Explainable machine learning framework to predict personalized physiological aging.

Attaining personalized healthy aging requires accurate monitoring of physiological changes and identifying subclinical markers that predict accelerated or delayed aging. Classic biostatistical methods most rely on supervised variables to estimate physiological aging and do not capture the full complexity of inter-parameter interactions. Machine learning (ML) is promising, but its black box nature eludes direct understanding, substantially limiting physician confidence and clinical usage. Using a broad population dataset from the National Health and Nutrition Examination Survey (NHANES) study including routine biological variables and after selection of XGBoost as the most appropriate algorithm, we created an innovative explainable ML framework to determine a Personalized physiological age (PPA). PPA predicted both chronic disease and mortality independently of chronological age. Twenty-six variables were sufficient to predict PPA. Using SHapley Additive exPlanations (SHAP), we implemented a precise quantitative associated metric for each variable explaining physiological (i.e., accelerated or delayed) deviations from age-specific normative data. Among the variables, glycated hemoglobin (HbA1c) displays a major relative weight in the estimation of PPA. Finally, clustering profiles of identical contextualized explanations reveal different aging trajectories opening opportunities to specific clinical follow-up. These data show that PPA is a robust, quantitative and explainable ML-based metric that monitors personalized health status. Our approach also provides a complete framework applicable to different datasets or variables, allowing precision physiological age estimation.

Explanations as a New Metric for Feature Selection: A Systematic Approach.

With the extensive use of Machine Learning (ML) in the biomedical field, there was an increasing need for Explainable Artificial Intelligence (XAI) to improve transparency and reveal complex hidden relationships between variables for medical practitioners, while meeting regulatory requirements. Feature Selection (FS) is widely used as a part of a biomedical ML pipeline to significantly reduce the number of variables while preserving as much information as possible. However, the choice of FS methods affects the entire pipeline including the final prediction explanations, whereas very few works investigate the relationship between FS and model explanations. Through a systematic workflow performed on 145 datasets and an illustration on medical data, the present work demonstrated the promising complementarity of two metrics based on explanations (using ranking and influence changes) in addition to accuracy and retention rate to select the most appropriate FS/ML models. Measuring how much explanations differ with/without FS are particularly promising for FS methods recommendation. While reliefF generally performs the best on average, the optimal choice may vary for each dataset. Positioning FS methods in a tridimensional space, integrating explanations-based metrics, accuracy and retention rate, would allow the user to choose the priorities to be given on each of the dimensions. In biomedical applications, where each medical condition may have its own preferences, this framework will make it possible to offer the healthcare professional the appropriate FS technique, to select the variables that have an important explainable impact, even if this comes at the expense of a limited drop of accuracy.

Gingival Orofacial Granulomatosis Clinical and 2D/3D Microscopy Features after Orthodontic Therapy: A Pediatric Case Report.

Orofacial granulomatosis (OFG) represents a heterogeneous group of rare orofacial diseases. When affecting gingiva, it appears as a chronic soft tissue inflammation, sometimes combined with the enlargement and swelling of other intraoral sites, including the lips. Gingival biopsy highlights noncaseating granulomatous inflammation, similar to that observed in Crohn's disease and sarcoidosis. At present, the etiology of OFG remains uncertain, although the involvement of the genetic background and environmental triggers, such as oral conditions or therapies (including orthodontic treatment), has been suggested. The present study reports the results of a detailed clinical and 2D/3D microscopy investigation of a case of gingival orofacial granulomatosis in an 8-year-old male patient after orthodontic therapy. Intraoral examination showed an erythematous hyperplasia of the whole gingiva with a granular appearance occurring a few weeks after the installation of a quad-helix. Peri-oral inspection revealed upper labial swelling and angular cheilitis. General investigations did not report ongoing extra-oral disturbances with the exception of a weakly positive anti-Saccharomyces cerevicae IgG auto-antibody. Two- and three-dimensional microscopic investigations confirmed the presence of gingival orofacial granulomatosis. Daily corticoid mouthwashes over a period of 3 months resulted in a slight improvement in clinical signs, despite an intermittent inflammation recurrence. This study brings new insights into the microscopic features of gingival orofacial granulomatosis, thus providing key elements to oral practitioners to ensure accurate and timely OFG diagnosis. The accurate diagnosis of OFG allows targeted management of symptoms and patient monitoring over time, along with early detection and treatment of extra-oral manifestations, such as Crohn's disease.

Characterization and Safety Profile of a New Combined Advanced Therapeutic Medical Product Platelet Lysate-Based Fibrin Hydrogel for Mesenchymal Stromal Cell Local Delivery in Regenerative Medicine.

Adipose-derived mesenchymal stromal cells (ASC) transplant to recover the optimal tissue structure/function relationship is a promising strategy to regenerate tissue lesions. Because filling local tissue defects by injection alone is often challenging, designing adequate cell carriers with suitable characteristics is critical for in situ ASC delivery. The aim of this study was to optimize the generation phase of a platelet-lysate-based fibrin hydrogel (PLFH) as a proper carrier for in situ ASC implantation and (1) to investigate in vitro PLFH biomechanical properties, cell viability, proliferation and migration sustainability, and (2) to comprehensively assess the local in vivo PLFH/ASC safety profile (local tolerance, ASC fate, biodistribution and toxicity). We first defined the experimental conditions to enhance physicochemical properties and microscopic features of PLFH as an adequate ASC vehicle. When ASC were mixed with PLFH, in vitro assays exhibited hydrogel supporting cell migration, viability and proliferation. In vivo local subcutaneous and subgingival PLFH/ASC administration in nude mice allowed us to generate biosafety data, including biodegradability, tolerance, ASC fate and engraftment, and the absence of biodistribution and toxicity to non-target tissues. Our data strongly suggest that this novel combined ATMP for in situ administration is safe with an efficient local ASC engraftment, supporting the further development for human clinical cell therapy.

Improvement of Mucosal Lesion Diagnosis with Machine Learning Based on Medical and Semiological Data: An Observational Study.

Despite artificial intelligence used in skin dermatology diagnosis is booming, application in oral pathology remains to be developed. Early diagnosis and therefore early management, remain key points in the successful management of oral mucosa cancers. The objective was to develop and evaluate a machine learning algorithm that allows the prediction of oral mucosa lesions diagnosis. This cohort study included patients followed between January 2015 and December 2020 in the oral mucosal pathology consultation of the Toulouse University Hospital. Photographs and demographic and medical data were collected from each patient to constitute clinical cases. A machine learning model was then developed and optimized and compared to 5 models classically used in the field. A total of 299 patients representing 1242 records of oral mucosa lesions were used to train and evaluate machine learning models. Our model reached a mean accuracy of 0.84 for diagnostic prediction. The specificity and sensitivity range from 0.89 to 1.00 and 0.72 to 0.92, respectively. The other models were proven to be less efficient in performing this task. These results suggest the utility of machine learning-based tools in diagnosing oral mucosal lesions with high accuracy. Moreover, the results of this study confirm that the consideration of clinical data and medical history, in addition to the lesion itself, appears to play an important role.

A Spectral Principal Component Analysis-Based Framework for Composite Hard/Soft Tissue Fluorescence Image Investigation.

Traditional thin sectioning microscopy of large bone and dental tissue samples using demineralization may disrupt structure morphologies and even damage soft tissues, thus compromising the histopathological investigation. Here, we developed a synergistic and original framework on thick sections based on wide-field multi-fluorescence imaging and spectral Principal Component Analysis (sPCA) as an alternative, fast, versatile, and reliable solution, suitable for highly mineralized tissue structure sustain and visualization. Periodontal 2-mm thick sections were stained with a solution containing five fluorescent dyes chosen for their ability to discriminate close tissues, and acquisitions were performed with a multi-zoom macroscope for blue, green, red, and NIR (near-infrared) emissions. Eigen-images derived from both standard scaler (Std) and Contrast Limited Adaptive Histogram Equalization (Clahe) pre-preprocessing significantly enhanced tissue contrasts, highly suitable for histopathological investigation with an in-depth detail for sub-tissue structure discrimination. Using this method, it is possible to preserve and delineate accurately the different anatomical/morphological features of the periodontium, a complex tooth-supporting multi-tissue. Indeed, we achieve characterization of gingiva, alveolar bone, cementum, and periodontal ligament tissues. The ease and adaptability of this approach make it an effective method for providing high-contrast features that are not usually available in standard staining histology. Beyond periodontal investigations, this first proof of concept of an sPCA solution for optical microscopy of complex structures, especially including mineralized tissues opens new perspectives to deal with other chronic diseases involving complex tissue and organ defects. Overall, such an imaging framework appears to be a novel and convenient strategy for optical microscopy investigation.

Periodontal Cell Therapy: A Systematic Review and Meta-analysis.

BACKGROUND: Periodontitis is a chronic inflammatory disease characterized by the loss of tooth-supporting tissues (or periodontium) leading to the formation of periodontal pocket then to tooth loss. Conventional therapies that involve tooth root debridement are still disappointing because they are more centered on periodontal repair than disease pathophysiology causes. The meta-analysis we present here focused on the results of experimental studies that investigated periodontal mesenchymal stromal cells (MSCs) therapy, a promising strategy to regenerate tissue, given to their immunomodulatory and trophic properties. METHODS: Using PubMed database and ICTRP search portal, 84 animal and 3 randomized human studies were analyzed. RESULTS: Overall, our results highlighted that MSCs grafting, regardless of their tissue origin, enhances periodontal regeneration. A defect morphology suitable for an initial clot stabilization increases the procedure efficacy, especially if cells are carried using a vehicle from natural origin. Nevertheless, methodological biases have been highlighted and still limit the translation to human with high prognosis and regulatory considerations. Besides, because only 2 randomized human trials demonstrated the efficacy of the procedure, further studies are needed to investigate periodontal regeneration procedures on experimental models closer to human pathophysiology. CONCLUSION: Although MSCs grafting in periodontal disease demonstrated therapeutic benefits in animal, it is critical to define more accurately protocols translatable to human and focus on the treatment of the pathology as a whole rather than on the restitution of the sole destroyed tissues.

Systemic Periodontal Risk Score Using an Innovative Machine Learning Strategy: An Observational Study.

Early diagnosis is crucial for individuals who are susceptible to tooth-supporting tissue diseases (e.g., periodontitis) that may lead to tooth loss, so as to prevent systemic implications and maintain quality of life. The aim of this study was to propose a personalized explainable machine learning algorithm, solely based on non-invasive predictors that can easily be collected in a clinic, to identify subjects at risk of developing periodontal diseases. To this end, the individual data and periodontal health of 532 subjects was assessed. A machine learning pipeline combining a feature selection step, multilayer perceptron, and SHapley Additive exPlanations (SHAP) explainability, was used to build the algorithm. The prediction scores for healthy periodontium and periodontitis gave final F1-scores of 0.74 and 0.68, respectively, while gingival inflammation was harder to predict (F1-score of 0.32). Age, body mass index, smoking habits, systemic pathologies, diet, alcohol, educational level, and hormonal status were found to be the most contributive variables for periodontal health prediction. The algorithm clearly shows different risk profiles before and after 35 years of age and suggests transition ages in the predisposition to developing gingival inflammation or periodontitis. This innovative approach to systemic periodontal disease risk profiles, combining both ML and up-to-date explainability algorithms, paves the way for new periodontal health prediction strategies.

Doctors in Medical Data Sciences: A New Curriculum.

Machine Learning (ML), a branch of Artificial Intelligence, which is competing with human experts in many specialized biomedical fields and will play an increasing role in precision medicine. As with any other technological advances in medicine, the keys to understanding must be integrated into practitioner training. To respond to this challenge, this viewpoint discusses some necessary changes in the health studies curriculum that could help practitioners to interpret decisions the made by a machine and question them in relation to the patient's medical context. The complexity of technology and the inherent criticality of its use in medicine also necessitate a new medical profession. To achieve this objective, this viewpoint will propose new medical practitioners with skills in both medicine and data science: the Doctor in Medical Data Sciences.

Place of a new radiological index in predicting pulp exposure before intervention for deep carious lesions.

BACKGROUND: During interventions for deep caries lesions without severe symptoms, preserving pulpal vitality is important to ensure treatment success, improve organ prognosis, and decrease cost-effectiveness. Current pre-operative radiographs allow visual estimation but not accurate measurement of lesion depth. PURPOSE: Investigate the ability of ratio 'remaining/total dentin thickness' (RDT/TDT, as determined on pre-operative radiographs) to predict pulp exposure during excavation. METHODS: This retrospective study (January 2018-June 2020) analyzed data on 360 patients. Four independent raters examined standard pre-operative radiographs and their contrasted versions. Lines put at the dentino-enamel junction, the floor of the carious lesion, and the pulp chamber wall allowed deriving RDT/TDT. Inter-rater agreements and concordance were assessed. A logistic regression accounting for measurement errors provided odds ratios that estimated the ability of the RDT/TDT to predict pulp exposure. RESULTS: The median RDT/TDT ratio ranges were 16.8-26.5% on standard and 16.2-24.6% on contrasted radiographs. Inter-rater agreements on RDT/TDT were rather poor and inter-rater reliability was low and similar in standard and contrasted radiographs: the concordance correlation coefficients (95% CIs) were estimated at 0.46 (0.40; 0.51) and 0.46 (0.40; 0.52), respectively. The risk of pulp exposure increased by 2.5 times [odds ratio (95% CI) 2.57 (2.06; 3.20)] per 10-point decrease of the ratio on standard radiographs vs. 4.15 (3.15; 5.46) on contrasted radiographs. CONCLUSION: RDT/TDT ratio is potentially helpful in predicting pulp exposure. However, the measurement errors on RDT and TDT being non-negligible and the interrater agreements poor, there is still place for advances through development of an automated process that will improve reliability and reproducibility of pulp exposure risk assessment. CLINICAL TRIAL: Trial registration number. ClinicalTrials.gov NCT04607395, October 29, 2020.

Machine Learning Analysis of the Anatomical Parameters of the Upper Airway Morphology: A Retrospective Study from Cone-Beam CT Examinations in a French Population.

The objective of this study is to assess, using cone-beam CT (CBCT) examinations, the correlation between hard and soft anatomical parameters and their impact on the characteristics of the upper airway using symbolic regression as a machine learning strategy. Methods: On each CBCT, the upper airway was segmented, and 24 anatomical landmarks were positioned to obtain six angles and 19 distances. Some anatomical landmarks were related to soft tissues and others were related to hard tissues. To explore which variables were the most influential to explain the morphology of the upper airway, principal component and symbolic regression analyses were conducted. Results: In total, 60 CBCT were analyzed from subjects with a mean age of 39.5 ± 13.5 years. The intra-observer reproducibility for each variable was between good and excellent. The horizontal soft palate measure mostly contributed to the reduction of the airway volume and minimal section area with a variable importance of around 50%. The tongue and the position of the hyoid bone were also linked to the upper airway morphology. For hard anatomical structures, the anteroposterior position of the mandible and the maxilla had some influence. Conclusions: Although the volume of the airway is not accessible on all CBCT scans performed by dental practitioners, this study demonstrates that a small number of anatomical elements may be markers of the reduction of the upper airway with, potentially, an increased risk of obstructive sleep apnea. This could help the dentist refer the patient to a suitable physician.

Driving regeneration, instead of healing, in adult mammals: the decisive role of resident macrophages through efferocytosis.

Tissue repair after lesion usually leads to scar healing and thus loss of function in adult mammals. In contrast, other adult vertebrates such as amphibians have the ability to regenerate and restore tissue homeostasis after lesion. Understanding the control of the repair outcome is thus a concerning challenge for regenerative medicine. We recently developed a model of induced tissue regeneration in adult mice allowing the comparison of the early steps of regenerative and scar healing processes. By using studies of gain and loss of function, specific cell depletion approaches, and hematopoietic chimeras we demonstrate here that tissue regeneration in adult mammals depends on an early and transient peak of granulocyte producing reactive oxygen species and an efficient efferocytosis specifically by tissue-resident macrophages. These findings highlight key and early cellular pathways able to drive tissue repair towards regeneration in adult mammals.

Comparison of different machine learning approaches to predict dental age using Demirjian's staging approach.

CONTEXT: Dental age, one of the indicators of biological age, is inferred by radiological methods. Two of the most commonly used methods are using Demirjian's radiographic stages of permanent teeth excluding the third molar (Demirjian's and Willems' method). The major drawbacks of these methods are that they are based on population-specific conversion tables and may tend to over- or underestimate dental age in other populations. Machine learning (ML) methods make it possible to create complex data schemas more simply while keeping the same annotation system. The objectives of this study are to compare (1) the capacity of ten machine learning algorithms to predict dental age in children using the seven left permanent mandibular teeth compared to reference methods and (2) the capacity of ten machine learning algorithms to predict dental age from childhood to young adulthood using the seven left permanent mandibular teeth and the four third molars. METHODS: Using a large radiological database of 3605 orthopantomograms (1734 females and 1871 males) of healthy French patients aged between 2 and 24 years, seven left permanent mandibular teeth and the 4 third molars were assessed using Demirjian's stages. Dental age estimation was then performed using Demirjian's reference method and various ML regression methods. Two analyses were performed: with the 7 left mandibular teeth without third molars for the under 16 age group and with the third molars for the entire study population. The different methods were compared using mean error, mean absolute error, root mean square error as metrics, and the Bland-Altman graph. RESULTS: All ML methods had a mean absolute error (MAE) under 0.811 years. With Demirjian's and Willems' methods, the MAE was 1.107 and 0.927 years, respectively. Except for the Bayesian ridge regression that gives poorer accuracy, there was no statistical difference between all ML tested. CONCLUSION: Compared to the two reference methods, all the ML methods based on the maturation stages defined by Demirjian were more accurate in estimating dental age. These results support the use of ML algorithms instead of using standard population tables.

Dental practice-based research networks (D-PBRN) worldwide: A scoping review.

OBJECTIVES: For more than twenty years, dental practice-based research networks (D-PBRN) have helped to structure clinical research in private practice. They bring together practitioners working in several structures and may include a greater number of subjects. The aims of this study were thus to systematically explore the scientific production from dental private practices in general and to map and describe the D-PBRN activity worldwide. DATA SOURCES: Two research procedures were carried out in parallel. The first was conducted as a scoping review to examine peer-reviewed literature indexed in the PubMed database and the second was performed on the World Wide Web to identify the main characteristics of the networks (location, scientific production…). STUDY SELECTION: 368 publications were identified among which 202 were published by PBRN members and the others by private practitioners not affiliated to any network. 210 (57 % of the included articles) were produced in the USA. A higher number of diverse centers are involved in each study when it is conducted by a PBRN (59.06 ± 66.59 vs. 13.51 ± 31.58 for networks and independent teams, respectively; p < 0.01). 24 D-PBRN were identified, a majority being based in the USA and 8 in Europe. CONCLUSIONS: Although dental practice-based research has grown over the years, the number of D-PBRN worldwide remains low. Even if it requires some investment to produce research in dental offices, this type of networks helps to fill the gap between private practice and research and to improve knowledge on oral health. RELEVANCE: The mapping of all the dental PBRN together with the research topics studied throughout the world make the relevance of this article. The ways to improve practice-based research in dentistry are also discussed in the paper.

Recent Advances in Cone-beam CT in Oral Medicine.

BACKGROUND: The cone-beam computed tomography (CBCT) technology has continuously evolved since its appearance in oral medicine in the early 2000s. OBJECTIVES: To present recent advances in CBCT in oral medicine: i) selection of recent and consensual evidence-based sources, ii) structured summary of the information based on an iterative framework and iii) compliance with ethical, public health and patient-centered concerns. MAIN FINDINGS: We will focus on technological advances, such as sensors and reconstruction algorithms used to improve the constant quality of the image and dosimetry. CBCT examination is now performed in almost all disciplines of oral medicine: currently, the main clinical disciplines that use CBCT acquisitions are endodontics and oral surgery, with clearly defined indications. Periodontology and ear, nose and throat medicine are more recent fields of application. For a given application and indication, the smallest possible field of view must be used. One of the major challenges in contemporary healthcare is ensuring that technological developments do not take precedence over admitted standards of care. The entire volume should be reviewed in full, with a systematic approach. All findings are noted in the patient's record and explained to the patient, including incidental findings. This presupposes the person reviewing the images is sufficiently trained to interpret such images, inform the patient and organize the clinical pathway, with referrals to other medical or oral medicine specialties as needed. CONCLUSION: A close collaboration between dentists, medical physicists, radiologists, radiographers and engineers is critical for all aspects of CBCT technology.

Broad-Spectrum Antibacterial Effects of Human Adipose-Derived Stromal Cells.

INTRODUCTION: Many pathological conditions may benefit from cell therapy using mesenchymal stromal cells, particularly from adipose tissue (ASCs). Cells may be grafted in an environment with a remnant polymicrobial component. The aim is to investigate the behavior of ASCs when brought in contact with a large panel of bacteria. MATERIALS AND METHODS: Carboxyfluorescein-labelled bacterial interaction with ASCs was followed by confocal time-lapse microscopy. Costaining with LAMP-1 was also analyzed. Viability of 4 gram-negative and 4 gram-positive bacterial strains after 6 h of coculture with ASCs was assessed by agar colony counting and by flow cytometry using SYTO-62®/propidium iodide (PI) for membrane permeabilization and DiOC6 for depolarization. A murine model of periodontitis was used to assess in vivo antibacterial capacities of ASCs. RESULTS: A significant increase of PI-positive events for all bacterial strains and an increase of the DiOC6 signal were obtained after contact with ASCs. The number of CFU was also significantly decreased for several bacterial strains. 0.4 µm transwell systems illustrated the necessary direct contact to induce maximal bacterial membrane damages. Some bacteria were observed into phagolysosomes, confirming macrophage-like properties of ASCs. In vivo, the bacterial load was significantly lower in the ASC-grafted side compared to the control. CONCLUSION: Our results highlight for the first time a broad range of antibacterial actions of ASCs, by phagocytosis, secretion of oxygenated free radicals and antibacterial molecules. These data are in line with the development of new therapeutic strategies based on ASC transplantation, appropriated in immune-dysbiotic tissue context such as periodontitis or chronic wounds.

Incidental Findings in Dental Radiology are Concerning for Family Doctors.

Cone Beam Computerized Tomography (CBCT) is an imaging technology increasingly used in dentistry. Depending on the size of the examination area, visualization of anatomical structures outside the indication area may reveal incidental findings (IF). The aims of this systematic review and meta-analysis were to 1) evaluate the frequency, location and different types of incidental findings (IF) revealed during CBCT examinations; 2) identify potential influencing factors such as gender or age; 3) highlight what the family doctor should know about CBCT and the benefits for medical care. 70 retrospective studies were included. 60% of IF are in the naso-oropharyngeal airway and paranasal sinuses. Carotid calcifications were observed with a mean prevalence of 9% CI95% [2-21]). Meta-regression showed a significant association of this prevalence with age, irrespective of gender. Given the high frequency of IF, with varying severity, the whole medical community is fully involved, and its opinion should be sought to ensure the best possible management for the patients. Physicians may also require CBCT examinations that would have been previously prescribed by a dentist, that may serve to better orientate investigations toward another imaging technique. The family doctor is therefore the dentist's main interlocutor and the main coordinator of the follow-up of IF.

The effect of periodontal treatment on patients with rheumatoid arthritis: The ESPERA randomised controlled trial.

OBJECTIVES: To assess the effect of periodontal treatment on clinical and biochemical parameters of rheumatoid arthritis (RA) and quality of life (QoL) in patients with moderately active RA who were diagnosed with periodontitis. METHODS: In this open-label randomised controlled trial, RA subjects (n = 22) were allocated to "immediate" or "delayed" periodontal treatment (full-mouth non-surgical scaling and root planing, systemic antibiotics, and oral hygiene instructions). The main outcome was the 3-month change on the Disease Activity Score 28 based on the Erythrocyte Sedimentation Rate (DAS28-ESR). The Health Assessment Questionnaire and the General Oral Health Assessment Index were used to assess general and oral health QoL, respectively. RESULTS: Periodontal health significantly improved after periodontal treatment (P = 0.03). Periodontal treatment appeared to be safe but led to no significant effects on the DAS28-ESR (adjusted mean difference with 95% confidence interval (aMD) of -0.03 [-0.98; 0.92]). There was no evidence of improvement in the general QoL after periodontal treatment and no significant effect was found for the oral health QoL, despite a positive trend in the "psychological impacts" domain (aMD of 0.13 [-0.07; 0.33], P = 0.20). CONCLUSIONS: Although no clinical effect of periodontal treatment on RA was identified, this trial provides important data to support periodontal care in RA patients. Periodontal treatment is safe and reduces oral inflammation with a possible effect on oral health QoL. Since both periodontitis and RA are complex and multifactorial chronic diseases, it is likely that patient-centred approaches involving both oral health professionals and rheumatologists will contribute to optimal patient care. ISRCTN79186420.