Deep learning for the screening of primary ciliary dyskinesia based on cranial computed tomography.

Objective: To analyze the cranial computed tomography (CT) imaging features of patients with primary ciliary dyskinesia (PCD) who have exudative otitis media (OME) and sinusitis using a deep learning model for early intervention in PCD. Methods: Thirty-two children with PCD diagnosed at the Children's Hospital of Fudan University, Shanghai, China, between January 2010 and January 2021 who had undergone cranial CT were retrospectively analyzed. Thirty-two children with OME and sinusitis diagnosed using cranial CT formed the control group. Multiple deep learning neural network training models based on PyTorch were built, and the optimal model was trained and selected to observe the differences between the cranial CT images of patients with PCD and those of general patients and to screen patients with PCD. Results: The Swin-Transformer, ConvNeXt, and GoogLeNet training models had optimal results, with an accuracy of approximately 0.94; VGG11, VGG16, VGG19, ResNet 34, and ResNet 50, which are neural network models with fewer layers, achieved relatively strong results; and Transformer and other neural networks with more layers or neural network models with larger receptive fields exhibited a relatively weak performance. A heat map revealed the differences in the sinus, middle ear mastoid, and fourth ventricle between the patients with PCD and the control group. Transfer learning can improve the modeling effect of neural networks. Conclusion: Deep learning-based CT imaging models can accurately screen for PCD and identify differences between the cranial CT images.

Auditory and Speech Outcomes of Cochlear Implantation in Children With Auditory Neuropathy Spectrum Disorder: A Systematic Review and Meta-Analysis.

OBJECTIVE: The aim of this meta-analysis was to critically assess the effect of cochlear implantation on auditory and speech performance outcomes of children with auditory neuropathy spectrum disorder (ANSD). MATERIAL AND METHODS: A systematic literature search was conducted on PubMed, EMbase, and Web of Science. The outcomes included speech recognition score, Categories of Auditory Performance (CAP), Speech Intelligibility Rating (SIR) score, and open-set speech perception. Results were expressed as standardized mean difference (SMD) or risk ratio (RR) with a 95% confidence interval (95% CI). RESULTS: A total of 15 studies was included in this meta-analysis. Pooled data showed that, there were no significant differences between ANSD and sensorineural hearing loss (SNHL) groups in terms of speech recognition score (SMD = 0.01, 95% CI: -0.45, 0.47; P = .959),CAP (SMD = 0.71, 95% CI: -0.13, 1.54; P = .098), SIR score (SMD = -0.09, 95% CI: -0.49, 0.32; P = .667), and open-set speech perception (RR = 0.85, 95% CI: 0.69, 1.05; P = .142). Sensitivity analysis by removing individual studies one at a time showed that the overall estimate and level of heterogeneity did not change substantially. CONCLUSION: The current evidence suggested that children with ANSD who underwent cochlear implants achieved comparable effects in auditory and speech performance as children with non-ANSD SNHL.

An in-depth discussion of cholesteatoma, middle ear Inflammation, and langerhans cell histiocytosis of the temporal bone, based on diagnostic results.

Objective: This study aimed to conduct an in-depth investigation of the learning framework used for deriving diagnostic results of temporal bone diseases, including cholesteatoma and Langerhans cell histiocytosis (LCH). In addition, middle ear inflammation (MEI) was diagnosed by CT scanning of the temporal bone in pediatric patients. Design: A total of 119 patients were included in this retrospective study; among them, 40 patients had MEI, 38 patients had histology-proven cholesteatoma, and 41 patients had histology-proven LCH of the temporal bone. Each of the 119 patients was matched with one-third of the disease labels. The study included otologists and radiologists, and the reference criteria were histopathology results (70% of cases for training and 30% of cases for validation). A multilayer perceptron artificial neural network (VGG16_BN) was employed and classified, based on radiometrics. This framework structure was compared and analyzed by clinical experts according to CT images and performance. Results: The deep learning framework results vs. a physician's diagnosis, respectively, in multiclassification tasks, were as follows. Receiver operating characteristic (ROC) (cholesteatoma): (0.98 vs. 0.91), LCH (0.99 vs. 0.98), and MEI (0.99 vs. 0.85). Accuracy (cholesteatoma): (0.99 vs. 0.89), LCH (0.99 vs. 0.97), and MEI (0.99 vs. 0.89). Sensitivity (cholesteatoma): (0.96 vs. 0.97), LCH (0.99 vs. 0.98), and MEI (1 vs. 0.69). Specificity (cholesteatoma): (1 vs. 0.89), LCH (0.99 vs. 0.97), and MEI (0.99 vs. 0.89). Conclusion: This article presents a research and learning framework for the diagnosis of cholesteatoma, MEI, and temporal bone LCH in children, based on CT scans. The research framework performed better than the clinical experts.

Apparent homozygosity for a novel splicing variant in EPS8 causes congenital profound hearing loss.

Autosomal recessive deafness-102 (DFNB102), a new profound prelingual non-syndromic hearing loss, is caused by mutations in the EPS8 gene. To date, only three such consanguineous families with three different homozygous variants in EPS8 have been reported. Here, we report the fourth case from a non-consanguineous Chinese family, an 11-month-old male infant presented with congenital profound non-syndromic hearing loss. Trio whole-exome sequencing initially identified the patient with a novel seemingly homozygous splicing variant NM_004447.5: c.1435-2A > T in intron 14 of the EPS8 gene and was inherited from his father; further CNVs analysis identified a novel 65.9 kb intragenic deletion and was inherited from his mother. The deletion is covering intron 14 that could account for the apparent homozygosity of the patient. In vitro splicing assay showed the variant c.1435-2A > T creates a new donor site at position c.1443, which is predicted to produce a stop codon after 14 additional amino acids (p.His479Cysfs*14). Furthermore, quantitative allele-specific expression assay showed that relative EPS8 gene expression in the patient significantly decreased (0-fold for the wild-type transcript and 0.25-0.27-fold for the mutant transcript) compared to the control (P < 0.05), indicating the pathogenicity of the identified variants. Overall, our study provides additional evidence that EPS8 is a causative gene for DFNB102 and highlights the clinical utility of simultaneous analysis of CNVs and SNVs to avoid potential errors in the diagnosis and interpretation of patients with apparent homozygosity.

Identification of key genes in allergic rhinitis by bioinformatics analysis.

OBJECTIVE: This study aimed to explore the potential molecular mechanism of allergic rhinitis (AR) and identify gene signatures by analyzing microarray data using bioinformatics methods. METHODS: The dataset GSE19187 was used to screen differentially expressed genes (DEGs) between samples from patients with AR and healthy controls. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were applied for the DEGs. Subsequently, a protein-protein interaction (PPI) network was constructed to identify hub genes. GSE44037 and GSE43523 datasets were screened to validate critical genes. RESULTS: A total of 156 DEGs were identified. GO analysis verified that the DEGs were enriched in antigen processing and presentation, the immune response, and antigen binding. KEGG analysis demonstrated that the DEGs were enriched in Staphylococcus aureus infection, rheumatoid arthritis, and allograft rejection. PPI network and module analysis predicted seven hub genes, of which six (CD44, HLA-DPA1, HLA-DRB1, HLA-DRB5, MUC5B, and CD274) were identified in the validation dataset. CONCLUSIONS: Our findings suggest that hub genes play important roles in the development of AR.

Recessive LOXHD1 variants cause a prelingual down-sloping hearing loss: genotype-phenotype correlation and three additional children with novel variants.

BACKGROUND: Biallelic mutations in LOXHD1 have been identified as the cause of DFNB77 (deafness, autosomal recessive 77). It is a new progressive, severe-to-profound, and late-onset nonsyndromic sensorineural hearing loss (NSHL), and is highly heterogeneous genetically and phenotypically. This study aimed to provide an additional three cases of DFNB77. METHODS: We presented three unrelated children diagnosed with prelingual mild-to-severe NSHL, and their audiograms showed mild hearing loss at 250 Hz before downsloping to a moderate-to-severe degree. Trio whole-exome sequencing (WES) was conducted to identify the pathogenic variants. Additionally, we reviewed the literature to further analyze the relationships between the genotype and audiology phenotype of LOXHD1. RESULTS: Six novel possible pathogenic LOXHD1 variants were identified, including three missense, one nonsense, and two splicing variants. The literature review showed that 68.5% of patients with DFNB77 onset before five years old; Most variants (62%) were associated with a down-sloping audiogram of mild-to-moderate hearing loss at low frequencies (200Hz, 500Hz), particularly variants in the protein domain of PLAT 9. We found that compared with homozygous LOXHD1 variants, individuals with heterozygous compound variants had a significantly milder phenotype, especially individuals carrying one missense and one splicing or bi-allelic missense variants (P < 0.05). Audiometric analysis at different ages showed that the hearing loss degree was aggravated at all frequencies by increasing age. CONCLUSIONS: We report three children with prelingual NSHL carrying six novel LOXHD1 variants. Furthermore, our work indicates that DFNB77 may be milder than previously reported and recommends considering the genotype combination and mutation location of LOXHD1 and race-specificity in DFNB77 molecular diagnoses and management.

miR-597-3p inhibits invasion and migration of thyroid carcinoma SW579 cells by targeting RAB23.

INTRODUCTION: Nowadays, it is indicated that miRNA was anomaly expressed in tumour. Previous studies have shown that miRNAs can regulate the proliferation, invasion, and migration of cancer cell-related processes. Meanwhile, current investigations show that RAB23 also plays an important role in cancer cell-related processes. But the potential mechanism remains unclear. MATERIAL AND METHODS: SW579 cells were selected and transfected with miR-597-3p mimics. Then the expression of miR-597-3p and RAB23 were measured by quantitative real-time polymerase chain reaction (PCR) and western blotting, respectively. Subsequently, the abilities of proliferation, invasion, and migration of SW579 cells were researched. For further study, the Luciferase reporter assay proved that miR-597-3p could target the expression of RAB23, and the proteins of invasion and migration were also measured to clear the mechanism. RESULTS: After being transfected with miR-597-3p mimics, the expression of miR-597-3p was remarkably increased and RAB23 was significantly decreased. The abilities of proliferation, invasion, and migration also decreased significantly. The miRTarase Database predicated and Luciferase reporter assay proved that RAB23 was the target gene of miR-597-3p. The expression of matrix metalloproteinase (MMP)-2, MMP-9, and N-cadherin was down-regulated, and the expression of E-cadherin was up-regulated. CONCLUSION: miR-597-3p could reduce the proliferation, invasion and migration abilities of SW579 cells, which may be related to the targeted inhibition of RAB23 expression and down-regulation of the expression levels of MMP-2, MMP-9 and N-cadherin proteins of SW579 cells.

Long noncoding RNAs SET-binding factor 2-antisense RNA1 promotes cell growth through targeting miR-431-5p/CDK14 axis in human papillary thyroid cancer.

Papillary thyroid cancer (PTC) is a frequent thyroid malignancy. With the significant regulatory role in tumor progression, more attention has been employed to investigate mechanism of long noncoding RNAs (lncRNAs) in progression of PTC. We prospectively explored the mechanism whereby lncRNA SET-binding factor 2-antisense RNA1 (SBF2-AS1) is implicated in pathogenesis of PTC. First, differentially expressed SBF2-AS1 between PTC and normal adjacent thyroid tissues was determined, and result indicated a higher SBF2-AS1 expression in PTC tissues than adjacent normal tissues. Moreover, highly SBF2-AS1 expression predicted a poor prognosis in PTC patients. Second, SBF2-AS1 overexpression promoted cell viability and cycle of PTC, while inhibited cell apoptosis. However, SBF2-AS1 downregulation reduced viability and cycle, while promoted cell apoptosis. Moreover, SBF2-AS1 could bind with miR-431-5p and showed negative correlation with miR-431-5p in PTC patients. Furthermore, miR-431-5p bind with cyclin-dependent kinase (CDK) 14 and showed negative correlation with CDK14 in PTC patients. Finally, overexpression of CDK14 counteracted with the inhibitory role of SBF2-AS1 downregulation on cell viability, cycle, and apoptosis of PTC. In conclusion, SBF2-AS1 exhibited oncogenic property in PTC, and knockdown of SBF2-AS1 could be a therapeutic strategy for PTC.

Novel heterozygous GATA3 and SLC34A3 variants in a 6-year-old boy with Barakat syndrome and hypercalciuria.

BACKGROUND: Barakat syndrome is an autosomal dominant disorder characterized by the triad of hypoparathyroidism, sensorineural deafness, and renal anomalies and is caused by mutations in GATA3 gene. SLC34A3 is the cause gene of hypophosphatemic rickets with hypercalciuria, and heterozygous carriers may have milder clinical symptoms. The aim of this study was to identify the underlying genetic cause of a patient who initially presented with renal failure, hypercalciuria, kidney stone, and bilateral sensorineural deafness. METHODS: A 6-year-old boy with complex clinical presentations was investigated. Comprehensive medical evaluations were performed including auditory function tests, endocrine function tests, metabolic studies, and imaging examinations. Molecular diagnoses were analyzed by trio whole-exome sequencing. RESULTS: One novel de novo deleterious variant (c. 324del) of the GATA3 gene was identified in the patient. The patient can be diagnosed with Barakat syndrome. In addition, one novel variant (c. 589A>G) of the SLC34A3 gene was detected, which was inherited from the father. This heterozygous variant can explain the hypercalciuria and kidney stone that occurred in both the patient and his father. CONCLUSION: This study provides a special case which is phenotype-driven dual diagnoses, and the two novel variants can parsimoniously explain the complex clinical presentations of this patient.

Systematic elucidation of the mechanism of geraniol via network pharmacology.

BACKGROUND: Geraniol is an acyclic monoterpene alcohol, which is extracted from the ethereal oils of aromatic plants. A systematic analysis of its mechanism of action has not yet been carried out. METHODS: In this study, the druggability of geraniol was assessed via Traditional Chinese Medicine Systems Pharmacology Database (TCMSP), and the potential targets of geraniol were identified using the Comparative Toxicogenomics Database (CTD). Additionally, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed using WebGestalt. Drug-target-pathway networks were constructed using Cytoscape to give a visual view. RESULTS: Our findings showed that geraniol has superb druggability with 38 putative identified target genes. GO, KEGG, and network analyses revealed that these targets were associated with cancer, inflammatory immunoreactions, and other physiological processes. CONCLUSION: Geraniol is predicted to target multiple proteins and pathways that shape a network which can exert systematic pharmacological effects.

Measurement of Thresholds Using Auditory Steady-State Response and Cochlear Microphonics in Children with Auditory Neuropathy.

BACKGROUND: The detection of precise hearing thresholds in infants and children with auditory neuropathy (AN) is challenging with current objective methods, especially in those younger than six months of age. PURPOSE: The aim of this study was to compare the thresholds using auditory steady-state response (ASSR) and cochlear microphonics (CM) in children with AN and children with normal hearing. RESEARCH DESIGN: The thresholds of CM, ASSR, and visual reinforcement audiometry (VRA) tests were recorded; the ASSR and VRA frequencies used were 250, 500, 1000, 2000, and 4000 Hz. STUDY SAMPLE: The participants in this study were 15 children with AN (27 ears) (1-7.6 years, median age 4.1 years) and ten children with normal hearing (20 ears) (1-8 years, median age four years). DATA COLLECTION AND ANALYSIS: The thresholds of the three methods were compared, and histograms were used to represent frequency distributions of threshold differences obtained from the three methods. RESULTS: In children with normal hearing, the average CM thresholds (84.5 dB) were significantly higher than the VRA thresholds (10.0-10.8 dB); in children with AN, both CM and VRA responses were seen at high signal levels (88.9 dB and 70.6-103.4 dB, respectively). In normal children, the difference between mean VRA and ASSR thresholds ranged from 17.5 to 30.3 dB, which was significantly smaller than the difference seen between the mean CM and VRA thresholds (71.5-72.3 dB). The correlation between VRA and ASSR in children with normal hearing ranged from 0.38 to 0.48, whereas no such correlation was seen in children with AN at any frequency (0.03-0.19). CONCLUSIONS: Our results indicated that ASSR and CM were poor predictors of the conventional behavioral threshold in children with AN.

Geraniol suppresses proinflammatory mediators in phorbol 12-myristate 13-acetate with A23187-induced HMC-1 cells.

BACKGROUND: Geraniol is a monoterpene alcohol that has anti-fungal, anti-cancer and anti-nociceptive properties, but its anti-allergic rhinitis (AR) property is unclear. METHODS: In this study, the anti-inflammatory role and its possible mechanisms of geraniol in human mast cell line (HMC-1) cells stimulated by inflammatory trigger phorbol 12-myristate 13-acetate plus A23187 (PMACI), as well as in ovalbumin (OVA)-induced AR mice models were investigated. RESULTS: PMACI results in a significant increase in the production of proinflammatory cytokines, such as TNF-α, IL-1ß, MCP-1, IL-6 and as well as histamine. Geraniol was found to inhibit both TNF-α, IL-1ß and IL-6 protein and mRNA expressions at concentrations of 40, 80, 160 µM. In OVA-induced AR models, geraniol treatment was able to suppress AR biomarkers (OVA-specific IgE and IL-1ß as well as histamine) and nasal rub scores. Interestingly, p38, a member of the mitogen-activated protein kinase (MAPK) signaling family, was found to be increasingly hypophosphorylated as geraniol dose was increased. Similar decreases in the nuclear level of p65, a member of the nuclear factor kappa B (NF-κB) signaling pathway, were also observed. CONCLUSION: Our data highlights that the anti-inflammatory properties of geraniol on AR-related markers in activated HCM-1 cells and OVA-induced AR models may be mediated through the regulation of the MAPK/NF-κB signaling pathway.

The outcomes of endoscopic management in young children with subglottic stenosis.

Subglottic stenosis (SGS) is a common cause of obstructed airway in children, and the treatment of pediatric SGS, especially congenital SGS, remains a challenge for the otolaryngologist. OBJECTIVE: To analyze the outcomes of endoscopic management in young children with SGS. METHODS: We performed a retrospective review of treatment with endoscopic balloon dilation (EBD) or EBD combined with endoscopic anterior cricoid split (EACS) for young SGS children, from December 2008 to December 2015. The ages of patients ranged from 2 days to 12 years, median age was 5 months. The grade of them ranged from II to IV. RESULTS: For acute acquired SGS, 19 cases received EBD alone and the other 3 cases received EBD and EACS, the success rate was about 95.5%; For chronic acquired SGS, EBD and EACS was performed in 6 patients with a success rate of 66.7%; For congenital SGS, EBD and EACS was performed in 28 patients with a success rate of 85.7%. Overall, the success rate of endoscopic management in 56 young children was about 87.5%. Besides, No procedure-related complications were observed in any patients. CONCLUSIONS: Endoscopic surgical technique offers a safe and effective approach for treatment of young children with SGS, especially in congenital SGS.

The homozygote p.V27I/p.E114G variant of GJB2 is a putative indicator of nonsyndromic hearing loss in Chinese infants.

The gap junction ß2 (GJB2) gene is associated with more than half of the recessive forms of hereditary hearing loss (HHL). However, the correlation between p.V27I and p.E114G variants of GJB2 and hearing phenotype remains controversial. This study aimed to clarify possible roles of these variants in Chinese infants with nonsyndromic hearing loss (NSHL). Hearing and gene tests were conducted in 300 infants (aged 0-3 months) with NSHL and 484 normal infants (aged 0-3 months). The p.V27I and p.E114G variants appeared frequently in both NSHL patients and normal controls. The allele and haplotype frequencies of p.V27I and p.E114G in patients and controls were compared, but no significant difference was observed (p=0.44 and p=0.26, respectively). Moreover, genotype frequencies of the p.V27I variant showed no significant difference between the two groups (p=0.66). Interestingly, more homozygote p.V27I/p.E114G subjects were found in NSHL infants than in controls (5/484 and 13/300, respectively), most of whom (61.54%) had mild or moderate hearing losses. Our results indicate that homozygote p.V27I/p.E114G is associated with mild and moderate HHL.

Prevalence of p.V37I variant of GJB2 among Chinese infants with mild or moderate hearing loss.

GJB2 accounts for more than 80% of recessive forms of hereditary hearing loss (HL); however, the correlation between the p.V37I variant of GJB2 and hearing phenotype is controversial. This study aimed to investigate the clinical and epidemiological characteristics of the p.V37I variant in sensorineural hearing loss in Chinese infants (0-3 months). Hearing and gene tests were conducted in 300 infants (aged 0-3 months) with sensorineural hearing impairment and 484 normal infants (aged 0-3 months). Among the 300 hearing-impaired infants, 16 (5.33%) exhibited homozygous p.V37I variation and 7 (2.34%) showed a compound-heterozygous p.V37I variation, whereas no homozygous p.V37I (0%) or compound-heterozygous p.V37I (0%) condition was found among the 484 normal infants. The hearing impairment ranged from mild to profound in all patients exhibiting the homozygous p.V37I or the compound-heterozygous p.V37I condition, although most patients (61.54%) exhibit mild or moderate HL. Our results indicated that the p.V37I variation of GJB2 mutation is mainly associated with mild or moderate hearing impairment. Therefore, otolaryngologists should also screen the p.V37I variant of GJB2 in patients with mild or moderate HL.

Prediction of frequency-specific hearing threshold using chirp auditory brainstem response in infants with hearing losses.

OBJECTIVES: To investigate the clinical usefulness of the LS-chirp auditory brainstem response for estimation of behavioral thresholds in young children with mild to severe hearing losses. METHODS: 68 infants (136 ears) aged 6-12 months (mean age=9.2 months) with bilateral mild to severe hearing losses were studied at Children's Hospital of Fudan University. In all cases, the children were referred for LS-chirp ABR and visual reinforcement audiometric (VRA) measurements. The low-frequency band chirp (LF-chirp) thresholds (frequency band=0.1-0.85kHz) were compared to the average VRA thresholds (frequency band=0.25-0.5kHz), whereas the high-frequency band chirp (HF-chirp) thresholds (frequency band=1-10kHz) were compared to the average VRA thresholds (frequency band=1-4kHz) using statistical correlation coefficient values. RESULTS: The LS-chirp ABR thresholds are very close to behavioral hearing levels. The mean differences between chirp-ABR and VRA thresholds were within 5dBHL for all measurements. The smallest mean threshold difference (<3dBHL) was obtained for the severe hearing loss group. The correlation coefficient values (r) were 0.97 at low-frequency and high-frequency bands. For each carrier frequency, the best correlations between chirp-ABR thresholds and VRA thresholds were obtained at VRA frequency of 0.25kHz/LF-chirp (r=0.98) and VRA frequency of 1kHz/HF-chirp (r=0.98). CONCLUSIONS: This study demonstrates the effectiveness using chirp-ABR predicted frequency-specific thresholds, especially of low and middle frequencies. LS-chirp ABR thresholds determined behavioral thresholds in patients with severe hearing losses were better than for mild hearing losses. The use of a chirp-ABR testing ensures higher sensitivity and accuracy than that of auditory stead-state evoked response (ASSR) for measuring frequency-specific thresholds in young children.

Performance of two hearing screening protocols in NICU in Shanghai.

OBJECTIVE: To study the sensitivity and specificity of targeted neonatal hearing screening for the single-session distortion product otoacoustic emissions (DPOAE) technique and the combined DPOAE/automated auditory brain-stem response (AABR) technique. METHODS: 3000 high-risk newborns were studied at Children's Hospital of Fudan University. They were required to take two different screening procedures separately. The first procedure consisted of DPOAE alone and the second consisted of DPOAE combined with the AABR. Based upon the etiology in high-risk babies, they were divided into four groups. In group I there were 670 very-low-birth-weight (VLBW) newborns (1340 ears), and in group II there were 890 preterm babies (1780 ears). 850 babies (1700 ears) suffered from hyperbilirubinemia in group III, whereas 790 babies (1580 ears) with asphyxia were in group IV. The babies in groups II, III, and IV came from the neonatal intensive-care unit (NICU) of our hospital. The study protocols consisted of the DPOAE alone and DPOAE combined with AABR hearing screening at an age of less than 1 month, and a diagnostic stage at the age of 2 months. RESULTS: With single-session DPOAE screening, the referral rate (8% of the NICU babies), the false-positive rate (4.96%) and the false-negative rate (0.8%) were higher. The different etiologies in NICU babies had significantly different referral rates (F-test, p<0.01). A 4.46% referral rate of hyperbilirubinemi babies was much lower. The combined DPOAE/AABR screening technique revealed a referral rate of 5.03%, a false-positive rate of 2% and a false-negative rate of 0.06%. The false-positive rate was well below the suggested 3% of the American Academy of Pediatric. Comparisons of the referral rate, false-positive rate and false-negative rate of two hearing screening protocols (DPOAE alone and combined DPOAE/AABR) revealed significant differences (t-test, p<0.05, p<0.01, p<0.01). 91 infants (3.03% of the NICU babies) who failed the combined DPOAE/AABR screening were confirmed on hearing impairment. Of 22 babies who passed DPOAE screening but failed the AABR screening had a severe to profound hearing loss based on classic ABR. These patients (24% of the NICU babies with hearing losses) with hyperbilirubinemia and asphyxia problems at newborn stage were diagnosed as auditory neuropathy based on evaluations of DPOAE screening passed, abnormal ABR and absent middle-ear muscle reflexes (MMR). CONCLUSION: Our study demonstrates the use of a combination of DPOAE and AABR testing ensures high sensitivity and acceptable specificity, and predict the AN profile in NICU babies. Our efforts identified 22 NICU babies with auditory neuropathy who hopefully will benefit from early remediation of their hearing deficit.