Crown morphology and additional canals in mandibular incisors using cone-beam computed tomography: A geometric morphometric study

Abstract

Context:

Mandibular incisors usually have a single root and canal, but additional canals may be present. Undetected canals can lead to endodontic failure.

Aims:

This study evaluated whether tooth shape and dimensions can predict additional canals.

Settings and Design:

This study evaluated the relationship between crown shape, buccolingual (BL) width, and overall tooth dimensions in mandibular incisors using cone-beam computed tomography and two-dimensional geometric morphometry.

Subjects and Methods:

Coronal morphology was documented using TPSutil (version 1.83) and TPSdig2 (version 2.31) in teeth with single and bifid canals. Landmark coordinates were analyzed in MorphoJ (version 1.07a) using Procrustes superimposition.

Statistical Analysis Used:

Chi-square test, Procrustes analysis of variance (ANOVA), and discriminant function analysis (DFA) were used.

Results:

Of 108 teeth assessed, 52.7% (n = 57) had single canals (Type I), and 47.2% (n = 51) had additional canals (98.03% Type II and 1.96% Type IV). The first five principal component analysis components accounted for 82% of variation. DFA classified 64.7% of additional canal teeth and 63.1% of single canal teeth (63.9% accuracy). Maximum variation occurred at cervical crest curvatures, with lower incisal edges in additional canal teeth. Procrustes ANOVA showed no size differences (P = 0.7301).

Conclusions:

Teeth containing additional canals displayed increased buccal and lingual crest widths and reduced incisal height compared to teeth with a single canal; however, the BL width differences were not statistically significant. Overall, crown dimensions did not show any meaningful association with the presence of extra canals in mandibular incisors.

INTRODUCTION

Understanding of tooth anatomy, root canal configuration, and morphological variations is crucial for achieving successful endodontic treatment outcomes. Although mandibular central and lateral incisors are generally described as single-rooted teeth with a single canal, studies show considerable variation.[1,2] Verma et al.[3] studied cone-beam computed tomography (CBCT) images of 200 patients with 800 permanent mandibular incisors in Indian subpopulation and demonstrated that Type 1 Vertucci’s classification (single root with single canal) was the most prevalent canal configuration (64.5%) in the mandibular anterior teeth in the Indian population. However, it is important to recognize that the remaining lower incisors studied can exhibit considerable anatomical variations. Failure to identify and appropriately manage these variations increases the risk of missed canals, which may ultimately compromise the success of endodontic treatment.[4,5]

CBCT provides detailed three-dimensional (3D) visualization of root canal morphology in axial, sagittal, and coronal planes, with relatively lower radiation doses than many conventional 3D imaging methods. CBCT has proven to be an excellent imaging modality for detection of different canal configurations of mandibular incisors.[6,7] Despite these advantages, routine use of CBCT is limited by cost and radiation considerations, and its application should adhere to the as low as reasonably achievable (ALARA) principle to ensure minimal exposure.[8,9] According to the recent joint position statement of the American Association of Endodontists and the American Academy of Oral and Maxillofacial Radiology, CBCT is not recommended for routine endodontic diagnosis or screening in the absence of clinical signs and symptoms and should be used only when diagnostic needs cannot be adequately fulfilled by lower-dose two-dimensional (2D) radiography.[10]

Studies have examined the relationship between crown dimensions and additional canals in mandibular incisors, canines, and premolars. Miyoshi et al.[11] examined 3010 mandibular incisors from a Japanese cohort and reported that teeth with two canals tended to have a larger buccolingual (BL) diameter. Similarly, Warren and Laws[12] observed that teeth with lower mesiodistal/faciolingual (M-D/F-L) crown indices more frequently had bifid canals. However, a study by Ghamari et al.[13] reported that although mandibular incisors with two canals and a Type III canal configuration showed greater mesiodistal and BL dimensions, their overall crown size (M-D/F-L index) did not differ significantly from that of single-canal incisors. None of these studies were conducted in Indian population. Accordingly, this study aimed to evaluate the BL width and coronal dimensions of mandibular anterior teeth using morphometric analysis of CBCT images from an Indian population and to determine whether these measurements can predict the presence of additional canals.

SUBJECTS AND METHODS

Study type and study setting

This retrospective cross-sectional study was conducted in the Department of Conservative Dentistry and Endodontics using CBCT scans obtained from the radiology archives of the Department of Oral Medicine and Radiology. After obtaining ethical clearance from the Institutional Ethics Committee (IEC no. 24161), the research work commenced.

Sample size calculation

According to the study by Warren and Laws,[12] which compared single and bifid canals in lower incisors with respect to coronal dimension, the required sample size for each group was 53 (1% alpha error, 99% power, and a clinically significant difference of 0.5 units). Geometric morphometry was employed for sample size calculation based on the number of landmarks and dimensions. Ten landmarks were measured in two dimensions, resulting in a total of 20 variables. After accounting for two degrees of freedom for rotation, translation, and scaling (in this case, four), the required sample size was 3 times the 16 variables, amounting to 54 samples per group. Consequently, 108 CBCT cases were included in the statistical analysis.

Inclusion and exclusion criteria

The inclusion criteria required the presence of mandibular central and lateral incisors with fully developed roots and closed apices, a high-quality CBCT image free of artifacts, and scans without any pathology in the mandible. CBCT scans exhibiting missing anterior teeth, anterior teeth with root canal treatments, postrestorations, coronal restorations, and defects such as internal and external root resorption, incomplete root development, open apices, and indications of attrition were excluded.

Cone-beam computed tomography image evaluation

CBCT scans were obtained from the radiology archives of the Department of Oral Medicine and Radiology, MCODS, Mangalore, which were acquired using Promax 3D Mid (Planmeca Oy. Helsinki, Finland) CBCT unit. The images were reconstructed using Romexis software version 4.6.2.R. (Planmeca Oy, Helsinki, Finland). Small FOV scan volumes that covered the incisors with a voxel size of 200 μm were selected for the study. Only patient demographic details, such as age and sex, were acquired from the dental records. The images were analyzed and assigned a number. The Romexis software (version 4.6.0 R) was used to evaluate the scans. To evaluate the canal morphology, axial sections were evaluated (in the explorer window). This was confirmed in the cross-sections (in the implant window). Cross section was aligned along the long axis of the tooth to obtain an image of both the canal and crown, in toto.

Canal configuration

The classification of diverse root canal morphologies was conducted based on the system described by Weine, which served as a standardized method for analyzing the structural variations within root canal systems. This approach allowed the categorization of canal configurations, facilitating a better understanding of their complexity.

Coronal morphology assessment

In the course of data collection, landmarks on the mesial aspect of the tooth were identified using TPS dig and TPSutil software (Rohlf F. J. [2025] tpsDig/tpsUtil, Department of Ecology and Evolution, State University of New York at Stony Brook). The TPSutil software (Version 1.83) produced 108 high-resolution images of the mandibular teeth in a *tps file format, which were subsequently used for landmark acquisition using the TPSdig2 software (Version 2.31). The landmark selection tool facilitated the definition of 10 for each mandibular anterior tooth among the 108 teeth [Figure 1].

Figure 1.

Ten landmarks on the mandibular anterior tooth for coronal morphology assessment

Buccolingual width calculation

The BL width was measured in millimeters on a sagittal section of the tooth. It is the distance from the buccal crest of curvature (the most convex point on the buccal surface) to the lingual crest of curvature (the most convex point on the lingual surface).

Geometric morphometric assessment

For the 2D annotation component of the study, the scale of the software was carefully calibrated to align with the ruler in each image. MorphoJ software version 1.07a (Christian P. Klingenberg, University of Manchester, UK). was used to analyze the landmark coordinates, employing statistical techniques such as Procrustes superimposition, principal component analysis (PCA), and discriminant function analysis (DFA). Transformation grids and visual plots were produced. The procedure involved standardizing the landmark coordinate data and performing Procrustes superimposition. Subsequently, a covariance matrix was constructed, and PCA was carried out.[14]

Statistical analysis

The association between the BL width at the level of the cementoenamel junction (CEJ) and the incidence of bifid canals in mandibular incisors was analyzed using statistical software (SPSS Version 20, IL, NY, USA). The test used was either an independent t-test or a Mann–Whitney U-test or Chi-square test, depending on the skewness of the data. P < 0.05 was considered statistically significant.

Procrustes analysis of variance (ANOVA) was used to assess the differences in coronal morphology between teeth with and without bifid canals. The ability of the shape data to classify teeth with extra root canals based on incisor shape was evaluated using DFA. To strengthen the analysis, the P value threshold was set to 0.003.

RESULTS

Sample and overall canal prevalence

A total of 108 mandibular incisors from 29 patients were evaluated using CBCT. Of these, 57 (52.7%) demonstrated a single canal (Weine Type I), whereas 51 (47.2%) exhibited an additional canal. Among teeth with additional canals, 50/51 (98.0%) were Weine Type II (two separate canals starting from the pulp chamber and joining as one, just short of the root apex) and 1/51 (2.0%) was Weine Type III (two separate canals starting from the pulp chamber to the root apex).[15]

Distribution by tooth and laterality

Additional canals were comparably distributed between the central (24/53; 45.3%) and laterals (27/55; 49.1%). Chi-square tests showed no significant difference by arch position (central: χ² = 0.182, df = 1, P = 0.669; laterals: χ² = 0.162, df = 1, P = 0.688). When examined by tooth number, the distributions were balanced (centrals: #31, 11/26 additional vs. #41, 13/27 additional; laterals: #32, 13/28 additional vs. #42, 14/27 additional) with no significant asymmetry on Chi-square testing.

Sex distribution

Females comprised 63 teeth (30 with additional canals and 33 with single canals) and males 45 teeth (21 with additional and 24 with single canals). The association between sex and the presence of an additional canal was not significant overall (χ² = 0.010, df = 1, P = 0.922).

Relation between buccolingual width and presence of additional canal

Mean BL width was numerically greater in teeth with an additional canal (5.994 mm) than in single-canal teeth (5.832 mm), with a mean difference of ~ 0.16 mm; however, the independent-samples t-test was not significant (t = 1.646, df = 106, P = 0.103). The group variances were 0.358 and 0.178, respectively.

Relation between shape and additional canals using geometric morphometrics

PCA of the 10-landmark configuration indicated that the first five principal components explained approximately 82.8% of the total shape variance. Procrustes ANOVA found no significant difference in centroid size between single- and double-canal teeth (F = 0.12, P = 0.7301) and no significant effect on shape (F = 1.2, P = 0.2588) [Table 1]. DFA yielded correct classification rates of 64.7% for additional canal teeth and 63.1% for single-canal teeth, with an overall accuracy of 63.9%; the associated distances were Procrustes = 0.0278 and Mahalanobis = 0.8937 [Table 2], and the Hotelling’s t²-test was nonsignificant (t² = 0.3187, P = 0.3206). Inspection of the DFA scatter and transformation grids suggested that landmarks 2 and 10 contributed most to the observed variation, followed by landmarks 1 and 9. In double-canal teeth, landmarks 5 and 6 tended to be positioned lower relative to single-canal teeth, consistent [Figure 2].

Table 1.

Procrustes analysis of variance to test the association of the centroid size and shape of the tooth in relation to the single and additional canals

Effect	SS	MS	df	F	P
Centroid size
Individual	13.94399	13.94399	1	0.12	0.7301
Residual	12349.89	116.5084	106
Shape, Procrustes ANOVA
Individual	0.020862	0.001304	16	1.2	0.2588
Residual	1.840794	0.001085	1696

SS: Sum of squared distances, MS: Mean of squared distances, ANOVA: Analysis of variance

Table 2.

Discriminant function analysis to evaluate the shape changes in crown to distinguish the single and additional canals

Group	Additional canal	Single canal	Total (%)
Additional canal	33	18	51 (64.7)
Single canal	21	36	57 (63.1)
Overall accuracy	63.9
Procrustes distance	0.02784018
Mahalanobis distance	0.8937

Figure 2.

Thin-plate spline deformation grids illustrating mesial crown outline shape variation of the mandibular incisor based on 10 landmarks, comparing double-canal (red) and single-canal (blue) configurations as derived from discriminant function analysis

DISCUSSION

The findings of our study revealed that the BL width at the CEJ was numerically greater in teeth with additional canals (5.994 mm) than in those with a single canal (5.832 mm), with a mean difference of 0.16 mm. Although this difference was not statistically significant, the observed trend aligns with the morphological literature on tooth dimensions and endodontic anatomy.[11,12,13] A study by Miyoshi et al.[11] demonstrated that the diameter of the crown in the two-canal group was significantly wider than that in the one-canal group. Similarly, in a study by Warren and Law,[12] the incidence of bifid canals was found to be slightly over 40% among 288 mandibular incisor teeth. By applying the M-D/F-L index, it was found that the incidence of bifid canals was greater among the lower score (i.e., more BL width) than among the high ones. Ghamari et al.[13] reported that mandibular incisors with two canals showed increased mesiodistal and BL dimensions, although their crown size was not significantly different from that of single-canal incisors. Apart from incisors, the relationship between the number of canals in a root and tooth dimensions has also been demonstrated in molars. Dibaji et al.[16] evaluated 311 maxillary first molars using CBCT and reported that teeth with a second mesiobuccal canal had significantly greater BL dimensions at the CEJ (11.9 mm vs. 10.8 mm; P < 0.001) compared with single-canal teeth.

The nonsignificant result in the present study may be attributed to several factors: heterogeneity in sample composition, inclusion of multiple tooth types, or methodological variations across studies. The unequal variances between the groups (0.358 vs. 0.178) suggest differential anatomical plasticity associated with canal configuration, indicating a potential biological mechanism linking pulp system complexity to crown enlargement.[17] The numerical trend observed is consistent with the biological principle that accommodating complex internal pulp systems may necessitate expanded external dimensions. Future research using larger, homogeneous samples, standardized tooth selection criteria, and multivariate analysis controlling for demographic and anatomical variables may clarify whether this dimensional difference achieves statistical significance while providing additional insights into the morphological determinants of endodontic complexity.

PCA of the 10-landmark configuration showed that the first five components accounted for 82.8% of the total shape variance, thereby capturing primary morphological variation. This variance aligns with the geometric morphometric studies of dental structures using similar protocols.[18] Despite this variance, the Procrustes ANOVA found no significant differences in centroid size between teeth with single and additional canals (F = 0.12, P = 0.7301) or overall shape (F = 1.2, P = 0.2588). These results suggest that gross external root morphology does not differ according to internal canal configuration. The lack of significant differences in centroid size aligns with studies on external tooth dimensions and internal canal anatomy. Ghamari et al.[13] evaluated 202 mandibular incisors using CBCT and found that BL (6.08 mm vs. 5.86 mm, P < 0.05) and mesiodistal (4.85 mm vs. 4.55 mm, P < 0.05) dimensions were slightly larger in two-canal teeth, but crown size indices showed no significant differences (P = 0.227). This pattern mirrors our results, suggesting that slight dimensional increases do not translate into detectable shape differences when analyzed through landmark-based approaches.

The DFA yielded classification accuracies of 64.7% for additional canal teeth and 63.1% for single-canal teeth, with an overall accuracy of 63.9%. The nonsignificant Hotelling’s t²-test (t² = 0.3187, P = 0.3206), with a small Procrustes distance (0.0278) and Mahalanobis distance (0.8937), confirms that morphometric separation between the groups is minimal and nonsignificant. However, the morphometric analysis did yield variation in the locations of the landmarks which can be used to distinguish the two types. Landmarks 2 and 10 [Figure 2] contributed the most to the variation, followed by landmarks 1 and 9. Teeth with additional canals showed landmarks 5 and 6 positioned lower than single-canal teeth, indicating wider cervico-labio-lingual contours. This finding supports Ghamari et al.,[13] who found that BL dimensions had stronger associations with canal multiplicity than mesiodistal measurements. The labio-lingual widening pattern may reflect developmental processes affecting transverse root expansion. However, these landmark shifts were not statistically significant. The clinical implication is that external crown morphology has limited predictive value in determining canal configuration in mandibular incisors.

CONCLUSIONS

This study evaluated whether the crown morphology of mandibular incisors could serve as a reliable indicator of additional root canals. Although minor variations in crown shape were observed, these differences did not form any consistent pattern that could assist in identifying more complex internal anatomy. Geometric morphometric analysis further demonstrated that the overall form of the crown does not reliably correspond to variations within the root canal system. These findings align with recent evidence, suggesting that external anatomical features alone cannot be depended upon for predicting internal canal configuration, particularly in mandibular anterior teeth, where subtle differences may not reflect the underlying complexity. Relying solely on crown morphology may therefore lead to missed canals and incomplete endodontic management. Given these limitations, adjunctive radiographic approaches such as multiple angulated periapical radiographs, and CBCT imaging in root-treated teeth with persistent symptoms, remain valuable tools in endodontic diagnosis and management.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.