Lirit Levi

All Publications

• Patency of the Cavernous Sinus After Medial Wall Resection: An Observational Study. Journal of neurological surgery. Part B, Skull base Renteria, A. E., Fieux, M., Castro, B., Levi, L., Hung, L. T., Azevedo, A., Lee, C. K., Chang, M. T., Hwang, P. H., Nayak, J. V., Fischbein, N., Fernandez-Miranda, J. C., Patel, Z. M. 2026; 87 (4): 462-468

  Abstract

  The cavernous sinus medial wall (CSMW) is a critical structure that can be resected in cavernous sinus (CS)-invading tumors. Intraoperative venous bleeding is inevitable upon entering the CS and necessitates hemostatic agents use to cease its flow. There is no literature on the fate of CS patency after such hemostasis. This study aims to address that gap.An observational study was conducted between 2018 and 2025, including patients undergoing endoscopic endonasal approach (EEA) with CSMW or transcavernous resection for CS-invading tumors. Pre- and postoperative magnetic resonance imaging (MRI) images were collected and analyzed for CS patency. Postoperative imaging was grouped into three intervals: 0 to 3, 3 to 12, and >12 months. Demographic data were analyzed to identify predictors of CS non-patency.Of the 186 patients included, postoperative CS patency was achieved in 87.1% after 3.72 months. Univariate analysis demonstrated that preoperative CS non-patency ( p  < 0.001), incomplete gross tumor resection ( p  = 0.015), meningioma diagnosis ( p  < 0.001), and previous head and neck radiation increased the postoperative CS non-patency risk. Multivariate analysis identified preoperative CS non-patency as the only independent factor for postoperative CS non-patency (odds ratio = 9.8, p  < 0.001). Persistent non-patency at 3 and 12 months increased long-term postoperative CS non-patency risk by 5.86 and 14.7 times, respectively.This is the first study to report that postoperative CS patency is preserved in most cases after EEA requiring CSMW. A strong predictor of postoperative CS non-patency was identified as preoperative CS non-patency. Further studies are warranted to assess if CS non-patency has significant clinical correlations.

  View details for DOI 10.1055/a-2649-2098

  View details for PubMedID 42404179

  View details for PubMedCentralID PMC13331650

• Clarity Without Credibility? Human Versus AI Abstracts in Otolaryngology WORLD JOURNAL OF OTORHINOLARYNGOLOGY-HEAD & NECK SURGERY Hack, S., Attal, R., Levi, L., Liu, D. T., Manavi, M., Locatello, L., Fieux, M., de Joya, E. M., Gutovitz, J. M., Eliahu, S., Hefetz, T., Boldes, T., Biadsee, A. 2026

  View details for DOI 10.1002/wjo2.70104

  View details for Web of Science ID 001720906300001

• Clarity Without Credibility? Human Versus AI Abstracts in Otolaryngology. World journal of otorhinolaryngology - head and neck surgery Hack, S., Attal, R., Levi, L., Liu, D. T., Manavi, M., Locatello, L. G., Fieux, M., de Joya, E. M., Gutovitz, J. M., Eliahu, S., Hefetz, T., Boldes, T., Biadsee, A. 2026

  Abstract

  This study evaluated whether otolaryngologists can distinguish between human- and machine-written abstracts. The primary question was whether large language models (LLMs) produce abstracts comparable in clarity and usefulness to human-authored work, and whether reviewers can identify authorship with accuracy.A blinded cross-sectional design was used. Forty-eight abstracts were evaluated, consisting of twenty-four human-authored abstracts and 24 generated by four LLMs. Human abstracts were drawn from articles published after July 2025 to minimize overlap with LLM training data. Twenty otolaryngologists independently reviewed all abstracts. Using a structured rubric, raters classified authorship, rated clarity, usefulness, and confidence on 5-point scales, and provided optional free-text explanations. Group comparisons were performed using chi-square and Mann-Whitney tests, with Kruskal-Wallis tests for model-level analyses.Overall recognition accuracy was 44.7%. Human-written abstracts were more often misclassified as AI than AI-generated abstracts were mistaken for human. Human abstracts received significantly higher clarity and usefulness scores than LLM abstracts, though effect sizes were small. Confidence did not correlate with correctness, indicating miscalibration of rater judgments. Model-level performance varied. Grok-generated abstracts were most easily identified as AI, whereas GPT-5 and Claude 3.5 more frequently resembled human writing. Free-text rationales commonly referenced style, vagueness, or lack of detail when AI authorship was suspected.LLMs generate abstracts that increasingly resemble human scientific writing, yet still lag in perceived usefulness and credibility. Clinicians were only moderately successful at detecting authorship and were frequently confident in incorrect classifications. These findings highlight both the promise and risks of AI-assisted scientific communication.

  View details for DOI 10.1002/wjo2.70104

  View details for PubMedID 42500058

  View details for PubMedCentralID PMC13398884

• Middle Turbinate Preservation Technique During Transpterygoid Approaches for Skull Base Surgery. The Laryngoscope Renteria, A. E., Fieux, M., Silva, B. C., Levi, L., Azevedo, A., Chang, M. T., Fernandez-Miranda, J. C., Nayak, J. V. 2026

  Abstract

  The expanded endonasal approach (EEA) to resect skull base lesions has classically included the resection of one or both middle turbinates (MT). Here, we describe for the first time a technique that spares the MT entitled middle turbinate release with preservation (MTRP) that can give an alternative and feasible approach to skull base surgeons that wish to maintain this endonasal landmark.

  View details for DOI 10.1002/lary.70381

  View details for PubMedID 41618707

• Artificial Intelligence for Endoscopic Surveillance Post-Treatment for Nasopharyngeal Carcinoma: Detection of Recurrence and Osteoradionecrosis. International forum of allergy & rhinology Levi, L., Huynh, J. D., Wang, Y., Fieux, M., Manavi, M. S., Renteria, A., Sethi, S. K., Ayoub, N., Patel, Z., Nayak, J. V., Hwang, P., Chang, M. T. 2025

  View details for DOI 10.1002/alr.70089

  View details for PubMedID 41420500

• Distinguishing Suspected Invasive Fungal Sinusitis From Its Mimics: A Clinicopathologic Analysis of Histopathology-Negative Cases. International forum of allergy & rhinology Idler, B. M., Renteria, A., Liu, D. T., Rubio-Jimenez, I., Levi, L., Nayak, J. V., Patel, Z. M., Chang, M. T., Rahman, M., Hwang, P. H., Ayoub, N. F. 2025

  View details for DOI 10.1002/alr.70088

  View details for PubMedID 41403233

• The Influence of Nasal Cavity Dimensions on Frequency of Treatment for Rhinitis Symptoms. International forum of allergy & rhinology Levi, L., Mohan, V., Ma, Y., Meliadis, C., Gopi, P., Nayak, J. V., Lin, B., Hwang, P. H., Johnson, J., Nadeau, K., Blaiss, M., Ray, T. S. 2025

  Abstract

  A wearable headband treats nasal congestion using acoustic resonance and AI-estimated anatomy. Individuals with higher nasal depth-to-height ratio used acoustic therapy more often. Depth-to-height ratio may be an anatomical marker for increased nasal congestion risk.

  View details for DOI 10.1002/alr.70060

  View details for PubMedID 41158109

• Utilizing a publicly accessible automated machine learning platform to enable diagnosis before tumor surgery. Communications medicine Hosseinzadeh, F., Liu, G., Tsai, E., Mahmoudi, A., Yang, A., Kim, D., Fieux, M., Levi, L., Abdul-Hadi, S., Adappa, N. D., Alt, J. A., Altartoor, K. A., Banyi, N., Challa, M., Chandra, R., Chang, M. T., Chen, P. G., Cho, D. Y., de Choudens, C. R., Chowdhury, N., Colon, C. M., DelGaudio, J. M., Del Signore, A., Dorismond, C., Dutra, D., Edalati, S., Edwards, T. S., Ferriol, J. B., Geltzeiler, M., Georgalas, C., Govindaraj, S., Grayson, J. W., Gudis, D. A., Harvey, R. J., Heffernan, A., Hwang, P. H., Iloreta, A. M., Knight, N. D., Kohanski, M. A., Lerner, D. K., Leventi, A., Lee, L. H., Lubner, R., Mahomva, C., Massey, C., McCoul, E. D., Nayak, J. V., Pak-Harvey, E., Palmer, J. N., Pandrangi, V. C., Psaltis, A. J., Raviv, J., Sacks, P., Sacks, R., Schaberg, M., Soudry, E., Sweis, A., Thamboo, A., Turner, J. H., Wang, S. X., Wise, S. K., Woodworth, B. A., Wormald, P. J., Patel, Z. M. 2025; 5 (1): 419

  Abstract

  In benign tumors with potential for malignant transformation, sampling error during pre-operative biopsy can significantly change patient counseling and surgical planning. Sinonasal inverted papilloma (IP) is the most common benign soft tissue tumor of the sinuses, yet it can undergo malignant transformation to squamous cell carcinoma (IP-SCC), for which the planned surgery could be drastically different. Artificial intelligence (AI) could potentially help with this diagnostic challenge.CT images from 19 institutions were used to train the Google Cloud Vertex AI platform to distinguish between IP and IP-SCC. The model was evaluated on a holdout test dataset of images from patients whose data were not used for training or validation. Performance metrics of area under the curve (AUC), sensitivity, specificity, accuracy, and F1 were used to assess the model.Here we show CT image data from 958 patients and 41099 individual images that were labeled to train and validate the deep learning image classification model. The model demonstrated a 95.8 % sensitivity in correctly identifying IP-SCC cases from IP, while specificity was robust at 99.7 %. Overall, the model achieved an accuracy of 99.1%.A deep automated machine learning model, created from a publicly available artificial intelligence tool, using pre-operative CT imaging alone, identified malignant transformation of inverted papilloma with excellent accuracy.

  View details for DOI 10.1038/s43856-025-01134-9

  View details for PubMedID 41062640

  View details for PubMedCentralID 9538572

• Patency of the Cavernous Sinus After Medial Wall Resection: An Observational Study JOURNAL OF NEUROLOGICAL SURGERY PART B-SKULL BASE Renteria, A. E., Fieux, M., Castro, B., Levi, L., Hung, L., Azevedo, A., Lee, C. K., Chang, M. T., Hwang, P. H., Nayak, J., Fischbein, N., Fernandez-Miranda, J., Patel, Z. M. 2025

  View details for DOI 10.1055/a-2649-2098

  View details for Web of Science ID 001530300900001

• Machine learning of endoscopy images to identify, classify, and segment sinonasal masses. International forum of allergy & rhinology Levi, L., Ye, K., Fieux, M., Renteria, A., Lin, S., Xing, L., Ayoub, N. F., Patel, Z. M., Nayak, J. V., Hwang, P. H., Chang, M. T. 2025

  Abstract

  We developed and assessed the performance of a machine learning model (MLM) to identify, classify, and segment sinonasal masses based on endoscopic appearance.A convolutional neural network-based model was constructed from nasal endoscopy images from patients evaluated at an otolaryngology center between 2013 and 2024. Images were classified into four groups: normal endoscopy, nasal polyps, benign, and malignant tumors. Polyps and tumors were confirmed with histopathological diagnosis. Images were annotated by an otolaryngologist and independently verified by two other otolaryngologists. We used high- and low-quality images to mirror real-world conditions. The models used for classification (EfficientNet-B2) and segmentation (nnUNet) were trained, validated, and tested at an 8:1:1 ratio. The performance accuracy was averaged across a 10-fold cross-validation assessment. Segmentation accuracy was assessed via Dice similarity coefficients.A total of 1242 images from 311 patients were used. The MLM was trained, validated, and tested on 663 normal, 276 polyps, 157 benign, and 146 malignant tumors images. Overall, the model performed at 84.1 ± 4.3% accuracy in the validation set and 80.4 ± 1.7% in the test set. The model correctly identified the presence of a sinonasal mass at 90.5 ± 1.2% accuracy rate. The MLM accuracy performance rates were 86.2 ± 1.0% for polyps and 84.1 ± 1.8% for tumors. Benign and malignant tumor subclassification achieved 87.8 ± 2.1% and 94.0 ± 2.4% accuracy, respectively. Segmentation accuracies for polyps were 72.3% and 72.8% for tumors.An MLM for nasal endoscopy images can perform with moderate to high accuracy in identifying, classifying, and segmenting sinonasal masses. Performance in future iterations may improve with larger and more diverse training datasets.

  View details for DOI 10.1002/alr.23525

  View details for PubMedID 39776302

• Association between US Wildfires and Health Care Utilization for Acute Rhinosinusitis International Forum of Allergy & Rhinology Levi, L., Levi, A., Fieux, M., Velasquez, E., Vo, R. H., Grimm, D., Hwang, P. H. 2025

  View details for DOI 10.1002/alr.23630

• Stepwise Empty Nose Syndrome Evaluation (SENSE) test-A modified cotton test for reduced bias in office diagnosis of empty nose syndrome. International forum of allergy & rhinology Levi, L., Yang, A., Tsai, E. F., Ma, Y., Ibrahim, N., Dholakia, S. S., Rao, V. K., Renteria, A., Cao, X., Chang, M. T., Nayak, J. V. 2024

  Abstract

  Diagnosis of empty nose syndrome (ENS) relies on the ENS six-item questionnaire (ENS6Q) with a score of ≥11, followed by a "positive" cotton test yielding seven-point reduction from baseline ENS6Q score via cotton placement to the inferior meatus (IM). Given the intricacies of diagnosing ENS and the propensity for false positives with the standard cotton test, we modified the classic single-step cotton test into a four-part Stepwise Empty Nose Syndrome Evaluation (SENSE) cotton test to reduce bias and evaluate the placebo effect.Individuals diagnosed with ENS underwent the SENSE test, a single-blinded, four-step, office-based cotton test, without topical anesthesia or decongestants. Conditions included: (1) placebo/no cotton placed; (2) complete cotton-blockade of nasal vestibule; (3) cotton placed medially against the nasal septum; and (4) cotton placed laterally in the IM (site of inferior turbinate tissue loss). With each condition, patients completed an ENS6Q.Forty-eight ENS patients were included. Twenty-nine percent demonstrated a placebo effect (p < 0.001), 40.4% had a positive response to complete cotton-blockade (p < 0.001), 64.4% to septum-placed cotton, and 79.1% to IM-placed cotton (p < 0.001), corresponding to a mean ENS6Q reduction of 11.9 points (p < 0.001). Notably, the mean difference in ENS6Q scores between septum and IM placement was 1.7 (p < 0.001).The SENSE test offers further insight into subtleties of nasal breathing experienced by ENS patients. The placebo effect can be prominent and important to consider with individual patients. While most ENS patients prefer any intranasal cotton placement over baseline, blinded testing reveals these patients can accurately discriminate minimal changes in nasal aerodynamics.

  View details for DOI 10.1002/alr.23442

  View details for PubMedID 39373717