Accepted Article

DR. ADHAM A AZIM (Orcid ID : 0000-0002-4603-5650)

PROF. LUCILA PIASECKI (Orcid ID : 0000-0003-0546-2928)

Article type : Original Scientific Article

The Buffalo study: outcome and associated predictors in endodontic microsurgery: a cohort study

Azim AA1, Albanyan H1,2, Azim KA3, Piasecki L1

1Division of Endodontics, University at Buffalo, Buffalo, NY, USA; 2College of Dentistry, King Saud bin
Abdulaziz University for Health Sciences, Riyadh, Saudi Arabia; 3Department of Psychology, University at
Buffalo, Buffalo, NY USA.

Keywords: Cone-beam Computed Tomography, Endodontic Microsurgery, Outcome predictors

Running Head: The Buffalo Study

Correspondence:
Dr. Adham A. Azim
Clinical Associate Professor – Division Head & Director, Advanced specialty Program in Endodontics
Address: School of Dental Medicine, University at Buffalo, 240 Squire Hall, Buffalo, NY 14214, USA
Email: [email protected]
Phone: +1 (716) - 829-3602

This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1111/iej.13419
This article is protected by copyright. All rights reserved
 Abstract
Accepted Article
Aim To 1) evaluate and compare the success rate of endodontic microsurgery (EMS) using periapical
radiographs (PAs) and cone-beam computed tomography (CBCT) scans; 2) identify prognostic factors
affecting the outcome; and 3) correlate the effect of guided tissue regeneration (GTR) on the pattern of apical
bone remodeling.
Methodology Eighty-two patients (101 teeth) who received EMS were included and followed-up using clinical
and radiographic examination (PAs and CBCT scans). Two calibrated endodontists evaluated the radiographic
healing (favorable or unfavorable) by assessing PAs and CBCT. The success (favorable radiographic outcome
with no clinical symptoms) and survival rates (tooth retention without clinical symptoms) were calculated, and
the cause of failure (diseased or fractured) was identified. Pretreatment (Age, sex, tooth type, position,
sequence of treatment, quality of root canal before surgery, presence/absence of through-and-through lesion,
presence/absence apico-marginal defect) and treatment (presence/absence of errors during surgery, type of
error (major or minor), retro-preparation depth, presence/absence of an isthmus, retro-filling material used,
presence/absence bone graft material and/or resorbable membrane) factors were recorded. Data were
statistically analyzed to determine the inter-observer, intra-observer, and inter-radiographic agreements.
Univariate, bivariate, and logistic regression analysis were used to determine prognostic factors affecting the
outcome, and the effect of GTR on the pattern of apical bone remodeling. The significance level was set at 5%.
Results Sixty-eight patients (83 teeth) presented for outcome evaluation (Recall rate: 84%). The survival rate
was 93%. The success rate was 88% using PA and 86% using CBCT when vertical root fracture (VRF) cases
were included. 94% using PAs, and 91% using CBCT when VRF cases were excluded. The intra- and inter-
observer agreements were substantial using CBCT, slight to a fair agreement using PA (P < 0.001), and slight
to moderate for inter-radiographic agreement. The presence of a major procedural error during the surgery was
the only negative predictor for the outcome of EMS (P = 0.013). GTR did not affect the success rate or the type
of healing using PA but affected the type of healing on CBCT (complete vs incomplete healing) and the pattern
of cortical plate remodeling (P < 0.001)
Conclusion The success and survival rate of endodontic microsurgery was very high, and the occurrence of a
major procedural error during surgery was the only factor affecting the outcome. GTR did not improve the
outcome but it affected the quality of apical bone remodeling following EMS.

This article is protected by copyright. All rights reserved

 Introduction
Accepted Article
Endodontic surgery is a procedure that seeks to treat patients who have clinical and radiographic signs and
symptoms of endodontic disease. Traditionally, it has been recommended as the last treatment resort when non-
surgical options are not feasible or are not expected to improve the outcome (Torabinejad & White 2016). Over
the past few decades, there have been advancements in the surgical endodontic armamentarium, which in part
improved the success of endodontic surgery, currently known as endodontic microsurgery (EMS) (Setzer et al.
2012, Tsesis et al. 2013). Previous studies often focused on comparing traditional endodontic surgical
techniques to EMS to highlight the necessity of using newer armamentarium (Setzer et al. 2012, Tsesis et al.
2013). Others attempted to correlate the outcome of EMS to patients’ age, tooth type, lesion size, quality of
endodontic treatment prior to surgery, and retro-filling material (Kim et al. 2008; Kruse et al. 2016, Safi et al.
2019, von Arx et al. 2019a). However, the effect of factors such as the sequence of treatment, retro-preparation
depth, procedural errors during surgery, and the use of guided tissue regeneration (GTR) on the success of
EMS are either limited or have not been well investigated (von Arx et al. 2019b).

Cone-beam computed tomography (CBCT) has been widely used in endodontics for diagnosis and treatment
planning (Mota de Almeida et al. 2014, Patel et al. 2019), and has been recommended as the appropriate image
modality for pre-surgical assessment and evaluation by the European Society of Endodontology (ESE) as well
as the American Association of Endodontics (AAE) (AAE/AAOMR 2016, ESE 2019). It has also been adopted
for outcome evaluation if used during pre-operative assessment (AAE/AAOMR 2016). CBCT has been used
for outcome evaluation with higher sensitivity and specificity in detecting post-treatment disease compared to
periapical radiographs (PAs) (Kanagasingham et al. 2017, Curtis et al. 2018). Only a few studies have reported
the outcome of endodontic surgery using CBCT and were often qualitative in nature (Schloss et al. 2017,
Curtis et al. 2018, Safi et al. 2019). Chen et al. (2015) proposed a quantitative method that allows three
different areas on the scan (cortical plate, resection plane, and apical area) to be scored when evaluating the
outcome of endodontic surgery. von Arx et al. (2016) later adopted the same scoring system and created an
index to evaluate the outcome of endodontic surgery called the “B index”. However, none of the previous
criteria included the interpretation for clinical cases with a significant amount of bone loss (through-and-
through and apico-marginal defects) or assessed the outcome following placement of radiopaque bone grafting
material that can mimic healthy bone structure radiographically.

Guided tissue regeneration (GTR) techniques have been proposed as a potential adjunct to enhance hard tissue
formation in surgical defects during EMS procedures (Taschieri et al. 2008a, 2008b). The techniques for GTR

This article is protected by copyright. All rights reserved

 vary between using grafting materials, resorbable collagen membranes, or a combination of both. The effect of
Accepted Article
GTR on the type of apical bone remodeling evaluated using CBCT has not been studied. Therefore, the aim of
this study was to 1) evaluate and compare the outcome of EMS using PA and CBCT; 2) identify prognostic
factors that can affect the success rate of EMS; and 3) evaluate the effect of GTR on apical bone remodeling.

Materials and Methods

Study design & patient population
This study followed the STROBE checklist and the institutional review board at the University at Buffalo, NY
(IRB no. STUDY00003107). Consecutive patients referred to the university clinics for endodontic surgery
between September 2015 and June 2018 were retrospectively identified. The follow-up visits were then
scheduled in a prospective manner after study approval. The patients were treatment planned for endodontic
surgery or re-surgery after the prognosis and the alternative treatment options (retreatment or extraction and
implant placement) had been explained and surgical intervention was deemed the most appropriate treatment
option for the patient by the endodontic resident and faculty.

Inclusion & exclusion criteria

All patients were between the ages of 18 - 85 years and had signs of apical periodontitis during the
clinical/radiographic examination. All the treated teeth had adequate coronal restoration, no mobility or
periodontal disease, and no confirmed signs of vertical root fracture (VRF) before treatment. If signs of VRF
were suspected (not confirmed), alternative treatment plans of extraction or root amputation were provided to
the patient and executed if VRF was identified during the surgery. These teeth were then further excluded.
Teeth with apical-marginal defects or through-and-through defects were included in the study and planned for
GTR. If the patient reported that the tooth/teeth had been extracted after the surgical procedure for restorative
reasons (e.g. to make a denture), these teeth were also excluded from the study.

Treatment planning & surgical procedure

Before each surgery, a thorough clinical and radiographic examination, using PAs, and limited field CBCT
scan using either a 4 cm x 4 cm Veraviewepocs 3D (J Morita, Osaka, Japan) or 5 cm x 5 cm Orthophos (Sirona
Group, Bensheim, Germany) with a voxel size of 125μm, were performed. All surgical procedures were
performed by endodontic residents and faculty using a surgical operating microscope (Opmi Pico, Zeiss,
Oberkochen, Germany). The EMS protocol involved a full-thickness flap reflection, osteotomy to expose the
root apices, and removal of the peri-radicular pathosis. Root-end resection was then performed using a high-

This article is protected by copyright. All rights reserved

 speed bur with minimal or no bevel followed by root-end preparation using ultrasonic instruments. In cases
Accepted Article
with inadequate root canal treatment or absence of a root canal filling at the time of the surgery, longer retro-
preparation surgical tips (6 or 9 mm) (Acteon, Satelec, Aquitaine, France) were used to prepare the root canal
space, followed by canal irrigation from the root-end using 2% chlorhexidine (CHX) before the placement of
the root-end filling material. During this step, the surgical suction was placed carefully close to the root-end to
avoid the irrigant from contacting the surgical site. The root-end filling materials used were either Pro Root
MTA (Dentsply Sirona, Charlotte, NC, USA) or BioCeramic (BC) root repair material (RRM) fast set putty
(Brasseler, Savanna, GA, USA), with or without BC sealer (Brasseler). In cases where bone graft alone, or in
combination with a barrier, was indicated/recommended by the supervising faculty, a freeze-dried allograft
(Alloss; ACE surgical supplies, Brockton, MA, USA) bone material and a resorbable collagen membrane
(conFORM; ACE surgical supplies) were used. The flap margins were then re-approximated and sutured, and
post-operative PAs were taken at the end of the surgical procedures. In a few cases, immediate post-operative
CBCT scans were taken for treatment assessment purposes; however, it was not part of the post-surgical
protocol. Post-operative instructions were given to the patient, and sutures were removed within two weeks. All
the surgical procedures were photographed and/or video-recorded, and stored in the department’s private
server.

Follow-up & outcome assessment

Follow-up visits were scheduled between one and three years after the surgery. All patients were called in and
requested to attend for a follow-up appointment. If the patient could not attend, the reason for not attending was
recorded. The follow-up visit included a clinical examination (percussion, palpation, and periodontal
examination), clinical pictures, PAs, and a CBCT scan. The outcome was assessed using both clinical and
radiographic examination (PA & CBCT). For the radiographic evaluation, each case was assessed
independently by two calibrated endodontists. The criteria set by Molven et al. (1987) were used to evaluate
the PAs (complete, incomplete, uncertain, or unsatisfactory healing). For 3D evaluation, modified criteria from
Chen et al. (2015) was used. Briefly, each reviewer was able to scroll through the entire scan to evaluate all
planes (coronal, axial, and sagittal) to evaluate the three indices: P (peri-radicular area), R (resection plane),
and C (cortical plate). Each plane received a score of 0, 1, or 2, that were summed to give a final score per
tooth from 0 to 6. For teeth with multiple roots, the root with the lowest score was used to evaluate the
outcome. According to the radiographic score, the healing was divided into four radiographic healing
categories, similar to that of Molven et al. (1987), complete (score 6), incomplete (score 5 or 4), uncertain
(score 3), or unsatisfactory healing (score 2, 1, or 0). The four healing categories identified from the PA and

This article is protected by copyright. All rights reserved

 CBCT images were further dichotomized into two radiographic outcome categories: favorable (complete and
Accepted Article
incomplete healing) or unfavorable (uncertain and unsatisfactory healing) outcome. The case was considered
successful when the tooth was clinically asymptomatic and had a favorable radiographic outcome. Any tooth
presenting with clinical symptoms and/or unfavorable radiographic outcome related to the treated tooth was
considered a failure. Teeth with clinical symptoms were radiographically evaluated to correlate the clinical
findings to the CBCT scores. They were also categorized based on the clinical presentation into either failure
due to presence of clinical or radiographic signs of a disease or due to VRF. The survival rate was defined as
the tooth being clinically asymptomatic and functioning in the patient’s mouth without the need for further
intervention, regardless of their radiographic outcome (favorable or unfavorable). (Table 1 and Fig. 1)

During the radiographic evaluation, the two endodontists independently reviewed all the PAs first, followed by
the CBCT DICOM files to prevent any bias. Each evaluator reviewed the cases and recorded their findings to
calculate the inter-observer agreement. After two months, the PAs and scans were re-evaluated to calculate the
intra-observer agreement. For the CBCT evaluation, DICOM files were exported to a 3D viewer software
(Horos, Horos project), and pre-operative, and follow-up CBCT images were reviewed simultaneously on two
27-inch iMac screens (Apple, Cupertino, CA, USA). In case of disagreement on any of the 2D or 3D readings,
both endodontists discussed the case until consensus was achieved.

Radiographical assessment of the hard tissue changes at the apical area

In this part of the study, only cases with a successful outcome on the CBCT scan were included (i.e.
radiographic score of 6, 5, or 4 and no clinical symptoms) to avoid any confounding factors that may affect the
healing pattern. The type of healing achieved [(complete (score 6) vs. incomplete (score 5 or 4)] as well as the
score achieved in each of the three indices (P, R, and C) (score 2 vs. score 1 and 0) were correlated to the
presence/absence of GTR (bone grafting placement with/without membrane).

Data collection
The following host and treatment variables were recorded from each procedure to evaluate their impact on the
treatment outcome:
1- Patient’s age
2- Patient’s sex
3- Tooth type (anterior, premolar or molar)
4- Tooth position (maxillary or mandibular)

This article is protected by copyright. All rights reserved

 5- Treatment sequence:
Accepted Article a) Retrograde root canal as a primary treatment (i.e. no history of a non-surgical (NS) endodontic
treatment/presence of root canal filling), according to Jonasson et al. (2017)
b) Retrograde root canal after a NS initial root canal treatment
c) Retrograde root canal after a NS retreatment
d) Re-surgery
6- Quality of endodontic treatment before surgery according to Tronstad et al. (2000)
a) Adequate: All visible root canals are obturated within 2 mm of the radiographic apex with no voids
b) Inadequate: Root filling is more than 2 mm short of the radiographic apex, overextension, at least
one canal is missing, or poor filling density.
7- Errors during surgery by the clinician
a) No error
b) Minor procedural error: An error that did not result in root perforation (retro-preparation being off-
angle without root perforation or enlarged retro-preparation on the mesio-distal or bucco-lingual plane)
(Fig. 2a-d) and all portals of exit are sealed.
c) Major procedural error: Excessive damage to the root (perforation during retro-preparation),
improper sealing of portals of exit (missing a canal during surgery or in-adequate retro-filling
placement) (Fig 2e-h), or excessive damage to the supporting structure (iatrogenic perforation of the
lingual/palatal plate resulting in a through-and-through osteotomy or complete removal of the
labial/buccal plate resulting in an apico-marginal defect).
8- Isthmus (present/absent)
9- Retro-preparation depth (< 3 mm, 3 mm, > 3 mm)
10- Retro-filling material (MTA, RRM putty, or BC sealer + RRM putty)
11- Presence/absence of alveolar bone over the buccal/labial root surface (apico-marginal defect) (Dietrich et
al. 2002).
12- Through–and–through defect (presence/absence)
13- GTR (Bone graft with/without membrane placement) (yes/no)

Statistical analysis
Cohen Kappa analysis was performed to calculate the inter-radiographic agreement between the 2D and the 3D
analysis using the 4 radiographic healing categories and the two radiographic outcome categories mentioned
above. Intra- and inter-observer agreements were calculated for the CBCT scores (0-6), and for the healing and

This article is protected by copyright. All rights reserved

 outcome categories on the PA and the CBCT radiographs. To examine the reliability and internal consistency
Accepted Article
of the scoring system (P + C + R), Cronbach’s alpha was calculated followed by principal components analysis
to investigate the factor structure of the scoring model with eigenvalue >1.00 for each extracted component.
Items were considered meaningful if loading onto the component was >.50. Given the variation in the follow-
up time between the treated teeth, an ordinal and binary logistic regression was performed on teeth that
survived asymptomatically to examine if the follow-up period was a predictive factor for the total score (0-6),
radiographic healing category (complete, incomplete, uncertain, or unsatisfactory), and the treatment outcome
(favorable or unfavorable).

The correlation between the prognostic factors and the success rate was analyzed using Pearsonճ chi-square

(2) tests at an -value of .05, with and without including cases with VRF. Initial analysis was followed by
posthoc testing using Bonferroni with adjusted p-value to protect from type I error. Binomial logistic regression
analysis was then performed to test the predictability of significant variables from the bivariate analysis on the

treatment outcome with an -value set at 0.05. Pearsonճ chi-square (2) analysis was also used with Yates

correction to establish a relationship between the utilization of GTR and a) the type of favorable radiographic
outcome using 2D and 3D imaging (complete (score 6) or incomplete (score 5 or 4) healing; b) the quality of
healing (complete vs. incomplete or no healing) within each index in cases with a favorable outcome on the 3D
scan. All the tests were performed at an -value of 0.05 using SPSS software (version 25, IBM, NY).

Results
Survival rate
A total of 82 patients (101 teeth) met the inclusion criteria. Sixty-eight patients (41 females and 27 males) with
a mean age of 54.7yrs and a median age of 57.6yrs presented for the follow-up examination (mean follow up
time: 18.8 months; median: 15 months) and were included to evaluate the outcome (Recall rate: 84%). 85 teeth
were clinically evaluated and six patients (6 teeth) presented with clinical signs of failure, two of which
presented before the 1-year recall time with clinical symptoms. The overall survival rate without further
intervention was 93%.

Qualitative analysis of cases with clinical signs of failure

Out of the six patients (6 teeth) that had clinical signs of failure, one patient had signs of procedure-related
failure due to a missed canal during the surgery, and the patient opted to extract the tooth. The other five

This article is protected by copyright. All rights reserved

 patients were diagnosed with VRF. Four patients presented with clinical signs and symptoms of VRF,
Accepted Article
including deep probing (> 6 mm). Two of these teeth (a mandibular anterior and a maxillary premolar) were
abutments for a 3-unit fixed partial denture, one of which showed signs of complete healing at the 1-year
follow up and signs of VRF at 2.5 yrs. after the surgery. One tooth (maxillary premolar) was the most distal
tooth in the arch and had clinical symptoms of VRF at the annual follow-up. One tooth (maxillary premolar),
had VRF that was not seen during the surgery but the symptoms started immediately following the surgical
intervention (within three weeks), and the tooth was extracted shortly afterward. In one case (a mandibular
molar), the patient initially presented with signs of procedure-related failure (major procedural error), and
during the re-surgery, the root appeared fractured. The patient then opted for root amputation.

Success rate & prognostic factors

Out of the sixty-eight patients (85 teeth) that presented for follow-up, one patient was further excluded from
evaluating the success rate due to the development of a generalized periodontal disease that was not associated
with the surgical procedure performed, but it did impact the radiographic interpretation. Another patient was
also excluded, despite evidence of radiographic healing, due to the presence of symptoms that were not
endodontic in nature. This left 83 teeth from 66 patients, including five teeth (5 patients) with VRF, to evaluate
the success rate. The success rate and the prognostic factors were evaluated twice, with and without including
VRF cases. When VRF cases were included, the success rate was 88% using PA and 86% using CBCT. After
excluding cases with vertical root fracture, the success rate was 94% using PA and 91% using CBCT (Fig 3).
The presence of a “major procedural error” during the surgery was the only factor associated with unfavorable
healing in the bivariate analysis (P< 0.001), with and without including VRF cases, and was the only negative
predictor for the treatment outcome in the logistic regression model (P = 0.013). The results are
displayed in Table 2.

Intra-observer & inter-observer agreements

In the 2D analysis, Cohen Kappa analysis revealed a “substantial” intra-observer agreement among the four
healing categories (K= .61; P < 0.001) and “moderate” agreement when dichotomized into favorable and
unfavorable radiographic outcome. (K= .52; P < 0.001). 3D analysis, on the other hand, had “substantial” to
“almost perfect” intra-observer agreement among all planes. Regarding the inter-observer reliability, the 2D
analysis showed a substantial agreement when the outcome was grouped into four healing categories. It
dropped to moderate when the radiographic outcome was dichotomized into favorable and unfavorable. On the

This article is protected by copyright. All rights reserved

 contrary, the agreements were “substantial” to “almost perfect” between the two reviewers during the 3D
Accepted Article
analysis. The results are displayed in Table 3.

Inter-radiographic agreement
Cohen Kappa analysis revealed a slight agreement between the PAs and CBCTs when the healing was divided
into four categories (K= .197; P < 0.001) and moderate agreement when dichotomized into favorable and
unfavorable (K= .416; P < 0.001). The agreement between the 2D and the 3D evaluations was the highest in
cases with complete healing (97%) or unsatisfactory healing (100%). The results are displayed in Table 4.

CBCT scores reliability & internal consistency

Cronbach’s alpha showed good internal consistency (.75) above the threshold of >.70 for newly developed
scales (Nunnally 1994). All the three indices indicated medium to large item-total correlation of r =.73 for P
index, r=.53 for R index, and r=.43 for C index. The principal components analysis identified only one factor
with an eigenvalue of 2.009, explaining 66.95% of the variance in the outcome model. All three indices loaded
meaningfully onto the outcome model at .910 for the P index, .824 for the R index and .708 for the C index.

A total of 77 asymptomatic and functional teeth were included to assess the effect of follow-up time on
predicting the total score. The ordinal regression model showed that the follow-up time did not significantly
predict the total score for every case (P = 0.891), or the radiographic healing category (P = 0.868). A binary
logistic regression revealed that time was also not predictive of a favorable or unfavorable radiographic
outcome (P = 0.360).

Hard tissue remodeling

A total of 71 teeth with successful apical surgery were included to assess the peri-radicular bone remodeling.
With regards to the healing score among the various indices, the P index exhibited the highest rate of complete
healing (94%), followed by the R index (90%). The C index exhibited the lowest rate of complete healing at
72%. There was no correlation between the application of GTR and the type of healing achieved (complete vs
incomplete) when periapical radiographs were used to evaluate the outcome (P = 0.398). There was,
however, a statistically significant correlation between the application GTR and the type of healing achieved
[complete (score 6) vs. incomplete (score 4 or 5)] on CBCT (P < 0.001). There were no correlations between
GTR and the scores within the R index (P = .137) or the P index (P = 0.707). There was, however, correlation
between GTR and the scores within the C index (P < 0.001). The results are displayed in Table 5.

This article is protected by copyright. All rights reserved

Accepted Article
Discussion
This study highlights the effect of several prognostic factors on the outcome of EMS and the impact GTR may
have on bone remodeling. The number of teeth included was also slightly larger than previously published data
with similar study design (Christiansen et al. 2009, Tanomaru et al. 2015, von Arx et al. 2016). The study
design had both prospective and retrospective elements to it. While all surgeries were conducted before
initiating the investigation, the study parameters were set before the follow-up appointments and outcome
evaluation. A limitation of this study would be the heterogenous follow-up times between the different cases.
While the regression model revealed that cases with longer follow-up time did not have better outcome, longer
recalls may allow further remodeling and improvement in the scores of some cases (von Arx et al. 2019c).
Another limitation in this study would be including more than one tooth/patient when evaluating prognostic
factors.

In this study, the scoring system proposed by Chen et al. (2015) for CBCT evaluation was adopted; however,
the interpretation was modified regarding two aspects: 1) the cortical index, to clarify various healing patterns
that can be encountered in cases with apico-marginal defects or through-and-through defects; and 2) the apical
index, which was modified to include the entire peri-radicular structure, now called the “P index” ( Fig. 1.) The
scoring system used herein was based on summing the results from each index ranging the scores from 0 to 6
(Fig. 4, 5, & 6). Statistical analysis revealed that the radiographic healing assessment relied on the scores from
all the three indices to predict the outcome with high loaded meaningful values for each index (>.50).

von Arx et al. (2016) proposed using the “B index” to interpret the individual scores from the three different
indices. The B index, however, lacks specificity as it combines all the cases that fall between complete healing
(score 6) and no healing (score 0) in one category. It was clear from the results that minor improvements in the
bony structure can be achieved at the follow-up visit, even in failed or fractured cases. In this study, cases that
received a score 3 were the most difficult to interpret. In some cases, a favorable radiographic outcome can be
achieved if the case is monitored for extended periods (Fig. 5a-d). In other cases, the outcome may not
improve further, but the case will not present with any clinical or radiographic signs of failure (i.e. scar
formation) (Fig. 5e-h). While, in other cases, the outcome can potentially become unfavorable. It should be
noted that none of the cases presenting with score 3 had any clinical symptoms and only cases with scores 0, 1,
or 2 did present with clinical symptoms of failure or fracture (6/7 cases with unsatisfactory healing presented
with clinical symptoms of failure).

This article is protected by copyright. All rights reserved

Accepted Article
Outcome assessment
The success and survival rates in this study were high and in-line with previously published data (Rubinstein &
Kim 2002, Torabinejad et al. 2015). A clear distinction was made between the success and survival rate of
surgically treated teeth to identify the cause/mode of failure, whether VRF (i.e. tooth failure), that will result in
tooth extraction/root amputation, or a failure due to persistence of a disease, that can be potentially corrected
via re-surgery. Due to the limited number of VRF cases and the presence of multiple confounding factors, a
quantitative analysis could not be determined, thus, qualitative assessment was performed. While this can be
considered a limitation in this study, a qualitative assessment may assist in understanding reasons for VRF
following surgical intervention, particularly with the lack of evidence on this subject. The qualitative analysis
suggests that the VRF could be:
1- present prior to initiating the surgery and was missed during the treatment. This was the interpretation of one
of the cases that had signs of VRF 3 weeks after the treatment;
2- induced by the surgical intervention during apical manipulation, or
3- initiated by other factors non-related to the surgical procedure, i.e. occlusal forces and oral
functions/parafunctional habits. This was the evaluators’ interpretation of three teeth with VRF that were either
abutment for a fixed partial denture or had no distal contact. These confounding factors have been associated
with a lower tooth survival rate following non-surgical and surgical endodontic treatment (Ng et al. 2011b,
Wang et al. 2017). Understanding the cause/mode of failure and presenting the outcome in a consistent manner
is essential in clinical reports. For instance, in non-surgical outcome studies, VRF is only attributed to the
survival rate of endodontically treated teeth, not the success rate (Farzaneh et al. 2004, Ng et al. 2011a).
However, in surgical outcome studies, there is no clear distinction on whether VRF cases are included or
excluded from evaluating the success rate (Safi et al. 2019, Truschnegg et al. 2020).

The success rate presented herein was 3% lower when CBCT was used for outcome evaluation. The findings
agree with previous studies (Tanomaru et al. 2015; von Arx et al. 2016; Safi et al. 2019). These studies,
however, had a larger difference in the results between the two radiographic methods, that ranged between 8%
to 27%. Differences in results may stem from the CBCT criteria used to interpret the radiographic findings, as
well as the categorization of theses interpretations to success and failure. While it was evident that more
radiographic defects were noted during the CBCT analysis, these defects are not clinical failure per se but
might rather represent incomplete or defective healing of a successful case, as illustrated in Fig. 4 (e – l). It was
noted that 85% of the cases with incomplete healing on the 3D scan were mistaken for complete healing during

This article is protected by copyright. All rights reserved

 the 2D evaluation (Table 3). This may stem from the inability of the PAs to properly evaluate the healing of the
Accepted Article
cortical plate (C index).

Prognostic factors
In this study, an attempted was made to study multiple outcome predictors with and without including cases
with VRF, to avoid statistical bias, since the cause of VRF may not be related to the surgical intervention as
discussed earlier. In both analyses, the presence of major procedural errors during surgical treatment was the
only factor affecting the outcome of EMS. Unlike non-surgical treatments where the outcome can be affected
by non-procedural factors such as remaining bacterial load, extra-radicular infection, or true cystic lesions
(Nair 2006), the outcome of EMS appears to be primarily operator dependent. Previous studies showed that the
tooth position, location, and the presence of an isthmus, were outcome predictors for EMS (Song et al. 2013,
Kim et al. 2016). This, however, was not evident in the present findings. The difference in results may stem
from the potential presence of a major procedural error in posterior teeth or teeth with isthmus resulting in
missing a canal, root perforation, or improperly sealing the root canal space (Fig. 2e-h). While such errors may
often be missed or questioned on PAs, they were well observed on the CBCT scan. Having a minor procedural
error, on the other hand, did not impact the treatment outcome. However, it was very prevalent and occurred in
23% of all the treated teeth.

Non-surgical retreatment has been often recommended as the treatment modality of choice following a failure
of an initial root canal treatment (Torabinejad & White 2016). The odds of failure have also been shown to be
higher when the quality of the root canal filling was inadequate at the time of the surgery (Safi et al. 2019). The
present results, however, disagree with these findings and revealed that neither the root canal quality before
surgery nor the treatment sequence had a significant impact on the surgical outcome (P > 0.05). That being
said, deeper retro-preparations (> 3 mm) were always considered when the quality of the root canal treatment
was inadequate. While the depth of retro-preparation did not result in any significant improvement in the
treatment outcome (P = 0.24), there was a lack of patient randomization. The results suggest that the decision
to proceed with a surgical or a non-surgical approach following the failure of a root canal treatment could be
made based on what is in the patient’s best interest rather than a particular treatment sequence. Clinicians,
however, may consider longer root-end retro-preparation if the root canal quality was poor or inadequate to
address the entire root canal space. The outcome in this study was also not affected by the type of retro-filling
material used, whether RRM or MTA, confirming the results from Safi et al. (2019). The combination of BC
sealer and RRM putty, also known as the “lid technique”, appeared to be a valid retro-filling technique with a

This article is protected by copyright. All rights reserved

 high success rate and can be particularly useful in cases with long retro-preparations. Furthermore, the results
Accepted Article
support Kim et al. (2018) that revealed no significant difference between the outcome of endodontic surgery
and re-surgery. All the re-surgery cases in this study were successfully managed and had a favorable outcome
(Table 2).

Previous studies often excluded cases that presented with a through-and-through lesion or an apico-marginal
defect (Christiansen et al. 2009, Kruse et al. 2016, Safi et al. 2019, von Arx et al. 2019a). These defects
represented around 20% of the cases that were referred for surgical treatment in this study, and excluding them
would not reflect the true outcome for cases requiring intervention by means of EMS. While historically such
cases have had a poor prognosis when surgically treated (Taschieri et al. 2008, Song et al. 2013), our results
showed that these defects did not affect the treatment outcome (P > 0.05). This can be attributed to the GTR
techniques used, as indicated in previous studies (Dietrich et al. 2002, 2003, Taschieri et al. 2008). Results,
however, should be interpreted with care due to the small patient sample size in these categories.

Apical bone remodeling

GTR did not improve the outcome of the surgical treatment nor it affected the type of favorable healing when
PA was used for assessment (P > 0.05). However, it significantly affected the type of favorable healing
achieved on the CBCT (complete vs incomplete) (P < 0.001), as well as the quality of the reestablished cortical
plate (C index) (P < 0.001). It had also no effect on the scores of the P or the R indices (P > 0.05). These
results highlight the limitation of PA to properly evaluate the healing of the cortical plate as mentioned before.
To properly assess the effect of GTR on bone remodeling, teeth with unfavorable outcome were excluded to
minimize confounding factors, since treatment failure will also affect bone remodeling. Out of the 71 cases
with a favorable outcome on the CBCT, 71% had complete reestablishment of the cortical plate (Score 2) at the
recall appointment. 24% had a concave appearance (score 1), and only 5% had no reestablishment (score 0).
These results were different from von Arx et al. (2019c) that revealed only 19% with complete reestablishment,
51% with concave appearance, and 30% with no cortical plate reestablishment. The difference in results may
stem from the exclusion criteria and the utilization of GTR in the present study.

There was a complete reestablished of the cortical plate (score 2) in around 86% of the cases with GTR,
compared to 38% in those without GTR (Fig. 7). While further improvement in the cortical plate may occur
with time, it is less likely to occur in cases where the cortical plate is reestablished in a concave/regressed
position (von Arx et al. 2019c). To avoid misinterpreting a healthy site for a site masked by the radiopaque

This article is protected by copyright. All rights reserved

 bone graft material, several cases received an immediate post-operative CBCT for treatment assessment
Accepted Article
purposes. This allowed viewing how a graft material appears on the CBCT immediately after placement, and at
the follow-up appointment, after remodeling. While sites receiving GTR can be often mistaken on a 2D
radiograph as completely healed, it can be easily identified on the CBCT.

Crossen et al. (2019) illustrated that the amount of linear cortical plate regression following apical surgery was
minimal with less than 0.25 mm, and thus questions the necessity of GTR in endodontic surgery. However,
their study had a small patient sample size (only 12 cases) with no information on the size of the defect at the
time of the surgery. The present observation disagrees with their findings, as the cortical plate regression varied
within the un-grafted cases reaching up to 3mm linear cortical plate regression, three years after treatment (Fig.
7h). The results, however, should be considered preliminary. Future research should explore more carefully the
recommendation for GTR beyond apico-marginal defects and through and through lesions to identify clinical
scenarios that may benefit from its application. Further studies with larger sample size and immediate post-
operative scans are needed to elucidate better a direct relationship between the post-surgical defect size and the
amount of cortical plate regression. In the meantime, clinicians may consider grafting surgical endodontic
defects to avoid regression in the bucco-lingual/palatal ridge thickness. This might be beneficial if the tooth is
lost, under any circumstances, and a dental implant needs to be placed.

Conclusion
Endodontic microsurgery appears to be a very predictable treatment with a high success and survival rate. The
success rate appears to be slightly higher when PA is used to evaluate the outcome, compared to CBCT.
However, CBCT evaluations were more consistent. The presence of a major procedural error during surgery
that compromised the apical seal or the integrity of the supporting structure was the primary cause of
unfavorable radiographic outcome and treatment failure. The sequence of treatment, the quality of endodontic
treatment before the surgery, and the use of GTR did not affect the success rate of EMS procedures. GTR,
however, affected the type of 3D healing (complete vs incomplete) and bone remodeling of the cortical plate
following EMS.

Conflict of Interest statement

The authors have stated explicitly that there are no conflicts of interest in connection with this article.

This article is protected by copyright. All rights reserved

 References
Accepted Article
AAE/AAOMR (2016): Use of Cone Beam Computed Tomography in Endodontics 2015/2016 Update.
https://www.aae.org/specialty/clinical-resources/cone-beam-computed-tomography/
Chen I, Karabucak B, Wang C et al. (2015) Healing after root-end microsurgery by using mineral trioxide
aggregate and a new calcium silicate-based bioceramic material as root-end filling materials in dogs. Journal of
Endodontics 41, 389-99.
Chong BS, Pitt Ford TR, Hudson MB (2003) A prospective clinical study of Mineral Trioxide Aggregate and
IRM when used as root-end filling materials in endodontic surgery. International Endodontic Journal 36, 520-
26.
Christiansen R, Kirkevang LL, Gotfredsen E, Wenzel A (2009) Periapical radiography and cone beam
computed tomography for assessment of the periapical bone defect 1 week and 12 months after root-end
resection. Dentomaxillofacial Radiology 38, 531-6.
Christiansen R, Kirkevang LL, Horsted-Bindslev P, Wenzel A (2009) Randomized clinical trial of root-end
resection followed by root-end filling with mineral trioxide aggregate or smoothing of the orthograde gutta-
percha root filling--1-year follow-up. International Endodontic Journal 42, 105-14.
Curtis DM, VanderWeele RA, Ray JJ, Wealleans JA (2018) Clinician-centered Outcomes Assessment of
Retreatment and Endodontic Microsurgery Using Cone-beam Computed Tomographic Volumetric Analysis.
Journal of Endodontics 44, 1251-6.
Dietrich T, Zunker P, Dietrich D, Bernimoulin JP (2002) Apicomarginal defects in periradicular surgery:
classification and diagnostic aspects. Oral Surgery Oral Medicine Oral Pathology Oral Radiology and
Endododontics 94, 233-9.
Dietrich T, Zunker P, Dietrich D, Bernimoulin JP (2003) Periapical and periodontal healing after osseous
grafting and guided tissue regeneration treatment of apicomarginal defects in periradicular surgery: results after
12 months. Oral Surgery Oral Medecine Oral Pathology Oral Radiology and Endodontics 95, 474-82.
ESE (2019) European Society of Endodontology position statement: Use of cone beam computed tomography
in Endodontics: International Endodontic Journal 52, 1675-8
Farzaneh M, Abitbol S, Lawrence HP, Friedman S, Toronto S (2004) Treatment outcome in endodontics-the
Toronto Study. Phase II: initial treatment. Journal of Endodontics 30, 302-9.
Jonasson P, Lennholm C, Kvist T (2017) Retrograde root canal treatment: a prospective case series.
International Endodontic Journal 50, 515-21.

This article is protected by copyright. All rights reserved

 Kanagasingam, S, Hussaini HM, Soo I., Baharin S, Ashar A, Patel S. (2017). Accuracy of single and parallax
Accepted Article
film and digital periapical radiographs in diagnosing apical periodontitis–a cadaver study. International
Endodontic Journal, 50, 427-36.
Kim E, Song JS, Jung IY, Lee SJ, Kim S (2008) Prospective clinical study evaluating endodontic microsurgery
outcomes for cases with lesions of endodontic origin compared with cases with lesions of combined
periodontal-endodontic origin. Journal of Endodontics 34, 546-51.
Kim S, Jung H, Kim S, Shin SJ, Kim E (2016) The Influence of an Isthmus on the Outcomes of Surgically
Treated Molars: A Retrospective Study. Journal of Endodontics 42, 1029-34.
Kim S, Song M, Shin SJ, Kim E (2016) A Randomized Controlled Study of Mineral Trioxide Aggregate and
Super Ethoxybenzoic Acid as Root-end Filling Materials in Endodontic Microsurgery: Long-term Outcomes.
Journal of Endodontics 42, 997-1002.
Kim D, Kim S, Song M, Kang DR, Kohli MR, Kim E (2018) Outcome of Endodontic Micro-resurgery: A
Retrospective Study Based on Propensity Score-matched Survival Analysis. Journal of Endodontics 44, 1632-
40.
Kruse C, Spin-Neto R, Christiansen R, Wenzel A, Kirkevang LL (2016) Periapical Bone Healing after
Apicectomy with and without Retrograde Root Filling with Mineral Trioxide Aggregate: A 6-year Follow-up
of a Randomized Controlled Trial. Journal of Endodontics 42, 533-537.
Molven O, Halse A, Grung B (1987) Observer strategy and the radiographic classification of healing after
endodontic surgery. International Journal of Oral & Maxillofacial Surgery 16, 432-9.
Mota de Almeida FJ, Knutsson K, Flygare L (2014) The effect of cone beam CT (CBCT) on therapeutic
decision-making in endodontics. Dentomaxillofacial Radiology 43, 20130137.
Nair PN (2006) On the causes of persistent apical periodontitis: a review. International Endodontic Journal 39,
249-81.
Ng YL, Mann V, Gulabivala K (2011a) A prospective study of the factors affecting outcomes of nonsurgical
root canal treatment: part 1: periapical health. International Endodontic Journal 44, 583-609.
Ng YL, Mann V, Gulabivala K (2011b) A prospective study of the factors affecting outcomes of non-surgical
root canal treatment: part 2: tooth survival. International Endodontic Journal 44, 610-25.
Nunnally JC (1994) Psychometric theory 3E: Tata McGraw-hill education.
Patel S (2009) New dimensions in endodontic imaging: Part 2. Cone beam computed tomography.
International Endodontic Journal 42, 463-75.
Rubinstein RA, Kim S (2002) Long-term follow-up of cases considered healed one year after apical
microsurgery. Journal of Endodontics 28, 378-83.

This article is protected by copyright. All rights reserved

 Safi C, Kohli MR, Kratchman SI, Setzer FC, Karabucak B (2019) Outcome of Endodontic Microsurgery Using
Accepted Article
Mineral Trioxide Aggregate or Root Repair Material as Root-end Filling Material: A Randomized Controlled
Trial with Cone-beam Computed Tomographic Evaluation. Journal of Endodontics 45, 831-9.
Schloss T, Sonntag D, Kohli MR, Setzer FC (2017) A Comparison of 2- and 3-dimensional Healing
Assessment after Endodontic Surgery Using Cone-beam Computed Tomographic Volumes or Periapical
Radiographs. Journal of Endodontics 43, 1072-9.
Setzer FC, Kohli MR, Shah SB, Karabucak B, Kim S (2012) Outcome of endodontic surgery: a meta-analysis
of the literature--Part 2: Comparison of endodontic microsurgical techniques with and without the use of higher
magnification. Journal of Endodontics 38, 1-10.
Song M, Kim SG, Lee SJ, Kim B, Kim E (2013) Prognostic factors of clinical outcomes in endodontic
microsurgery: a prospective study. Journal of Endodontics 39, 1491-7.
Song M, Kim SG, Shin SJ, Kim HC, Kim E (2013) The influence of bone tissue deficiency on the outcome of
endodontic microsurgery: a prospective study. Journal of Endodontics 39, 1341-5.
Tanomaru FM, Jorge EG, Guerreiro-Tanomaru JM, Reis JM, Spin-Neto R, Goncalves M (2015) Two- and
tridimensional analysis of periapical repair after endodontic surgery. Clinical Oral Investigation 19, 17-25.
Taschieri S, Del Fabbro M, Testori T, Saita M, Weinstein R (2008) Efficacy of guided tissue regeneration in
the management of through-and-through lesions following surgical endodontics: a preliminary study.
International Journal of Periodontics & Restorative Dentistry 28, 265-71.
Torabinejad M, Landaez M, Milan M et al. (2015) Tooth retention through endodontic microsurgery or tooth
replacement using single implants: a systematic review of treatment outcomes. Journal of Endodontics 41, 1-
10.
Torabinejad M, White SN (2016) Endodontic treatment options after unsuccessful initial root canal treatment:
Alternatives to single-tooth implants. Journal of the American Dental Association 147, 214-20.
Tronstad L, Asbjornsen K, Doving L, Pedersen I, Eriksen HM (2000) Influence of coronal restorations on the
periapical health of endodontically treated teeth. Endodontics & Dental Traumatology 16, 218-21.
Truschnegg A, Rugani P, Kirnbauer B, Kqiku L, Jakse N, Kirmeier R (2020) Long-term Follow-up for Apical
Microsurgery of Teeth with Core and Post Restorations. Journal of Endodontics 46, 178-83.
Tsesis I, Rosen E, Taschieri S, Telishevsky Strauss Y, Ceresoli V, Del Fabbro M (2013) Outcomes of surgical
endodontic treatment performed by a modern technique: an updated meta-analysis of the literature. Journal of
Endodontics 39, 332-9.

This article is protected by copyright. All rights reserved

 von Arx T, Janner SF, Hanni S, Bornstein MM (2016) Evaluation of New Cone-beam Computed Tomographic
Accepted Article
Criteria for Radiographic Healing Evaluation after Apical Surgery: Assessment of Repeatability and
Reproducibility. Journal of Endodontics 42, 236-42.
von Arx T, Jensen SS, Janner SFM, Hanni S, Bornstein MM (2019a) A 10-year Follow-up Study of 119 Teeth
Treated with Apical Surgery and Root-end Filling with Mineral Trioxide Aggregate. Journal of Endodontics
45, 394-401.
von Arx T, Marwik E, Bornstein MM (2019b) Effects of Dimensions of Root-End Fillings and Peripheral Root
Dentine on the Healing Outcome of Apical Surgery. European Endodontic Journal 4, 49-56.
von Arx T, Janner SFM, Hanni S, Bornstein MM (2019c) Radiographic Assessment of Bone Healing Using
Cone-beam Computed Tomographic Scans 1 and 5 Years after Apical Surgery. Journal of Endodontics 45,
1307-13.
Wang ZH, Zhang MM, Wang J, Jiang L, Liang YH (2017) Outcomes of Endodontic Microsurgery Using a
Microscope and Mineral Trioxide Aggregate: A Prospective Cohort Study. Journal of Endodontics 43, 694-8

This article is protected by copyright. All rights reserved

 Figure Legends
Accepted Article
Figure 1 Illustrations of the different patterns of healing for the three indices (C, R, and P index), and the
scores received accordingly.
Figure 2 PAs and CBCT images of different types of minor (a–d) and major (e–h) Procedural errors.
(a & b): PAs showing retro-preparation being off angle from the root canal space. (c & d) PA and CBCT
showing an over enlarged retro-preparation. (e) PA showing root perforation during long retro-preparation. (f)
Follow-up CBCT of a mandibular molar showing improper sealing of the apical 3mm. Arrow pointing to an
area where there is no retro-filling material. (g) Follow-up CBCT showing a completely off-angle retro-
preparation Arrow pointing to a missed ML canal with no retro-filling. (h) Follow-up CBCT of a maxillary
molar. Arrow pointing to the missed MB2 which is located at a more coronal position compared to MB1.
Figure 3 Flow chart of participants and recall rate
Figure 4 Pre-op and follow-up CBCT images in the sagittal and coronal view illustrating three cases with
different patterns of favorable radiographic outcome. Case 1 (a – d): Maxillary lateral incisor showing
complete healing at all the 3 indices, each receiving a score of 2. Arrows pointing to the missing palatal plate
(2b) and the reestablished buccal and palatal plates (2d) at the follow-up. Case 2 (e –h): Maxillary premolar
with arrow pointing to the R index on the palatal root receiving a score of 1 while the P and C indices receiving
a score of 2 each. Case 3 (i – l): Mandibular molar with arrow pointing to the R and P indices on the mesial
root where both areas shows partial reestablishment receiving a score of 1 and the C index receiving a score of
2. b, buccal; p, palatal
Figure 5 Pre-op and follow-up PA and CBCT images of varies cases of uncertain healing. Case 1 (a – d):
Mandibular molar where non-surgical treatment was attempted without success due to canal calcification. (a)
Pre-op PA showing calcified canal with periapical radiolucency associated with both roots (b) Pre-op CBCT
image showing periapical radiolucency associated with the mesial root. (c) PA showing 1-yr. follow-up with
improvement compared to the pre-op PA. (d) Follow-up CBCT image showing complete reestablishment of the
C index (Score 2) and slight improvement in the R index (Score 1), while the P index did not improve
compared to the Pre-op CBCT (Score 0). Case 2 (e – h): Mandibular anterior tooth which iatrogenically
became through-and-through during the surgery (major procedural error) and did not receive GTR. (e, f) Pre-op
PA and CBCT showing small apical radiolucency associated with the mandibular incisor. (g) 1-yr. follow-up
showing scar formation on the PA. (h) CBCT follow-up shows improvement of the P index (Score 1) and the R
index (Score 2) but no reestablishment of the C index (Score 0). Arrow points to the reestablished resection
plane.

This article is protected by copyright. All rights reserved

 Figure 6 Pre-op and follow-up PA and CBCT images of varies cases of unsatisfactory healing. Case 1 (a – e):
Accepted Article
A Mandibular molar that presented at 1 yr. after surgery with clinical signs of failure. (a- c) Shows Pre-op PA
and CBCT images. (d, e) follow-up PA and CBCT images showing partial improvement in the P and R indices
(Score 1) but no reestablishment of the C index (Score 0). Case 2 (f – j): Maxillary premolar that presented
with signs of VRF 1 yr. after surgery. (f, g) Pre-op PA & CBCT (h) Immediate post-op PA after placement of
bone graft. (i) 1-yr follow-up PA showing recurrence of a radiolucency in the apical area. (J) CBCT showing
improvement in the P index (Score 1) but no improvement on the C and R indices (score 0). Case 3 (k – o): A
maxillary central incisor with a failed surgical treatment. (k) Pre-op PA showing a large periapical
radiolucency associated with the lateral side of the root and lateral canal exit. (l) CBCT showing a through-and-
through defect. (m) Immediate post-op PA showing retro-filling in place, lateral canal not addressed and bone
graft in the apical area. (n, o) follow-up PA & CBCT images showing no bone reestablishment of any of the
three indices.
Figure 7 Pre-op and follow up CBCT images showed different patterns of cortical bone healing in cases with
and without GTR.
(a) Pre-op CBCT showing the loss of the cortical plate in one area and minor expansion in another. (b) 1 year
follow up after GTR showing slight bulging of the buccal plate. (c) Pre-op CBCT image with an arrow pointing
to natural dehiscence of the cortical plate at the mid-root level. (d) 1 year follow after GTR with an arrow
pointing to the newly gained bone in the same area. (e) Pre-op CBCT showing a small lesion associated with
the apical area of an upper anterior tooth. (f) 1 year follow up with an arrow pointing to the regression noticed
in the cortical plate of a case without GTR. (g) Pre-op CBCT image of an upper anterior tooth with no lesion
but patient presented with symptoms and surgery was indicated. (h) 1 year follow up with an arrow pointing to
a major regression of the cortical plate of a case without GTR.

This article is protected by copyright. All rights reserved

 Table 1 assessment of the different radiographic planes
Accepted Article
Index Definition Score 0 Score 1 Score 2
R Resection plane No bone deposition (no Partial bone deposition Complete bone
formation of PDL) (partial formation of PDL deposition (complete
space) formation of PDL space
of normal width)
P Peri-radicular No apparent bone Partial bone formation in Complete bone
area formation around the the peri-radicular area formation in the peri-
peri-radicular area that appears to be smaller radicular area
regardless of the R or C than the lesion size in the
indices pre-operative CBCT
regardless of the R or C
indices
C Cortical plate Cortical plate is not re- Re-established, but Re-established and flat
(access bony established or has been concave or with minor
window) partially re-established indentations
with/without areas of
communication between
the buccal and
lingual/palatal plate
Outcome evaluation (R + P + C)
Favorable radiographic outcome Unfavorable radiographic outcome
Complete Incomplete healing Uncertain healing Unsatisfactory healing
healing
Score 6 Score 5 or 4 Score 3 Score 2, 1, or 0
Success Favorable radiographic outcome + No clinical symptoms
Failure Unfavorable radiographic outcome and/or presence of clinical symptoms
Survival Favorable/unfavorable radiographic outcome + No clinical symptoms
PDL, Periodontal ligament.

This article is protected by copyright. All rights reserved

Accepted Article
Table 2 Univariate distribution and bivariate analysis of prognostic factors.
Prognostic factors n Success Failure Success rate % P value Success rate % P value
Diseased VRF Total (Without VRF cases) (All cases)

Sex 0.520 0.066

Male 29 22 3 4 7 88 75.9
Female 54 49 4 1 5 92.5 90.7
Tooth position 0.468 0.404
Maxillary 57 50 4 3 7 92.6 87.7
Mandibular 26 21 3 2 5 87.5 80.7
Tooth type 0.389 0.335
Anterior 44 40 3 1 4 93 90.9
Premolar 15 12 0 3 3 100 80
Molar 24 19 4 1 5 82.6 79.2
Sequence of treatment: Retrograde root canal as a: 0.109 0.167
Primary treatment 2 1 1 0 1 50 50
After NS initial treatment 56 50 3 3 6 94.3 89.2
After NS retreatment 18 13 3 2 5 81.3 72.2
Re-surgery 7 7 0 0 0 100 100
Quality of RCT prior to surgery 0.541 0.395
Good 43 39 3 2 5 95.1 90.7
Poor/no treatment 39 32 4 3 7 88.9 82.1
Error during surgery <0.001* <0.001*
No error 53 49 1 3 4 98 92.4
Minor error 22 20 1 1 2 95.2 90.9
Major error 8 2 5 1 6 28.5 25

This article is protected by copyright. All rights reserved

Accepted Article Retro-preparation depth 0.746 0.967
<3mm 3 3 0 0 0 100 100
3mm 61 51 5 5 10 91.1 83.6
6mm 11 9 2 0 2 81.8 81.8
9mm 8 8 0 0 0 100 100
Isthmus present 0.324 0.433
yes 33 27 4 2 6 87.1 81.8
No 50 44 3 3 6 93.6 88
Retro-filling material 0.593 0.054
MTA 29 23 2 4 6 92 79.3
RRM putty 22 18 3 1 4 85.7 81.8
BC sealer + RRM putty 32 30 2 0 2 93.8 93.8
GTR 0.135 0.162
yes 56 50 3 3 6 94.3 89.3
No 27 21 4 2 6 84 77.8
Through - Through defect 0.375 0.415
yes 14 11 2 1 3 84.6 78.6
No 69 60 5 4 9 92.3 87
Apical-marginal defect 0.803 0.817
Yes 6 6 0 0 0 100 100
No 77 65 7 5 12 92.9 84.4

n = number of teeth; NS = Non-surgical; * indicates significance

This article is protected by copyright. All rights reserved

Accepted Article Table 3 Intra-observer and inter-observer agreement.
Intra-observer agreement Inter-observer agreement
Index Observer 1 Observer 2 Combined results 1st 2nd observation
(N=81) (N=81) (N=162) observation (2 months later)
2D analysis
PA healing categories 88.6% 90.9% 89.2% 91.1% 89.8%
(complete, incomplete, uncertain, unsatisfactory) K=.585 K=.627 K=.61 K= .641 K=.619
PA outcome categories 92.4% 90.9% 91.7% 93.6% 89.8%
(favorable/unfavorable) K=.538 K=.505 K=.52 K= .586 K=.459
3D analysis
R index 88.6% 92.4% 91.9% 89.9% 92.4%
K= .619 K= .771 K=.70 K=.686 K=.777
P index 96.1% 97.5% 96.8% 97.5% 94.9%
K=.861 K=.911 K=.89 K=.866 K=.816
C index 94.9% 94.9% 93.7% 92.4% 97.4%
K=.837 K=.837 K=.83 K= .740 K=.947
0 – 6 scores 81% 89.8% 85.4% 83.5% 92.4%
K=.696 K=.839 K=.76 K= .689 K= .802
Healing categories 84.8% 93.6% 89.2% 84.8% 94.9%
(complete, incomplete, uncertain, unsatisfactory) K=.734 K=.890 K= .81 K=.685 K=.848
Outcome categories 100% 100% 100% 98.7% 98.7%
(favorable/unfavorable) K=1.00 K=1.00 K=1.00 K=.934 K=.934
Kappa values: <0, no agreement; 0–0.2, slight agreement; 0.21–0.40, fair agreement; 0.41– 0.60, moderate agreement; 0.61–0.80, substantial agreement; 0.81–1,
almost perfect agreement.

This article is protected by copyright. All rights reserved

Accepted Article
Table 4 Inter-radiographic agreement between 2D and 3D evaluations among the different healing groups.

3D favorable radiographic outcome 3D unfavorable radiographic outcome

complete healing incomplete healing uncertain healing unsatisfactory healing Total
complete healing 43 23 2 0 68
2D
favorable 73
radiographic
outcome
incomplete healing 0 3 2 0 5

uncertain healing 1 1 1 1 4
2D
unfavorable unsatisfactory 0 0 0 6 (5 are VRF) 6 10
radiographic healing
outcome
44 27 5 7
Total
83
71 12

This article is protected by copyright. All rights reserved

Accepted Article Table 5 Distribution of cases receiving GTR within the “favorable outcome” on 2D and 3D imaging as well as the different healing pattern
scores within the various indices on the 3D scans.
Favorable healing n GTR No GTR P Value
Favorable healing in 2D
Complete healing 68 47 21 0.398
Incomplete healing 5 2 3
Total 73 49 24

Favorable healing in 3D
Complete healing (score 6) 44 39 5
Incomplete healing (score 5 or 4) 27 11 16 <0.001*
Total 71 50 21

3D Healing pattern scores Score 2 Score 1 Score 0 Score 2 Score 1 Score 0

Peri-radicular area (P index) 71 48 2 0 19 2 0 0.720
Resection plane (R index) 71 46 2 2 18 3 0 0.707
Cortical plate (C index) 71 43 4 3 8 13 0 <0.001*
The P values refer to the effect of GTR application on the healing pattern. n = number of cases; * indicates significance

This article is protected by copyright. All rights reserved

Accepted Article

iej_13419_f1.tiff

This article is protected by copyright. All rights reserved

Accepted Article

iej_13419_f2.tiff

This article is protected by copyright. All rights reserved

Accepted Article

iej_13419_f3.tiff

This article is protected by copyright. All rights reserved

Accepted Article

iej_13419_f4.tiff

This article is protected by copyright. All rights reserved

Accepted Article

iej_13419_f5.tiff

This article is protected by copyright. All rights reserved

Accepted Article

iej_13419_f6.tiff

This article is protected by copyright. All rights reserved

Accepted Article

iej_13419_f7.tiff

This article is protected by copyright. All rights reserved