An introduction to machine learning and generative artificial intelligence for otolaryngologists-head and neck surgeons: a narrative review.

PURPOSE: Despite the robust expansion of research surrounding artificial intelligence (AI) and machine learning (ML) and their applications to medicine, these methodologies often remain opaque and inaccessible to many otolaryngologists. Especially, with the increasing ubiquity of large-language models (LLMs), such as ChatGPT and their potential implementation in clinical practice, clinicians may benefit from a baseline understanding of some aspects of AI. In this narrative review, we seek to clarify underlying concepts, illustrate applications to otolaryngology, and highlight future directions and limitations of these tools. METHODS: Recent literature regarding AI principles and otolaryngologic applications of ML and LLMs was reviewed via search in PubMed and Google Scholar. RESULTS: Significant recent strides have been made in otolaryngology research utilizing AI and ML, across all subspecialties, including neurotology, head and neck oncology, laryngology, rhinology, and sleep surgery. Potential applications suggested by recent publications include screening and diagnosis, predictive tools, clinical decision support, and clinical workflow improvement via LLMs. Ongoing concerns regarding AI in medicine include ethical concerns around bias and data sharing, as well as the "black box" problem and limitations in explainability. CONCLUSIONS: Potential implementations of AI in otolaryngology are rapidly expanding. While implementation in clinical practice remains theoretical for most of these tools, their potential power to influence the practice of otolaryngology is substantial.

NOX1 is essential for TNFα-induced intestinal epithelial ROS secretion and inhibits M cell signatures.

OBJECTIVE: Loss-of-function mutations in genes generating reactive oxygen species (ROS), such as NOX1, are associated with IBD. Mechanisms whereby loss of ROS drive IBD are incompletely defined. DESIGN: ROS measurements and single-cell transcriptomics were performed on colonoids stratified by NOX1 genotype and TNFα stimulation. Clustering of epithelial cells from human UC (inflamed and uninflamed) scRNASeq was performed. Validation of M cell induction was performed by immunohistochemistry using UEA1 (ulex europaeus agglutin-1 lectin) and in vivo with DSS injury. RESULTS: TNFα induces ROS production more in NOX1-WT versus NOX1-deficient murine colonoids under a range of Wnt-mediated and Notch-mediated conditions. scRNASeq from inflamed and uninflamed human colitis versus TNFα stimulated, in vitro colonoids defines substantially shared, induced transcription factors; NOX1-deficient colonoids express substantially lower levels of STAT3 (signal transducer and activator of transcription 3), CEBPD (CCAAT enhancer-binding protein delta), DNMT1 (DNA methyltransferase) and HIF1A (hypoxia-inducible factor) baseline. Subclustering unexpectedly showed marked TNFα-mediated induction of M cells (sentinel cells overlying lymphoid aggregates) in NOX1-deficient colonoids. M cell induction by UEA1 staining is rescued with H2O2 and paraquat, defining extra- and intracellular ROS roles in maintenance of LGR5+ stem cells. DSS injury demonstrated GP2 (glycoprotein-2), basal lymphoplasmacytosis and UEA1 induction in NOX1-deficiency. Principal components analyses of M cell genes and decreased DNMT1 RNA velocity correlate with UC inflammation. CONCLUSIONS: NOX1 deficiency plus TNFα stimulation contribute to colitis through dysregulation of the stem cell niche and altered cell differentiation, enhancing basal lymphoplasmacytosis. Our findings prioritise ROS modulation for future therapies.

Inflamed Ulcerative Colitis Regions Associated With MRGPRX2-Mediated Mast Cell Degranulation and Cell Activation Modules, Defining a New Therapeutic Target.

BACKGROUND & AIMS: Recent literature has implicated a key role for mast cells in murine models of colonic inflammation, but their role in human ulcerative colitis (UC) is not well established. A major advance has been the identification of mrgprb2 (human orthologue, MRGPX2) as mediating IgE-independent mast cell activation. We sought to define mechanisms of mast cell activation and MRGPRX2 in human UC. METHODS: Colon tissues were collected from patients with UC for bulk RNA sequencing and lamina propria cells were isolated for MRGPRX2 activation studies and single-cell RNA sequencing. Genetic association of all protein-altering G-protein coupled receptor single-nucleotide polymorphism was performed in an Ashkenazi Jewish UC case-control cohort. Variants of MRGPRX2 were transfected into Chinese hamster ovary (CHO) and human mast cell (HMC) 1.1 cells to detect genotype-dependent effects on ß-arrestin recruitment, IP-1 accumulation, and phosphorylated extracellular signal-regulated kinase. RESULTS: Mast cell-specific mediators and adrenomedullin (proteolytic precursor of PAMP-12, an MRGPRX2 agonist) are up-regulated in inflamed compared to uninflamed UC. MRGPRX2 stimulation induces carboxypeptidase secretion from inflamed UC. Of all protein-altering GPCR alleles, a unique variant of MRGPRX2, Asn62Ser, was most associated with and was bioinformatically predicted to alter arrestin recruitment. We validated that the UC protective serine allele enhances ß-arrestin recruitment, decreases IP-1, and increases phosphorylated extracellular signal-regulated kinase with MRGPRX2 agonists. Single-cell RNA sequencing defines that adrenomedullin is expressed by activated fibroblasts and epithelial cells and that interferon gamma is a key upstream regulator of mast cell gene expression. CONCLUSION: Inflamed UC regions are distinguished by MRGPRX2-mediated activation of mast cells, with decreased activation observed with a UC-protective genetic variant. These results define cell modules of UC activation and a new therapeutic target.

Cough Sounds in Screening and Diagnostics: A Scoping Review.

OBJECTIVE: The aim of the study was to examine applications of cough sounds towards screening tools and diagnostics in the biomedical and engineering literature, with particular focus on disease types, acoustic data collection protocols, data processing and analytics, accuracy, and limitations. DATA SOURCES: PubMed, EMBASE, Web of Science, Scopus, Cochrane Library, IEEE Xplore, Engineering Village, and ACM Digital Library were searched from inception to August 2021. REVIEW METHODS: A scoping review was conducted on screening and diagnostic uses of cough sounds in adults, children, and animals, in English peer-reviewed and gray literature of any design. RESULTS: From a total of 438 abstracts screened, 108 articles met inclusion criteria. Human studies were most common (77.8%); the majority focused on adults (57.3%). Single-modality acoustic data collection was most common (71.2%), with few multimodal studies, including plethysmography (15.7%) and clinico-demographic data (7.4%). Data analytics methods were highly variable, with 61.1% using machine learning, the majority of which (78.8%) were published after 2010. Studies commonly focused on cough detection (41.7%) and screening of COVID-19 (11.1%); among pediatric studies, the most common focus was diagnosis of asthma (52.6%). CONCLUSION: Though the use of cough sounds in diagnostics is not new, academic interest has accelerated in the past decade. Cough sound offers the possibility of an accessible, noninvasive, and low-cost disease biomarker, particularly in the era of rapid development of machine learning capabilities in combination with the ubiquity of cellular technology with high-quality recording capability. However, most cough sound literature hinges on nonstandardized data collection protocols and small, nondiverse, single-modality datasets, with limited external validity. Laryngoscope, 134:1023-1031, 2024.

An Evaluation of Quality, Reliability, and Accuracy of Vocal Health Content on TikTok.

OBJECTIVES: Social media plays an ever-growing part in modern life and is a frequent source of health information for patients. Singers are particularly likely to receive health information solely from fellow musicians and may experience barriers to seeking vocal healthcare. However, there have been no studies to date evaluating vocal health information on social media. Our study aimed to assess the quality, reliability, and accuracy of vocal health content on TikTok. STUDY DESIGN: Cross-sectional analysis. METHODS: Three searches were carried out on Tiktok, using the terms "vocal health," "vocal injury," and "voice tips." The top 50 videos in each category were cataloged for extraction of data and for analysis on three discrete scales. Two independent reviewers rated each video using the Global Quality Scale (GQS), modified DISCERN scale, and Accuracy in Digital-health Instrument (ANDI). RESULTS: After the removal of duplicates and unavailable content, 146 videos were analyzed. The mean (range) length was 59.8 seconds (5-239), and number of views per video was 886,265 (432-36,700,000). The vast majority of videos (94.5%) were created by non-clinicians; only two videos (1.37%) were posted by otolaryngologists. The mean (SD) GQS score was 2.34 (0.75) out of a maximum of five, the DISCERN score was 0.97 (0.56) out of five, and the ANDI score was 2.85 (0.87) out of four. Video length was positively correlated with GQS and DISCERN scores, but views, likes, and shares were either not associated or negatively associated with GQS, DISCERN, and ANDI. CONCLUSIONS: Most videos were of low quality and reliability and moderate accuracy. Measures of popularity were either uncorrelated or negatively correlated with quality, reliability, and accuracy, suggesting that TikTok users are more likely to engage with lower-quality content online. This implies a potential role for vocal health professionals to fill a crucial gap with reliable information on social media.

National Cohort Data Suggests an Association Between Serious Mental Illness and Audiometric Hearing Loss.

OBJECTIVE: To explore whether there is an association between serious mental illness (SMI) and hearing loss (HL) among US Hispanic adults. STUDY DESIGN AND SETTING: Cross-sectional epidemiological study (Hispanic Community Health Study), including multicentered US volunteers. METHODS: Multivariable linear regressions were conducted to study the association between SMI and HL. Adjustments were made for potential confounders including age, sex, education, vascular disease (hypertension or diabetes mellitus), and cognition. SMI was defined by (1) antipsychotic medication classification and (2) the use of at least 1 antipsychotic medication specifically used to treat SMI in clinical psychiatric practice. HL was measured by pure tone audiometry. RESULTS: A total of 7581 subjects had complete data. The mean age was 55.2 years (SD = 7.5 years) and the mean pure tone average in the better ear was 16.8 dB (SD = 10.7 dB). A total of 194 (2.6%) subjects were taking a HCHS-defined antipsychotic and 98 (1.3%) were taking at least 1 antipsychotic specifically used to treat SMI. On multivariable regression, use of HCHS's classified antipsychotics was associated with 3.75 dB worse hearing (95% confidence interval [CI] = 2.36-5.13, P < .001) and use of antipsychotics specific for SMI was associated with 4.49 dB worse hearing (95% CI = 2.56-6.43, P < .001) compared to those not using antipsychotics. CONCLUSION: SMI, as defined by either the use of HCHS-defined antipsychotics or the use of antipsychotic medication specific for SMI, is associated with worse hearing, controlling for potential confounders. Whether SMI contributes to HL, antipsychotic medication (through ototoxicity) contributes to HL, or whether HL contributes to SMI is unknown and warrants further investigation.

Speech Quality Perception in Unilateral Cochlear Implant Users With Single-Sided Deafness.

OBJECTIVE: Cochlear implant (CI) users frequently complain about speech quality perception (SQP). In patients undergoing cochlear implantation for single-sided deafness, there is concern that poor SQP from the implanted ear will negatively impact binaural (CI + normal hearing [NH]) SQP. In this study, we investigate if binaural SQP is measurably different than unimplanted NH alone. STUDY DESIGN: Cross-sectional study. SETTING: Tertiary care center. METHODS: Fifteen unilateral CI users with NH in the contralateral ear completed the validated Columbia Speech Quality Instrument. This instrument consists of 9 audio clips rated across 14 specific speech qualities using a 10-point visual analog scale. SQP was assessed in 3 conditions: CI only, NH only, and CI + NH. RESULTS: Median speech quality scores were worse in the CI only condition compared to the NH only (50.0 vs 72.6, P = .0003) and binaural (50.0 vs 71.0, P = .007) conditions. Median speech quality scores were not significantly different between the NH only and binaural conditions (72.6 vs 71, P = .8). Compared to NH, CI speech quality sounded less clear, less natural, and more mechanical. CONCLUSION: Compared to NH, SQP is poorer with a CI alone. However, in contrast to expectation, there is no significant difference between NH and binaural SQP. This suggests poorer CI speech perception does not negatively impact binaural SQP in patients undergoing cochlear implantation for single-sided deafness.

Mannose Supplementation Curbs Liver Steatosis and Fibrosis in Murine MASH by Inhibiting Fructose Metabolism.

Metabolic dysfunction-associated steatohepatitis (MASH) can progress to cirrhosis and liver cancer. There are no approved medical therapies to prevent or reverse disease progression. Fructose and its metabolism in the liver play integral roles in MASH pathogenesis and progression. Here we focus on mannose, a simple sugar, which dampens hepatic stellate cell activation and mitigates alcoholic liver disease in vitro and in vivo . In the well-validated FAT-MASH murine model, oral mannose supplementation improved both liver steatosis and fibrosis at low and high doses, whether administered either at the onset of the model ("Prevention") or at week 6 of the 12-week MASH regimen ("Reversal"). The in vivo anti-fibrotic effects of mannose supplementation were validated in a second model of carbon tetrachloride-induced liver fibrosis. In vitro human and mouse primary hepatocytes revealed that the anti-steatotic effects of mannose are dependent on the presence of fructose, which attenuates expression of ketohexokinase (KHK), the main enzyme in fructolysis. KHK is decreased with mannose supplementation in vivo and in vitro, and overexpression of KHK abrogated the anti-steatotic effects of mannose. Our study identifies mannose as a simple, novel therapeutic candidate for MASH that mitigates metabolic dysregulation and exerts anti-fibrotic effects.

Outcomes of Gender-Affirming Voice and Communication Modification Training for Non-binary Individuals: A Case Series.

OBJECTIVES: There is currently no research reporting solely on outcomes of voice and communication modification training (VCMT) in individuals who identify as non-binary and genderqueer (NBGQ) in the English literature. This study aimed to describe the objective and subjective impact of VCMT on the voice of NBGQ individuals undergoing a 12-week gender-affirming VCMT program. METHODS: A retrospective consecutive case series of NBGQ individuals enrolled in a VCMT program was performed. Demographics, Transgender Self-Evaluation Questionnaire (TSEQ), fundamental frequency (F0), and frequency range were collected before and after the program. RESULTS: Four NBGQ individuals enrolled between January 2019 and June 2021; the mean age was 27.0 years. While all four participants represented in this case series showed improvement in at least one of their initial goals, only one improved both their F0 and TSEQ scores; the other three participants had mixed results. CONCLUSION: NBGQ individuals experienced improvements in self-reported outcomes and changes in acoustic measures after completing VCMT in our case series. Individuals experienced significant improvement in subjective outcomes despite small changes in acoustic measures, and vice versa. More research is needed to better understand the voice and communication needs of NBGQ individuals, along with their outcomes with VCMT. LEVEL OF EVIDENCE: Level 4.

Single-cell transcriptomics reveals conserved cell identities and fibrogenic phenotypes in zebrafish and human liver.

The mechanisms underlying liver fibrosis are multifaceted and remain elusive with no approved antifibrotic treatments available. The adult zebrafish has been an underutilized tool to study liver fibrosis. We aimed to characterize the single-cell transcriptome of the adult zebrafish liver to determine its utility as a model for studying liver fibrosis. We used single-cell RNA sequencing (scRNA-seq) of adult zebrafish liver to study the molecular and cellular dynamics at a single-cell level. We performed a comparative analysis to scRNA-seq of human liver with a focus on hepatic stellate cells (HSCs), the driver cells in liver fibrosis. scRNA-seq reveals transcriptionally unique populations of hepatic cell types that comprise the zebrafish liver. Joint clustering with human liver scRNA-seq data demonstrates high conservation of transcriptional profiles and human marker genes in zebrafish. Human and zebrafish HSCs show conservation of transcriptional profiles, and we uncover collectin subfamily member 11 (colec11) as a novel, conserved marker for zebrafish HSCs. To demonstrate the power of scRNA-seq to study liver fibrosis using zebrafish, we performed scRNA-seq on our zebrafish model of a pediatric liver disease with mutation in mannose phosphate isomerase (MPI) and characteristic early liver fibrosis. We found fibrosis signaling pathways and upstream regulators conserved across MPI-depleted zebrafish and human HSCs. CellPhoneDB analysis of zebrafish transcriptome identified neuropilin 1 as a potential driver of liver fibrosis. Conclusion: This study establishes the first scRNA-seq atlas of the adult zebrafish liver, highlights the high degree of similarity to human liver, and strengthens its value as a model to study liver fibrosis.