The best radiographic method for determining root canal morphology in mandibular first premolars: A study of Chinese descendants in Taiwan.

BACKGROUND/
PURPOSE: There is large variation in root canal morphology and undetected canals and incomplete instrumentation are reasons for root canal treatment failure. The purpose of this study was to determine the best radiographic method for determining root canal morphology in mandibular first premolars in Chinese descendants in Taiwan.
MATERIALS AND METHODS: Mandibular first premolars extracted due to caries, periodontal diseases, trauma, or for orthodontic reasons were used. Four indices were examined: (1) root canal bifurcation observed in the buccolingual view; (2) root canal continuity in the buccolingual view; (3) double root outline in the buccolingual view; and (4) Vertucci canal classification in the mesiodistal view.
RESULTS: A total of 82 left and right mandibular first premolars were included, a complicated root canal was confirmed in 38 (46.3%) by cross-sectional imaging and a single root canal was found in 44 (53.7%). Bifurcation identified on the mesiodistal view exhibited the highest sensitivity (94.7%) and second highest specificity (88.6%) for identifying a complicated root canal; however, this view is not possible to obtain clinically. Canal bifurcation on the buccolingual view was the most specific (93.2%), but had the lowest sensitivity (73.7%). Canal continuity on the buccolingual view had a sensitivity of 94.7%, and specificity of 70.5%.
CONCLUSION: Combined X-ray analyses, such as performing the buccolingual view for identification of canal bifurcation and canal continuity, may increase the accuracy of identifying complex root canal morphology.

Introduction

Root canal treatment depends on complete canal debridement and filling of the root canal system. Undetected canals and incomplete instrumentation are reasons for root canal treatment failure. Variation in root canal anatomy is common, and has been shown to be associated with race and gender.2, 3, 4 To this end, knowledge of the morphology and variations of root canals is essential for successful endodontic treatment.

Canal anatomy of mandibular premolars has been shown to vary greatly. Studies have shown that mandibular first premolars with two or more canals occur in 13.7% of Caucasians in the United States and 46% of Chinese individuals.2, 5 Another study has shown that the incidence of three canals in mandibular first premolars ranges from 0.4% to 2%.6, 7 Furthermore, Fan et al8, 9 reported that 24% or more of mandibular first premolars have C-shaped root canals.

Cone-beam computed tomography (CBCT) and micro-CT are commonly used in dentistry, and can accurately determine root canal morphology.11, 12, 13, 14, 15, 16 However, CBCT and micro-CT are not available in all parts of the world, especially underdeveloped and developing countries, and due to the cost and dose of radiation periapical X-rays are still the most widely used method for determining root canal morphology prior to endodontic treatment. Accurate detection of complex canal morphology on X-ray is necessary to avoid missing root canals during treatment, but this can be difficult. Thus, the purpose of this study was to determine the best radiographic method for determining root canal morphology of mandibular first premolars.

Materials and methods

This study utilized mandibular first premolars extracted due to caries, periodontal disease, trauma, and other reasons at the Department of Oral and Maxillofacial Surgery of Taipei-Veteran General Hospital. Only first premolars with a mature and intact root structure and devoid of fractures were used. The details of the collection, preparation, and anatomical examination of the teeth have been previously published. This study was approved by the Institutional Review Board of the hospital, and all patients had previously provided informed consent for the dental procedures performed.

Both buccolingual and mesiodistal parallel radiographs were obtained for each tooth using a Size 2, E speed radiograph cassette (CEA AB, Strangnas, Sweden), and a Heliodent DS X-ray system (Sirona Dental Company, Bensheim, Germany). The source–object distance was 55 mm, and other parameters were 7 mA, 60 kVp, and a 0.12-second exposure time. Radiographs were developed with an automatic X-ray film processor (Dent-X 810 Basic Auto Film Processor; Dentx Visionary Imaging, Elmsford, NY, USA) and mounted. Each radiograph was examined independently by two endodontists, each with > 10 years' experience, using a 2 × magnifying viewer (JS Dental Manufacturing, Inc., Ridgefield, CT, USA) with a bright view box (King bright box, 5000 D × 10 W × 2; Asanuma & Co., Ltd., Hamamatsu, Shizuoka, Japan). The final analysis and findings of the radiographs were reached by consensus of the two reviewers.

Four indices were examined for their ability to identify complex root canal morphology: (1) root canal bifurcation observed in the buccolingual view; (2) root canal continuity in the buccolingual view; (3) double root outline in the buccolingual view; and (4) Vertucci canal classification in the mesiodistal view.

Root canal continuity was classified as previously described. Briefly, the classifications were as follows: (A) large canal becoming less obvious, and may logically determine there is a bifurcation; (B) large canal becoming thinner and deviating towards one side, there may be one small and one large canal or furcated roots, and may logically determine that there is a bifurcation; (C) medium root canal, gradual tapering, cannot logically determine that there are two canals but proximal view may display a second root canal; and (D) buccolingual view shows a direct bifurcation.

Statistical analysis

Counts and percentages were calculated for the X-ray characteristics of the root canal configuration and outline of the root in both the bucco-lingual and mesio-distal views. To test the efficacy of the four indices, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and positive and negative likelihood ratios (LRs) with 95% confidence intervals were calculated using the findings of standard of cross-sectional imaging as the gold standard. The indices were calculated by the following formulas.Sensitivity (true positive rate) = (number of positive results / total number of complicated canals) × 100.Specificity (true negative rate) = (number of negative results / total number of single canals) × 100.PPV = (number of confirmed complicated canals / total number of positive results) × 100.NPV = (number of confirmed single canals / total number of negative results) × 100.Positive LR = ratio of the true positive to false positive rate = sensitivity / (1−specificity).Negative LR = ratio of the false negative to true negative rate = (1−sensitivity) / specificity.

The index with relatively higher values in terms of the sensitivity, specificity, PPV, NPV, accuracy, and positive LR along with a lower negative LR was defined as the best index to identify a complicated root canal. The difference in accuracy among the four indices was examined by Cochran's Q test, followed by McNemar's test with Bonferroni correction for multiple post hoc comparisons. The significance level was defined as 0.05, with an adjusted significance level of 0.008 (0.05/6). All statistical analyses were two-sided, and performed with SAS version 9.2 software (SAS Inc., Cary, NC, USA).

Results

A total of 82 left and right mandibular first premolars were included in this study. A complicated root canal was confirmed in 38 (46.3%) by cross-sectional imaging, and a single root canal was found in 44 (53.7%) (Table 1). The three most common mesio-distal features identified on X-rays were root tapering (47, 57.3%), roots with a C-shaped configuration (13, 15.9%), and blunt roots (11, 13.4%). Of the 38 premolars with complicated root canals, 21 (55.3%) were Vertucci Type IV, six (15.8%) Type I, five (13.2%) Type III, four (10.5%) Type II, and two (5.2%) Type V.

Table 1

Canal configuration and X-ray characteristics of study teeth from the buccolingual and mesiodistal views.

	No.	Percentage of total
Cross-sectional image
Single root canal	44	53.7
Complicated root canal	38	46.3
Outline of root in mesiodistal view
Continuous tapering	47	57.3
Blunt	11	13.4
Bulge	6	7.3
Bifurcated root	4	4.9
C-shaped/fissured root	13	15.9
Uncertain	1	1.2
Radiographic features of complicated canals in mesiodistal view a
Type I	6	15.8
Type II	4	10.5
Type III	5	13.2
Type IV	21	55.3
Type V	2	5.2
Outline of root in buccolingual view
Continuous tapering	67	81.7
Blunt	1	1.2
Bulge	6	7.3
C-shaped/fissured root	8	9.8
Canal configuration in buccolingual view
Normal	33	40.2
Disappeared root	12	14.6
Tapered and skewed root	15	18.3
Tapered and small root with a suspected second canal specified from P-view	13	15.9
Bifurcated root	8	9.8
Uncertain	1	1.2
Specified bifurcation in buccolingual view
No	51	62.2
Yes	31	37.8
Radiographic feature of complicated canal in buccolingual view b
Sudden disappearance of root canal	17	54.9
Sudden disappearance with canal deviation	11	35.5
Type IV canal	3	9.6
Double root outline in buccolingual view
No	33	40.2
Yes	49	59.8
Bifurcation in mesiodistal view
No	41	50.0
Yes	41	50.0

Data were from 38 complicated canals identified by cross-sectional imaging.

Data were from 31 bifurcated roots identified by the buccolingual view.

Buccolingual X-rays showed that 67 (81.7%) premolars had tapering roots, eight (9.8%) had fissured roots, and seven (8.5%) had blunt or bulging roots. Root tapering and skew was found in 18.3% of canals, root tapering and shrinkage in 15.9%, root disappearance in 14.6%, and root bifurcation in 9.8%. Thirty-one (37.8%) bifurcated root canals were identified by the endodontists based on buccolingual features, among which 54.9% of the root canals were disappeared, 35.5% were deviated, and 9.6% were Vertucci Type IV. Root canal bifurcation was also identified in 49 premolars based on a double root outline in the buccolingual view, and in 41 premolars from the mesiodistal view.

The diagnostic efficacies of the four indices for detecting a complicated root canal are summarized in Table 2. The analysis indicated that bifurcation identified in the mesiodistal view exhibited the best diagnostic accuracy of 91.5%, the lowest negative LR, and higher values of the other measures. Additionally, the four indices showed varying degrees of accuracy (P = 0.011), and a difference between the canal continuity in buccolingual view and the bifurcation in mesiodistal view was noted (81.7% vs. 91.5%, P = 0.004). Likewise, the presence of a double root outline displayed lower accuracy than the morphological reading of the mesiodistal view (74.4% vs. 91.5%, P = 0.001).

Table 2

Diagnostic performance of X-ray characteristics to detect complicated root canal morphology.

	Buccolingual view: canal bifurcation	Buccolingual view: canal continuity	Buccolingual view: double root outline (PDL double shadow)	Mesiodistal view: vertucci Type II–V
Sensitivity	73.7% (56.9%, 86.6%)	94.7% (71.9%, 95.6%)	86.8% (82.3%, 99.4%)	94.7% (82.3%, 99.4%)
Specificity	93.2% (81.3%, 98.6%)	70.5% (47.8%, 77.6%)	63.6% (54.8%, 83.2%)	88.6% (75.4%, 96.2%)
PPV	90.3% (74.3%, 98.0%)	73.5% (52.5%, 80.1%)	67.3% (58.9%, 85.1%)	87.8% (73.8%, 96.2%)
NPV	80.4% (66.9%, 90.2%)	93.9% (68.1%, 94.9%)	84.8% (78.0%, 99.3%)	95.1% (83.5%, 99.4%)
Accuracy	84.1% (80.5%, 87.2%)	81.7% (63.6%, 83.4%)	74.4% (71.6%, 89.4%)	91.5% (83.2%, 96.5%)
LR+	10.81 (3.57, 32.75)	3.21 (2.02, 5.09)	2.39 (1.58, 3.60)	8.34 (3.64, 19.29)
LR−	0.28 (0.16, 0.48)	0.07 (0.02, 0.29)	0.21 (0.09, 0.48)	0.06 (0.02, 0.23)

Data presented with (95% confidence interval).

LR− = negative likelihood ratio; LR+, positive likelihood ratio; NPV = negative predicted value; PDL = periodontal ligament; PPV = positive predicted value.

Representative tooth radiographs are shown in Figure 1, Figure 2, Figure 3, Figure 4.

Figure 1

Representative Case 1; (A) buccolingual view. Canal reading: continuity A, root canal becomes unclear in the middle third of tooth root. Root reading: root outline in the middle and apical third is observed as a double root outline (PDL double shadow), apex bifurcation; (B) mesiodistal view shows a Vertucci Type IV canal.

Figure 2

Representative Case 2; (A) buccolingual view. Canal reading: continuity B, shift to one side in the middle third of the tooth root. Root reading: root outline in the middle third is observed as a double root outline (PDL double shadow), apex bifurcation; (B) mesiodistal view: Vertucci Type II canal, fused in the middle third of the tooth.

Figure 3

Representative Case 3; (A) buccolingual view. Canal reading: continuity C, medium root canal, gradual tapering, cannot logically determine that there are two canals but proximal view may display a second root canal. Root reading: root outline in the middle and apical third is observed as a double root outline (PDL double shadow), no bifurcation at the apex; (B) mesiodistal view: Vertucci Type II canal.

Figure 4

Representative Case 4; (A) buccolingual view. Canal reading: continuity E, direct bifurcation. Root reading: root outline in the middle third is observed as a double root outline (PDL double shadow), apex bifurcation; (B) mesiodistal view: Vertucci Type IV canal, bifurcation in the middle third of the tooth root; (C) section showing a C-shaped canal.

Discussion

This study investigated the diagnostic accuracy of different X-ray views for identifying complicated root canals of mandibular first premolars. Each premolar tooth was assessed in four different X-ray images by two endodontists, and the results were compared against those of anatomic pathological evaluation. The analysis indicated that bifurcation identified on the mesiodistal view exhibited the best diagnostic accuracy (91.5%) with the highest sensitivity (94.7%) and second highest specificity (88.6%) for identifying a complicated root canal. The reason that bifurcation identified on the mesiodistal view had the best diagnostic accuracy is likely because mandibular premolars have two root canals, which are in a buccolingual direction. However, the mesiodistal view is clinically impossible and X-ray films from different angles are necessary for diagnosis.

Canal bifurcation identified on the buccolingual view was the most specific (93.2%) and had highest PPV (90.3%); however, it had the lowest sensitivity (73.7%). Canal continuity identified on the buccolingual view also had a high sensitivity (94.7%, identical to that of the mesiodistal view), but a lower specificity (70.5%). Taken together the results suggest that combined X-ray analyses, such as performing the buccolingual view for identification of canal bifurcation and canal continuity, may increase the accuracy of identifying complex root canal morphology.

Whereas the skills and knowledge of the endodontist are paramount for successful root canal treatment, radiographic analysis is important for determining canal morphology. Although micro-CT is highly accurate, in a typical practice X-ray is the most common tool used for examining root canals due to the cost and dose of radiation. Two-dimensional X-rays, however, often cannot effectively display the full anatomy of the root canal system, though a combination of X-rays from different angles may be useful. England et al suggested that two X-rays with a 20° difference in angle combined with zooming can assist in determining multicanal morphology of mandibular first premolars. Martínez-Lozano et al examined X-rays taken from different angles to view premolars, and found 20° and 40° angles to be effective in identifying multicanal morphology of mandibular first premolars. In a study by Yoshioka et al, four dentists interpreted the X-rays of 139 mandibular first premolars followed by cross-sectioning of the teeth to observe the root canal morphology. The authors found that a sudden narrowing of the main canal accurately predicted the existence of multicanal morphology and there was no statistically significant difference among the four dentists with respect to the coincidence rate (93%–96%) of the canal number evaluated on radiographs (P > 0.05). Literature reviews of human mandibular first and second premolar root canal morphology concluded that using a second X-ray with a 40° offset and zooming an image of sudden narrowing of the canal may help diagnose mandibular first premolar variations.23, 24

Numerous studies have illustrated variation in root canal morphology. The aforementioned review of mandibular first premolars included data of > 6700 teeth, and found that ∼98% were single-rooted, 1.8% had two roots, 0.2% had three roots, and < 0.1% had four roots. A single canal was present in 75.8% of the teeth, two or more canals were found in 24.2% of the teeth, a single apical foramen was found in 78.9% of the teeth, and 21.1% had two or more apical foramina. With respect to racial variations, in an early study Amos assessed 1000 patients and found that 16% of Caucasians and 21.6% of African Americans had bifurcated canals of the lower first premolar. Zillich and Dowson performed X-rays followed by opening the pulp chamber for inspection and reported that 23.1% of mandibular first premolars had multiple root canals. Trope et al studied 400 African Americans and 400 Caucasians who received full mouth X-rays, and based on obvious changes in the density of the root canal space, a second periodontal ligament (PDL) contour, and root furcation, they reported that 13.7% and 32.8%, respectively, had multicanal morphology. In a study in Hong Kong, Walker examined 1000 mandibular first premolars and found multicanal morphology in 36%. Serman and Hasselgren reported that 15.7% of dental patients had multicanal morphology, and Sabala et al reported that 22.8% of patients had lower first premolar canal variation, the highest percentage of all teeth. A study in Kuwait that examined X-rays of teeth that had previous root canal treatment found that 40% of mandibular first premolars had more than one root canal. In our prior study, we cross-sectioned 82 mandibular first premolars of the Chinese ethnicity and found 44% had complex root canal morphology, of which 22% had two canals, 18% had C-shaped canals, and 6% were circumferential at the apical area.

Studies of multicanal morphology with respect to gender have provided inconclusive findings. For example, Sert and Bayirli studied Turks and found 44% of men and 35% of women had multicanal morphology. However, Serman and Hasselgren found no significant difference in the occurrence of multicanal morphology between males and females. Some recent genetic studies have found multicanal morphology of first premolars to be associated with X-chromosome mutations.28, 29

In this study the sensitivity of a double root outline on the buccolingual view for identifying complex root canal morphology was 86.8%; however, its specificity was 63.6% resulting in the lowest diagnostic accuracy of 74.4%. As determined by sectioning, there were 10 cases with a radicular groove, and the X-rays of those teeth all exhibited a double root outline. Two of 10 cases had one root canal, and the other eight cases had complicated canals. No relationship between a double root outline and radicular grooves was noted. Radicular groves are developmental depressions on the surface of the root, and when they are present the area of the PDL may be increased. Furthermore, X-ray images of teeth with radicular grooves may be misdiagnosed as having an extra root canal, and radicular grooves are associated with complicated root canal morphology.9, 16, 31 Studies have shown that radicular grooves are present in 24–41% of mandibular first premolars.9, 15, 16 Chen et al also reported that in Chinese mandibular first premolars the prevalence of radicular grooves in multiple and complex canals was 90.2%, which was much greater than the prevalence in single canals (17.4%).

There are limitations of this study that should be considered. Although considered highly accurate, the cross-sectioning method of evaluating root canal morphology may be inaccurate as a result of technical cone cut error. However, the use of teeth with pathologically confirmed root canal morphology as the gold standard is a strength of this analysis. Radiographic findings were obtained by consensus of the two readers, and thus no evaluation of inter- or intraobserver reliability was performed. This study used film X-rays, and currently digital X-rays are commonly performed. There may be differences in the interpretation of the two types of images,32, 33, 34, 35 and this was not evaluated in the current study. An ideal method of performing the study would have been to perform digital radiographs prior to extraction of the teeth. Lastly, the number of teeth studied was limited; further study with a larger sample number should be performed to confirm the results.

In conclusion, although bifurcation identified on mesiodistal radiographs exhibits high diagnostic accuracy (91.5%) for identifying a complicated root canal, these images are not possible to obtain clinically. Combined X-ray analyses, however, such as performing the buccolingual view for identification of canal bifurcation and canal continuity, may increase the accuracy of identifying complex root canal morphology.

Conflicts of interest

The authors have no conflicts of interest relevant to this article.