Summary Background Deep learning offers considerable promise for medical diagnostics. We aimed to evaluate the diagnostic accuracy of deep learning algorithms versus health-care professionals in classifying diseases using medical imaging. Methods In this systematic review and meta-analysis, we searched Ovid-MEDLINE, Embase, Science Citation Index, and Conference Proceedings Citation Index for studies published from Jan 1, 2012, to June 6, 2019. Studies comparing the diagnostic performance of deep learning models and health-care professionals based on medical imaging, for any disease, were included. We excluded studies that used medical waveform data graphics material or investigated the accuracy of image segmentation rather than disease classification. We extracted binary diagnostic accuracy data and constructed contingency tables to derive the outcomes of interest: sensitivity and specificity. Studies undertaking an out-of-sample external validation were included in a meta-analysis, using a unified hierarchical model. This study is registered with PROSPERO, CRD42018091176. Findings Our search identified 31 587 studies, of which 82 (describing 147 patient cohorts) were included. 69 studies provided enough data to construct contingency tables, enabling calculation of test accuracy, with sensitivity ranging from 9·7% to 100·0% (mean 79·1%, SD 0·2) and specificity ranging from 38·9% to 100·0% (mean 88·3%, SD 0·1). An out-of-sample external validation was done in 25 studies, of which 14 made the comparison between deep learning models and health-care professionals in the same sample. Comparison of the performance between health-care professionals in these 14 studies, when restricting the analysis to the contingency table for each study reporting the highest accuracy, found a pooled sensitivity of 87·0% (95% CI 83·0–90·2) for deep learning models and 86·4% (79·9–91·0) for health-care professionals, and a pooled specificity of 92·5% (95% CI 85·1–96·4) for deep learning models and 90·5% (80·6–95·7) for health-care professionals. Interpretation Our review found the diagnostic performance of deep learning models to be equivalent to that of health-care professionals. However, a major finding of the review is that few studies presented externally validated results or compared the performance of deep learning models and health-care professionals using the same sample. Additionally, poor reporting is prevalent in deep learning studies, which limits reliable interpretation of the reported diagnostic accuracy. New reporting standards that address specific challenges of deep learning could improve future studies, enabling greater confidence in the results of future evaluations of this promising technology. Funding None.

A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis

This work provides a systematic review of the literature from January 2003 to April 2014 pertaining to the incidence, pathophysiology, diagnosis, and treatment of osteonecrosis of the jaw (ONJ), and offers recommendations for its management based on multidisciplinary international consensus. ONJ is associated with oncology-dose parenteral antiresorptive therapy of bisphosphonates (BP) and denosumab (Dmab). The incidence of ONJ is greatest in the oncology patient population (1% to 15%), where high doses of these medications are used at frequent intervals. In the osteoporosis patient population, the incidence of ONJ is estimated at 0.001% to 0.01%, marginally higher than the incidence in the general population (<0.001%). New insights into the pathophysiology of ONJ include antiresorptive effects of BPs and Dmab, effects of BPs on gamma delta T-cells and on monocyte and macrophage function, as well as the role of local bacterial infection, inflammation, and necrosis. Advances in imaging include the use of cone beam computerized tomography assessing cortical and cancellous architecture with lower radiation exposure, magnetic resonance imaging, bone scanning, and positron emission tomography, although plain films often suffice. Other risk factors for ONJ include glucocorticoid use, maxillary or mandibular bone surgery, poor oral hygiene, chronic inflammation, diabetes mellitus, ill-fitting dentures, as well as other drugs, including antiangiogenic agents. Prevention strategies for ONJ include elimination or stabilization of oral disease prior to initiation of antiresorptive agents, as well as maintenance of good oral hygiene. In those patients at high risk for the development of ONJ, including cancer patients receiving high-dose BP or Dmab therapy, consideration should be given to withholding antiresorptive therapy following extensive oral surgery until the surgical site heals with mature mucosal coverage. Management of ONJ is based on the stage of the disease, size of the lesions, and the presence of contributing drug therapy and comorbidity. Conservative therapy includes topical antibiotic oral rinses and systemic antibiotic therapy. Localized surgical debridement is indicated in advanced nonresponsive disease and has been successful. Early data have suggested enhanced osseous wound healing with teriparatide in those without contraindications for its use. Experimental therapy includes bone marrow stem cell intralesional transplantation, low-level laser therapy, local platelet-derived growth factor application, hyperbaric oxygen, and tissue grafting.

https://asbmr.onlinelibrary.wiley.com/doi/pdf/10.1002/jbmr.2405

Diagnosis and Management of Osteonecrosis of the Jaw: A Systematic Review and International Consensus

The ability to monitor blood flow is critical to patient recovery and patient outcomes after complex reconstructive surgeries. Clinically available wired implantable monitoring technology requires careful fixation for accurate detection and needs to be removed after use. Here, we report the design of a pressure sensor, made entirely of biodegradable materials and based on fringe-field capacitor technology, for measuring arterial blood flow in both contact and non-contact modes. The sensor is operated wirelessly through inductive coupling, has minimal hysteresis, fast response times, excellent cycling stability, is highly robust, allows for easy mounting and eliminates the need for removal, thus reducing the risk of vessel trauma. We demonstrate the operation of the sensor with a custom-made artificial artery model and in vivo in rats. This technology may be advantageous in real-time post-operative monitoring of blood flow after reconstructive surgery.

Biodegradable and flexible arterial-pulse sensor for the wireless monitoring of blood flow

From times immemorial manual palpation served as a source of information on the state of soft tissues and allowed detection of various diseases accompanied by changes in tissue elasticity. During the last two decades, the ancient art of palpation gained new life due to numerous emerging elasticity imaging (EI) methods. Areas of applications of EI in medical diagnostics and treatment monitoring are steadily expanding. Elasticity imaging methods are emerging as commercial applications, a true testament to the progress and importance of the field.In this paper we present a brief history and theoretical basis of EI, describe various techniques of EI and, analyze their advantages and limitations, and overview main clinical applications. We present a classification of elasticity measurement and imaging techniques based on the methods used for generating a stress in the tissue (external mechanical force, internal ultrasound radiation force, or an internal endogenous force), and measurement of the tissue response. The measurement method can be performed using differing physical principles including magnetic resonance imaging (MRI), ultrasound imaging, X-ray imaging, optical and acoustic signals.Until recently, EI was largely a research method used by a few select institutions having the special equipment needed to perform the studies. Since 2005 however, increasing numbers of mainstream manufacturers have added EI to their ultrasound systems so that today the majority of manufacturers offer some sort of Elastography or tissue stiffness imaging on their clinical systems. Now it is safe to say that some sort of elasticity imaging may be performed on virtually all types of focal and diffuse disease. Most of the new applications are still in the early stages of research, but a few are becoming common applications in clinical practice.

An overview of elastography - an emerging branch of medical imaging.

The prefrontal cortex (PFC) is not only important in executive functions, but also pain processing. The latter is dependent on its connections to other areas of the cerebral neocortex, hippocampus, periaqueductal gray (PAG), thalamus, amygdala, and basal nuclei. Changes in neurotransmitters, gene expression, glial cells, and neuroinflammation occur in the PFC during acute and chronic pain, that result in alterations to its structure, activity, and connectivity. The medial PFC (mPFC) could serve dual, opposing roles in pain: (1) it mediates antinociceptive effects, due to its connections with other cortical areas, and as the main source of cortical afferents to the PAG for modulation of pain. This is a 'loop' where, on one side, a sensory stimulus is transformed into a perceptual signal through high brain processing activity, and perceptual activity is then utilized to control the flow of afferent sensory stimuli at their entrance (dorsal horn) to the CNS. (2) It could induce pain chronification via its corticostriatal projection, possibly depending on the level of dopamine receptor activation (or lack of) in the ventral tegmental area-nucleus accumbens reward pathway. The PFC is involved in biopsychosocial pain management. This includes repetitive transcranial magnetic stimulation, transcranial direct current stimulation, antidepressants, acupuncture, cognitive behavioral therapy, mindfulness, music, exercise, partner support, empathy, meditation, and prayer. Studies demonstrate the role of the PFC during placebo analgesia, and in establishing links between pain and depression, anxiety, and loss of cognition. In particular, losses in PFC grey matter are often reversible after successful treatment of chronic pain.

/pdf/role-of-the-prefrontal-cortex-in-pain-processing-3wpwaw78cx.pdf

Role of the Prefrontal Cortex in Pain Processing.

Experiments have previously demonstrated the therapeutic potential of mobilized dental pulp stem cells (MDPSCs) for complete pulp regeneration. The aim of the present pilot clinical study is to assess the safety, potential efficacy, and feasibility of autologous transplantation of MDPSCs in pulpectomized teeth. Five patients with irreversible pulpitis were enrolled and monitored for up to 24 weeks following MDPSC transplantation. The MDPSCs were isolated from discarded teeth and expanded based on good manufacturing practice (GMP). The quality of the MDPSCs at passages 9 or 10 was ascertained by karyotype analyses. The MDPSCs were transplanted with granulocyte colony-stimulating factor (G-CSF) in atelocollagen into pulpectomized teeth. The clinical and laboratory evaluations demonstrated no adverse events or toxicity. The electric pulp test (EPT) of the pulp at 4 weeks demonstrated a robust positive response. The signal intensity of magnetic resonance imaging (MRI) of the regenerated tissue in the root canal after 24 weeks was similar to that of normal dental pulp in the untreated control. Finally, cone beam computed tomography demonstrated functional dentin formation in three of the five patients. Human MDPSCs are safe and efficacious for complete pulp regeneration in humans in this pilot clinical study.

/pdf/pulp-regeneration-by-transplantation-of-dental-pulp-stem-48r81q0cyl.pdf

Pulp regeneration by transplantation of dental pulp stem cells in pulpitis: a pilot clinical study

Objective Recently, 3-dimensional–computed tomography (3D-CT) imaging has been used in the diagnosis and surgical treatment planning of patients with craniofacial deformities. The present authors have developed a 3D-CT imaging procedure for a 3-dimensional coordinate point evaluation system to assess and diagnose patients with facial asymmetry. Study design The CT data of 16 subjects was selected retrospectively as the control group from patients who had undergone CT examinations to diagnose conditions other than maxillofacial deformities. Anatomical landmarks modified from orthodontic craniometric (cephalometric) points were defined on the 3D-CT images and the asymmetry index of each point was calculated in millimeters. A diagrammatic chart with a baseline indicating the mean asymmetry indices plus the standard deviation in the control group was designed. The resulting diagrammatic chart was used to evaluate the degree of deformity in facial asymmetry patients. Results and conclusions The topography of facial asymmetry was assessed. The 3D-CT imaging technique as described herein is a practical method of evaluating the morphology of facial asymmetry.

3D-CT evaluation of facial asymmetry.

Objective The aims of this study were to characterize the symmetrical features of patients with facial deformities and to suggest a classification system for facial asymmetry based on 3-dimensional-computed tomography (3D-CT) evaluation. Study design Preoperative CT images were investigated for 49 patients with maxillofacial deformities. Asymmetry indices were calculated for some landmarks in the maxilla, mandibular body and the mandibular ramus regions. Results Asymmetry was observed most frequently in the mandibular body region. The subjects were divided into 3 groups. Subjects without any asymmetry in the 3 regions were classified as Group I (44.9%). Groups II or III were defined for subjects without or with maxillary region asymmetry, respectively, and they were subcategorized. In Group IIA (8.2%), asymmetry was shown solely in the mandibular body region, and in Group IIB (18.4%), additional asymmetry was shown in the mandibular ramus region. Group IIIA (6.1%) showed asymmetry only in the maxillary region, and Group IIIB (22.4%) showed asymmetry in all 3 regions. Conclusion The 3D-CT classification for facial asymmetry has the potential to replace the conventional cephalometric classification.

http://clinicasabbadini.com/download.php?codigo=e0f1sokfv69u0yibe95mjdvll14numoyetpw8r3vn2fdqogwpb

3D-CT evaluation of facial asymmetry in patients with maxillofacial deformities.

Objective. To observe changes in the pharyngeal airway and the hyoid bone position after mandibular setback osteotomy in 30 patients with mandibular prognathism by means of 3-dimensional computed tomography (3DCT). Study design. Preoperative and postoperative computed tomography (CT) examinations were performed on 17 patients treated by sagittal split ramus osteotomy with rigid osteosynthesis and on 13 patients treated by intraoral vertical ramus osteotomy without osteosynthesis. The amount of mandibular setback was measured by the preoperative to postoperative difference of the mandibular position in axial CT images. The sizes of the preoperative and postoperative pharyngeal airway were evaluated from semitransparent and crosscut 3DCT images. Postoperative displacement of the hyoid bone was evaluated by a technique to superimpose a postoperative hard tissue 3DCT image on the preoperative image. The helical scan technique was used in the CT examination. The volume rendering technique was used to create 3DCT images. Results. The mean mandibular setback was 7.8 ± 2.1 mm with a range of 5 to 11 mm. Three months after surgery, the lateral and frontal widths of the pharyngeal airway had decreased significantly in comparison with the preoperative width. The mean reduction rates of the lateral and frontal width were 23.6% and 11.4%, respectively. The diminished airway did not recover by either 6 months or 1 year after surgery in most cases. Downward and posterior displacement of the hyoid bone was seen postoperatively. There were positive correlations between the amount of mandibular setback and reduction of the lateral width of the pharyngeal airway (r = 0.54) and the amount of hyoid bone displacement (r = 0.42). There were no significant differences between the two surgical techniques. Conclusion. Three-dimensional computed tomography was a practical imaging technique to evaluate the morphologic airway changes. The pharyngeal airway may have irreversible narrowing after mandibular setback surgery. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2000;89:278-87)

Three-dimensional computed tomographic evaluation of morphologic airway changes after mandibular setback osteotomy for prognathism

Objectives: The distal root of the mandibular first molar occasionally has an extra root, which can directly affect the outcome of endodontic therapy. In this study, we examined the diagnostic perf...

/pdf/a-deep-learning-artificial-intelligence-system-for-370a1ver9b.pdf

Yoshiko Ariji

Papers

Pulp regeneration by transplantation of dental pulp stem cells in pulpitis: a pilot clinical study

3D-CT evaluation of facial asymmetry.

3D-CT evaluation of facial asymmetry in patients with maxillofacial deformities.

Three-dimensional computed tomographic evaluation of morphologic airway changes after mandibular setback osteotomy for prognathism

A deep-learning artificial intelligence system for assessment of root morphology of the mandibular first molar on panoramic radiography