Elliott Sina

Mr. Elliott Sina

orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
27 November 2023 in Thyroid®
Verified
Journal Article
Closed access
DCDaniel J. Campbell
Daniel J. Campbell
LELeonard E. Estephan
Leonard E. Estephan
ESElliott Sina
Elliott Sina
EMEric V. Mastrolonardo
Eric V. Mastrolonardo
RARahul Alapati
Rahul Alapati
ECElizabeth E. Cottrill
Elizabeth E. Cottrill
Background: ChatGPT, an artificial intelligence (AI) chatbot, is the fastest growing consumer application in history. Given recent trends identifying increasing patient use of Internet sources for self-education, we seek to evaluate the quality of ChatGPT-generated responses for patient education on thyroid nodules. Methods: ChatGPT was queried 4 times with 30 identical questions. Queries differed by initial chatbot prompting: no prompting, patient-friendly prompting, 8th-grade level prompting, and prompting for references. Answers were scored on a hierarchical score: incorrect, partially correct, correct, or correct with references. Proportions of responses at incremental score thresholds were compared by prompt type using chi-squared analysis. Flesch–Kincaid grade level was calculated for each answer. The relationship between prompt type and grade level was assessed using analysis of variance. References provided within ChatGPT answers were totaled and analyzed for veracity. Results: Across all prompts ( n = 120 questions), 83 answers (69.2%) were at least correct. Proportions of responses that were at least partially correct ( p = 0.795) and correct ( p = 0.402) did not differ by prompt; responses that were correct with references did ( p < 0.0001). Responses from 8th-grade level prompting were the lowest mean grade level (13.43 ± 2.86) and were significantly lower than no prompting (14.97 ± 2.01, p = 0.01) and prompting for references (16.43 ± 2.05, p < 0.0001). Prompting for references generated 80/80 (100%) of referenced medical publications within answers. Seventy references (87.5%) were legitimate citations, and 58/80 (72.5%) provided accurately reported information from the referenced publication. Conclusion: ChatGPT overall provides appropriate answers to most questions on thyroid nodules regardless of prompting. Despite targeted prompting strategies, ChatGPT reliably generates responses corresponding to grade levels well-above accepted recommendations for presenting medical information to patients. Significant rates of AI hallucination may preclude clinicians from recommending the current version of ChatGPT as an educational tool for patients at this time.

1 July 2024 in OTO Open
Verified
Journal Article
Open access
ESElliott Sina
Elliott Sina
DCDaniel J. Campbell
Daniel J. Campbell
ADAlexander Duffy
Alexander Duffy
SMShreya Mandloi
Shreya Mandloi
PBPeter Benedict
Peter Benedict
GNGurston Nyquist
Gurston Nyquist
Objective: While most patients with COVID‐19‐induced olfactory dysfunction (OD) recover spontaneously, those with persistent OD face significant physical and psychological sequelae. ChatGPT, an artificial intelligence chatbot, has grown as a tool for patient education. This study seeks to evaluate the quality of ChatGPT‐generated responses for COVID‐19 OD.Study Design: Quantitative observational study.Setting: Publicly available online website.Methods: ChatGPT (GPT‐4) was queried 4 times with 30 identical questions. Prior to questioning, Chat‐GPT was “prompted” to respond (1) to a patient, (2) to an eighth grader, (3) with references, and (4) no prompt. Answer accuracy was independently scored by 4 rhinologists using the Global Quality Score (GCS, range: 1‐5). Proportions of responses at incremental score thresholds were compared using χ2 analysis. Flesch‐Kincaid grade level was calculated for each answer. Relationship between prompt type and grade level was assessed via analysis of variance.Results: Across all graded responses (n = 480), 364 responses (75.8%) were “at least good” (GCS ≥ 4). Proportions of responses that were “at least good” (P < .0001) or “excellent” (GCS = 5) (P < .0001) differed by prompt; “at least moderate” (GCS ≥ 3) responses did not (P = .687). Eighth‐grade level (14.06 ± 2.3) and patient‐friendly (14.33 ± 2.0) responses were significantly lower mean grade level than no prompting (P < .0001).Conclusion: ChatGPT provides appropriate answers to most questions on COVID‐19 OD regardless of prompting. However, prompting influences response quality and grade level. ChatGPT responds at grade levels above accepted recommendations for presenting medical information to patients. Currently, ChatGPT offers significant potential for patient education as an adjunct to the conventional patient‐physician relationship.

14 April 2025 in The Laryngoscope
Verified
Journal Article
Open access
ESElliott Sina
Elliott Sina
ACAndrew Corr
Andrew Corr
LELeonard E. Estephan
Leonard E. Estephan
PHPatrick Hunt
Patrick Hunt
MBMaurits Boon
Maurits Boon
CHColin Huntley
Colin Huntley
Objective: Drug‐induced sleep endoscopy (DISE) is used to identify airway obstruction patterns in patients with obstructive sleep apnea (OSA). This procedure is typically performed in either the traditional operating room (OR) or an outpatient surgery center (SC). There is limited data regarding the cost and efficiency of DISE in the endoscopy suite (ES). This study compares the cost, efficiency, and safety of DISE performed in the ES compared to the SC.Methods: This retrospective review included patients who underwent DISE between July 2023 and April 2024 at a single tertiary care center. Consecutive patients with CPAP‐intolerant OSA undergoing DISE in the ES or SC were included. Patient demographics, sleep studies, procedure‐related costs, procedure duration characteristics, and anesthesia documentation were analyzed.Results: Eighty‐four patients (36 ES, 48 SC) were included. Patient sex, age, BMI, and preoperative AHI/REI were similar in both cohorts (p > 0.05). DISE in the ES significantly reduced total facility time (48 min, p < 0.0001), time in procedure room (6.6 min, p < 0.0001), DISE procedure time (1.7 min, p = 0.0034), total recovery time (30.4 min, p < 0.0001), and procedure room turnover time (7.1 min, p = 0.0012). Costs for ES procedures were 39.35% lower, specifically in costs associated with anesthesia, operating room, and recovery room (all comparisons, p < 0.001). No complications occurred in either group.Conclusions: Performing DISE in the ES increases efficiency and reduces costs compared to the outpatient SC while maintaining safety for patients with CPAP‐intolerant OSA.Level of Evidence: 3

11 June 2025 in Retina
Verified
Journal Article
Closed access
ESElliott Sina
Elliott Sina
JPJose Bs Pena
Jose Bs Pena
SZSidra Zafar
Sidra Zafar
NBNikhil K Bommakanti
Nikhil K Bommakanti
AKAjay E. Kuriyan
Ajay E. Kuriyan
YYYoshihiro Yonekawa
Yoshihiro Yonekawa
Purpose: Automated machine learning (AutoML) is an artificial intelligence (AI) tool that streamlines image recognition model development. This study evaluates the diagnostic performance of Google VertexAI AutoML in differentiating age-related macular degeneration (AMD), diabetic macular edema (DME), epiretinal membrane (ERM), retinal vein occlusion (RVO), and healthy controls using optical coherence tomography (OCT) images. Methods: A publicly available, validated OCT dataset of 1965 de-identified images from 759 patients was used. Images were labeled and uploaded to VertexAI. A single-label classification model was trained, validated, and tested using an 80%-10%-10% split. Diagnostic metrics included area under the precision-recall curve (AUPRC), sensitivity, specificity, and positive and negative predictive value (PPV, NPV). A sub-analysis evaluated neovascular versus non-neovascular AMD. Results: The AutoML model achieved high accuracy (AUPRC = 0.991), with sensitivity, specificity, and PPV of 95.9%, 96.9%, and 95.9%, respectively. AMD classification performed best (AUPRC = 0.999, precision = 98.4%, recall = 99.2%). ERM (AUPRC = 0.978, precision = 92.9%, recall = 86.7%) and DME (AUPRC = 0.895, precision = 81.3%, recall = 86.7%) followed. RVO recall was 80% despite 100% precision. Neovascular AMD outperformed non-neovascular AMD (AUPRC = 0.963 vs. 0.915). Conclusion: Our AutoML model accurately classifies OCT images of retinal conditions, demonstrating performance comparable or superior to traditional ML methods. Its user-friendly design supports scalable AI-driven clinical integration.

21 January 2025 in Otolaryngology–Head and Neck Surgery
Verified
Journal Article
Open access
ESElliott Sina
Elliott Sina
SSSomya Shankar
Somya Shankar
MBMaurits S. Boon
Maurits S. Boon
CHColin T. Huntley
Colin T. Huntley
Objective: To compare the incidence of motor vehicle accidents (MVAs) among patients with obstructive sleep apnea (OSA) undergoing continuous positive airway pressure (CPAP) therapy and sleep surgery.Study Design: Retrospective cohort study using the TriNetX national clinical database.Setting: Analysis of a nationwide patient cohort.Methods: The database was queried for patients with OSA (International Classification of Diseases, 10th Revision: G47.33) treated with CPAP (Current Procedural Terminology [CPT]: 94660), sleep surgery (CPT: 42145, 64568, 64582), or no treatment. Sleep surgeries included uvulopalatopharyngoplasty and hypoglossal nerve stimulation. 1:1 Propensity score matching (PSM) ensured cohort comparability by age, sex, and race. MVA incidence rates were analyzed, along with a subanalysis of OSA‐related comorbidities as a surrogate of OSA severity.Results: A total of 2,832,437 patients with OSA were identified. Patients with OSA undergoing sleep surgery demonstrated a significantly lower incidence of MVAs (3.403%) compared to the OSA + CPAP cohort (6.072%) and the no‐treatment group (4.662%). The odds ratio (OR) for MVA incidence in the no‐treatment group compared to the OSA + sleep surgery cohort was 1.214 (95% confidence interval [CI]: 1.060‐1.391, P = .0051). The OR in the OSA + sleep surgery cohort compared to the CPAP cohort was 0.545 (95% CI: 0.480‐0.618, P < .0001). Patients with OSA who experienced MVAs were more likely to have comorbidities such as hypertension, diabetes, and heart failure following the accident.Conclusion: Sleep surgery significantly reduces the risk of MVAs in patients with OSA compared to CPAP and no treatment. In appropriate candidates, surgery should be considered to mitigate personal and public health risks associated with OSA.

10 October 2023 in Clinical Case Reports
Verified
Journal Article
Open access
ESElliott Sina
Elliott Sina
CHChihun J. Han
Chihun J. Han
ECElizabeth E. Cottrill
Elizabeth E. Cottrill
Aberrant migration of parathyroid glands from their embryologic origin may result in undescended parathyroid glands. We present a case of an ectopic parathyroid adenoma at the level of the pyriform sinus. A 41‐year‐old female was evaluated for primary hyperparathyroidism. Following non‐localizing ultrasound and planar sestamibi imaging, the patient underwent SPECT/CT and 4‐D computed tomography demonstrating evidence of an ectopic parathyroid adenoma. The surgical approach was modified based on the location. Following extirpation, PTH fell from 80 to 16 pg/mL, and the 15‐min post‐excision level remained stable at 14pg/mL, indicating a biochemical cure. While rare, undescended parathyroid adenoma should be considered when preoperative imaging fails to identify a target adenoma or after unsuccessful surgery. The combined use of 99m Tc‐MBI or 4D CT and other anatomical scans may improve diagnostic accuracy. Due to the potential need to perform a second incision to conduct a four‐gland exploration, preoperative patient discussion regarding surgical risks may differ from that of a standard parathyroidectomy.