Scientific Abstract Presentations: Research & Technology Session 2 Friday, October 16 Fri, Oct 16 8:00-9:30 a.m. CDT

Background: Following the transition of USMLE Step 1 to P/F in 2022, more applicants have pursued dedicated research years, emphasizing research productivity as a key factor. Prior studies have described broad trends in research output and match success, but there is limited data examining where research years were completed relative to the matched program or how sub-Is, regional preference, and institutional factors have evolved. This study examines recent shifts in these factors among matched applicants across the last three cycles, comparing two post–Step 1 P/F cohorts (2024–2025) with the preceding 2023 cohort.

Methods: Data on matched applicants, sub-Is, research years, and institutional characteristics, were obtained from a publicly available, applicant-reported Google Sheet for the 2023–2025 integrated plastic surgery match cycles. Region and medical school rankings were verified using Doximity and U.S. News & World Reports. Categorical variables were compared across years using Chi-square or Fisher's exact tests, with two-sample z-tests applied for pairwise comparisons.

Results:A total of 207, 214, and 223 applicants successfully matched in 2023, 2024, and 2025, respectively. The proportion of applicants matching to their home institution remained stable across all three years (18.4%, 18.2%, 17.9%; p=0.94). In contrast, the proportion completing a sub-I at their matched program increased significantly from 30.9% in 2023 to 38.8% in 2024 and 38.6% in 2025 (p=0.048). The proportion completing a research year also rose markedly from 9.7% in 2023 to 30.4% in 2024 and 29.6% in 2025 (p<0.001). Although not statistically significant, there was a shift in where research years were conducted. More applicants completed research at their eventual matched program (7.5% in 2024 vs. 11.7% in 2025; p=0.21) and fewer at other institutions (22.9% vs. 17.9%; p=0.18). Most matched applicants graduated from medical schools with an associated plastic surgery residency, with rates remaining stable over time (89.4%, 83.6%, 85.2%; p=0.19). Similarly, over one-third of matched applicants came from top 40 medical schools (38.6%, 43.5%, 37.2%; p=0.32). Same-region patterns were also consistent across cycles (50.7%, 45.8%, 52.0%; p=0.39), with the highest rates observed in the Northeast (69%, 57.1%, 66.1%) and South (56.9%, 47.5%, 61.9%), followed by the Midwest (45.8%, 39.7%, 45.6%) and West (21.6%, 35.9%, 24.4%).

Conclusions: Following the Step 1 P/F transition, more matched applicants completed research years and sub-Is at the institutions where they later matched, even as overall home program match rates remained stable. This trend suggests that engagement through research or sub-Is, rather than home program status alone, has become increasingly influential in match success, highlighting the growing importance of networking and program familiarity over sheer research productivity in the absence of a numerical Step 1 score.

Background: Preclinical studies of autologous fat graft (AFG) retention have traditionally relied on athymic nude mice, which lack functional T lymphocytes but retain innate immune function. While permissive for xenograft survival, this model may inadequately capture adaptive immune–mediated regulation of angiogenesis and inflammatory remodeling that influence graft survival in humans. Humanized murine models, engrafted with human hematopoietic stem cells, support reconstitution of human lymphoid and myeloid populations and provide a more physiologically representative immune microenvironment. To our knowledge, humanized murine models have not previously been used to study autologous fat graft retention. We hypothesized that immune phenotype significantly influences fat graft remodeling and volumetric retention.

Methods: Female athymic (n=4) and humanized (n=4) mice were studied. Humanized mice were engrafted with human CD34⁺ hematopoietic stem cells prior to fat grafting. To control donor–level variability, lipoaspirate from each human donor was injected into matched athymic–humanized mouse pairs. A single bolus (≤1 mL) of human lipoaspirate was injected subcutaneously into the dorsum of each animal. Grafts were evaluated using high-resolution 7-Tesla MRI. Three-month volumetric retention was quantified using standardized segmentation and normalized to injected volume. Donor-matched retention was compared using a two-sided paired t-test (α=0.05).

Results: Mean volumetric retention was higher in humanized mice compared to athymic mice (44.4% ± 20.2 vs 21.8% ± 14.5), representing an approximately two-fold increase in retained volume. In donor-matched paired analysis, humanized mice demonstrated significantly greater retention (mean paired difference 22.6%, 95% CI 9.3–36.0; p=0.013). Retention in humanized mice ranged from 27–75.6%, compared to 11.3–44.1% in athymic controls.

Conclusions: Immune phenotype significantly influences preclinical fat graft retention. Enhanced persistence in humanized mice suggests that adaptive immune–mediated inflammatory remodeling contributes to graft survival. These findings indicate that commonly used athymic models may underestimate volumetric retention and position the humanized murine model as a novel, more physiologically representative platform for investigating adipose graft biology and improving translational relevance in fat graft research.

Background: Reduction mammoplasty remains one of the most frequently performed procedures in plastic surgery. Though effective at reducing symptoms of macromastia, wound healing complications are relatively common, with complications most frequently seen along incision junction sites, areas that experience the greatest tension and ischemia.(1) Altrazeal Transforming Powder Dressing (TPD) is a patented, advanced wound care product consisting of non-resorbable methacrylate-based granules that transform upon contact with moisture, producing a non-occlusive, flexible film that can conform to complex wound topography. The application of TPD in the setting of postoperative reduction mammoplasty has not been previously described.

Methods: This case series includes 10 patients who presented with delays or complications in wound healing following bilateral reduction mammoplasty in a single-surgeon, single-center institutional setting. Wound complications were identified at a range of 7–14 days postoperatively, presenting with superficial dehiscence and delayed healing requiring local wound care. Prior to TPD application, patients were managed with standard local wound care. Once selected for treatment, per product application protocol, Altrazeal TPD was applied to cleansed, moistened wound beds. Upon transformation, the resulting flexible matrix conformed to the irregular wound topography characteristic of reduction mammoplasty incisions. Outcomes were assessed clinically and photographically, with scar quality evaluated using the validated Scar Cosmesis Assessment and Rating (SCAR) scale.

Results: Treatment with Altrazeal TPD yielded consistent wound closure and favorable aesthetic outcomes across all patients. Complete wound closure was achieved in all wounds within approximately 4-8 weeks of initiating TPD therapy, without the need for surgical revision or secondary procedures. Aesthetic results were encouraging, with preservation of breast shape and contour. Final scars exhibited favorable pigmentation, vascularization, and suppleness on assessment. The commonly reported sequela of increased scar width was observed, given the pathophysiology of healing by secondary intention, though overall cosmetic outcomes were not compromised. No adverse reactions to TPD were reported.

Conclusion: This case series describes the first reported application of Altrazeal TPD in the management of wound healing complications following reduction mammoplasty. It demonstrated reliable wound closure, favorable aesthetic outcomes, and an excellent safety profile across all patients without the need for surgical intervention. The unique ability to conform to complex three-dimensional wound geometry, maintain a moist healing environment, and minimize the burden of frequent dressing changes makes it particularly well-suited for postoperative breast wounds. These findings suggest a promising role for TPD in post-operative care and warrant further investigation for standardization into wound care protocols in breast surgery.

References: 1. American Society of Plastic Surgeons. (2025). 2024 Procedural Statistics Release. https://www.plasticsurgery.org/plastic-surgery-statistics

Purpose: Comprehensive patient assessment extends beyond objective clinical parameters to include a patient's perception of their health and recovery. Patient reported outcome measures (PROMs) have emerged as a cornerstone of patient centered care, providing clinicians with insight into patient's subjective experiences.(1–3) Despite the value, the collection of PROM data has been limited in plastic surgery practice by logistical barriers, often due to the labor intensive process of patient recruitment, survey distribution, and data collection.(3)The integration of PROM data collection directly within the electronic medical record (EMR) may help address these challenges by enabling automation of the more cumbersome processes. In this study, we present our early experience integrating the Vanderbilt Mini-PROM-Breast (VMP-B), a validated short-form patient reported outcome measure for breast surgery, into our institutional EMR at Vanderbilt University Medical Center (VUMC).

Methods: The VMP-B survey was integrated into Epic in close collaboration with our institutional information technology (IT) team. Automated patient identification was completed via a scheduled weekly process that queried the EMR for ICD-10 breast cancer diagnoses, and an upcoming surgery containing a mastectomy or breast reconstruction CPT code. Patents received the survey and subsequent notifications if not completed, at predefined time points: 7 days pre-operation, 1-, 3-, 6-, and 12-months post-operation. Survey responses remained securely stored in the EMR and available for extraction using Epic's SlicerDicer tool. The primary outcome was the overall survey response rate, defined as the proportion of eligible patients who completed at least one VMP-B survey.

Results: Between January 2024 and October 2025, a total of 694 patients scheduled for breast procedures were automatically identified by the EMR as eligible to receive the VMP-B survey. Among these, 379 patients completed at least one survey, yielding an overall response rate of 55%. Our results show that patient response rate for subsequent surveys decreased, with 33.5% of patients completing more than one survey.

Conclusion: Integration of the VMP-B into Epic was successfully achieved, with response rate exceeding outcomes previously reported for web-based surveys which range from 46 - 51%.(4) This implementation establishes a reproducible and adaptable framework for embedding validated, short-form PROMs into the EMR to increase survey response rate while decreasing burden to clinical staff.

References 1. Tan YH, Siew JX, Thomas B, Ng KC. Patient-reported outcome measures and value-based medicine in paediatrics: a timely review. Singapore Med J. 2023;64(5):285-293. doi:10.11622/SMEDJ.2021102 2. Deshpande P, Rajan S, Sudeepthi Bl, Abdul Nazir C. Patient-reported outcomes: A new era in clinical research. Perspect Clin Res. 2011;2(4):137. doi:10.4103/2229-3485.86879 3. Biswas S, Mullikin A, Phillips BT. Trends in Patient-Reported Outcomes Reporting in Breast Reconstruction: A Scoping Literature Review. Ann Plast Surg. 2023;90(5):501-505. doi:10.1097/SAP.0000000000003545 4. Meyer VM, Benjamens S, El Moumni M, Lange JFM, Pol RA. Global Overview of Response Rates in Patient and Health Care Professional Surveys in Surgery: A Systematic Review. Ann Surg. 2022;275(1):E75-E81. doi:10.1097/SLA.0000000000004078

Background:
 Manual extraction of operative details from electronic health records (EHRs) is time-consuming, error-prone, and inconsistent between reviewers(1). Generative large language models (LLMs) offer the ability to convert unstructured operative reports into structured datasets and may provide a scalable alternative to traditional chart abstraction(2,3). Breast reconstruction using the deep inferior epigastric perforator (DIEP) flap relies on detailed operative documentation, including perforator selection, flap configuration, recipient vessel selection, anastomotic technique, venous outflow strategy, and intraoperative events. These variables are critical for outcomes research, quality assessment, and registry development, yet their narrative documentation makes manual data extraction resource-intensive (4). This study evaluated the accuracy of a HIPAA-compliant LLM workflow for automated extraction of operative variables into a structured DIEP flap registry.

Methods:
 A single-center retrospective study included 230 consecutive bilateral DIEP flap breast reconstructions (460 breasts) performed between March 2017 and January 2024. Operative reports were obtained from the Epic EHR and processed using UCSF's HIPAA-secure VERSA platform, which provides protected access to OpenAI models without external data retention or training. GPT-3.5-Turbo and GPT-4o were evaluated using an API-based workflow with temperature set to 0 to ensure deterministic outputs. A structured prompt was iteratively developed to extract 33 predefined operative variables per breast, formatted as one breast per row, with bilateral cases generating two rows. Two trained reviewers independently performed manual abstraction using a predefined data dictionary as the reference standard, with discrepancies resolved through adjudication. The primary outcome was per-variable concordance between LLM output and manual chart review, and paired McNemar tests were used to compare model performance across identical cases.

Results:
 A total of 15,180 data points (33 variables x 460 breasts) were evaluated. Overall exact-match accuracy was 96.9% for GPT-4 compared with 88.4% for GPT-3.5, with significantly higher concordance for GPT-4 (p < 0.001). GPT-4 demonstrated high accuracy for routinely documented operative variables, including laterality, lymph node procedures, mesh use, and regional blocks. Microvascular variables, including recipient vessel identification and anastomotic characteristics, were also extracted with high accuracy. Performance declined for rare intraoperative events: arterial thrombosis demonstrated sensitivity of 0.50 with specificity of 0.986, and venous congestion demonstrated sensitivity of 0.304 with specificity of 0.986, reflecting high specificity with occasional missed true events. Automated extraction reduced abstraction time from approximately 38 hours of manual review to 3 hours of end-to-end processing, representing a 12.7-fold reduction in workload.

Conclusions:
 A HIPAA-compliant LLM workflow can accurately and efficiently extract detailed DIEP flap operative data from unstructured operative reports. GPT-4 demonstrated significantly higher concordance than GPT-3.5 while substantially reducing manual abstraction burden. Automated extraction performed best for standardized operative variables and detailed microvascular fields, whereas rare intraoperative events may require structured documentation and selective human verification.

1.Wong GC, Chung KC. Natural Language Processing.Hand Clin. 2026;42(1):27-36. 2.Usuyama N, Wong C, Zhang S, et al. Biomedical Natural Language.Annu Rev Biomed Data Sci. 2025;8:471-490. 3.Gu B, Shao V, Liao Z, et al. Scalable information extraction.BMC Med Res Methodol. 2025;25:23. 4.Penev YP, Buchanan TR, Ruppert MM, et al. Electronic Health Record.JMIR Med Inform. 2024;12:e58130.

Purpose: Biomaterials are capable of eliciting biological responses when implanted into living tissues. Among them, Bioglass has demonstrated promising bioactive properties, including stimulation of cellular activity and enhancement of tissue regeneration (1). This study aimed to evaluate the effects of Bioglass on cutaneous wound healing in a murine model.

Methods: An initial in vitro analysis was conducted to determine optimal Bioglass concentrations based on cell viability. Concentrations of 1% and 2%, associated with higher viability rates, were selected for in vivo evaluation. Sixteen male Wistar rats (200–250 g) underwent standardized surgical creation of four full-thickness dorsal skin wounds per animal. The animals were allocated into three groups: Control (no treatment), 1% Bioglass, and 2% Bioglass. Each group was subdivided according to evaluation time points (7 and 14 postoperative days). At euthanasia, tissue samples were collected for histological and molecular analyses, including inflammatory cell count, vascular density, follicular appendage quantification, granuloma percentage, and angiogenesis-related gene expression (VEGF and ICAM-1).

Results: On postoperative day 7, the 1% Bioglass group exhibited a 1.4-fold increase in angiogenesis compared to controls (55.5 vs. 76; p = 0.04), while the 2% Bioglass group demonstrated a 1.5-fold increase (55.5 vs. 86; p = 0.02). No significant difference was observed between the 1% and 2% groups at this time point. By day 14, the control group showed higher angiogenesis compared to the 1% Bioglass group (106.5 vs. 76; p = 0.008). However, the 2% Bioglass group demonstrated significantly greater vascular density than the 1% group (109.5 vs. 86; p = 0.004). Gene expression analysis revealed significant modulation of VEGF and ICAM-1 across groups and time points, suggesting regulation of angiogenic and inflammatory pathways.

Conclusion: Bioglass enhanced early neovascularization without triggering excessive inflammatory response, indicating its potential as a therapeutic adjunct in wound healing. Further controlled studies are warranted to confirm these findings and to elucidate long-term regenerative outcomes.

References: 1- Filip GA, Achim M, Mihalte P, Miclaus MO, Cristea C, Melinte G, Gheban B, Munteanu DM, Cadar O, Simon I, Pana O, Barbu Tudoran L, Clichici S, Stefan R. Design, in vitro bioactivity and in vivo influence on oxidative stress and matrix metalloproteinases of bioglasses in experimental skin wound. J Trace Elem Med Biol. 2021 Dec;68:126846. doi: 10.1016/j.jtemb.2021.126846. Epub 2021 Aug 20. PMID: 34438314.

PURPOSE: Nonthermal plasma (NTP) has emerged as a surface-modification strategy to increase surface hydrophilicity, protein adsorption, and cellular adhesion [1]. However, the extent to which NTP treatment translates to improved osseointegration outcomes of implant surfaces requires further investigation. Therefore, this study evaluated whether NTP treatment improves osseointegration of endosteal implants in type IV bone in an ovine model.

MATERIALS AND METHODS: Titanium implants were categorized as untreated controls (FGM) or NTP-treated implants (FGMP) and placed into the iliac crest of n=10 adult sheep. In treated implants, NTP (using argon gas) was applied immediately prior to surgical placement as previously described [2], and animals were allowed to heal for either 6 or 12 weeks post-implantation (n=5 per time point). Histological sections were analyzed for bone-to-implant contact (BIC) and bone area fractional occupancy (BAFO), expressed as percentages.

RESULTS: All implants demonstrated successful osseointegration at both healing time points. Peri-implant trabecular bone networks were observed in both groups at 6 weeks; however, FGMP implants exhibited more continuous bone apposition along the implant surface. However, there were no significant differences in %BIC or %BAFO (p=0.071 and p=0.171, respectively). By 12 weeks, FGMP implants demonstrated a denser, more interconnected peri-implant trabecular network in direct contact with the implant surface, along with more mature woven bone formation and bridging across healing chambers. Quantitative histomorphometric analysis supported these findings, with FGMP implants demonstrating significantly greater %BIC and %BAFO compared with untreated controls (p=0.046 and p=0.007, respectively).

CONCLUSIONS: Brief NTP treatment immediately prior to implantation significantly improved long-term osseointegration at 12 weeks. Because NTP can be applied intraoperatively without altering implant design, it represents a practical and scalable adjunct for surgical applications requiring stable fixation. This approach may be particularly relevant for craniofacial fixation and orthopedic hardware, where reliable bone-implant integration is critical for long-term stability and prevention of fixation failure and warrants further evaluation.

REFERENCES: 1. Nayak VV, Mirsky NA, Slavin BV, Witek L, Coelho PG, Tovar N. Non-Thermal Plasma Treatment of Poly(tetrafluoroethylene) Dental Membranes and Its Effects on Cellular Adhesion. Materials (Basel). 2023;16(20):6633. Published 2023 Oct 10. doi:10.3390/ma16206633 2. Panariello BHD, Denucci GC, Tonon CC, et al. Tissue-Safe Low-Temperature Plasma Treatment for Effective Management of Mature Peri-Implantitis Biofilms on Titanium Surfaces. ACS Biomater Sci Eng. 2024;10(12):7647-7656. doi:10.1021/acsbiomaterials.4c01413

Background Operating rooms are significant contributors to environmental waste, pollution, and healthcare expenses. Despite compelling arguments questioning the necessity of full gowning and draping during small skin and superficial procedures, many surgeons continue to adhere to traditional practices, generating excess waste. Disposable gowns and drapes made from Spunbond-Meltblown-Spunbond (SMS) polypropylene plastic are particularly problematic due to their carbon-intensive production and non-biodegradability. Our study aims to assess the effectiveness of minimal draping techniques and minimal surgical packs in reducing OR waste.

Methods After receiving Institutional Review Board approval, a prospective observational analysis was conducted on patients who underwent superficial excisional repairs by four plastic and reconstructive surgeons from July 2023 to August 2025 at a children's hospital. Participating surgeons were stratified into two groups based on their use of either standard or minimal draping and minimal packs techniques. Waste was sorted and weighed post-operatively using standardized protocols. Clinical outcomes were recorded on chart review. Independent t-tests were performed to compare waste outputs between the cohorts. Life-cycle analysis was conducted to evaluate the environmental impact of Spunbond-Meltblown-Spunbond (SMS) polypropylene plastic drapes and gowns from resource extraction to disposal. A cost analysis was also performed using robust regression to account for variability and outliers in cost data, providing reliable estimates of savings associated with reduced gown and drape utilization.

Results Data was collected from 40 cases, with 20 cases using minimal draping. Cases performed with a minimal draping protocol generated significantly less waste compared to standard draping methods, with a mean reduction of 3.1 kg per case (p<0.05). Waste reduction was predominantly driven by eliminating disposable gowns and decreasing drape-related material use, saving 158g of gowns per person scrubbed and 0.47 kg of drapes per case (p<0.05). Based on the life cycle analysis of SMS fabric, this reduction translates to an estimated 2.9-5.3 kg CO2e (carbon dioxide equivalents) saved per case. Furthermore, minimal draping significantly reduced costs, saving about $7,000 per case. There were no surgical site infections within 30 days for both groups.

Conclusion Minimal draping protocols and minimal surgical packs may be an effective and safe way to reduce surgical waste. Implementing environmentally conscious strategies in the operating room can also significantly reduce associated costs. These findings highlight the critical need to reassess surgical protocols to better align with sustainable healthcare practices. To achieve meaningful and lasting change, it's imperative to provide comprehensive education to clinical staff and to initiate large-scale institutional reforms.

Purpose: Peer review is fundamental to maintaining scientific rigor, yet challenges such as quality variability, implicit bias, and reviewer burden can hinder fair and timely manuscript evaluation. [1-3] These barriers disproportionately impact authors from low- and middle-income countries and non-English-speaking regions with limited access to expert editorial guidance. [4,5] Advancements in generative AI offer a scalable opportunity to improve review consistency and accessibility by generating structured, actionable feedback to support authors in improving their work prior to submission. This study evaluates the feasibility and quality of AI-assisted manuscript review in plastic surgery research by comparing baseline AI, iteratively refined AI approaches, and human peer reviews using the Review Quality Index (RQI).

Methods: Fifteen plastic surgery manuscripts (5-clinical studies, 5-clinical trials, 5-systematic reviews) were evaluated using three progressively refined AI approaches: AI1 (baseline GPT-5), AI2 (custom GPT with retrieval-augmented generation (RAG) integrating peer review methodology papers), and AI3 (iteratively optimized custom GPT with meta-prompting and journal-specific guidelines). Human peer reviews and corresponding manuscript first drafts were obtained from journals publishing open peer review. Two blinded raters independently scored each review on a 5-point Likert scale using the validated RQI tool (maximum total score=40). Scores were averaged across raters. Analyses included descriptive statistics, inter-rater reliability (ICC), and Friedman tests with post hoc Wilcoxon signed-rank tests using Holm correction for multiple comparisons.

Results: AI3 (iteratively refined custom GPT) demonstrated significantly higher scores than all other review types (AI3=28.53±3.76, Human=25.17±4.87, AI2=24.30±5.02, AI1=24.00±1.77; Friedman X2(3)=19.29, p<0.001). Post-hoc comparisons showed AI3 significantly outperformed Human(p=0.028), AI2(p=0.008), and AI1(p=0.008) reviews. Question-level analysis demonstrated significant differences across review types for 7 of 8 dimensions (all p<0.05), with the largest effects observed for methodological assessment (Kendall's W=0.69), interpretation of results (W=0.63), and evidence provision (W=0.59). AI3 maintained consistent performance across all manuscript types (score range=28.0–29.1), while human reviews showed greater variability (range=22.1–27.0). Inter-rater reliability was moderate for the standard GPT model and lower for human and other AI configurations.

Conclusions: AI-assisted manuscript review shows strong potential to improve the clarity, efficiency, and consistency of scientific manuscript evaluation. Iterative refinement of AI-assisted manuscript review through meta-prompting and retrieval-augmented generation integrating domain-specific knowledge significantly improves review quality. The findings suggest that AI performance in manuscript review is not static but can be systematically enhanced through targeted optimization and may serve as an effective adjunct to human peer review by generating structured feedback while reducing reviewer burden and supporting authors in refining their work prior to submission. These findings support the integration of carefully optimized AI-based assistance into manuscript review workflows to enhance scientific rigor and promote broader equity in research publication.

References: 1. Haffar S, Bazerbachi F, Murad MH. Peer Review Bias: A Critical Review. Mayo Clin Proc. 2019;94(4):670-676. doi:10.1016/j.mayocp.2018.09.004 2. Stahel PF, Moore EE. Peer review for biomedical publications: we can improve the system. BMC Med. 2014;12:179. Published 2014 Sep 26. doi:10.1186/s12916-014-0179-1 3. Loonen MP, Hage JJ, Kon M. Who benefits from peer review? An analysis of the outcome of 100 requests for review by Plastic and Reconstructive Surgery. Plast Reconstr Surg. 2005;116(5):1461-1475. doi:10.1097/01.prs.0000178796.82273.7c 4. Abbas Z, Adesokun F, et al. How to overcome barriers to publication in low- and middle-income countries: Recommendations from early career psychiatrists and researchers from around the world. Asia Pac Psychiatry. 2021;13(4):e12495. doi:10.1111/appy.12495 5. Smith OM, Davis KL, Pizza RB, et al. Peer review perpetuates barriers for historically excluded groups. Nat Ecol Evol. 2023;7(4):512-523. doi:10.1038/s41559-023-01999-w

Background Cosmetic surgery is performed by physicians from diverse training pathways.(1) In many states, physician licensure alone does not restrict the performance of cosmetic surgery to plastic surgeons.(2,3) Prior studies have demonstrated that a substantial portion of physicians advertising cosmetic surgical services perform procedures that fall outside the scope of their accredited residency training.(1) These practice patterns highlight ongoing concerns regarding patient safety, training adequacy, and regulatory oversight. Florida is one state where there are limited regulations governing which specialties can perform cosmetic surgeries and where disciplinary action cases are publicly available for reveiw.(2–4) This study presents a state-level review of disciplinary cases in Florida involving cosmetic procedures to better characterize physician specialty backgrounds, procedures performed, and reported adverse outcomes, stratified by American Board of Plastic Surgery (ABPS) certification. Methods The Florida Department of Health Discipline and Administrative Actions Database was queried to identify physicians classified as "Medical Doctor" who incurred disciplinary actions, including revocation, suspension, probation, or fine between 2015 and 2026. Each record was individually reviewed. Cases were included if they involved a plastic or cosmetic surgical procedure performed by a physician who was or was not board-certified in plastic surgery by the American Board of Plastic Surgery (ABPS). Data collected included physician specialty background, procedure performed, and adverse outcomes. Results We identified 14 disciplinary actions (2018–2025) involving 11 unique physicians. Actions included revocation (n=4), suspension (n=4), and probation (n=6). One physician accounted for four probation actions. Physicians without ABPS board certification accounted for 13/14 actions across specialties, including general surgery, obstetrics and gynecology, family medicine, internal medicine/cardiology, and occupational preventive medicine; 1/14 actions involved an ABPS-certified plastic surgeon. Common procedures included liposuction (n=8) and gluteal fat grafting/augmentation (n=5). Reported adverse outcomes included death (n=7 total; n=6 among non-ABPS cases), sepsis (n=1), major wound complications (n=1), and internal organ perforation (n=3). Conclusion In Florida (2018–2025), 14 cosmetic surgery–related disciplinary actions involving 11 physicians were identified. Most actions involved physicians without ABPS board certification (13/14) and commonly followed liposuction and/or gluteal fat grafting procedures. Reported outcomes included 7 patient deaths and 3 internal organ perforations. Further state-level analyses are warranted to determine whether similar patterns exist across other jurisdictions. References 1. Long EA, Gabrick K, Janis JE, Perdikis G, Drolet BC. Board Certification in Cosmetic Surgery: An Evaluation of Training Backgrounds and Scope of Practice. Plast Reconstr Surg. 2020;146(5):1017-1023. doi:10.1097/PRS.0000000000007242 2. Choudhry S, Kim NA, Gillum J, et al. State medical board regulation of minimally invasive cosmetic procedures. J Am Acad Dermatol. 2012;66(1):86-91. doi:10.1016/j.jaad.2011.01.009 3. States Lax in Regulating Cosmetic Surgery. https://www.namd.org/journal-of-medicine/1001-states-lax-in-regulating-cosmetic-surgery.html 4. Florida Department of Health - Discipline & Administrative Actions. https://mqa-internet.doh.state.fl.us/MQASearchServices/EnforcementActionsPractitioner

Background: Chronic wounds remain a major clinical and economic burden in the United States, affecting roughly 30 million patients and generating close to $100 billion in annual healthcare costs. Identifying affordable, biologically targeted therapies is therefore a priority. Prior work from our group demonstrated that locally injected Avena sativa-derived molecules, Avenanthramide (AVN) and β-glucan (BG), accelerate wound closure and improve scar quality, yet the mechanisms underlying these benefits remain poorly understood. This study integrates immunohistochemical and single-cell transcriptomic analyses to investigate the cellular and molecular effects of AVN and BG during wound repair.

Methods: Excisional wounds were created on the dorsal surface of C57BL/6 mice. Each mouse received subcutaneous injections of either 1% AVN, β-glucan, or PBS (control). Wounds were harvested on day 14 for histologic evaluation and single-cell RNA sequencing. Bioinformatic analyses were performed using Seurat and CellChat.

Results: AVN treatment accelerated healing and significantly altered immune cell populations. Myeloid cells comprised 42% of all cells in control wounds but only 18% in AVN treated wounds. Macrophages and neutrophils decreased from 60% and 15% in controls to 40% and 5% with AVN, respectively. In contrast, monocytes and dendritic cells increased from 4% and 5% to 10% and 20%. AVN treated macrophages had significantly elevated IGF1 expression (1.8-fold) with reduced CXCL12 expression, which is typically associated with inflammation resolution and enhanced tissue regeneration. These findings suggest that AVN promotes a transition from inflammatory to reparative immune states.

β-glucan treatment also accelerated healing but instead produced more fibroblast focused effects. Fibroblasts had the largest transcriptional shift, with a 1.5-fold increase in upregulated genes compared with control. BG treated fibroblasts upregulated Plac8, Ly6c1, Myc, and Mmp3, which suggests Wnt/β-catenin pathway involvement. Cox-2 expression was also elevated, which is consistent with activation of NF-κB mediated inflammatory signaling. Histologic and transcriptomic analyses showed higher COL7A1 expression in control compared to BG treated fibroblasts, suggesting that reduced COL7A1 may contribute to the increased scarring phenotype observed in β-glucan wounds.

Conclusion: AVN and β-glucan exert complementary yet distinct effects on wound healing. AVN reduces myeloid cell infiltration and promotes IGF1 macrophages that support inflammation resolution and regenerative repair. β-glucan drives fibroblast activation, migration, and pro-inflammatory signaling while reducing COL7A1 expression, which may explain its association with increased scarring. Thus, AVN may be beneficial in patients with chronic inflammation or impaired macrophage function (e.g., older adults), whereas β-glucan may enhance healing in immunocompromised or fibroblast deficient wounds. Future work is needed to investigate therapeutic strategies that selectively promote AVN induced immune cell recruitment and polarization or modulate BG induced fibroblast states to optimize human wound healing.

Background: Non-communicable diseases and injuries now account for the majority of global disability-adjusted life years lost, and plastic and reconstructive surgery (PRS) is positioned to address nearly two-thirds of this burden. Despite this, only 41% of U.S. residency programs incorporate structured global surgery education. Ethical complexities in international collaboration-including scope of practice, informed consent, standards of care, and sustainable partnership-necessitate deliberate training. To address this gap, the University of Massachusetts PRS residency integrated the Surgeons in Humanitarian Alliance for Reconstruction, Research and Education (SHARE) curriculum as a mandatory component of training.

Methods: In 2024, UMass PRS incorporated SHARE's virtual global surgery curriculum, which includes case-based assemblies, ethical discussions, and African surgeon-led teaching emphasizing LMIC leadership. Pre- and post-intervention surveys were administered using 5-point Likert scales to assess residents' global health exposure, confidence in culturally competent care, and ethical decision-making. Likert responses were analyzed both by mean score and collapsed into low (1–2), moderate (3), and high (4–5) confidence categories.

Results: Eleven residents completed the pre-survey; post-survey response rates ranged from 9–10 participants per item. At baseline, 75% of residents reported 0–2 prior global health experiences, confirming limited exposure. Pre-intervention, most residents rated their confidence in providing culturally competent care as moderate (8/11), with fewer reporting high confidence (2/11). Following participation in SHARE sessions, high-confidence responses increased to 5 residents, while moderate responses decreased to 4, indicating a shift toward greater perceived competence.

Residents demonstrated strong baseline ethical awareness; 100% endorsed the importance of upholding identical standards of care internationally (pre: 8 "almost always true"; post: 7). Notably, after curriculum exposure, more residents selected "almost never true" when asked whether it is appropriate to practice outside their U.S.-based scope (increase from 5 to 7 responses), suggesting heightened ethical caution. Similarly, fewer residents endorsed advising or interrupting local physicians inappropriately, reflecting improved awareness of professional boundaries in global settings.

When ranking core competencies, respect for cultural practices remained consistently prioritized, while post-survey responses demonstrated increased recognition of advocacy and systems-level awareness. Across multiple items addressing consent, operational workflow, and cultural understanding, residents maintained strong agreement that these principles are essential to ethical global engagement.

Conclusion: Mandatory integration of the SHARE curriculum was associated with measurable increases in resident confidence in culturally competent care and strengthened ethical discernment regarding scope, collaboration, and standards of practice. By centering LMIC surgeon leadership and structured ethical reflection, this model moves beyond traditional "medical mission" paradigms toward sustainable partnership. These findings support formal incorporation of global surgery education into PRS residency curricula and suggest scalability to programs nationwide seeking to standardize ethical, culturally grounded global surgical training.

Background: The discovery of the osteogenic properties of BMP-2 has translated to the current clinical use of exogenously manufactured, recombinant human BMP-2 (rhBMP-2) for bone grafting in Craniofacial defects, orthopedic instrumentation, oral/maxillofacial surgery, and fracture fixation. To provide a clinical benefit, supraphysiological concentrations of rhBMP-2 are utilized intraoperatively in patients, which has been associated with a range of potential complications in the literature including ectopic bone formation and in some patients a significant local inflammatory response. Surprisingly, intralesional gene transfer or mRNA-LNP transfection of BMP-2 has demonstrated efficient defect closure in preclinical models with several orders of magnitude less BMP-2 expression. Previous work has demonstrated that the Wnt and Hedgehog family of morphogens exert their influence far beyond their in situ source of production by associating with secreted nanoscale extracellular vesicles (EVs). Engineered EVs containing BMP-2 have demonstrated osteogenic potential both in vitro and in vivo, suggesting that EVs may be acting as a delivery vehicle to potentiate the downstream signaling and superior osteogenesis of transfected BMP-2 compared to clinically approved rhBMP-2.

Methods: HEK293T cells were transfected with DNA constructs encoding various BMP-2 designs and known exosomal tetraspanin proteins. 72h post-transfection, cell cultures were pelleted at 1,000×g for 10 min, supernatants were passed through 0.45 μm filters, and concentrated using centrifugal filters with 100 kDa molecular weight cut-off. Supernatant was further purified by ultracentrifugation in a TLA100.3 rotor at 50,000 rpm (135,000×g) for 2h at 4°C on a 20% w/v sucrose cushion in polycarbonate centrifuge tubes. Supernatant and sucrose were aspirated, and the remaining pellets were resuspended in 200 μL sterile pH 7.4 PBS at 4°C overnight. To remove residual cell debris and large apoptotic bodies, resuspended samples were transferred to microcentrifuge tubes and centrifuged at 10,000×g for 10 min; clarified supernatants were collected for subsequent ELISAs, western blots, and stem cell differentiation assays.

Results: Western blots confirmed the presence of BMP-2 in the exosomal fractions of transfected HEK293T cell culture supernatants, and the BMP-2 yield in this fraction was increased by the co-expression of CD63, a tetraspanin protein commonly found in exosomes. Indirect ELISA of plated exosomal fractions demonstrated a more than 4-fold increase in BMP-2 release when co-transfecting CD63 and BMP-2 compared to BMP-2 alone, with no change in the BMP-2 yield isolated from transfected cell lysate (p<0.05). Free, soluble BMP-2 in these cell culture supernatants was undetectable in the filtrate of the 100kDa molecular weight concentrator. Appending a short lysine/arginine-rich peptide motif found in BMP-2 to cytosolic proteins promotes their secretion on extracellular vesicles.

Conclusions: Our findings suggest that the BMP-2 that is over-expressed by transfection and secreted is largely associated with EVs, and a lysine/arginine-rich peptide motif found in the mature domain of BMP-2 mediates this association. Appending BMP-2's exosome-targeting motif to bioactive proteins provides a new approach for exosome engineering. Further functional assays with osteogenic stem cell lines will confirm the potency of exosomal BMP-2. Improving BMP-2 secretion by engineering EV-association may lead to a superior therapeutic modality for bone healing.

Purpose: Facial plastic surgeons routinely face the same dilemma: a compelling postoperative outcome to share, but no reliable way to protect patient identity without compromising facial details. Conventional methods such as black boxes, blurring, and cropping disrupt facial proportions, obscure aesthetic context, and introduce visual artifacts that may bias interpretation. Additionally, modern facial recognition software can often re-identify inadequately masked images. Generative AI offers a fundamentally different approach to frontal photograph deidentification. Rather than concealing identity, it replaces selected facial regions with realistic synthetic features, while preserving a surgical region of interest (ROI). This study evaluates the feasibility and effectiveness of generative AI-enabled regional facial deidentification, as assessed by lay observers.

Methods: Eight standardized facial identities representing four races (White, Black, Asian, Latinx) and two genders were digitally modified using generative AI tools in Adobe Photoshop. Five anatomical zones were defined for selective regeneration: upper face, midface, lower face, hairline, and jaw/neck. All 26 possible single- and multi-region edit combinations were generated for each identity, plus unaltered and fully altered controls, yielding 208 unique stimuli. A total of 550 lay participants evaluated 52 randomized image pairs (original versus stimulus). Participants indicated whether the images appeared to depict the same or different person and rated their confidence on a 5-point scale. Adequate deidentification was defined as ≥70% of respondents judging the modified image as a different person. Responses were analyzed using generalized linear mixed-effects models accounting for repeated measures across raters and identities.

Results: Each edit pattern received approximately 1,120 independent responses. Single-region edits were largely ineffective, with midface-only (12.2%) and lower face-only (12.8%) rarely preventing recognition. In contrast, multi-region edits substantially improved deidentification. The most effective pattern was regeneration of the upper face + lower face + hair + jaw/neck while preserving the midface, achieving 70.2% deidentification. Similar multi-region patterns preserving either the lower face alone (64.9%) or midface and jaw/neck (64%) also performed well. Male faces were more likely to be judged as different (β=0.784, p<.001), although confidence did not differ by gender (p=.282). White (β =0.117, p=.002) and Latinx (β=0.179, p<0.001) faces were more frequently perceived as different, and White faces were associated with higher rater confidence ratings (p<.001). Confidence increased with the number of regenerated regions and was significantly higher when images were judged as different people (p<.001).

Conclusion: Generative AI-enabled regional facial regeneration can achieve effective deidentification while preserving a surgical ROI. Multi-region edits that retain key operative areas, particularly the midface and lower face, demonstrated the strongest performance, suggesting certain regions may be reliably preserved without local distortion introduced by traditional masking techniques. This approach represents a potential shift from obscuring identity to replacing it, enabling image sharing that maintains clinical integrity while protecting privacy. Observed differences by race and gender highlight the importance of diverse validation as this technology moves toward clinical implementation. Ongoing work is applying this framework to clinical photographs to evaluate its utility in surgical outcome sharing.