Artificial intelligence in diagnosis of maxillary sinusitis: A clinical study

Significance: Pressure injury is one of the most common health problems among hospitalized patients worldwide, and accurate and timely diagnosis is crucial for its treatment. Research on the application of artificial intelligence in the diagnosis of pressure injury is increasing, but there is currently no comprehensive meta-analysis to evaluate the accuracy of artificial intelligence in diagnosing different pressure injury stages. Recent Advances: This study synthesizes evidence on artificial intelligence diagnosis of pressure injury, focusing on evaluating diagnostic performance across different stages using core metrics including sensitivity, specificity, and the area under the summary receiver operating characteristic (SROC) curve. Critical Issues: Key findings from 21 included studies (12 contributing 47 eligible datasets) indicate high overall diagnostic accuracy of artificial intelligence for pressure injury, with sensitivity of 0.74 (95% confidence interval [CI]: 0.69-0.78), specificity of 0.93 (95% CI: 0.91-0.94), and area under the SROC curve of 0.92 (95% CI: 0.90-0.94). Moreover, the area under the SROC curve varies across different stages of pressure injury, with area under the curve values for stage 1, stage 2, stage 3, stage 4, unstageable, and deep tissue pressure injury of 0.95 (0.93-0.97), 0.85 (0.82-0.88), 0.88 (0.84-0.90), 0.94 (0.92-0.96), 0.96 (0.94-0.97), and 0.98 (0.96-0.99), respectively. Future Directions: Artificial intelligence models based on pressure injury image data show substantial potential for clinical application in pressure injury diagnosis. However, the need for high-quality studies with rigorous reporting and external validation remains critical to address current limitations and advance clinical translation.

This systematic review evaluated the application of artificial intelligence (AI) in the imaging-based diagnosis of cholesteatoma. A comprehensive search of PubMed, Scopus, and Web of Science up to September 2025 identified 8,160 records, of which 7 studies met the inclusion criteria. Five studies used temporal bone CT, one investigated otoscopic images, and one assessed CT-based staging of mastoid extension. Reported internal validation accuracies were generally high (>90%), particularly with convolutional neural network (CNN) architectures, while external validations showed moderate performance (78-88%). A multicenter 3D convolutional neural network (CNN) with automated region-of-interest detection demonstrated consistent generalizability and aided surgical planning in prospective testing. The otoscopy-based model achieved high accuracy for differentiating cholesteatoma from normal membranes (>98%), though performance decreased when distinguishing it from other middle ear pathologies. Across studies, AI models often performed comparably to human readers, and explainability tools highlighted relevant diagnostic features. Most studies were retrospective and single-center, with limited external validation. Overall, AI shows strong potential for cholesteatoma diagnosis using CT and otoscopic imaging, but multicenter prospective studies with standardized evaluation and clinical impact assessment are needed before routine implementation.

To investigate the use of transfer learning when applying a deep learning source model from one institution (institution A) to another institution (institution B) for creating effective models (target models) for the detection of maxillary sinuses and diagnosis of maxillary sinusitis on panoramic radiographs. In addition, to determine appropriate numbers of training data for the transfer learning. Source model was created using 350 panoramic radiographs from institution A as training data. Transfer learning was performed by adding 25, 50, 100, 150, or 225 panoramic radiographs as training data from institution B to the source model; this yielded the target models T25, T50, T100, T150 and T225. Each model was then evaluated using test data that comprised 40 images from institution A, 30 images from institution B. The performance indices (recall, precision and F1 score) for detecting the maxillary sinuses by the source model exceeded 0.98 when using test data A from institution A, but they deteriorated when using test data B from institution B. In the evaluation of target models using test data B, model T25 showed improved detection performance (recall of 0.967). The diagnostic performance of model T50 for maxillary sinusitis exceeded 0.9 in sensitivity. Transfer learning, which involves applying a small amount of data to the source model, yielded high performances in detecting the maxillary sinuses and diagnosing the maxillary sinusitis on panoramic radiographs. This study serves as a reference when adapting source models to other institutions.

Craniopharyngioma is a congenital brain tumor with clinical characteristics of hypothalamic-pituitary dysfunction, increased intracranial pressure, and visual field disorder, among other injuries. Its clinical diagnosis mainly depends on radiological examinations (such as Computed Tomography, Magnetic Resonance Imaging). However, assessing numerous radiological images manually is a challenging task, and the experience of doctors has a great influence on the diagnosis result. The development of artificial intelligence has brought about a great transformation in the clinical diagnosis of craniopharyngioma. This study reviewed the application of artificial intelligence technology in the clinical diagnosis of craniopharyngioma from the aspects of differential classification, prediction of tissue invasion and gene mutation, prognosis prediction, and so on. Based on the reviews, the technical route of intelligent diagnosis based on the traditional machine learning model and deep learning model were further proposed. Additionally, in terms of the limitations and possibilities of the development of artificial intelligence in craniopharyngioma diagnosis, this study discussed the attentions required in future research, including few-shot learning, imbalanced data set, semi-supervised models, and multi-omics fusion.

The application of artificial intelligence in the diagnosis of obstructive lung diseases is an exciting phenomenon. Artificial intelligence algorithms work by finding patterns in data obtained from diagnostic tests, which can be used to predict clinical outcomes or to detect obstructive phenotypes. The purpose of this review is to describe the latest trends and to discuss the future potential of artificial intelligence in the diagnosis of obstructive lung diseases. Machine learning has been successfully used in automated interpretation of pulmonary function tests for differential diagnosis of obstructive lung diseases. Deep learning models such as convolutional neural network are state-of-the art for obstructive pattern recognition in computed tomography. Machine learning has also been applied in other diagnostic approaches such as forced oscillation test, breath analysis, lung sound analysis and telemedicine with promising results in small-scale studies. Overall, the application of artificial intelligence has produced encouraging results in the diagnosis of obstructive lung diseases. However, large-scale studies are still required to validate current findings and to boost its adoption by the medical community.

The use of ultrasonography in the diagnosis of maxillary sinusitis in pediatric patients has been reported recently because of the improvement of the accuracy of ultrasound technology. We thus compared B-mode ultrasonography and computed tomography in the diagnosis of maxillary sinusitis in pediatric patients. Thirty-six maxillary sinuses in 18 patients (10 females, 8 males, ages ranging from 7-15 years with an average age of 10.4 years) were examined. Ultrasonography of the maxillary sinus was performed in the horizontal and the vertical direction. Paranasal computed tomography and B-mode ultrasonography were performed within a few days. In some of these patients the maxillary sinuses were examined with a fiberscope. Sensitivity, specificity, false-positive, false-negative, positive predictive value and negative predictive value of B-mode ultrasonography compared with computed tomography were 92.6%, 100%, 0%, 7.4%, 100% and 81.8%, respectively. It appeared that ultrasonography was more sensitive than X-ray imaging, because the sensitivity and specificity of X-ray imaging of the maxillary sinus in pediatric patients compared with CT was reportedly 70-80%. A meaningful correlation of ultrasonography and CT was accepted as an assessment of desease severity. There are some problems with diagnosis by ultrasonography. There is no differentiation of mucosal thicking, cyst and discharge and imaging are less useful in pediatric patients. Because of these reasons, clinical sign and views in the nose are important for a correct diagnosis in pediatric patients. Furthermore, the most suitable age range to diagnose maxillary sinusitis correctly in pediatric patients must be examined.

Ultrasound is one of the methods proposed by researchers to investigate paranasal sinus disease conditions like rhinosinusitis. This method, unlike Computed Tomographic (CT) scan, does not involve exposure to radiation, and it is cost effective. Moreover, an ultrasound scan is safe, readily available, and non-invasive compared to the sinus puncture method, and it is easily accepted by patients. However, there is no enough information on the utilization and accuracy of this method in the diagnosis of maxillary sinusitis. The aim of this review is to investigate and elaborate on the role of ultrasound scan in the diagnosis of maxillary sinusitis. We conducted a comprehensive literature search across multiple research databases, including PubMed, Scopus, Web of Science, Google Scholar, Embase, EBSCO and Cochrane database. The result indicated that both A-mode and B-mode ultrasound scans were accurate in the diagnosis of maxillary sinusitis, with excellent sensibility, specificity, positive and negative predictive values compared with Magnetic Resonant Imaging (MRI), CT scan and sinus endoscopy. Therefore, we recommend the use of ultrasound scan for the diagnosis of maxillary sinusitis, especially in our environment where there are factors that limit the utilization of CT scan and MRI (such as repeated breakdown of the machines, high cost, and lack of good electricity).

Real-time Intraoperative Computed Tomography Assessment of Quality of Permanent Interstitial Seed Implantation for Prostate Cancer

As the most common cancer after lung cancer presently, breast cancer needs to gain extensive awareness of both general public and healthcare professionals. With continuous developments of technology researchers on medical science began to attach great significance to the integration of artificial intelligence (AI) and clinical diagnostics. With the purpose of enhancing early diagnosis and prognosis and implementing more effective and timely treatment for breast cancer, increased research efforts are currently aimed at applying AI to breast cancer detection and classification, opening up new avenues for improving patient care. The paper initially discusses the progress of AI in the diagnosis and prognosis of breast cancer and presents an overview of the existing literature of research in this area. It describes the main methods and procedures employed in AI-assisted diagnosis and response prediction with particular emphasis on radiological as well as pathological approaches. Comparative analysis is also provided comparing differences in sensitivity and specificity between clinicians and AI systems when interpreting images. Furthermore, the paper discusses existing challenges in the introduction of AI in clinical practice to assist clinicians’ diagnosis and carry out treatment. The major challenges include requirements for data volumes, standardization and ethical problems. Then, it predicted the future progress of AI technologies in diagnostics and prognosis of breast cancer, providing practical suggestions on promoting applications of AI in diagnosis, therapy and general use.

BackgroundSince it is difficult for expert endoscopists to diagnose early gastric cancer in post‐eradication status, it may be critical to evaluate the stratification of high‐risk groups using the advance of gastric atrophy or intestinal metaplasia. We tried to determine whether the combination of endoscopic artificial intelligence (AI) diagnosis for the evaluation of gastric atrophy could be a useful tool in both pre‐ and post‐eradication status.Methods270 Helicobacter pylori‐positive outpatients (Study I) were enrolled and Study II was planned to compare patients (n = 72) with pre‐eradication therapy with post‐eradication therapy. Assessment of endoscopic appearance was evaluated by the Kyoto classification and Kimura‐Takemoto classification. The trained neural network generated a continuous number between 0 and 1 for gastric atrophy.ResultsThere were significant associations between the severity of gastric atrophy determined by AI endoscopic diagnosis and not having a regular arrangement of collecting venules in angle, visibility of vascular pattern, and mucus using Kyoto classification in H. pylori‐positive gastritis. There were significant differences (p = 0.037 and p = 0.014) in the severity of gastric atrophy between the high‐risk group and low‐risk group based on the combination of Kimura‐Takemoto classification and endoscopic AI diagnosis in pre‐ and post‐eradication status. The area under the curve values of the severity of gastric atrophy (0.674) determined by the combination of Kimura‐Takemoto classification and gastric atrophy determined by AI diagnosis was higher than that determined by Kimura‐Takemoto classification alone in post‐eradication status.ConclusionA combination of gastric atrophy determined by AI endoscopic diagnosis and Kimura‐Takemoto classification may be a useful tool for the prediction of high‐risk groups in post‐eradication status.

Objectives:The optimal goal of radiography is to provide high-quality diagnostic images with the least patient radiation dose. The aim of this study was to evaluate the accuracy of cone-beam computed tomography (CBCT) and intraoral photostimulable phosphor plate (PSP) digital and film-based conventional radiography for detection of artificial cancellous bone defects.Materials and Methods:Five dry human mandibles were used in this study. The mandibles were placed inside a water bath made of plexiglass plates; then PSP and CBCT scans were obtained. The mandibles were cut by a coping saw in buccolingual dimension and oval defects measuring 6.1×6.1 mm, 3×6.1 mm and 4×4 mm were created by a milling machine in the spongy bone. After fixing the two parts together (buccal and lingual), radiographs were repeated. Presence or absence of defects on images was evaluated and recorded by the two observers. Using SPSS 16, compatibility level, sensitivity, specificity and receiver operating curve (ROC) analysis were determined for each observer.Results:The intraobserver agreement in all three imaging modalities was low to moderate (kappa≤0.613). The inter-observer agreement in all the three imaging modalities was moderate (kappa=0.406). The area under the curve (AUC) of the imaging modalities in each observation was not significantly different. The area under the curve based on defect size for the two observers was not significantly different either.Conclusion:Defects confined to spongy bone can be identified on film and PSP radiographs and CBCT scans. However, interpretation of PSP images and CBCT scans needs greater expertise and skills.

Traumatic Brain Injury (TBI) has been increasingly recognized as a leading cause of death and disability worldwide.ObjectiveTo summarize clinical applications of artificial intelligence, including machine learning and deep learning, in the diagnosis and prognosis of traumatic brain injury.MethodsThe authors conducted a scoping review of original clinical research studies on humans published in English after January 1, 2014. A search was performed using PubMed, including PMC, MEDLINE, and Bookshelf. The search terms were applied to the title field and included: (TBI) AND (Artificial Intelligence OR Machine Learning OR Deep Learning). Studies meeting inclusion criteria were screened and selected for review. The reference lists of the included studies were also screened to identify any additional eligible articles.ResultsOf 493 studies identified, seven met the inclusion criteria and were included in the analysis, which summarizes study title, publication year, study objective, key findings, and conclusions.ConclusionArtificial intelligence shows promise in aiding diagnosis and improving prognostic insights in traumatic brain injury. Although few clinical trials have been conducted, early results are encouraging. Future progress will require more clinical studies and efforts to address the current limitations of AI tools in medicine.

Comparison between effective radiation dose of CBCT and MSCT scanners for dentomaxillofacial applications

Our aim was to evaluate the effects of balloon sinuplasty on the size of the ostium in the maxillary sinuses in patients with chronic rhinosinusitis from cone beam computer tomography (CBCT) scans of the sinus. This is a blinded retrospective trial in patients who had undergone balloon sinuplasty of the maxillary sinus. CBCT scans were taken and SNOT-22 Quality of Life questionnaire completed before and 12 months after the operation. The size of the maxillary ostium was measured from the CBCT scans three-dimensionally. The association of changes in the SNOT-22 scores of the ostium was analysed. We discovered that the balloon sinuplasty increased the size of the maxillary ostium in all dimensions. The changes were statistically significant (p<0.05) in the axial diameter and the ostium area. The number of patent ostia increased after the intervention. The association between SNOT-22 score and ostium patency were statistically significant before the operation. Our conclusion is that the threedimensional measuring technique provides a reliable method to evaluate ostium dimensions. Balloon sinuplasty increased the size of the maxillary ostium and the result was maintained for 12 months after the operation.

BackgroundThe emergence of robotic Cone Beam Computed Tomography (CBCT) imaging systems in trauma departments has enabled 3D anatomical assessment of musculoskeletal injuries, supplementing conventional 2D fluoroscopic imaging for examination, diagnosis, and treatment planning. To date, the primary focus has been on trauma sites in the extremities.PurposeTo determine if CBCT images can be used during the treatment planning process in spinal instrumentation and laminectomy procedures, allowing accurate 3D‐printed pedicle screw and laminectomy drill guides to be generated for the cervical and thoracic spine.MethodsThe accuracy of drill guides generated from CBCT images was assessed using animal cadavers (ovine and porcine). Preoperative scans were acquired using a robotic CBCT C‐arm system, the Siemens ARTIS pheno (Siemens Healthcare, GmbH, Germany). The CBCT images were imported into 3D‐Slicer version 4.10.2 (www.slicer.org) where vertebral models and specific guides were developed and subsequently 3D‐printed. In the ovine cadaver, 11 pedicle screw guides from the T1–T5 and T7–T12 vertebra and six laminectomy guides from the C2–C7 vertebra were planned and printed. In the porcine cadaver, nine pedicle screw guides from the C3–T4 vertebra were planned and printed. For the pedicle screw guides, accuracy was assessed by three observers according to pedicle breach via the Gertzbein–Robbins grading system as well as measured mean axial and sagittal screw error via postoperative CBCT and CT scans. For the laminectomies, the guides were designed to leave 1 mm of lamina. The average thickness of the lamina at the mid‐point was used to assess the accuracy of the guides, measured via postoperative CBCT and CT scans from three observers. For all measurements, the intraclass correlation coefficient (ICC) was calculated to determine observer reliability.ResultsCompared with the planned screw angles for both the ovine and porcine procedures (n = 32), the mean axial and sagittal screw error measured on the postoperative CBCT scans from three observers were 3.9 ± 1.9° and 1.8 ± 0.8°, respectively. The ICC among the observes was 0.855 and 0.849 for the axial and sagittal measurements, respectively, indicating good reliability. In the ovine cadaver, directly comparing the measured axial and sagittal screw angle of the visible screws (n = 14) in the postoperative CBCT and conventional CT scans from three observers revealed an average difference 1.9 ± 1.0° in axial angle and 1.8 ± 1.0° in the sagittal angle. The average thickness of the lamina at the middle of each vertebra, as measured on‐screen in the postoperative CBCT scans by three observes was 1.6 ± 0.2 mm. The ICC among observers was 0.693, indicating moderate reliability. No lamina breaches were observed in the postoperative images.ConclusionHere, CBCT images have been used to generate accurate 3D‐printed pedicle screw and laminectomy drill guides for use in the cervical and thoracic spine. The results demonstrate sufficient precision compared with those previously reported, generated from standard preoperative CT and MRI scans, potentially expanding the treatment planning capabilities of robotic CBCT imaging systems in trauma departments and operating rooms.