Comparative Performance of ChatGPT 3.5 and GPT4 on Rhinology Standardized Board Examination Questions.

Objective: Advances in deep learning and artificial intelligence (AI) have led to the emergence of large language models (LLM) like ChatGPT from OpenAI. The study aimed to evaluate the performance of ChatGPT 3.5 and GPT4 on Otolaryngology (Rhinology) Standardized Board Examination questions in comparison to Otolaryngology residents. Methods: This study selected all 127 rhinology standardized questions from www.boardvitals.com, a commonly used study tool by otolaryngology residents preparing for board exams. Ninety-three text-based questions were administered to ChatGPT 3.5 and GPT4, and their answers were compared with the average results of the question bank (used primarily by otolaryngology residents). Thirty-four image-based questions were provided to GPT4 and underwent the same analysis. Based on the findings of an earlier study, a pass-fail cutoff was set at the 10th percentile. Results: On text-based questions, ChatGPT 3.5 answered correctly 45.2% of the time (8th percentile) (P = .0001), while GPT4 achieved 86.0% (66th percentile) (P = .001). GPT4 answered image-based questions correctly 64.7% of the time. Projections suggest that ChatGPT 3.5 might not pass the American Board of Otolaryngology Written Question Exam (ABOto WQE), whereas GPT4 stands a strong chance of passing. Discussion: The older LLM, ChatGPT 3.5, is unlikely to pass the ABOto WQE. However, the advanced GPT4 model exhibits a much higher likelihood of success. This rapid progression in AI indicates its potential future role in otolaryngology education. Implications for Practice: As AI technology rapidly advances, it may be that AI-assisted medical education, diagnosis, and treatment planning become commonplace in the medical and surgical landscape. Level of Evidence: Level 5.

The Use of Artificial Intelligence to Improve Readability of Otolaryngology Patient Education Materials.

OBJECTIVE: The recommended readability of health education materials is at the sixth-grade level. Artificial intelligence (AI) large language models such as the newly released ChatGPT4 might facilitate the conversion of patient-education materials at scale. We sought to ascertain whether online otolaryngology education materials meet recommended reading levels and whether ChatGPT4 could rewrite these materials to the sixth-grade level. We also wished to ensure that converted materials were accurate and retained sufficient content. METHODS: Seventy-one articles from patient educational materials published online by the American Academy of Otolaryngology-Head and Neck Surgery were selected. Articles were entered into ChatGPT4 with the prompt "translate this text to a sixth-grade reading level." Flesch Reading Ease Score (FRES) and Flesch-Kincaid Grade Level (FKGL) were determined for each article before and after AI conversion. Each article and conversion were reviewed for factual inaccuracies, and each conversion was reviewed for content retention. RESULTS: The 71 articles had an initial average FKGL of 11.03 and FRES of 46.79. After conversion by ChatGPT4, the average FKGL across all articles was 5.80 and FRES was 77.27. Converted materials provided enough detail for patient education with no factual errors. DISCUSSION: We found that ChatGPT4 improved the reading accessibility of otolaryngology online patient education materials to recommended levels quickly and effectively. IMPLICATIONS FOR PRACTICE: Physicians can determine whether their patient education materials exceed current recommended reading levels by using widely available measurement tools, and then apply AI dialogue platforms to modify materials to more accessible levels as needed. LEVEL OF EVIDENCE: Level 5.

Resection of Intracochlear Schwannomas With Immediate Cochlear Implantation.

INTRODUCTION: Intralabyrinthine schwannomas, including the intracochlear variety, are rare benign tumors. They can cause a number of symptoms and have the potential to grow to involve other critical structures of the inner ear and skull base. While surgical resection is feasible, there is typically permanent hearing dysfunction as a result of resection and subsequent fibrosis. Here, we present 2 cases of intracochlear schwannomas (ICS) that were successfully resected with simultaneous cochlear implant placement. METHODS: Patient 1 presented with an intravestibulocochlear schwannoma. This patient underwent a translabyrinthine approach. Endoscopic assistance was used to dissect the tumor from the vestibule and basal turn of the cochlea, through an enlarged round window approach. A cochlear implant was placed via a round window cochleostomy. Patient 2 presented with an intracochlear schwannoma involving the basal and middle turns of the cochlea. The patient underwent a trans-otic approach for resection. A large portion of the cochlear promontory required unroofing for complete exposure of the tumor. A cochlear implant was then placed via a round window cochleostomy. RESULTS: Upon cochlear implant activation, Patient 1's sound field thresholds using the implant were near the normal range of hearing, ranging from 25 to 50 dB HL from 250 to 6000 Hz. Patient 2's 6-month post-operative cochlear implant sound field testing ranged from 20 to 30 dB HL from 250 to 6000 Hz and speech recognition testing revealed 59% on AZ Bio sentences compared to 0% pre-operatively. CONCLUSION: Simultaneous cochlear implantation after resection of intracochlear schwannomas is safe and successful in restoring hearing. Attention to adequate exposure and endoscopic assistance, when required, allow for gross total resection while minimizing trauma to cochlear structures. In such cases, immediate cochlear implantation allows for hearing rehabilitation before likely cochlear fibrosis can occur.

Evaluation of galectin-3 and intestinal fatty acid binding protein as serum biomarkers in autosomal recessive polycystic kidney disease.

BACKGROUND: Autosomal recessive polycystic kidney disease (ARPKD) causes fibrocystic kidney disease, congenital hepatic fibrosis, and portal hypertension. Serum galectin-3 (Gal-3) and intestinal fatty acid binding protein (I-FABP) are potential biomarkers of kidney fibrosis and portal hypertension, respectively. We examined whether serum Gal-3 associates with kidney disease severity and serum I-FABP associates with liver disease severity in ARPKD. METHODS: Cross-sectional study of 29 participants with ARPKD (0.2-21 years old) and presence of native kidneys (Gal-3 analyses, n = 18) and/or native livers (I-FABP analyses, n = 21). Serum Gal-3 and I-FABP were analyzed using enzyme linked immunosorbent assay. Kidney disease severity variables included estimated glomerular filtration rate (eGFR) and height-adjusted total kidney volume (htTKV). Liver disease severity was characterized using ultrasound elastography to measure liver fibrosis, and spleen length and platelet count as markers of portal hypertension. Simple and multivariable linear regression examined associations between Gal-3 and kidney disease severity (adjusted for liver disease severity) and between I-FABP and liver disease severity (adjusted for eGFR). RESULTS: Serum Gal-3 was negatively associated with eGFR; 1 standard deviation (SD) lower eGFR was associated with 0.795 SD higher Gal-3 level (95% CI - 1.116, - 0.473; p < 0.001). This association remained significant when adjusted for liver disease severity. Serum Gal-3 was not associated with htTKV in adjusted analyses. Overall I-FABP levels were elevated, but there were no linear associations between I-FABP and liver disease severity in unadjusted or adjusted models. CONCLUSIONS: Serum Gal-3 is associated with eGFR in ARPKD, suggesting its value as a possible novel biomarker of kidney disease severity. We found no associations between serum I-FABP and ARPKD liver disease severity despite overall elevated I-FABP levels.

Outcomes of curative treatment for head and neck squamous cell carcinoma in very elderly adults ≥80 years old.

BACKGROUND: Head and neck cancer treatment can be difficult and advancing age is associated with greater frailty. It is unclear if curative treatment for very elderly patients is beneficial. This study compared outcomes to curative treatment in patients ≥80 aged with HNSCC to patients aged 70-79. METHODS: Retrospective study of 114 patients diagnosed with HNSCC. Overall survival (OS), Disease-Free Survival (DFS), and local-regional control (LRC) were compared and adjusted for confounders. RESULTS: Patients aged 70-79 had a higher median OS (35 months [95% CI, 19.58-50.42]) compared with patients aged ≥80 (19 months [11.72-26.28]; p = 0.008) but similar DFS and LRC. KPS < 90 was the stand-alone independent prognostic factor for OS (HR = 2.14 [1.05-4.38]). CONCLUSION: Very elderly HNSCC patients (aged ≥80) can have favorable outcomes with curative therapy and advanced chronological age alone should not prohibit patients from receiving treatment. Performance status may be a greater predictor of survival outcome than age alone. LEVEL OF EVIDENCE: Level III.

Paclitaxel use in pregnancy: neonatal follow-up of infants with positive detection of intact paclitaxel and metabolites in meconium at birth.

Paclitaxel is often excluded during pregnancy for women with breast cancer due to limited neonatal follow-up. We confirmed in utero fetal Paclitaxel exposure for 8 newborns. Birth details and follow-up to 36 months of age is reported. Meconium samples from newborns exposed to chemotherapy were screened by liquid chromatography-high resolution mass spectrometry while blinded to maternal treatment during pregnancy. Newborn information at birth and annually was obtained. Mean gestational age (GA) at cancer diagnosis and start of chemotherapy was 8.7 + 6.2 weeks and 17.1 ± 3.5 weeks. Paclitaxel was started at a mean GA of 27.0 ± 5.8 weeks. Paclitaxel followed Doxorubicin/Cyclophosphamide in 6 cases, 5-Fluouracil/Doxorubicin/Cyclophosphamide in 1, and was used alone in 1. Mean number of days between Paclitaxel and birth was 23 ± 15. Identification of Paclitaxel and/or metabolites was made in all meconium from paclitaxel-exposed fetuses. Birthweight was < 10% for GA in 3 infants. Three anomalies occurred: mild hip dysplasia without further treatment and mitral valve stenosis. The third child was diagnosed with Cleidocranial Dysostosis, a familial anomaly. Mean age at pediatric follow-up is 18.7 + 9.3 months. Pediatricians report eczema and recurrent otitis media in 1 child, iron deficiency anemia and upper respiratory infection in 2. One child is < 10% for height and weight at 15 months. All are meeting developmental milestones at median age of 18.7 months, range: 6-36 months. CONCLUSION: Up to 3 years of age, follow-up of neonates exposed to Paclitaxel in utero is reassuring. Continued observation of neonatal development is essential. WHAT IS KNOWN: • Chemotherapy during the second and third trimester of pregnancy does not result in an increase in congenital malformations or developmental delay. • In non-human primate studies by Van Calsteren et al., variable plasma and/or tissue concentrations of taxanes, carboplatin, and trastuzumab were encountered in the fetal compartment. • Pilot data reported by the current investigators proved that paclitaxel crosses the human placenta. WHAT IS NEW: • This current article provides medical and developmental follow up on the newborns from this exposure for 3 years after birth.

A Virtual Curriculum to Prepare Medical Students to Achieve Accreditation Council for Graduate Medical Education Level-1 Milestones in Orthopaedic Surgery.

As a part of the American healthcare system's response to the Coronavirus Disease 2019 (COVID-19) global pandemic, the Association of American Medical Colleges recommended that medical schools temporarily remove students from clinical settings and transition to an entirely online learning environment. This posed an unprecedented challenge to students in the clinical years of their medical education. To address this unexpected shift, we modified an in-person workshop to teach orthopaedic trauma basics to 5-week virtual course for third year medical students from several schools in New Jersey and Pennsylvania. We focused on moving students toward the Level-1 milestones for basic fracture care with a combination of weekly lectures and virtual interactive small group sessions, all conducted via WebEx and proctored by an orthopaedic attending or resident. The course was well received by students. Participation in the course was completely voluntary and did not count for credit at any institution. The course was valuable to students because the students who registered chose to fully complete the 5-week course and no student missed more than one small group session. On a postcourse survey, 100% of students said they would be highly likely to recommend the course to a future student, and the average rating for educational value of the course was 4.98 of 5. Given the current limitations in clinical education because of the COVID-19 pandemic, our course provides a reasonable alternative to clinical experience and prepares students with the knowledge and many of the skills that are required to succeed as orthopaedic interns. Furthermore, the success of our course this year suggests that similar programing may be a useful adjunct to clinical experiences even when it is safe to return to more traditional medical school scheduling.