Table 6.

Summary of findings from the fourteen diagnostic efficacy studies identified in the review

Clinical context(s)	Caries	Purpose of imaging	Diagnosis of proximal caries cavitation
Sansare et al. (2014)
Diagnostic efficacy level	Level 2: diagnostic accuracy
Imaging (index tests)	1: CBCT [Kodak 9000; small FoV; 0.07-mm voxel] 2: Bitewing radiography Two observers performed independent assessments of index tests, blinded to the true diagnosis
Reference standard	Elective temporary tooth separation
Patient sample description	34 adults for whom there was suspicion of cavitation on visual examination. Prevalence of condition: 61% cavitated Ages: 18–63 years. Gender: 17F/17M. Setting: secondary care healthcare facility in India
Key outcomes	Using CBCT gave statistically significant greater sensitivity and accuracy. CBCT: Se = 75% and 79%, Sp = 77% and 77%, Acc = 76% and 78% Bitewing: Se = 46% and 42%, Sp = 84% and 87%, Acc = 61% and 59%
Study strengths	Blinding of assessors to diagnostic truth Reference standard appears robust
Study weaknesses	Adult population (weakness in the context of the review) Recruitment process unclear Only two assessors of images, both radiologists not dentists Unclear about the time interval between assessing CBCT and bitewing images

Clinical context(s)	Dental trauma	Purpose of imaging	Diagnosis of horizontal root fracture position and angulation in permanent incisors
Bornstein et al. (2009)
Diagnostic efficacy level	Level 3: diagnostic thinking efficacy
Imaging (index tests)	1. CBCT [3 DX Accuitomo XYZ Tomograph; 4 cm × 4 cm FoV; 0.125 mm voxel] 2. Periapical radiograph (analogue; paralleling technique] 3. Occlusal radiograph (analogue)
Reference standard	Not applicable
Patient sample description	38 patients presenting as emergencies with single or multiple horizontal root fractures of permanent teeth, with 44 fractured maxillary incisors within this sample Ages: mean age 24 years (8–52 years), Gender: 12F/26 M, Setting: University-based secondary healthcare facility in Switzerland
Key outcomes	Key outcome measure was fracture location on facial and palatal surfaces The fracture location on the palatal surface was more coronal on CBCT than on radiographs. In particular, a cervical fracture was more common on CBCT, potentially influencing management
Study strengths	Consecutive patients Clearly described methods
Study weaknesses	Retrospective study Selection bias possible: only cases with true-positive diagnoses on both CBCT and conventional imaging included Single observer performed the study assessments

Clinical context(s)	Dental anomalies	Purpose of imaging	Localisation of unerupted and supernumerary teeth in the maxilla
Ziegler and Klimowicz (2013)
Diagnostic efficacy level	Level 2: diagnostic accuracy
Imaging (index tests)	1: CBCT [iCAT—model not specified. FoV size not specified; 0.3-mm voxel probably used, but not clear] 2: Intraoral or panoramic radiographs; localisation using the “magnification method” (no explanation of what this means) Three surgeons made preoperative assessments of images independently
Reference standard	Intraoperative findings by surgeon
Patient sample description	61 mixed age group patients Ages: mean age 15 years (9–57  years), gender: not specified, setting: secondary care healthcare facility in Norway, prior testing: unknown
Key outcomes	Higher proportion of correct pre-operative localisation of bucco-palatal position using CBCT CBCT: 96.7% correct Radiographic “magnification method”: 39.3% correct
Study strengths	Prospective study Surgical reference standard
Study weaknesses	Recruitment process unclear Uncertain time gap between the index tests Inadequate conventional radiography (single panoramic or intraoral); the “magnification method” used is not explained; so unfair comparison against CBCT The index test assessments by the surgeons not clearly explained No statistics presented but “statistically significant” used in text Numerous text errors in the publication

Clinical context(s)	Dental anomalies and pathological conditions	Purpose of imaging	Diagnosis and treatment planning of impacted maxillary canines, including diagnosis of root resorption in permanent incisors
Haney et al. (2010)
Diagnostic efficacy level	Level 3: Diagnostic thinking efficacy and Level 4: Therapeutic efficacy
Imaging (index tests)	1. CBCT: Hitachi MercuRay; CBWorks software used to produce ”3D” selected images for study 2. Panoramic radiograph 3. Occlusal radiograph 4. Two periapical radiographs Seven faculty members were assessors in the study: four orthodontists and three oral surgeons
Reference standard	Not applicable
Patient sample description	18 consecutive patients with impacted maxillary canines. sample included 25 canines with 7 bilateral impactions Ages: mean age 16.9 years (12.3–34.6  years), gender: 12F/6M, setting: University-based secondary healthcare facility in USA
Key outcomes	Assessors make different decisions on aspects of diagnosis and treatment plans using CBCT for a minority of cases, e.g.:  16% disagreement for labiopalatal position of canine  50% disagreement when localising the cusp in the vertical dimension  36% agreement regarding root resorption between the two methods  Clinicians’ confidence in the accuracy of diagnosis and the treatment plan was statistically greater using CBCT
Study strengths	Prospective study on consecutive patient sample Full range of conventional radiographs available Seven assessors Statistical analyses clear
Study weaknesses	Presentation of images as print-outs on paper Risk of recognition of cases by assessors; radiographs and CBCT images were viewed on same session, including repeat cases for reliability assessment, with no “washout” period CBCT images viewed only as “3D” reconstructions, unlike normal practice Combining observations and decisions of all assessors for data analysis “Confidence” in diagnosis and treatment plans measured together

Clinical context(s)	Dental anomalies and pathological conditions	Purpose of imaging	Diagnosis and treatment planning of impacted and supernumerary teeth
Katheria et al. (2010)
Diagnostic efficacy level	Level 3: Diagnostic thinking efficacy and level 4: therapeutic efficacy
Imaging (index tests)	1. CBCT: 3D Sirona Galileos; FoV = full facial bone scan; voxel size not stated 2. Panoramic-like image and maxillary occlusal-like image synthesised from CBCT dataset (“Traditional Radiographs”, TR) Ten paediatric residents and ten paediatric dentists served as observers. Each observer viewed four cases only
Reference standard	Not applicable
Patient sample description	Eight patients’ radiographic records, each with one impacted canine or supernumerary tooth in the anterior maxilla. Ages: not specified: only “paediatric”, gender: not specified, setting: University-based secondary healthcare facility in USA
Key outcomes	No significant difference in “pathology diagnosis” using TR or CBCT Significantly greater proportion of decisions on location of pathology classed as “correct” using CBCT compared with TR (but no information on how “correct” was identified) Significantly higher proportion of observer decisions that root resorption was present using CBCT compared with TR Significantly higher proportions of observer decisions on “usefulness of CBCT” in the “very useful” category for both diagnosis and treatment planning
Study strengths	Multiple observers in the study Observers reviewed four cases using TR and four cases using CBCT, but not both imaging types of the same case, so no risk of recognition of case Missing information in the publication relevant to the review
Study weaknesses	Retrospective design High risk of selection bias Small number of cases Radiographs synthesised from CBCT Missing information in the publication relevant to the review (e.g. patient data) Use of terms such as “correct diagnosis” in the absence of any diagnostic truth Combining observations and decisions of assessors for data analysis

Clinical context(s)	Dental anomalies and pathological conditions	Purpose of imaging	Localisation and other imaging aspects of impacted maxillary canines
Alqerban et al. (2011)
Diagnostic efficacy level	Level 3: diagnostic thinking efficacy
Imaging (index tests)	1. CBCT: 3D Accuitomo-XYZ; FoV = 30 × 40 mm; voxel size = 0.125 mm 2. CBCT: Scanora 3D CBCT; FoV = 75 × 100 mm; voxel size = 0.2 mm 3. Panoramic radiograph: Cranex Tome (Soredex) Sample had either 1 or 2, but all had 3 Eleven assessors for study (3 experienced dental practitioners and 8 postgraduate (PG) students) The PG students only assessed a limited number of aspects of the imaging
Reference standard	Not applicable
Patient sample description	Sixty consecutive patients with impacted or ectopically erupting maxillary canines seeking orthodontic treatment Ages: Mean age 13.2 years (± 4.2 years); (6.3–28.9 years), Gender: 37F/23M, Setting: University-based secondary healthcare facility in Belgium
Key outcomes	Greater agreement between observers for all variables was achieved when using CBCT Observers’ decisions based on CBCT and panoramic radiography were significantly different for:  Canine location  Detection of the presence or absence of root resorption of the lateral incisor  Detection of the presence or absence of root resorption in the central incisor (Accuitomo group only)  Severity of lateral incisor root resorption
Study strengths	Consecutive patient sample Inter-observer agreement assessed thoroughly
Study weaknesses	Retrospective Major weakness was no intraoral radiographs, only panoramic radiographs, so how could position and resorption be assessed properly from a single panoramic image? Only partial assessments by PG students No intra-observer repeatability assessment Combining observations and decisions of assessors for data analysis

Clinical context(s)	Dental anomalies	Purpose of imaging	Diagnosis and treatment planning of impacted maxillary canines
Botticelli et al. (2011)
Diagnostic efficacy level	Level 3: diagnostic thinking efficacy and Level 4: therapeutic efficacy
Imaging (index tests)	1. CBCT: NewTom 3G (Quantitative Radiology s.r.l., Verona, Italy); FoV = not specified; Voxel size = not specified) 2. Conventional imaging: panoramic radiograph, periapical radiograph and lateral cephalogram Eight dentists acted as observers (3 specialists and 5 PG trainees: 2 at end of training and 3 early in training)
Reference standard	Not applicable
Patient sample description	Twenty-seven patients with 39 ectopic maxillary canines undergoing orthodontic treatment Ages: Mean age 11.8 years, Gender: 17F/10M, Setting: University-based secondary healthcare facility in Denmark
Key outcomes	Observers’ decisions based on CBCT and conventional radiography were statistically significantly different for:  Mesio-distal localization of the apex  Vertical level of the clinical crown  Overlap with the lateral incisor  Labio-palatal position of the crown  Labio-palatal position of the apex  Root resorption of neighbouring incisor/s  Treatment strategy (more observational strategy with conventional imaging; more interventional with CBCT)  Treatment assessed as more difficult with CBCT  Image quality (CBCT better) However, for all except treatment difficulty and image quality the majority of the decisions (≥ 64%) were the same when using CBCT or conventional radiography
Study strengths	Prospective Comprehensive conventional radiographic series Eight observers Clear written and visual presentation of findings
Study weaknesses	Lack of detail about conduct of index tests Images presented as Powerpoint presentations, with pre-selected CBCT images No intra-observer repeatability assessment Combining observations and decisions of assessors for data analysis

Clinical context(s)	Dental anomalies	Purpose of imaging	For imaging in the context of three different clinical contexts and in four countries (cost analysis)
Christell et al. (2012a)
Diagnostic efficacy level	Level 6: societal efficacy
Imaging (index tests)	CBCT: Four different scanners in four different centres No comparator imaging NewTom 3G [Quantitative Radiology (QR),Verona, Italy] Scanora (Soredex, Helsinki, Finland) Accuitomo MCT-1 (Morita, Kyoto, Japan) NewTom CVT 9000 (QR)
Reference standard	Not applicable
Patient sample description	One hundred and sixty patients referred for a CBCT examination during one calendar year for one of the following: imaging of maxillary canines with eruption disturbances, of an area with tooth loss prior to implant treatment or of a lower third molar planned for removal (cost analysis) Ages: Means and age ranges of the twelve combinations of country/clinical indication all presented in the paper. A paediatric group is not separately presented, but forms most of the maxillary canine group in two settings. Gender: not specified, settings: Four University-based secondary care specialist centres, in Romania, Belgium, Sweden and Lithuania
Key outcomes	Estimates for direct and indirect costs varied among the healthcare systems Estimates for direct and indirect costs varied according to clinical application Variation in direct costs was mainly owing to different capital costs Variation in indirect costs mainly owing to differences in examination fees Cost-efficacy established in one healthcare system might not be so in a different system
Study strengths	Input of health economist to research team Consecutive patients
Study weaknesses	No assessment of outcomes for patients No comparator imaging method (no incremental cost calculations) Patient’s or accompanying person’s average earnings used to calculate indirect costs, not real earnings

Clinical context(s)	Dental anomalies	Purpose of imaging	Imaging of maxillary canines with eruption disturbances (cost analysis)
Christell et al. (2012b)
Diagnostic efficacy level	Level 6: societal efficacy
Imaging (index tests)	1. New imaging method: CBCT [Accuitomo (Morita, Kyoto, Japan); FoV not specified; voxel size not specified] + panoramic radiograph [Planmeca Pro Max, (Helsinki, Finland)] 2. Conventional imaging method: panoramic radiograph [Planmeca Pro Max, (Helsinki, Finland)] + intraoral radiographs [Planmeca Intra (Helsinki, Finland)] Both methods included one panoramic radiograph per examination. The new method based on a recorded mean of 1.4 CBCT examinations per examination. The conventional method based on a recorded mean of 2.9 intraoral radiographs per examination
Reference standard	Not applicable
Patient sample description	Forty-seven patients referred for examination of maxillary canines with eruption disturbances during one calendar year Ages: mean age 14 years (10–19 years), gender: not specified, setting: University-based secondary care specialist centre in Sweden
Key outcomes	Framework for performing a cost analysis developed Adoption of “new” imaging method resulted in an incremental cost per examination of €46.58 (cost per examination for the new method = 128.38€ and for the conventional method = €81.80
Study strengths	Novel framework for cost analysis of diagnostic methods Input of health economist to research team Consecutive patients
Study weaknesses	No assessment of outcomes for patients Based on single clinic: specific costs not generalisable

Clinical context(s)	Developmental disorders	Purpose of imaging	Diagnosis and treatment plan related to teeth next to alveolar clefts and imaging of the cleft itself
Wriedt et al. (2017)
Diagnostic efficacy level	Level 3: Diagnostic thinking efficacy and Level 4: Therapeutic efficacya
Imaging (index tests)	1. CBCT: Accuitomo, Morita, Japan; FoV 40 × 40 mm; Voxel size: not specified 2. Panoramic radiographs [+ Study casts] Twelve clinicians made the assessments (6 maxillofacial surgeon or orthodontic specialists and 6 PG students)
Reference standard	Not applicable
Patient sample description	20 patients with 22 alveolar clefts, undergoing (late primary) secondary bone grafting of the alveolar cleft(s) Ages: mean age 12.5 years (± 5.5 years) (8–32 years), gender: 4F/16M, setting: University-based secondary care specialist centre in Germany
Key outcomes	In 74% of decisions, the cleft type was the same using CBCT and radiographic imaging All cleft borders were rated as “clearly visible” using CBCT, but over half were assessed as “unclear” on radiography Decisions on “clearly defined root” and not clearly defined root” were the same using the two imaging methods in about half of cases The majority of treatment proposals “alignment possible” or “not possible” were unchanged when using CBCT For the lateral incisor, the given proposals differed by up to 43.9% At long-term review after treatment, 65.9% to 92% of the proposals concerning the alignment of teeth were correct using radiographs, and 68.2% to 94.7% of the proposals were correct using CBCT A small-volume CBCT may be justified only as supplement to a routine panoramic X-ray in selected special cases
Study strengths	Consecutive patients Twelve examiners Inclusion of study casts, not only imaging
Study weaknesses	Retrospective design No intraoral occlusal radiograph, only panoramic Combining observations and decisions of assessors for data analysis No intra-observer repeatability assessment

Clinical context(s)	Pathological conditions	Purpose of imaging	Detection of resorption in association with unerupted teeth
Mak (2015)
Diagnostic efficacy level	Level 2: Diagnostic accuracy
Imaging (index tests)	1: CBCT [iCAT Next Generation. FoV varied; voxel size varied] 2: At least two conventional radiographs (intraoral or panoramic radiographs Nine residents, either paediatric dentistry or dental and maxillofacial radiology, served as observers
Reference standard	“Silver standard”: opinion of one Dental Radiologist using CBCT and radiographic images
Patient sample description	34 paediatric patients with impacted teeth. Mainly supernumerary teeth and mainly in the anterior maxilla Ages: mean 11.7 years (± 2.3  years); age range not given; maximum permissible age 18 years; gender: 18F/16M, setting: University-based secondary care specialist centre for paediatric dentistry
Key outcomes	No significant differences in diagnostic accuracy between imaging. Using conventional radiographs gave similar specificity to when CBCT was used. Trend to higher sensitivity using CBCT CBCT: Mean Se = 47%; Mean Sp = 85% Radiographs: Mean Se = 73%; Mean Sp = 87% Resorption prevalence = 15%
Study strengths	Nine observers, with assessment of intra-observer reliability
Study weaknesses	Small sample size and low resorption prevalence Retrospective study Potential selection bias: convenience sample of patients who had CBCT and radiographs. Possible inclusion of more difficult cases Use of expert-based reference standard, which was based on the viewing of the index test images No standardised CBCT imaging format No standardised conventional radiograph combination

Clinical context(s)	Pathological conditions	Purpose of imaging	Diagnosis of root resorption in permanent incisors in relation to canine impactions.
Jawad et al. (2016)
Diagnostic efficacy level	Level 3: Diagnostic thinking efficacy
Imaging (index tests)	CBCT: “The majority of the CBCT images were taken with an OP300 machine” with “small volume” Conventional radiography varied between patients but most only had a panoramic radiograph
Reference standard	Not applicable
Patient sample description	35 patient cases, over a 1-year period, in which CBCT imaging was taken to assess root resorption associated with impacted canines. 42 canines in sample, 40 in maxilla and 2 in the mandible Ages: not specified, gender: not specified, setting: University-based secondary care specialist centre in the UK
Key outcomes	Root resorption observed on 63% of cases using CBCT and 19% of cases using radiographs Of 14 cases judged not to be resorbed on radiographs, 5 had root resorption on viewing CBCT
Study strengths	Assessment of intra-observer repeatability made
Study weaknesses	Retrospective study Potential selection bias (inclusion of patients who had been a priori chosen for CBCT) Lack of detail on the patient sample Variable conventional imaging Lack of detail on conduct of index tests

Clinical context(s)	Pathological conditions	Purpose of imaging	As an aid to treatment planning for external cervical resorption (ECR)
Goodell et al. (2018)
Diagnostic efficacy level	Level 4: therapeutic efficacy
Imaging (index tests)	CBCT [3D Accuitomo 170; FoV = 40 × 40 mm; voxel size 0.08 mm] Intraoral radiographs (digital, CCD sensor); unclear how many and which type Six examiners (2 specialist endodontists, 2 senior endodontic residents, 2 junior endodontic residents)
Reference standard	Not applicable
Patient sample description	25 patients with 30 teeth referred for management of ECR and who had also undergone a CBCT examination. Ten “control” teeth of unspecified origin. Ages: not specified, gender: not specified, setting: specialist centre for endodontics in a United States Army facility
Key outcomes	All 30 ECR cases were identified using CBCT imaging and 29 using periapical radiography Inter-rater agreement higher for CBCT Individual treatment plans changed in 56.7% of cases using CBCT Consensus decisions on dichotomised treatment plan (“repair” versus “no repair”) changed in six out of the 30 cases (20%)
Study strengths	Clinical scenario provided to observers Specified “wash-out” time period between viewing radiographs and CBCT
Study weaknesses	Retrospective design Risk of selection bias “Composite” consensus scores presented from examiners, for some aspects, but method of consensus unclear Presentation of diagnostic accuracy results in absence of any reference standard being specified (results not presented here as this failed a review inclusion criterion)

Clinical context(s)	Other uses	Purpose of imaging	Forensic identification by recording teeth present and absent, dental restorations (extent and material), impacted teeth, any pathosis
Murphy et al. (2012)
Diagnostic efficacy level	Level 2: diagnostic accuracy
Imaging (index tests)	CBCT [iCAT Classic. FoV varied; voxel size varied] One forensic odontologist performed assessments, including repeated assessment for intra-observer reliability. A second person assessed a small sub-sample for inter-observer reliability.
Reference standard	Panoramic radiograph [Sirona Orthophos CD]
Patient sample description	30 patients who had both panoramic radiographs and CBCT examinations, consisting of 10 in each of three age cohorts (data for the ≤ 17 years cohort only considered here) Ages: not specified other than which age cohort (≤ 17  years), Gender: not specified, Setting: University-based secondary care facility in the UK
Key outcomes	Information could be collected accurately and reliably using CBCT, compared with using panoramic radiographs CBCT: Se = 83.3% (95% CI 78.3–88.3); Sp = 100%; PPV = 100%; NPV = 99.5% (93.5–100)
Study strengths	Clear statistical presentation Assessment of inter- and intra-observer reliability
Study weaknesses	Retrospective study Small sample size Possible selection bias Low prevalence of positive findings to record Variable time gap between the index tests Use of a panoramic radiograph as the reference standard. May be justified as it is an existing clinical standard, but not an adequate diagnostic “truth” for diagnostic accuracy Single observer provided the main data

FoV field of view of CBCT, M male, F female, Se sensitivity, Sp specificity, Acc accuracy, PPV positive predictive value, NPV negative predictive value, PG postgraduate, sd standard deviation

^aWriedt et al. (2017) presented results of actual treatment which might best be described as “prognostic efficacy”, but which are included here as an aspect of therapeutic efficacy