A content-aware chatbot based on GPT 4 provides trustworthy recommendations for Cone-Beam CT guidelines in dental imaging.

OBJECTIVES: To develop a content-aware chatbot based on GPT-3.5-Turbo and GPT-4 with specialized knowledge on the German S2 Cone-Beam CT (CBCT) dental imaging guideline and to compare the performance against humans. METHODS: The LlamaIndex software library was used to integrate the guideline context into the chatbots. Based on the CBCT S2 guideline, 40 questions were posed to content-aware chatbots and early career and senior practitioners with different levels of experience served as reference. The chatbots' performance was compared in terms of recommendation accuracy and explanation quality. Chi-square test and one-tailed Wilcoxon signed rank test evaluated accuracy and explanation quality, respectively. RESULTS: The GPT-4 based chatbot provided 100% correct recommendations and superior explanation quality compared to the one based on GPT3.5-Turbo (87.5% vs. 57.5% for GPT-3.5-Turbo; P = .003). Moreover, it outperformed early career practitioners in correct answers (P = .002 and P = .032) and earned higher trust than the chatbot using GPT-3.5-Turbo (P = 0.006). CONCLUSIONS: A content-aware chatbot using GPT-4 reliably provided recommendations according to current consensus guidelines. The responses were deemed trustworthy and transparent, and therefore facilitate the integration of artificial intelligence into clinical decision-making.

Vertical head rotation as major source of differences between time-separated digital cephalometric radiographs of patients acquired in one cephalostat X-ray device.

BACKGROUND: The purpose of the retrospective study was to analyze the reliability and repeatability of specific landmark-positions used in cephalometry to determine the major sources of absolute landmark position differences for repeated, time-separated (time-point I and II) digital cephalometric radiographs (CEPH) of the same patients. METHODS: 100 pairs of CPEHs from the database of a sample of adult patients (18 to 28 years) were analyzed by one calibrated observer and three landmark points (Sella: S, Nasion: N, Subspinale: A) were digitally marked using ImageJ-software. The coordinates of these points entered the evaluation using displacement vectors as primary endpoints between the coordinates of the landmarks in the two images as well as SNA-angles and the angle ω of SN relative to the floor. RESULTS: Displacement vectors between CEPHI and CEPHII were rather large (N: 7.95 ± 4.85 mm, S: 5.34 ± 3.50 mm, A: 4.81 ± 3.95 mm. SNA was rather stable between the two sequential radiographs (mean difference: 0.002° ± 1.85°). and did not correlate with age of the patient (SNAI: spearman-Rho: 0.0239, p = 0.8134; SNAII : spearman-Rho: 0.0244, p = 0.8096). Although the vertical angle ω did not differ between CEPHI and CEPHII (mean difference: 0.4° ± 4.7°, pwilcoxon = 0.8155), it showed a quadratic relationship (pF-statistic: < 2.2e-16) with the length of the displacement vector N. CONCLUSION: The significantly varying location of the reference points S, N and A between time-separated CEPHs of one patient can largely be explained by different angulation (head rotation within the sagittal plane) of the Frankfurt plane to the floor (horizontal plane).

Fractal dimension in CBCT images as predictor for MRONJ: a retrospective cohort study.

OBJECTIVE: To estimate boxcounting fractal dimension in a standardized approach in CBCT images of the mandible and maxilla in a group of patients with MRONJ in comparison to a healthy control group. MATERIAL AND METHODS: From our records, a study group consisting of 80 maxillary and mandibular CBCTs in 77 cases of patients having MRONJ was collected. The control group consisted of 84 mandibular and maxillary CBCTs in a total of 78 patients. Using the boxcounting method, fractal dimension (FD) was estimated in a standardized fashion either cranially to the apex of the canine (maxilla) or beneath the tooth apices of the lower molars in the bone area above the mandibular canal (mandible). Intra-observer reproducibility of the FD-measurements was assessed by 6 repeated measurements in 10 individuals. FD values were correlated to age and sex of the individuals as well as to region of interest (ROI) sizes. RESULTS: FD in the study group (1.684 ± 0.051) was roughly 3.5 % lower than in the control group (1.745 ± 0.026, p < 0.0001). Sex and age had a significant (p < 0.001) influence on FD values in the study group, yet not in the control group. FD values increased with age (study group, spearman-rho: 0.2895, p < 0.05) and also ROI size (both groups, p < 0.0001). Reproducibility was good (intra-class correlation coefficient (ICC): 0.87). CONCLUSIONS: Fractal dimension as assessed by boxcounting seems to be a good descriptor for MRONJ in jaw bones. Influence of age and sex on the outcome values needs to be further investigated in future studies. CLINICAL RELEVANCE: CBCTs could be assessed with respect to FD to obtain an overview of the disease status of MRONJ patients.

Accuracy in Detecting Artificial Root Resorption in Panoramic Radiography versus Tomosynthetic Panoramic Radiographs.

INTRODUCTION: The aim of this study was to compare the detection accuracy of panoramic radiography (PAN) and tomosynthetically reconstructed panoramic radiography (TPAN) for the detection of artificial external root resorption in a multiobserver approach. METHODS: Thirty-six teeth in 5 dry human mandibles were prepared with artificial root resorption by means of diamond bur defects (0.8, 1.0, and 1.2 mm). The samples were exposed in a digital panoramic radiographic unit in a predetermined appropriate position using a wax layer (6.25 mm thickness) as soft tissue scatter equivalent and a water-filled plastic bottle to mimic the absorption of the cervical spine. This resulted in 5 panoramic radiographs and 5 tomosynthetically reconstructed panoramic radiographs, which were evaluated by 13 observers regarding the visibility of artificial root resorption by means of a 5-point confidence scale. Seven of these observers repeated the process after a minimum interval of 30 days. A receiver operating characteristic analysis was conducted with the area beneath the receiver operating characteristic curves (Az) as the main accuracy parameter. Inter- and intrarater reproducibility were calculated by means of the intraclass coefficient using a 2-way random effects model. RESULTS: The mean Az for TPAN (0.76; median = 0.77; range, 0.70-0.85) was slightly yet significantly higher (P < .05, Wilcoxon test) than for PAN (0.75; median = 0.75; range, 0.69-0.82). The Az values for both methods were highest in the premolar and lowest in the molar region. The mean sensitivity for TPAN was 0.54 (specificity = 0.96) and 0.50 (specificity = 0.96) for PAN. Intraclass coefficient values indicated that intra- (PAN: mean = 0.53 ± 0.088; TPAN: mean = 0.55 ± 0.102; P < .05,Wilcoxon test) and interrater (PAN: mean = 0.47 [0.43-0.51], TPAN: mean = 0.47 [0.42-0.51]) reproducibility were both moderate. CONCLUSIONS: From our ex vivo study, we observed slightly higher accuracy in the detection of artificial root resorption from tomosynthetically reconstructed panoramic radiographs compared with conventional digital panoramic radiographs.

Surgical evaluation of panoramic radiography and cone beam computed tomography for therapy planning of bisphosphonate-related osteonecrosis of the jaws.

OBJECTIVES: The purpose of this retrospective study was to compare the clinical significance of panoramic radiography (orthopantomography [OPTG]) and cone beam computed tomography (CBCT) for therapy planning of bisphosphonate-related osteonecrosis of the jaw (BRONJ) by surgeons. STUDY DESIGN: Using standardized questionnaire, eight maxillofacial surgeons evaluated intraoral photographs of the clinical situation of 14 patients with BRONJ as well as the corresponding radiographic images (OPTG, CBCT). The presence of five typical BRONJ signs (bone-remodeling, periosteal reaction, osteosclerosis, sequestra, and continuity of cortical bone) was evaluated with OPTG and CBCT. The influence of radiologic information on therapy decision was examined as well. RESULTS: On the basis of the information from intraoral photographs only, seven of the eight surgeons indicated that an additional radiographic examination was necessary for further therapy planning. For evaluation of the five radiographic BRONJ signs, CBCT provided significantly better values compared with OPTG (all P < .05). A before-and-after comparison showed that four of the eight surgeons changed their therapy concept after having three-dimensional CBCT information. The majority (6 of 8) of the surgeons considered that an additional CBCT was required for therapy planning, even after having studied the clinical photographs and OPTG images. CONCLUSION: These data demonstrate a significant advantage of CBCT over OPTG for surgeons with regard to therapeutic planning for BRONJ.

Implant treatment planning regarding augmentation procedures: panoramic radiographs vs. cone beam computed tomography images.

OBJECTIVES: To evaluate the impact of cone beam computed tomography (CBCT) imaging on treatment planning regarding augmentation procedures for implant placement. MATERIAL AND METHODS: Panoramic radiographs and CBCT images of 40 patients requesting single-tooth implants in 59 sites were retrospectively analyzed by six specialists in implantology, and treatment planning was performed. Therapeutic recommendations were compared with the surgical protocol performed initially. RESULTS: Bone height estimation from panoramic radiographs yielded to higher measures and greater variability compared to CBCT. The suggested treatment plan for lateral and vertical augmentation procedures based on CBCT or panoramic radiographs coincided for 55-72% of the cases. A trend to a more invasive augmentation procedure was seen when planning was based on CBCT. Panoramic radiography revealed 57-63% (lateral) vs. 67% (vertical augmentation) congruent plans in agreement with surgery. Among the dissenting sites, there was a trend toward less invasive planning for lateral augmentation with panoramic radiographs, while vertical augmentation requirements were more frequently more invasive when based on CBCT. CONCLUSIONS: Vertical augmentation requirements can be adequately determined from panoramic radiographs. In difficult cases with a deficient lateral alveolar bone, the augmentation schedule may better be evaluated from CBCT to avoid underestimation, which occurs more frequently when based on panoramic radiographs only. However, overall, radiographic interpretation and diagnostic thinking accuracy seem to be mainly depending on the opinion of observers.

Automated detection of patient movement during a CBCT scan based on the projection data.

OBJECTIVES: To develop an automated procedure to detect patient motion on the projection images acquired during a cone beam computed tomography (CBCT) scan and to evaluate the method's feasibility on small real-world CBCT images in relation to visual assessment. METHODS: Based on optical flow theory, software was developed using the sequence of the projection images of a CBCT machine for automated detection of patient motion. Averaged acceleration vectors were used as measurement data and compared with visual assessment of the projection images displayed as video. Seventy-nine CBCT data sets (small field-of-view: 40 mm) from our patient database were selected in a sequential fashion and evaluated with the software. RESULTS: 10 out of 79 (13%) were allocated to a patient movement. A threshold of 0.4 pixel/frame transition was empirically determined as indicating motion by visual assessment of the image sequence. Relative to this standard of reference, the software reached 80% sensitivity versus 67% specificity. CONCLUSIONS: Optical flow seems to be an efficient concept for automated detection of patient motion on the projection images acquired during a CBCT scan.

A contact-free volumetric measurement of facial volume after third molar osteotomy: proof of concept.

OBJECTIVE: The present study tested the reliability of an optical scanning device for the objective assessment of postoperative facial swelling. STUDY DESIGN: Twenty control subjects bearing a defined volume of water (10-30 mL) in an intraorally carried balloon were tested to assess the measurement accuracy of the device. As a proof of concept, facial volumes of 59 surgical cases were recorded before osteotomy and 1 and 7 days after intervention with the use of a structured light scanner. RESULTS: The median difference between the applied and the measured volumes was 0.67 mL for the control test with the artificial swelling simulated using water balloons. For subjects having third molar osteotomy, extraoral volume increased to 5.29 cm(3) 1 day after surgery (95% CI 5.22-8.52) and decreased to 0.00 mL (95% CI 0.85-2.55) after 7 days. CONCLUSIONS: Contact-free visible-light 3-dimensional scanning is reliable for the objective assessment of postoperative facial swelling.

Effects of minute misregistrations of prefabricated markers for image-guided dental implant surgery: an analytical evaluation.

OBJECTIVES: The goal of the present study was to develop a theoretical analysis of errors in implant position, which can occur owing to minute registration errors of a reference marker in a cone beam computed tomography volume when inserting an implant with a surgical stent. MATERIAL AND METHODS: A virtual dental-arch model was created using anatomic data derived from the literature. Basic trigonometry was used to compute effects of defined minute registration errors of only voxel size. The errors occurring at the implant's neck and apex both in horizontal as in vertical direction were computed for mean ±95%-confidence intervals of jaw width and length and typical implant lengths (8, 10 and 12 mm). RESULTS: Largest errors occur in vertical direction for larger voxel sizes and for greater arch dimensions. For a 10 mm implant in the frontal region, these can amount to a mean of 0.716 mm (range: 0.201-1.533 mm). Horizontal errors at the neck are negligible, with a mean overall deviation of 0.009 mm (range: 0.001-0.034 mm). Errors increase with distance to the registration marker and voxel size and are affected by implant length. CONCLUSION: Our study shows that minute and realistic errors occurring in the automated registration of a reference object have an impact on the implant's position and angulation. These errors occur in the fundamental initial step in the long planning chain; thus, they are critical and should be made aware to users of these systems.

Diagnostic yield of ink-jet prints from digital radiographs for the assessment of approximal carious lesions: ROC-analysis.

AIMS: To investigate the diagnostic quality of different quality, individually calibrated ink-jet printers for the very challenging dental radiographic task of approximal carious lesion detection. MATERIALS AND METHODS: A test-pattern evaluating resolution, contrast and homogeneity of the ink-jet prints was developed. 50 standardized dental radiographs each showing two neighbouring teeth in natural contact were printed on glossy paper with calibrated, randomly selected ink-jet printers (Canon S520 and iP4500, Epson Stylus Photo R2400). Printing size equalled the viewing size on a 17″ cathode-ray-tube monitor daily quality-tested according to German regulations. The true caries status was determined from serial sectioning and microscopic evaluation. 16 experienced observers evaluated the radiographs on a five-point confidence scale on all prints plus the viewing monitor with respect to the visibility of a carious lesion. A non-parametric Receiver-Operating Characteristics (ROC-) analysis was performed explicitly designed for the evaluation of readings stemming from identical samples but different modality. Significant differences are expressed by a critical ratio z exceeding ±2. Diagnostic accuracy was determined by the area (Az) underneath the ROC-curves. RESULTS: Average Az-values ranged between 0.62 (S520 and R2400) and 0.64 (monitor, iP4500), with no significant difference between modalities (P=0.172). Neither significant (range mean z: -0.40 (S520) and -0.11 (iP4500)) nor clinically relevant differences were found between printers and viewing monitor. CONCLUSIONS: Our results for a challenging task in dental radiography indicate that calibrated, off-the-shelf ink-jet printers are able to reproduce (dental) radiographs at quality levels sufficient for radiographic diagnosis in a typical dental working environment.

Epulis granulomatosa as an oral manifestation of Klippel-Trénaunay syndrome.

The Klippel-Trénaunay syndrome (KTS) was first described by Klippel and Trénaunay in 1900. It is characterized by the triad of hemihypertrophy of soft and hard tissue, naevus flammeus and venous varicosity in the affected area. Though all oral tissues may be affected, only 5% of KTS show manifestations in the head and neck region. Only three cases are described with an oral manifestation, showing gingival overgrowth clinically and histologically corresponding to a pyogenic granuloma. It is still uncertain whether the combination of gingival fibromatosis and KTS is significant or coincidental. We report about a 25-year-old patient with KTS and recidivous gingival fibromatosis, clinically and histologically corresponding to an epulis fibromatosa in a case report. It is suggested that this occurrence is significant.

B-mode versus A-mode ultrasonographic measurements of mucosal thickness in vivo.

OBJECTIVE: The objective of this study was to compare the measurement accuracy of A-mode and B-mode ultrasonographic assessments of palatal masticatory mucosal thickness in vivo. STUDY DESIGN: Thickness of palatal masticatory mucosa in the molar region of 50 volunteers was measured once with a B-mode ultrasound device (10 MHz) combined with a standoff for intraoral applications. At identical reading points thickness was also assessed with an A-mode device (5 replicates) and by needle probing for determination of true thickness. RESULTS: On average true thickness ranged between 2.9 and 3.3 mm (95% confidence intervals; median, 3.1 mm). B-mode ultrasonography overestimated truth by 0.19 mm (median) and differed significantly from truth (P =.008), whereas A-mode ultrasonography slightly underestimated true thickness by -0.05 mm (median). Limits of agreement were wider for A-mode ultrasonography (2.00 mm) than for B-mode ultrasonography (0.88 mm). CONCLUSIONS: Because accuracy of 0.2 mm is acceptable for many clinical indications, B-mode and A-mode yield sufficient accuracy for measurements of mucosal thickness.