Performance of three artificial intelligence (AI)-based large language models in standardized testing; implications for AI-assisted dental education.

INTRODUCTION: The emerging rise in novel computer technologies and automated data analytics has the potential to change the course of dental education. In line with our long-term goal of harnessing the power of AI to augment didactic teaching, the objective of this study was to quantify and compare the accuracy of responses provided by ChatGPT (GPT-4 and GPT-3.5) and Google Gemini, the three primary large language models (LLMs), to human graduate students (control group) to the annual in-service examination questions posed by the American Academy of Periodontology (AAP). METHODS: Under a comparative cross-sectional study design, a corpus of 1312 questions from the annual in-service examination of AAP administered between 2020 and 2023 were presented to the LLMs. Their responses were analyzed using chi-square tests, and the performance was juxtaposed to the scores of periodontal residents from corresponding years, as the human control group. Additionally, two sub-analyses were performed: one on the performance of the LLMs on each section of the exam; and in answering the most difficult questions. RESULTS: ChatGPT-4 (total average: 79.57%) outperformed all human control groups as well as GPT-3.5 and Google Gemini in all exam years (p < .001). This chatbot showed an accuracy range between 78.80% and 80.98% across the various exam years. Gemini consistently recorded superior performance with scores of 70.65% (p = .01), 73.29% (p = .02), 75.73% (p < .01), and 72.18% (p = .0008) for the exams from 2020 to 2023 compared to ChatGPT-3.5, which achieved 62.5%, 68.24%, 69.83%, and 59.27% respectively. Google Gemini (72.86%) surpassed the average scores achieved by first- (63.48% ± 31.67) and second-year residents (66.25% ± 31.61) when all exam years combined. However, it could not surpass that of third-year residents (69.06% ± 30.45). CONCLUSIONS: Within the confines of this analysis, ChatGPT-4 exhibited a robust capability in answering AAP in-service exam questions in terms of accuracy and reliability while Gemini and ChatGPT-3.5 showed a weaker performance. These findings underscore the potential of deploying LLMs as an educational tool in periodontics and oral implantology domains. However, the current limitations of these models such as inability to effectively process image-based inquiries, the propensity for generating inconsistent responses to the same prompts, and achieving high (80% by GPT-4) but not absolute accuracy rates should be considered. An objective comparison of their capability versus their capacity is required to further develop this field of study.

Placental TLR recognition of salivary and subgingival microbiota is associated with pregnancy complications.

BACKGROUND: Pre-term birth, the leading cause of neonatal mortality, has been associated with maternal periodontal disease and the presence of oral pathogens in the placenta. However, the mechanisms that underpin this link are not known. This investigation aimed to identify the origins of placental microbiota and to interrogate the association between parturition complications and immune recognition of placental microbial motifs. Video Abstract METHODS: Saliva, plaque, serum, and placenta were collected during 130 full-term (FT), pre-term (PT), or pre-term complicated by pre-eclampsia (PTPE) deliveries and subjected to whole-genome shotgun sequencing. Real-time quantitative PCR was used to measure toll-like receptors (TLR) 1-10 expression in placental samples. Source tracking was employed to trace the origins of the placental microbiota. RESULTS: We discovered 10,007 functionally annotated genes representing 420 taxa in the placenta that could not be attributed to contamination. Placental microbial composition was the biggest discriminator of pregnancy complications, outweighing hypertension, BMI, smoking, and maternal age. A machine-learning algorithm trained on this microbial dataset predicted PTPE and PT with error rates of 4.05% and 8.6% (taxonomy) and 6.21% and 7.38% (function). Logistic regression revealed 32% higher odds of parturition complication (95% CI 2.8%, 81%) for every IQR increase in the Shannon diversity index after adjusting for maternal smoking status, maternal age, and gravida. We also discovered distinct expression patterns of TLRs that detect RNA- and DNA-containing antigens in the three groups, with significant upregulation of TLR9, and concomitant downregulation of TLR7 in PTPE and PT groups, and dense correlation networks between microbial genes and these TLRs. 70-82% of placental microbiota were traced to serum and thence to the salivary and subgingival microbiomes. The oral and serum microbiomes of PTPE and PT groups displayed significant enrichment of genes encoding iron transport, exosome, adhesion, quorum sensing, lipopolysaccharide, biofilm, and steroid degradation. CONCLUSIONS: Within the limits of cross-sectional analysis, we find evidence to suggest that oral bacteria might translocate to the placenta via serum and trigger immune signaling pathways capable of inducing placental vascular pathology. This might explain, in part, the higher incidence of obstetric syndromes in women with periodontal disease.

Clinical, immunological and microbiological evaluation of experimental peri-implant mucositis and gingivitis in subjects with Grade C, stage III/IV periodontitis background.

AIM: To compare individuals with a periodontitis background (Grade C, stage III/IV-formerly generalized aggressive periodontitis) (H-GAP) with periodontally healthy subjects (H-Health) in terms of molecular changes (immunological/microbiological) accompanying experimental peri-implant mucositis and gingivitis. MATERIALS AND METHODS: H-GAP and control (H-Health) subjects were recruited, and experimental mucositis/gingivitis was induced around a single screw-retained implant and one contralateral tooth. Participants refrained from oral hygiene for 21 days in the selected areas, followed by professional prophylaxis and hygiene instructions for 21 days. Clinical parameters, immunological markers (multiplex analysis) and microbial data (16S rRNA gene sequencing) were collected at baseline, during induction (7, 14 and 21 days) and following remission (42 days). RESULTS: Clinically, no significant differences were observed between the groups (n = 10/each group) (H-GAP vs. H-Health) (p > .05, Mann-Whitney test) and the type of site (tooth vs. implant) (p > .05, Wilcoxon test) at the time of onset and resolution, or severity of gingival/mucosal inflammation. H-GAP displayed lower concentrations of the cytokines interleukin (IL)-1B, IL-4, IL-17, tumor necrosis factor-α and interferon-γ around implants than H-Health at baseline and during induction of mucositis (p < .05, Mann-Whitney test). In both groups, implants showed significantly higher inflammatory background at baseline and all subsequent visits when compared with teeth (p < .05, Wilcoxon test). Alpha and ß-diversity metrics showed a significant shift in the microbiome composition and abundances of core species during induction and resolution of peri-implant mucositis and gingivitis (p < .05, restricted maximum likelihood method of Shannon and Bray-Curtis indices, respectively). Differences were not significant for these parameters between the H-Health and H-GAP groups when the periodontal and peri-implant microbiomes were compared separately; however, at each time point, the peri-implant microbiome differed significantly from the periodontal microbiome. CONCLUSIONS: Within the limitations of this pilot study (e.g. low power), it can be concluded that different microbial shifts contribute to the onset and progression of inflammatory responses around teeth and implants and that history of periodontal disease experience plays an additional role in modulating the immune response of peri-implant and periodontal tissues to biofilm accumulation.

Biome-microbiome interactions in peri-implantitis: A pilot investigation.

BACKGROUND: Dental implants replace missing teeth in at least 100 million people, yet over one million implants fail every year due to peri-implantitis, a bacterially induced inflammatory disease. Our ability to treat peri-implantitis is hampered by a paucity of information on host-microbiome interactions that underlie the disease. Here, we present the first open-ended characterization of transcriptional events at the mucosal-microbial interface in the peri-implant crevice. METHODS: We simultaneously sequenced microbial and human mRNA from five pairs of healthy and diseased implants from the same patient and used graph theoretics to examine correlations between microbial and host gene expression in the peri-implant crevice. RESULTS: We identified a transcriptionally active peri-implant microbiome surrounding healthy implants. Microbial genes encoding phenylalanine, tyrosine, and tryptophan biosynthesis, cysteine, methionine, arginine, proline, and histidine metabolism correlated to human genes encoding cell development, metabolism, morphogenesis, adhesion, gap junctions, cell-cell signaling, and immunoinflammatory pathways, suggesting a role for commensals in protecting epithelial integrity. In disease, we found 4- to 200-fold upregulation in microbial genes encoding biofilm thickness, heme transport and utilization, and Gram-negative cell membrane synthesis. These genes correlated with mucosal zinc finger proteins, apoptosis, membrane transport, inflammation, and cell-cell communication. CONCLUSIONS: Within the limitations of a small sample size, our data suggest that microbial dysbiosis in the peri-implant sulcus might promote abandonment of host-bacterial transactions that dictate health and instead drive a move towards chronic programming of a non-healing wound.

Predicted functional and taxonomic analysis of subgingival biofilm of Grade C periodontitis in young patients under maintenance therapy.

BACKGROUND: In Grade C periodontitis in young patients (PerioC-Y), the functional roles of the subgingival community after years of periodontal treatment are still underexplored. This study evaluated the taxonomic and predicted functional content of the subgingival microbiome of PerioC-Y patients under supportive periodontal therapy (SPT). METHODS: Clinical and microbiological data from subgingival biofilm were assessed from 10 PerioC-Y patients at two time points: at baseline and after 5.7 ± 1.3 years of SPT. This was compared with 15 patients without a history of periodontitis. The V1-V3 and V4-V5 regions of the 16S rRNA were sequenced using the Illumina Miseq. Microbial composition was evaluated by the core microbiome, and alpha- and beta-diversity. The microbiome functional content was predicted using Picrust2, and the gene differential abundance was analyzed with DESeq2. RESULTS: Clinical improvements were seen in PerioC-Y-SPT. Differences in ß-diversity between PerioC-Y and health were observed (health x PerioC-Y-baseline, P = 0.02; health x PerioC-Y-SPT, P = 0.05). Moreover, although ß-diversity did not statistically change between baseline and SPT in PerioC-Y, the microbial correlation evidenced increased Streptococcus and decreased Treponema network contributions during SPT. Based on predicted functional data, treatment induced a reduction in genes related to flagellar protein and signal transduction in PerioC-Y. However, compared with healthy individuals, some genes remained more highly abundant in PerioC-Y-SPT, such as quorum sensing and efflux pump transporters. CONCLUSION: Despite clinical improvements and a shift in taxonomic composition, the PerioC-Y patients' periodontal treatment was not enough to reach a similar microbiome to patients without disease experience. Some functional content in this biofilm remained altered in PerioC-Y regardless of disease control.

An in vitro model for studies of attenuation of antibiotic-inhibited growth of Aggregatibacter actinomycetemcomitans Y4 by polyamines.

Polyamines are ubiquitous polycationic molecules that are present in all prokaryotic and eukaryotic cells, and they serve as important modulators of cell growth, stress, and cell proliferation. Polyamines are present at high concentrations in the periodontal pocket and could potentially affect the stress response of periodontal bacteria to antibiotics. The effects of polyamines on inhibition of growth by amoxicillin (AMX), azithromycin (AZM), and doxycycline (DOX) were investigated with the Y4 strain of Aggregatibacter actinomycetemcomitans (Aa). Bacteria were grown in brain heart infusion broth under the following conditions: (1) Aa only, (2) Aa + polyamine mix (1 mM putrescine, 0.4 mM spermidine, and 0.4 mM spermine), (3) Aa + antibiotic, and (4) Aa + antibiotic + polyamines. Growth curve analysis, minimal inhibitory concentration determination, and transcriptomic studies were conducted. The presence of exogenous polyamines produced a small, but significant increase in Aa growth, and polyamines attenuated the inhibitory effects of AMX, AZM, and DOX on growth. Transcriptomic analysis revealed that polyamines upregulate expression of ribosomal biogenesis proteins and small subunits, attenuate the bacterial stress response to antibiotics, and modulate bacterial nutritional pathways in a manner that could potentially increase the virulence of Aa. In summary, the polyamine-rich environment found in periodontal pockets appears to protect Aa and reduce its susceptibility to several antimicrobial agents in this in vitro model.

Anna Karenina and the subgingival microbiome associated with periodontitis.

BACKGROUND: Although localized aggressive periodontitis (LAP), generalized aggressive periodontitis (GAP), and chronic periodontitis (CP) are microbially driven diseases, our inability to separate disease-specific associations from those common to all three forms of periodontitis has hampered biomarker discovery. Therefore, we aimed to map the genomic content of, and the biological pathways encoded by, the microbiomes associated with these clinical phenotypes. We also estimated the extent to which these biomes are governed by the Anna Karenina principle (AKP), which states that eubiotic communities are similar between individuals while disease-associated communities are highly individualized. METHODS: We collected subgingival plaque from 25 periodontally healthy individuals and diseased sites of 59 subjects with stage 3 periodontitis and used shotgun metagenomics to characterize the aggregate of bacterial genes. RESULTS: Beta-dispersion metrics demonstrated that AKP was most evident in CP, followed by GAP and LAP. We discovered broad dysbiotic signatures spanning the three phenotypes, with over-representation of pathways that facilitate life in an oxygen-poor, protein- and heme-rich, pro-oxidant environment and enhance capacity for attachment and biofilm formation. Phenotype-specific indicators were more readily evident in LAP microbiome than GAP or CP. Genes that enable acetate-scavenging lifestyle, utilization of alternative nutritional sources, oxidative and nitrosative stress responses, and siderophore production were unique to LAP. An attenuation of virulence-related functionalities and stress response from LAP to GAP to CP was apparent. We also discovered that clinical phenotypes of disease resolved variance in the microbiome with greater clarity than the newly established grading system. Importantly, we observed that one third of the metagenome of LAP is unique to this phenotype while GAP shares significant functional and taxonomic features with both LAP and CP, suggesting either attenuation of an aggressive disease or an early-onset chronic disease. CONCLUSION: Within the limitations of a small sample size and a cross-sectional study design, the distinctive features of the microbiomes associated with LAP and CP strongly persuade us that these are discrete disease entities, while calling into question whether GAP is a separate disease, or an artifact induced by cross-sectional study designs. Further studies on phenotype-specific microbial genes are warranted to explicate their role in disease etiology. Video Abstract.

Parents with periodontitis impact the subgingival colonization of their offspring.

Early acquisition of a pathogenic microbiota and the presence of dysbiosis in childhood is associated with susceptibility to and the familial aggregation of periodontitis. This longitudinal interventional case-control study aimed to evaluate the impact of parental periodontal disease on the acquisition of oral pathogens in their offspring. Subgingival plaque and clinical periodontal metrics were collected from 18 parents with a history of generalized aggressive periodontitis and their children (6-12 years of age), and 18 periodontally healthy parents and their parents at baseline and following professional oral prophylaxis. 16S rRNA amplicon sequencing revealed that parents were the primary source of the child's microbiome, affecting their microbial acquisition and diversity. Children of periodontitis parents were preferentially colonized by Filifactor alocis, Porphyromonas gingivalis, Aggregatibacter actinomycetemcomitans, Streptococcus parasanguinis, Fusobacterium nucleatum and several species belonging to the genus Selenomonas even in the absence of periodontitis, and these species controlled inter-bacterial interactions. These pathogens also emerged as robust discriminators of the microbial signatures of children of parents with periodontitis. Plaque control did not modulate this pathogenic pattern, attesting to the microbiome's resistance to change once it has been established. This study highlights the critical role played by parental disease in microbial colonization patterns in their offspring and the early acquisition of periodontitis-related species and underscores the need for greater surveillance and preventive measures in families of periodontitis patients.

Methods to mitigate infection spread from aerosol-generating dental procedures.

Infection control measures play a critical role in preventing the spread of disease in healthcare settings. Concerns that SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2), the virus that causes Coronavirus Disease 2019, may be transmitted through droplets and aerosols from both symptomatic and asymptomatic individuals has turned the spotlight on healthcare interventions that involve aerosol generation in the oral cavity, such as many dental and periodontal procedures. This commentary seeks to familiarize the dental practitioner with various infection control methods that may be implemented to mitigate spread of infection in dental settings through aerosol-generating dental procedures.

Microbiomics: Were we all wrong before?

Periodontal microbiology has historically been based on an "us against them" paradigm, one that focuses mainly on identifying microbes and viruses that cause disease. However, such a bottom-up approach limits our appreciation of the incredible diversity of this ecosystem and the essential ways in which microbial interactions contribute to health and homeostasis of the subgingival niche. Microbiomics-the science of collectively characterizing and quantifying molecules responsible for the structure, function, and dynamics of a microbial community-has enabled us to study these communities in their natural habitat, thereby revolutionizing our knowledge of host-associated microbes and reconceptualizing our definition of "human." When this systems-biology approach is combined with ecologic principles, it explicates the complex relationship that exist between microbiota and between them and us, the human. In this volume of Periodontology 2000, a group of 12 female scientists take the lead in investigating how metagenomics, genomics, metatranscriptomics, proteomics, metaproteomics, and metabolomics have achieved the following: (a) widened our view of the periodontal microbiome; (b) expanded our understanding of the evolution of the human oral microbiome; (c) shone a light on not just bacteria, but also other prokaryotic and eukaryotic members of the community; (d) elucidated the effects of anthropogenic behavior and systemic diseases on shaping these communities; and (e) influenced traditional patterns of periodontal therapeutics.

Probing periodontal microbial dark matter using metataxonomics and metagenomics.

Our view of the periodontal microbial community has been shaped by a century or more of cultivation-based and microscopic investigations. While these studies firmly established the infection-mediated etiology of periodontal diseases, it was apparent from the very early days that periodontal microbiology suffered from what Staley and Konopka described as the "great plate count anomaly", in that these culturable bacteria were only a minor part of what was visible under the microscope. For nearly a century, much effort has been devoted to finding the right tools to investigate this uncultivated majority, also known as "microbial dark matter". The discovery that DNA was an effective tool to "see" microbial dark matter was a significant breakthrough in environmental microbiology, and oral microbiologists were among the earliest to capitalize on these advances. By identifying the order in which nucleotides are arranged in a stretch of DNA (DNA sequencing) and creating a repository of these sequences, sequence databases were created. Computational tools that used probability-driven analysis of these sequences enabled the discovery of new and unsuspected species and ascribed novel functions to these species. This review will trace the development of DNA sequencing as a quantitative, open-ended, comprehensive approach to characterize microbial communities in their native environments, and explore how this technology has shifted traditional dogmas on how the oral microbiome promotes health and its role in disease causation and perpetuation.

Interventions to prevent periodontal disease in tobacco-, alcohol-, and drug-dependent individuals.

Substance abuse affects more than one sixth of the world's population. More importantly, the nature of the abuse and the type of addictive substances available to individuals is increasing exponentially. All substances with abusive potential impact both the human immuno-inflammatory system and oral microbial communities, and therefore play a critical role in the etiopathogenesis of periodontal diseases. Evidence strongly supports the efficacy of professionally delivered cessation counseling. Dentists, dental therapists, and hygienists are ideally placed to deliver this therapy, and to spearhead efforts to provide behavioral and pharmacologic support for cessation. The purpose of this review is to examine the biologic mechanisms underlying their role in disease causation, to understand the pharmacologic and behavioral basis for their habituation, and to investigate the efficacy of population-based and personalized interventions in prevention of periodontal disease.

Demystifying the mist: Sources of microbial bioload in dental aerosols.

The risk of transmitting airborne pathogens is an important consideration in dentistry and has acquired special significance in the context of recent respiratory disease epidemics. The purpose of this review, therefore, is to examine (1) what is currently known regarding the physics of aerosol creation, (2) the types of environmental contaminants generated by dental procedures, (3) the nature, quantity, and sources of microbiota in these contaminants and (4) the risk of disease transmission from patients to dental healthcare workers. Most dental procedures that use ultrasonics, handpieces, air-water syringes, and lasers generate sprays, a fraction of which are aerosolized. The vast heterogeneity in the types of airborne samples collected (spatter, settled aerosol, or harvested air), the presence and type of at-source aerosol reduction methods (high-volume evacuators, low volume suction, or none), the methods of microbial sampling (petri dishes with solid media, filter paper discs, air harvesters, and liquid transport media) and assessment of microbial bioload (growth conditions, time of growth, specificity of microbial characterization) are barriers to drawing robust conclusions. For example, although several studies have reported the presence of microorganisms in aerosols generated by ultrasonic scalers and high-speed turbines, the specific types of organisms or their source is not as well studied. This paucity of data does not allow for definitive conclusions to be drawn regarding saliva as a major source of airborne microorganisms during aerosol generating dental procedures. Well-controlled, large-scale, multi center studies using atraumatic air harvesters, open-ended methods for microbial characterization and integrated data modeling are urgently needed to characterize the microbial constituents of aerosols created during dental procedures and to estimate time and extent of spread of these infectious agents.

Adverse effects of electronic cigarettes on the disease-naive oral microbiome.

Six percent of Americans, including 3 million high schoolers, use e-cigarettes, which contain potentially toxic substances, volatile organic compounds, and metals. We present the first human study on the effects of e-cigarette exposure in the oral cavity. By interrogating both immunoinflammatory responses and microbial functional dynamics, we discovered pathogen overrepresentation, higher virulence signatures, and a brisk proinflammatory signal in clinically healthy e-cigarette users, equivalent to patients with severe periodontitis. Using RNA sequencing and confocal and electron microscopy to validate these findings, we demonstrate that the carbon-rich glycol/glycerol vehicle is an important catalyst in transforming biofilm architecture within 24 hours of exposure. Last, a machine-learning classifier trained on the metagenomic signatures of e-cigarettes identified as e-cigarette users both those individuals who used e-cigarettes to quit smoking, and those who use both e-cigarettes and cigarettes. The present study questions the safety of e-cigarettes and the harm reduction narrative promoted by advertising campaigns.

Systemic Risk Factors for the Development of Periimplant Diseases.

PURPOSE: Although periimplantitis results from the tissue destructive effects of a dysbiotic periimplant microbiome, several factors may either contribute to the dysbiosis or influence the host response to this bacterial challenge and thereby increase the risk of disease. The goal of this narrative review is examine extrinsic factors that might increase the risk at both subject and site levels. MATERIALS AND METHODS: The PubMed (MEDLINE) database was searched for articles examining the influence of systemic conditions on periimplantitis or implant failure. Key search terms included "systemic," "medications," "periodontitis," "dental implant," "periimplantitis," "implant failure" and related terms. Manual searches were also performed for the following journals: Clinical Oral Implants Research, International Journal of Periodontics and Restorative Dentistry, Journal of Clinical Periodontology, International Journal of Oral and Maxillofacial Implants, Implant Dentistry, and Journal of Periodontology. The inclusion criteria were cohort studies and case-control studies with at least 10 participants per group and with at least 6 months of follow-up. RESULTS: Certain systemic diseases, medications, radiotherapy, and behavioral factors, such as oral hygiene and compliance with periodontal maintenance therapy, appear to significantly increase the risk of disease.

Exploring a temporal relationship between biofilm microbiota and inflammatory mediators during resolution of naturally occurring gingivitis.

BACKGROUND: This study uses multiple, contemporary methodologies to expand our knowledge of the temporal relationship between host-microbial interactions and clinical signs of gingivitis. METHODS: Subgingival plaque and crevicular fluid samples were collected from 31 systemically healthy adults with naturally occurring plaque-induced gingivitis. Professional prophylaxis was administered and participants were followed over 7 weeks. Microbial characterization was performed using a bead-based hybridization assay and cytokine analysis using bead-based flow cytometry. RESULTS: The provision of sequential interventions, oral hygiene instruction, and subsequent professional prophylaxis brought about significant reduction of plaque and resolution of gingivitis at all post baseline time points (P < 0.0001). Candidate cytokines that increased significantly (95% level) were interleukin (IL)-1ß, matrix metalloproteinases (MMP)-1, MMP-3, MMP-8, MMP-9, from baseline to week 2; regulated on activation, normal T cell expressed and secreted (RANTES) at week 4 and week 8; macrophage inflammatory protein (MIP)-1α and MIP-1ß at week 8. Resolution of inflammation was accompanied by a shift in the microbiological flora toward those species associated with health. CONCLUSIONS: This study provides further evidence of the dynamic relationships that exist between the overt clinical signs, the microbial biofilm, and the host response in gingivitis and upon resolution following clinical interventions. Understanding the interactions between the host immune system and subgingival microbial communities during the resolution of established gingivitis continues to evolve as additional knowledge is achieved through using new analytical technologies. The present study confirms a critical effect of oral hygiene measures on restoration of microbial eubiosis in subgingival communities, confirming the important role for home care and professional intervention in maintaining oral health.