Nelson, Darlene R.

Abstract Rationale: Airway-stenosing disease is a significant clinical problem in multiple populations. While the immediate endoscopic and clinical response to bronchoscopic interventions are high, rates of clinically significant recurrence are elevated, and often require multiple and frequent reinterventions. Newer strategies to prevent recurrence of central and proximal airway stenosis are needed. Methods: Retrospective review of bronchoscopic balloon bronchoplasties performed at the Mayo Clinic utilizing paclitaxel coated balloons (PCB) in 2023-2024 for recurrent airway stenosis. IN.PACT Admiral and AV balloons (sizes 5-12 mm, 3.5 mcg/mm2 paclitaxel dose) were used with an inflation time of 2-3 minutes. Procedural success was defined as <50% airway stenosis or resolution of the procedural indication. Reintervention was considered only if a therapeutic bronchoscopy was subsequently required to address the same lesion treated with PCB. Only patients with follow-ups >90 days were included in the reintervention inter-procedural time analysis. Results: Thirty-three procedures were performed in 18 patients. Seventeen (51%) for lung transplantation-associated stenosis (29% anastomotic/perianastomotic, 76% lobar and segmental stenoses), 8 (24%) for granulomatous polyangiitis-associated stenosis, 3 for postintubation tracheal stenosis and 2 for chemical/pill injury. Eighteen cases (54%) utilized an additional dilation modality (12 non-coated balloon dilation, 4 rigid dilation), 5 spray cryotherapy, 6 other ablative therapies (4 CO2-laser, 2 electrocautery cuts), and 6 received an airway stent during the same procedure. 57% percent of interventions occurred in the lobar, 27% mainstem/bronchus intermedius, 9% tracheal, 7% segmental airways. Thirty-one (93%) cases had a satisfactory initial endoscopic result. The median number of bronchoscopic therapeutic interventions required before the use pf PCB was 5 (IQR 2.5-11.5), which significantly decreased to 0.1 (IQR 0-1, range 0-9) after PCB (p <0.001). The reintervention-free survival significantly increased after PCB dilation from a median 35 days (IQR 28.7-80.5) to 132 days (IQR 64-197; p=0.004). Twenty-one cases required subsequent bronchoscopies, sixteen (76%) of which demonstrated improvement in lesion severity when compared to the results of prior interventions. Only 16 (48%) required therapeutically addressing the same lesion. Only 2 minor complications occurred (6%, self-limited bleeding and unplanned admission), none related to paclitaxel use. Six-month mortality was 3% from an unrelated cause. Conclusions: The locoregional antiproliferative effects of paclitaxel show promise in preventing recurrence, minimizing the need for repeat procedures, and extending intervals between interventions in patients with recurrent, predominantly benign airway stenosis. Its use is feasible, safe, and future research is warranted, particularly comparing PCB with the currently available strategies.

Abstract Large language models are becoming ubiquitous in the editing and generation of written content and are actively being explored for their use in medical education. The use of artificial intelligence (AI) engines to generate content in academic spaces is controversial and has been meet with swift responses and guidance from academic journals and publishers regarding the appropriate use or disclosure of use of AI engines in professional writing. To date, there is no guidance to applicants of graduate medical education programs in using AI engines to generate application content—primarily personal statements and letters of recommendation. In this Perspective, we review perceptions of using AI to generate application content, considerations for the impact of AI in holistic application review, ethical challenges regarding plagiarism, and AI text classifiers. Finally, included are recommendations to the graduate medical education community to provide guidance on use of AI engines in applications to maintain the integrity of the application process in graduate medical education.

Background: CO2 Laser (CO2L) technology deployable through flexible endoscopes now allows for their use throughout the airway, although published data are limited. Methods: Retrospective analysis of CO2L bronchoscopic procedures, excluding glottic and subglottic interventions. Procedural success was defined as >50% visual reduction in airway obstruction in the area treated or resolution of the procedural indication. Results: Seventy-two procedures were performed on 36 patients. Nonmalignant indications comprised 66%: stent-associated granulation tissue (28%), granulomatosis with polyangiitis lesions (23%), and lung transplant-related granulation tissue (16%) were the most common. Bronchoscopic access was flexible only in 81% and primarily rigid (combined with flexible) in 18%. The site of intervention was the trachea at 19%, the mainstem at 56%, and lobar/segmental airways at 45%. Procedural success was 89%. CO2L was used exclusively in 19%; in 81%, additional techniques were required, most commonly balloon dilation (59%), cryo-debulking (23%), and rigid dilation (16%). Malignant indications had a nonsignificant trend toward requiring adjuvant techniques (P=0.05). Seventy-six percent of the patients required more than 1 procedure. CO2L exclusive cases had no statistically different needs for subsequent therapeutic bronchoscopies (P=0.10) or time to reintervention (109 vs. 41 days, P=0.07), and reintervention-free survival was similar (P=0.10) and difficult to predict. The complication rate attributable to CO2L was 2.7%. Conclusion: CO2L is a safe and useful tool when precise cutting and vaporization are desired. Its use in multi-modality approaches has high levels of success in adequately selected lesions, adding an ablative potential to dilation techniques. Vasculitis-associated scars/webs and granulation tissue (including stent-associated) appear to be ideal targets.