

Quality assessment of the methodology

The methodological quality of the selected studies was evaluated independently and in parallel by three reviewers (ER, RS, and EK), ²⁸ who contacted the authors of the studies with requests for clarification or to provide missing information if required.

In order to summarize the validity of studies, they were dichotomized as:

1. Low risk of bias—studies that were evaluated as having a “low risk of bias.” ²²

2. High risk of bias—studies that were evaluated as having “some concerns of bias” or a “high risk of bias.”

In cases in which there was a discrepancy between the scoring of the three reviewers, a consensus was reached by discussion. ^11,12

Data synthesis and analysis

All cases were classified according to the 3-item outcome: success, uncertain, or failure. Outcome evaluation was based on clinical and radiographic findings:

Clinical evaluation

Clinical evaluation considered the presence of clinical signs or symptom such as pain, swelling, tenderness to percussion or palpation, or presence of a sinus tract. Cases having any clinical sign or symptom were classified as a “failure” regardless of the radiographic evaluation.

Radiographic evaluation

Radiographic evaluation of PA radiographs was based the criteria suggested by Rud et al ²³ and/or Molven et al ²⁴ (i.e. “complete healing”, “incomplete healing”, “uncertain healing”, and “unsatisfactory healing”).

Radiographic evaluation of CBCT scans was based on criteria suggested by Schloss et al ²⁵ (i.e. “complete healing”, “limited healing,” and “unsatisfactory healing”) and/or von Arx et al ²⁶ (i.e. according to B-index—“fully healed,” “partially healed,” or “not healed”).

Radiographic outcome evaluation was defined as follows:

1. “Success” by PA evaluation was defined as “complete healing” or “incomplete healing” ^23,24 and as “complete healing” ²⁵ /“fully healed” ²⁶ by CBCT evaluation.

2. “Uncertain” by PA evaluation was defined as “uncertain healing” ^23,24 and as “limited healing” ²⁵ /“partially healed” ²⁶ by CBCT evaluation.

3. “Failure” by PA evaluation was defined as “unsatisfactory healing” ^23,24 and as “unsatisfactory healing” ²⁵ /“not healed” ²⁶ by CBCT evaluation.

Figure 1 presents the classification algorithm for outcome assessment.

Figure 1.

Classification of the outcome.

Statistical evaluation

The statistical analysis was conducted using IBM SPSS Statistics v. 27 (SPSS Inc, Chicago, IL). The Kruskal–Wallis test was used to test for differences between the distributions of more than two groups for a non-normal continuous dependent variable—the proportions of an outcome. For only two groups, the Mann–Whitney U test was used. The significance level was p ≤ 0.05. ^11,12

Results

The results of the electronic databases search and the manual search are illustrated in Figure 2. ²⁹ Nineteen articles met the inclusion criteria, of which 18 ^{8,10,14,30–44} used PA evaluations and 2 ^14,45 used CBCT evaluations. These 19 articles were subjected to data extraction, methodological quality assessment, and data analysis. Nine articles concerned RCTs while the others reported prospective cohort studies. A total number of 2200 teeth in 1834 patients were included in the analysis. Table 1 presents the 19 selected articles.

Figure 2.

The results of the literature search. GTR = Guided tissue regeneration.

Table 1.

Included studies

Study	Study design	No. of patients	No. of teeth	Radiographic outcome evaluation method	Outcome
					Success	Uncertain	Failure
Chong et al, 6 2003	RCT	122	122	PA	98	7	7
Maddalone & Gagliani, 1 2003	PCS	79	120	PA	111	4	5
Lindeboom et al, 3 2005	RCT	90	100	PA	91	8	1
Taschieri et al, 2 2005	PCS	32	46	PA	42	1	3
Taschieri et al, 17 2006	RCT	53	71	PA	66	3	2
Taschieri et al, 15 2007	PCS	21	28	PA	26	1	1
von Arx et al, 16 2007	PCS	106	106	PA	88	11	7
Taschieri et al, 18 2008	RCT	61	100	PA	91	3	6
Kim et al, 8 2008* (Only lesions of endodontic origin)	PCS	N/A	148	PA	141	0	7
Saunders, 7 2008	PCS	276	276	PA	193	17	19
Wälivaara et al, 4 2009** (Only cases that used IRM root-end filling)	RCT	N/A	69	PA	60	5	4
Christiansen et al, 9 2009	RCT	25	25	PA	25	0	0
von Arx et al, 36 2010	PCS	173	173	PA	158	10	5
Wälivaara et al, 5 2011	RCT	153	194	PA	167	0	27
Song & Kim, 11 2012	RCT	192	192	PA	181	4	7
Çalışkan et al, 44 2016	PCS	90	90	PA	30	3	7
Wang & Liang, 12 2017	PCS	59	74	PA	67	3	1
Safi et al, 13 2019*** (Parentheses include evaluation by CBCT)	RCT	120	120	PA & CBCT	112 (62)	2 (40)	6 (18)
von Arx, 45 2019	PCS	47	47	CBCT	9	31	7

CBCT, cone beam computed tompgraphy; IRM, intermediate restorative material; PA, periapical ; PCS, prospective case series; RCT, randomized controlled study.

Study design: RCT; PCS. Radiographic outcome evaluation method: PA; CBCT

Considering patient-related factors, 13 articles provided information about the gender of the participants: 524 males and 761 females. Nine articles reported the mean age of patients, which ranged from 32 to 54.6 years (minimum, maximum age range was 11–83 years). Table 2 presents the patient-related factors.

Table 2.

Patient-related factors

Article	Mean age	Male	Female
Maddalone & Gagliani, 1 2003	40	28	51
Lindeboom et al, 3 2005		33	47
Taschieri et al, 2 2005	40.8	12	20
Taschieri et al, 17 2006	39.1	20	33
Taschieri, et al, 15 2007	42.1	9	12
Saunders, 2 2008	45	100	176
Christiansen et al, 9 2009	54.6	20	24
von Arx et al, 10 2010	50.2	81	92
Song & Kim, 11 2012		69	123
Calıskan et al, 44 2016		52	38
Wang & Liang, 12 2017	32	24	47
Safi et al, 13 2019		51	69
von Arx et al 14 2019	54.4	25	29

There were no significant differences in outcome between treatments performed with lower magnification (loupes) and those performed using higher magnification (endoscope/microscope) when evaluated using PA radiographs (p > 0.05). No study that employed loupes evaluated the outcome using CBCT.

Regarding the retrofilling material, 15 articles compared different materials: 4 articles used IRM, 10 used MTA, 8 used EBA, and 1 used a bioceramic filling material. Nine studies addressed post-restoration as a parameter that could affect the outcome and considered the presence or absence of a post/screw. Table 3 presents the treatment-related factors.

Table 3.

Treatment-related factors

Magnification	Article	IRM	EBA	MTA	Other
Loupes	Maddalone & Gagliani, 1 2003		EBA
	Taschieri et al, 2 2005		EBA
	Lindeboom et al, 3 2005	IRM		MTA
	Walivaara et al, 4 2009* (Only cases that used IRM root-end filling)	IRM
	Wälivaara et al, 5 2011	IRM	EBA
Microscope	Chong et al 6 2003	IRM		MTA
	Saunders et al, 7 2008			MTA
	Kim et al, 8 2008
	Christiansen et al, 9 2009			MTA
	von Arx et al, 10 2010			MTA
	Song & Kim, 2012		EBA	MTA
	Wang & Liang, 12 2017			MTA
	Safi et al, 13 2019			MTA	Endo sequence
	von Arx et al, 45 2019
Endoscope	Taschieri et al, 15 2007		EBA
	von Arx et al, 16 2007		EBA	MTA
Loupes or  Endoscope	Taschieri et al, 17 2006		EBA
Endoscope or  Microscope	Taschieri et al, 18 2008		EBA
	Calıskan et al, 44 2016			MTA

EBA, ethoxybenzoic acid; IRM, intermediate restorative material; MTA, mineral trioxide aggregate.

With respect to tooth-related parameters, the tooth type/location was one of the prognostic factors assessed in the study. Nine articles distinguished between maxillary and mandibular anterior teeth (incisors/canines), and posterior teeth (premolars/molars). The other studies combined all types of teeth when evaluating the outcome. Only one article provided outcome data regarding lesion size. Table 4 presents the tooth-related factors.

Table 4.

Tooth-related factors

Article	Upper	Lower	Anterior	Posterior
Maddalone & Gagliani, 1 2003			62	58
Taschieri et al, 2 2005	27	19	32	14
Taschieri et al, 17 2006	34	37	49	22
Taschieri et al, 15 2007	15	13	22	6
Taschieri et al, 18 2008	57	43	62	38
Wälivaara et al, 5 2011	117	75	54	138
Calıskan et al, 44 2016	70	20	90
Wang & Liang, 12 2017			50	21

The radiographic evaluation modality used (PA or CBCT) significantly affected the outcome assessment (p < 0.05). The average percentages of successful cases as assessed by PA or CBCT evaluation were found to be 90 and 35% respectively, with average values for uncertain cases of 4 and 48%, and 6 and 14% for failed cases as evaluated by PA and CBCT respectively (Figure 3).

Figure 3.

Success proportions. CBCT = Cone beam computed tomography; PA = Periapical.

None of the other patient, treatment- or tooth-related factors except the radiographic evaluation modality used (PA or CBCT), had any significant effect on the outcome assessment.

According to the predefined quality assessment of methodology (Figure 4), four studies (all RCTs) were considered to have a low risk of bias. ^10,33,39,42 15 studies were considered to have a high risk of bias. Five of these studies were RCTs ^{8,14,31,36,43} and 10 were PCS. ^{30,32,34,35,37,38,41,44,45 44}

Figure 4.

Methodological quality assessment.

The estimates of success in low risk of bias studies ranged between 86 and 92%, with values between 19 and 100% for high risk of bias studies.

Discussion

Our systematic review of the literature found that the only factor that significantly affected the evaluation of the outcome of mSET was the imaging modality: the success rates assessed by CBCT were significantly lower when compared to PA radiography (35% vs  90% respectively). Furthermore, it was found that these discrepancies in outcome assessment between these imaging modalities are manifested predominantly in the proportion of uncertain cases (48% for CBCT and only 4% for PA).

The results of the current systematic review also revealed once again that with adequate case selection, and by using the reliable histologically validated criteria of Rud et al, ²³ mSET is a predictable procedure, with 90% of cases evaluated as “success” more than 1 year following the surgery. Furthermore, it was found that with adherence to the strict protocols of the mSET technique, patient, tooth, or treatment-related factors do NOT affect the outcome. These findings are in accordance with previous systematic reviews ^11,12 that found comparable outcome results with little or no effects of these factors on the outcome.

Since the publication of a previous systematic review, ¹² there have been many additional studies of the outcomes of mSET. Furthermore, the past 10 years have witnessed a growing shift towards the use of CBCT evaluation-based outcome assessment, and away from traditional outcome evaluation methods, that were based on PA evaluation and criteria suggested by Rud et al ²³ and/or Molven et al. ²⁴ Thus, published reports may be inconsistent and confusing. ^8,46,47 The current systematic review is the first to compare the outcome assessment results of the common PA based outcome evaluation to more recently introduced CBCT-based evaluation. In order to identify the best available scientific evidence regarding the effect of the imaging modality on the outcome, the present study involved an exhaustive literature search according to the PRISMA (preferred reporting items for systematic reviews and meta-analyses) statement ²⁹ (Figure 2).

In the present review, the methodological quality of the included studies was appraised, and was described as the risk of bias in included studies. A relatively high proportion of the included studies were judged as having a high risk of bias, where the non-randomized nature of the case series study design had the most negative impact on the methodological quality of the included studies (Figure 4). Furthermore, the success rates reported in studies whose methodological quality was considered lower, were more variable than those appearing in high quality studies. As a consequence, care should be taken in interpreting the results and implementing them into clinical practice.

To support clinical decision-making, the present systematic review divided treatment outcomes into three categories: success, uncertain, and failure. Cases categorized as “success” required no further intervention, those categorized as “uncertain” required further follow-up, while cases classified as “failure” required further intervention such as re-surgery, root amputation, or extraction. This approach not only makes clinical sense, but has also been previously validated against a histological gold-standard, ²³ albeit only for PA radiography. ²³

It is important to note that 18 of the included studies used PA radiography with only two employing CBCT radiography for outcome assessment, and that the CBCT-based studies were judged to have lower methodological quality and significant higher outcome variability than those using PA radiography. The differences between PA and CBCT in outcome assessment may have significant implications on the decision-making following endodontic surgery: while most cases may be judged successful by PA evaluation and require no further intervention or follow-up, approximately half (48%) of cases are classified as uncertain when evaluated by CBCT and require further follow-up. Thus, the clinical implication of evaluating the outcome of endodontic surgery by means of CBCT is the need for further follow-up due to uncertainty regarding the outcome of the case.

Our results correspond with two recent studies that provided an important look into the effects of the recent trend towards CBCT evaluation on the outcome assessment and ultimately on the decision-making following endodontic surgery. In the first study, Kruse et al evaluated how additional information gained from CBCT affects the diagnosis and treatment planning that was previously based on PA evaluation and clinical examination. They reported that the use of CBCT for follow-up following endodontic surgery led to more cases diagnosed with apical periodontitis and consequently to the choice of a more aggressive treatment plan.

An incorrect conclusion from these results is that CBCT is more efficient in the diagnosis of the periapical diagnosis following endodontic surgery, and that its use will provide a more appropriate treatment plan. This common error in the understanding of the results of studies comparing the diagnostic efficacy of PA to CBCT evaluation is related to the lack of a suitable gold-standard. Although the gold-standard in endodontic studies is histological assessment, ²¹ this is not always employed, which endangers the validity of the results of studies lacking this control. ²¹

Another study by the same authors, again addressing this clinical dilemma, examined the diagnostic validity of CBCT- and PA-based assessments in endodontic surgical cases that required further surgery. The unique aspect of this study was that histology on periapical tissues removed during the re-surgery served as a valid gold-standard for the radiographic evaluations. The observers assessed the CBCT and the PA radiographs, and re-surgery was suggested for non-healed teeth. All re-operated teeth evaluated histologically were diagnosed as non-healed by CBCT scans, while 58% of them were diagnosed as successfully healed by PA. Biopsies taken during the second surgeries to verify the presence of periapical disease revealed that 42% were free from periapical disease, 16% had mild inflammation, and 42% had moderate to intense inflammation. Importantly, an accurate diagnosis was obtained in 63% of cases evaluated with PA, and in 58% cases with CBCT. As a result the researchers concluded that 42% of the re-operated teeth had no periapical disease at all and thus no gain from re-surgery, and therefore that not all “lesions” seen by CBCT actually represent true periapical disease. Our finding that the success rates assessed by CBCT were significantly lower and uncertain when compared to PA radiography supports these previous reports.

Moreover, these unique studies together with our results provide an interesting view regarding the decision-making process following endodontic surgery: the use of CBCT following endodontic surgery to assess the surgical outcome can lead to overdiagnosis and overtreatment.

While this is the best available evidence as of now, the number of studies that evaluated the outcome of mSET using CBCT was low. Future research on the evaluation of outcome of endodontic surgery is necessary. Well-controlled studies using a larger sample size comparing the evaluation of the outcome using CBCT and PA radiography are needed.

Since we have identified a need for additional research to assess the actual efficacies of CBCT for these purposes, we would recommend that until such results are available, the use of CBCT for decision-making following endodontic surgery should be assessed with caution.

Conclusion

The evaluation of clinical success rates for mSET depends on the radiographic method used for outcome evaluation and, specifically, an outcome assessment based on CBCT may lead to significantly lower estimates of the success rate than an evaluation based on PA radiography. It should be noted that the success rates of CBCT-based studies were of lower methodological quality and were more variable than those in the higher quality studies that used PA radiography.

Our research indicates that the use of CBCT to assess the outcome following endodontic surgery may lead to uncertain conclusions regarding the outcome in about half of the cases.

We therefore conclude that the use of CBCT may adversely affect the outcome evaluation and decision-making following the surgery and presents a clinical dilemma regarding the need for further follow-up/intervention.