Scientific Abstract Presentations: Research & Technology Session 1 Thursday, September 26 Thu, Sep 26 9:45-11:45 a.m. PDT

Background: The recent advancements in artificial intelligence (AI) has shifted the landscape of telehealth in which AI chatbots, such as ChatGPT, have demonstrated numerous potential applications in medicine (1). While text-based systems are efficient and increase the availability of health information, they are not a personal approach to patient interaction and lack human-like qualities that increase user effectiveness, usability, and trust (2). In the field of plastic and reconstructive surgery (PRS), ChatGPT has shown to be capable of providing basic, general patient education on topics such as rhinoplasty and breast augmentation (3, 4). However, preferences for a traditional chatbot versus a human AI VideoBot in the context of PRS have yet to be assessed.

Methods: Synthesia, an AI video generation platform, was utilized to develop a human AI avatar that was integrated with ChatGPT 3.5 to create the VideoBot. The VideoBot was then integrated into Tolstoy, a platform for creating interactive videos, to develop an interactive VideoBot whereby the human AI avatar was able to answer four of the most commonly asked questions related to breast reconstruction. We utilized Zapier, a workflow automation software, to develop a ChatGPT 3.5-integrated text-based chatbot. An anonymous 20 item survey adapted from Corritore et al.'s 2005 validated Measurement of Online Trust (5) was distributed online via Amazon Mechanical Turk to female participants. Participant demographic data was collected, and participants were asked to rank their preferences between the VideoBot and text-based chatbot on a 1-7 Likert scale (1=Strongly Disagree, 7= Strongly Agree).

Results: A total of 396 responses were gathered. Participants were 18-64 years old, 97% had received some form of plastic or reconstructive surgery, and 95% of those patients underwent a breast-related medical procedure. Perceptions of truthfulness, believability, content expertise, ease of use, and safety were similar between the VideoBot and chatbot. Most participants preferred the VideoBot over the traditional chatbot (63.5% versus 28.1%) as they found it more captivating than the text-based chatbot. 77% of participants would have preferred to see someone that they identified with as the human AI avatar in terms of gender and race.

Conclusion: There is no difference in user effectiveness, usability, and trust between the VideoBot and text-based chatbot. However, the human-like quality of the VideoBot allows for a more interactive experience than the traditional chatbot. The race and gender effect should be further explored in providing a more personable telehealth experience for patients.

References: 1. Thirunavukarasu AJ, Ting DSJ, Elangovan K, Gutierrez L, Tan TF, Ting DSW. Large language models in medicine. Nat Med. Aug 2023;29(8):1930-1940. doi:10.1038/s41591-023-02448-8 2. F. Dsouza, R. Shaharao, Y. Thakur, P. Agwan, G. Sakarkar and P. Gupta, "Advancement in Communication using Natural Language based VideoBot System," 2022 IEEE Bombay Section Signature Conference (IBSSC), Mumbai, India, 2022, pp. 1-5, doi: 10.1109/IBSSC56953.2022.10037380. 3. Xie Y, Seth I, Hunter-Smith DJ, Rozen WM, Ross R, Lee M. Aesthetic Surgery Advice and Counseling from Artificial Intelligence: A Rhinoplasty Consultation with ChatGPT. Aesthetic Plast Surg. Oct 2023;47(5):1985-1993. doi:10.1007/s00266-023-03338-7 4. Seth I, Cox A, Xie Y, et al. Evaluating Chatbot Efficacy for Answering Frequently Asked Questions in Plastic Surgery: A ChatGPT Case Study Focused on Breast Augmentation. Aesthet Surg J. Sep 14 2023;43(10):1126-1135. doi:10.1093/asj/sjad140 5. Corritore CL, Marble RP, Wiedenbeck S, Kracher B, Chandran A. Measuring online trust of websites: Credibility, perceived ease of use, and risk.

Purpose: The paucity of underrepresented in medicine (URiM) plastic surgeons is well recognized; Black and Hispanic individuals comprise 3.6% and 5.7% of the workforce, compared to 12% and 15% of the population, respectively.(1,2) Despite recognition of this discrepancy, the racial and ethnic composition of trainees has not changed significantly over the past several decades.(3) To investigate this inconsistency, we performed a qualitative study aimed at gaining deeper insights into how residency program directors (PDs) perceive the importance of racial and ethnic diversity in training. Our primary outcome included identification of PD views and deliberate initiatives around the recruitment of URiM residents.

Methods: Integrated plastic surgery residency PDs and assistant PDs were invited to participate in individual semi-structured interviews. Interview questions focused on applicant selection for interview, interview day, and strategies to improve resident diversity. Interviews were deidentified, transcribed, and analyzed inductively to generate themes based on patterns observed across narratives.

Results: In this ongoing study, we have interviewed eight PDs, all from university-based programs in the West (n=2), Midwest (n=2), and Northeast (n=4) regions of the US. We identified three key themes: 1) Creating Exposure and Mentorship Opportunities for URiM Candidates, 2) Unique Qualities and Contribution of URiM Candidates, and 3) Building a Culture of Inclusion and Acceptance of URiM Candidates. With respect to theme one, PDs describe how the disproportionately low number of URiM plastic surgery residents is a pipeline issue that originates well before medical school. There is a consensus among PDs that intentional exposure and mentorship efforts are necessary to facilitate early recruitment to the specialty and residency preparation. Regarding theme two, PDs overwhelmingly agree that scholastic achievements be retained while enhancing diversity. In other words, equity and excellence can be achieved simultaneously. Additionally, attributes such as grit, distance-travelled, and leadership warrant consistent consideration. Furthermore, PDs recognize that diversity is important for patient care and innovation, and it is possible that the unique qualities that URiM candidates can offer are not captured in current application processes. Finally, related to theme three, PDs consistently acknowledge that explicit commitment to a culture of diversity and inclusion is important to attract and maintain URiM candidates in their programs. They understand that the presence of multiple URiM residents in a program is a poignant indication of a welcoming culture, while the presence of one or no URiM residents may raise concern for tokenism.

Conclusions: We identified several themes critical to the recruitment and retention of URiM trainees to plastic surgery residency. Early exposure and mentorship, including funded research opportunities, could facilitate preparation for the competitive residency application process. Program-specific and institution-wide efforts are both necessary to demonstrate support for diversity and inclusion. Additional initiatives are needed to generate interest among URiM students, enhance review of applications, and adopt innovative recruitment strategies to support diversity, equity, and inclusion in plastic surgery residencies.

References: 1. Butler PD, Britt LD, Longaker MT. Ethnic Diversity Remains Scarce in Academic Plastic and Reconstructive Surgery. Plast Recontr Surg 2009; 123: 1618-1627. 2. Chawla S, Chawla A, Hassain M, et al. The State of Diversity in Academic Plastic Surgery Across North America. Plast Reconstr Surg Glob Open 2021; e3928 3. Santosa KB, Preist CR, Oliver JD, et al. Influence of faculty diversity on resident diversity across surgical subspecialties. Am J Surg. 2022; 224:273-281.

Introduction Burn injuries often require immediate assistance and specialized care for optimal management and outcomes. Especially in severe burns, a prompt referral to an appropriate healthcare facility is crucial in ensuring timely and effective treatment. The emergence of artificial intelligence (AI) has just recently started being applied to healthcare decision-making. However, its role in healthcare delivery is still under scrutiny. This study aims to evaluate whether ChatGPT provides safe and accurate medical answers to commonly asked questions regarding acute burn injuries and their management by comparing ChatGPT's responses to those provided by the American Burn Association (ABA).

Methods Twelve commonly asked questions available on the ABA official website regarding burn injuries, management, and patient referral were prompted into ChatGPT. Each response was compared with the ABA Guidelines for Burn First Aid and Patient Referral, the gold standard for accurate and evidence-based burn care recommendations. Three burn specialists independently compared the appropriateness and completeness of each ChatGPT response based on the ABA guidelines by giving a score from 1 (poor concordance) to 5 (excellent concordance) to each answer according to the Global Quality Score (GQS).

Results The average score for ChatGPT-generated responses was 4.56 ± 0.65. This indicates that raters found the answers were good to excellent quality with the most important topics covered and in high concordance with the ABA guidelines. Overall, 58% of the questions were focused on first aid and management of acute burns, while 42% focused on patient referral criteria. Of the twelve questions asked, two (16.7%) ChatGPT-generated responses scored a perfect 5.00 ± 0.00, demonstrating perfect concordance with the ABA guidelines. Nine (75.0%) ChatGPT answers were between good and excellent quality, with average GQS scores ranging from 4.33 to 4.67. One (8.3%) response describing chemical burn referral recommendations scored 3.33 ± 0.58, indicating somewhat concordance with the ABA guidelines.

Conclusion This study presents an initial comparison of commonly asked burn injury and management questions between ChatGPT-generated responses and the gold standard ABA guidelines. Based on the feedback of three experienced burn specialists, ChatGPT's answers are accurate and in high concordance with the ABA recommendations. Our results demonstrate that ChatGPT can adequately, comprehensibly, and thoroughly describe the most appropriate treatment and referral for burn injuries. We foresee that ChatGPT may play a role as a complementary tool in patient education, having a profound impact on clinical practice, research, and education.

Background Surgeons in Humanitarian Alliance for Reconstruction, Research and Education (SHARE) is an innovative program from the Plastic Surgery Foundation designed to address global disparities in plastic surgery. SHARE uniquely provides medical students opportunities to learn about plastic surgery from a global surgery and cultural humility perspective. To refine our curriculum, we developed a survey to investigate current global and plastic surgery education among medical students who participate in SHARE and their attitudes toward SHARE's current medical student engagement opportunities.

Methods A cross-sectional survey was designed to assess barriers to involvement in plastic surgery and global surgery, motivations for joining SHARE, SHARE's impact in alleviating barriers to involvement in plastic surgery and global surgery, and whether medical students' involvement in SHARE affected interest in pursuing a plastic surgery or global surgery career. The survey was distributed via listserv to medical students involved in SHARE as of February 2024.

Results Thirty-seven out of 79 medical students completed the survey (47%). Forty-nine percent were between the ages of 26-30 years old and 49% were first-generation immigrants to the United States or Canada or were a member of a first-generation family. Racial and ethnic composition included: 32% White/Caucasian, 29% Asian/Asian American, 25% Black/African American, 11% Middle-Eastern/North African, and 4% preferred not to say. While most participants (81%) were interested in pursuing plastic surgery residency, 61% attended a medical school without an integrated plastic surgery program and 31% had never had an opportunity to discuss plastic surgery cases in any setting prior to joining SHARE. Among participants who had any previous global health involvement, only 44% of these global health opportunities focused on global surgery specifically. Primary motivations for participating in SHARE were to network with plastic surgeons, learn about the experience of surgeons outside of North America, and to learn more about global surgery. The activities that medical students participated in within SHARE that enabled exposure and involvement in global plastic surgery included: clinical assemblies (36%), small groups (27%), panel at Plastic Surgery the Meeting 2023 (23%), and serving on a subcommittee (14%). The majority of participants (57%) were more interested in plastic surgery after being involved with SHARE (p < 0.001). Ninety-three percent of participants stated that their interest in incorporating global surgery into their career was moderately or greatly increased after participating in SHARE, compared to 7% unchanged and 0% decreased (p < 0.001).

Conclusion Our study highlights the critical role that SHARE plays in alleviating barriers medical students face regarding plastic surgery exposure and global surgery education. These results will be used to target opportunities to further optimize medical students' role in the SHARE program.

Purpose Soft tissue loss, whether resulting from surgery, trauma, or infection, often necessitates restoration to improve appearance and function. While autogenous fat transfer (AFT) is a popular method for soft tissue augmentation, it requires a surgical procedure and has well known challenges such as inconsistent graft take and the need for specialized equipment. Decellularized tissues have emerged as promising scaffolds for tissue engineering, with adipose tissue being a readily available resource. In this study, we investigated a novel decellularization method using supercritical carbon dioxide (scCO2), which holds the promise to remove cells and cellular debris while leaving the bioactive components within the ECM relatively undisturbed, especially compared to traditional methods. Our goal was to assess the efficacy of this environmentally conscious and biocompatible alternative method in decellularization using patient-derived adipose tissue.
Methods and Materials Under an active IRB protocol, adipose tissue was obtained from patients undergoing breast reduction and/or abdominoplasty. Following removal of the skin, the adipose tissue was stored at -20°C until processing. The tissue was either minced, or further reduced using a blender or a cryo-mill. Delipidation was performed by treating the tissue with isopropanol at 37°C with agitation. This was followed by treatment with scCO2 at 35°C and 1,450 psi for 1 hour. For decellularization, the samples were immersed in a decellularization solution containing ethanol, NovaKill™ (peracetic acid-based solution), and DI water, followed by exposure to scCO2 at 35°C and 2,300 psi for 4 hours. After a wash with PBS, the samples underwent an additional scCO2 treatment at 35°C and 2,300 psi for 1 hour before being stored at –20°C. The processed adipose tissue was lyophilized for DNA quantification using the DNeasy Blood and Tissue Kit and cryo-embedded for histochemical analysis. Results Different processing methods of human derived adipose tissue had varying effects on the efficacy of scCO2 delipidation and delipidation. Minced adipose tissue maintained high levels of lipid and intact adipose ECM matrix comparable to non-processed tissue. Further processing of adipose with blending and cryo-milling exhibited a significant reduction of lipid content in delipidated samples compared to unprocessed tissue, with 77-90% further reduction in decellularized samples compared to those only delipidated. DNA content in delipidated and decellularized tissue ranged from 275-295 ng/mg (above the standard for decellularization). However, there was a significant decrease (41-53%) in DNA content in delipidated and decellularized tissue compared to only delipidated samples. On H&E staining, we observed compact adipose ECM matrix and no obvious nuclear components in processed tissue. This finding was supported by lack of observed DAPI positive nuclei.
Conclusions Current delipidation and decellularization with scCO2 can significantly reduce lipid and DNA content while preserving ECM in human derived adipose tissue. Further modification of treatment parameters such as additional rounds of scCO2 and higher temperatures will be considered to improve the efficacy and efficiency of this process, to produce an adequately delipidated and decellularized soft tissue injectable for animal studies and eventual clinical use.

Background: Teaching spatial anatomy using conventional learning tools remains challenging. Virtual reality (VR) applications can be effective tools in learning anatomy in three dimensions. Methods: A randomized controlled study assessed user understanding of spatial carpal anatomy using a VR Wrist Game module intervention. The 3D Wrist Game was developed out of a computerized tomography (CT) scan of a right hand & wrist by using segmentation technique. The model was then processed and used for building the game in the virtual environment. A user is provided with a headset and two hand-held controllers. As the user gestures or manipulates the controllers, corresponding actions occur in the virtual realm. Eleven first year medical students at the University of Illinois at Chicago completed the study. The VR group underwent training using the VR Wrist Game. All participants completed pre- and post-intervention questionnaires, a textbook reading, and assessment of carpal anatomy knowledge using a standardized physical model test. Results: Visual-based learning was preferred by all participants, while kinesthetic learning was favored by 83% of the VR group and 80% of the control group (Pearson Chi-square = 0.020; p = 0.887). Prior experience with video gaming and VR was similar among participants in the control and VR groups. For the VR group, prior video game experience and prior VR experience had no effect on Wrist Game scores (Spearman's rho = 0.494 and -0.395; p = 0.320 and 0.439, respectively). On the physical model test, the VR group placed a significantly higher number of correctly positioned and oriented bones (Mann-Whitney U = 4.5; p = 0.039), and they were 20 times more likely to perform this task better than the control group (OR = 20.0; p = 0.055). The VR group was 8 times more likely to complete the test faster (OR = 8.0; p = 0.140). There was no statistically significant difference in the number of carpal bones that were correctly placed but incorrectly oriented. The Wrist Game score directly correlated with the physical model test score (r2= 0.82; p = 0.019). VR group participants agreed that the Wrist Game module can help practice transferable motor skills and that they would recommend the VR module to colleagues to learn and practice carpal bone anatomy. Conclusions: These results suggest that VR effectively provides transferable spatial skills useful in real-life anatomy knowledge application. Faster task completion highlighted VR group's relative comfort with the three-dimensional physical model. We believe that spatial understanding, a transferable skill trained by 3D tools such as VR, is the most clinically relevant and this knowledge is the most applicable in clinical clerkships.

Purpose Plastic and Reconstructive Surgery is one of the least funded medical subspecialties by the National Institutes of Health (NIH) and has been for some time. One of the stated goals of the Plastic Surgery Foundation (PSF) is to fund new plastic surgery researchers and projects and improve the proliferation of future studies in the field. By doing this, the PSF is creating a foundation that can support future grant applications to the NIH for larger projects. The purpose of this study is to conduct a comparison of the grants the PSF has funded to plastic surgery related grants the NIH has funded. Aided by this analysis, plastic surgery researchers can use the findings from this study to improve the odds of accessing further funding from the NIH. Methods For the NIH grants, the NIH RePORTER database was queried with the names of all American Society of Plastic Surgery (ASPS) members between 2003 to 2024. The abstracts from the resulting grants were then manually read and identified as to whether the grant was related to plastic and reconstructive surgery. The abstracts were then manually categorized into the same categories the PSF uses. A second pass was completed by another researcher to ensure validity of manual data collection and categorization. For the PSF grants, the topic data as well as the number of grants under each topic was manually extracted from the PSF archive of funded grants from 2003 to 2024. Chi square tests were then performed between the number of grants awarded by the PSF and NIH. Results NIH and PSF funded grant topics were compared with Chi Square tests. The NIH was 0.606 times as likely to fund reconstructive or cosmetic breast related grants as the PSF, 0.204 times as likely to fund composite tissue allotransplantation related grants; 4.951 times as likely to fund cranio/maxillofacial/head and neck related grants; 2.747 times as likely to fund fat grafting related grants; 0.577 times as likely to fund hand or upper extremity grants; 0.089 times as likely to fund technology based grants; and 5.687 times as likely to fund wound or scar related grants. Conclusion Researchers who wish to use PSF grants as a stepping stone to gain more funding from the NIH may consider focusing their grants on cranio/maxillofacial/head and neck, fat grafting, and wound/scar related topics.

Purposes Though widely used in plastic and reconstructive surgery autologous fat grafts are often limited by their unpredictable and often inadequate volume retention. A major cause of graft loss is the insufficient rate and degree of revascularization of the transplanted fat, resulting in ischemic necrosis. Previous work in our lab aimed at improving the survival of transplanted fat demonstrated that human umbilical vein endothelial cells (HUVEC) transduced with E4ORF1 region of human adenovirus Type 5 (E4s) promoted adipocyte survival in both in vitro and in vivo models. Here, we analyzed if a novel cell line, reset vascular cells (R-VEC), were capable of yielding similar results. This cell population consists of HUVEC adult endothelial cells that have been transduced with the transcription factor ETV2 which is responsible for promoting and maintaining microvascular network development. This transcription factor is normally only expressed during early angiogenesis and is no longer expressed after 20 weeks in utero. This unique cell line has demonstrated promising angiogenetic capabilities in in vitro culture, suggesting they might promote adipocyte survival after fat transplantation.

Methods Primary mature adipocytes were isolated from patient-derived fat tissue and cultured alone, with E4s, or with R-VECs within neutralized 0.6% (w/v) collagen. 200uL of each experimental group was plated in triplicate with an R-VEC density of 250,000 cells/mL. In adipocyte conditions, a concentration of 11.5% adipocyte (V/V) was mixed within collagen alone or with ECs in addition to collagen. Constructs were harvested on days 0, 4, and 7. Perilipin immunofluorescent staining around lipid droplets was used to verify live adipocytes. An entire cross-section of each construct was imaged and stitched using imaging software. Spaces between 70-120um (the average size of adipocytes), without perilipin expression were counted as non-viable adipocytes. The percentage of viable adipocytes was calculated.

Results Perilipin staining confirmed the presence of live adipocytes in all constructs and yielded consistent adipocyte survival rates at Day 0 across all groups. After 7 days, the R-VEC group exhibited significantly greater adipocyte viability when compared to both adipocytes in coculture with E4s and adipocytes in monoculture. On day 7, the R-VEC group demonstrated the highest adipocyte survival at 87.5%, with significantly lower amount in E4 at 63.4% and adipocyte alone at 49.5% (p<0.05).

Conclusions In a biomimetic three-dimensional in vitro model of fat transplantation, co-culture of adipocytes with R-VECs resulted in superior adipocyte viability compared to culture with either E4 or HUVEC. These findings may have translational implications as R-VEC cells may be considered for future FDA authorization. Further studies are underway to determine if this pro-survival effect is resultant from the R-VEC capability to form microvascular networks, elaboration of pro-survival chemokines, or some combination of both.

Introduction: In 2005, Pew Research reported that 80% of Americans used the internet to search for medical information.1 Medical information can be difficult for patients to understand, especially when they require surgical intervention. A survey conducted by the American Health Information Management Association Foundation found that 11% of Americans leave their appointment more confused and 25% do not feel comfortable asking their doctor certain health-related questions.2 Following a doctor's visit 42% research their provider's suggestions, with 80% reporting that they use the internet to do so.2 Although traditional search engines like Google, Bing, and Yahoo remain popular among Americans for locating information, Chat Generative Pre-Trained Transformer (ChatGPT) (Open AI, San Francisco, CA) has emerged as a new source to gather information. This study aimed to evaluate its applicability in medical and surgical information gathering in the lay public.

Methods: An 18-item questionnaire was distributed using Amazon Mechanical Turk (abbreviated as MTurk) to assess the lay public's views and experience using OpenAI's Chat Generative Pre-Trained Transformer (ChatGPT) as a means of obtaining health information related to surgical procedures. To assess knowledge and utilization of ChatGPT, chi-squared analyses were carried out to compare questionnaire answers between younger participants and participants 65 years and older. The same analysis was used to compare demographic information between participants who have used ChatGPT to those who have not. Statistical significance was defined as a 2 tailed P < 0.05. All statistical analyses were carried out using Stata 18 (StataCorp LLC, College Station, TX).

Results: A total of 667 respondents completed the questionnaire. A majority of respondents (93.7%, n = 625) reported searching the internet for health information, with 68.4% reporting that they have used ChatGPT to search for health-related information. Roughly 95% of respondents reported having heard of ChatGPT (94.8%, n = 632) and almost 82% of participants had used it in the past (81.6%, n = 544). Interestingly, the most common surgical information searched was "procedure details" (27.7%), followed by "surgeon education" (19.6%). When comparing ChatGPT users to those that had not tried the chatbot, age varied significantly with higher proportions of individuals aged 25-34 having tried the chatbot than other groups (p = 0.002). More than 60% of participants who used ChatGPT reported that they would prefer to use ChatGPT to search for medical information instead of traditional search engines due to its convenience and comprehensiveness.

Conclusions: ChatGPT is a rapidly growing platform for patients to obtain medical information. The Internet is still a major source of information prior to surgery and our data demonstrate that patients are most interested in information like procedure details and surgeon's education. Surgeons should be aware of the rapid evolution of medical information delivery and recognize that providing this information preemptively will help with patients' counselling processes.

REFERENCES

1. Susannah F. Health Information Online. Pew Research Center. Updated 2014-02-13. https://policycommons.net/artifacts/620412/health-information-online/

2. Understanding, Access and Use of Health Information in America. AHIMA Foundation. 2023. https://ahimafoundation.org/media/ngfbggsk/oct2021understandingaccessusehealthinformationamericaahimafoundation.pdf

Purpose: As matching into plastic and reconstructive surgery becomes more competitive every year, applicants feel pressured to be productive in all aspects. With the recent change of Step 1 to pass/fail, there is an even greater emphasis on research publications, with matched applicants in the 2022-2023 cycle averaging at around 28.4 total abstracts, presentations, and publications. Such high expectations may lead to temptations to falsely embellish parts of the application, especially in areas that are difficult or time-consuming to verify on the new Plastic Surgery Common Application (PSCA). Previous rates of falsified publications in plastic surgery applications vary widely, ranging from 2–34%.1,2 We aim to quantify rates of publication falsification and inflation in the most recent PSCA applicant pool.

Methods: A retrospective review of 336 applications to our institution's integrated plastic and reconstructive surgery residency program was performed for the 2023–2024 cycle. To protect applicant identities, a randomized sample was selected to collect the following data: applicant demographics, "number of total peer-reviewed publications" reported on PSCA, and basic publication-related information from CVs. A search the Internet for PubMed indexing of any manuscripts that applicants marked as "published" was conducted. Publications not found were flagged and validated by a third party member unrelated to the 2023-2024 application cycle. Data was collected on a confidential and encrypted spreadsheet.

Results: A total of 27 out of 102 (26.5%) reviewed applicants had falsified information regarding their publication list, including incorrect author lists and unsearchable manuscripts. More common was inflation of the "number of total peer-reviewed publication number" self-reported on PSCA, done by 33 applicants (32.4%), via inclusion of "submitted" or "in-progress" manuscripts, abstracts from poster/oral presentations, surgical videos, and non-peer-reviewed articles. IMG status, Step 2 score, and number of claimed publications did not significantly differ between the groups that lied or inflated their numbers and those that did not.

Conclusion: The rate of research falsification and inflation in the PSCA is within the range found in the literature for previous application cycles using ERAS. However, any amount of falsification and inflation is still of concern. In the future, there may be more national discourse on measures needed to prevent professional misconduct and have greater standardization for research productivity reporting in the PSCA.

References: 1. Rodriguez-Unda NA, Webster ND, Verheyden CN. Trends in Academic "Ghost Publications" in Plastic Surgery Residency Applications: A 3-Year Study. Plast Reconstr Surg Glob Open. 2020;8(1):e2617. Published 2020 Jan 17. doi:10.1097/GOX.0000000000002617 2. Larson JD, Sisolak BB, King TW. The presence of "ghost" citations in an applicant pool of an integrated plastic surgery residency program. Plast Reconstr Surg. 2010;126(4):1390-1394. doi:10.1097/PRS.0b013e3181ead0d1

Purpose: Artificial intelligence (AI) models are already being extensively applied in medicine; however, recent studies have revealed the existence of significant gender and racial gaps with the utilization of AI in the care and education of patients. Resultantly, there is a growing concern that these gaps may lead to unintended biases and inequalities in patient care (1). Furthermore, demographic disparities have been established in many surgical subspecialties, including hand surgery, with women and people of color often in the minority (2). This paper intends to analyze the demographic representation of hand surgeons in AI-generated visuals models in order to shed light on any disparities and analyze the consequential implications for both the medical community and broader society.

Methods: We assessed three of the most popular and publicly available AI text-to-image generators, including DALL-E 3, Midjourney, and DreamStudio. Images were generated using the prompt "a photo of the face of a hand surgeon." Three reviewers independently evaluated over 300 AI-generated images, categorizing them according to gender (female and male) and race (non-White, defined as any race other than non-Hispanic White, and White). Inter-rater reliability was determined using Cohen's Kappa. Chi-square was performed to compare the distribution of female and non-White hand surgeons in the AI-generated images with current demographic data of hand surgeons in the United States. Statistical significance was established at alpha = 0.01.

Results: Cohen's kappa for racial agreement across three AI platforms was 0.608 (moderate to substantial agreement), and for gender agreement was 1 (perfect agreement). Cohen's kappa did not differ when comparing each AI platform for gender or racial agreement. DALL-E 3 showed a significant difference between percentage of rater identified whites and non-whites when compared to the national average of PR (plastic and reconstructive) surgeons (76.6% white, p<0.01)-- image output showed 64% white PR surgeons. On the contrary, DALL-E 3 did not show a significant difference between image output percent males (91%) and the national average of PR male surgeons (83%, p=0.03). Midjourney image outputs favored white (100%), male (100%) PR surgeons, and this was significantly higher than the national average (p<0.01). DreamStudio showed outputs reflective of the national average of male PR surgeons (81%, p=0.59), but showed significantly more white PR surgeons (97%) than the national average.

Conclusion: As AI technologies continue to shape healthcare, our study aims to underscore the urgency of cultivating more inclusive AI datasets that accurately reflect the growing diversity within the hand surgery profession. Addressing this gap is crucial for fostering equitable advancements in AI applications, enhancing medical education, and ensuring a comprehensive understanding of hand surgery.

1. Cirillo D, Catuara-Solarz S, Morey C, et al. Sex and gender differences and biases in artificial intelligence for biomedicine and healthcare. NPJ Digit Med. 2020;3:81. Published 2020 Jun 1. doi:10.1038/s41746-020-0288-5
2. Dacus AR, Behar B, Washington K. Advocacy for Diversity in Hand Surgery. Hand Clin. 2023;39(1):25-31. doi:10.1016/j.hcl.2022.08.011

Purpose: A common postoperative complication associated with mastectomies is skin necrosis (1). Our previous studies have shown that topical tacrolimus (TT) decreases rates of skin necrosis in an ischemic flap rodent model (2)(3). Clinically, we treat our patients who have breast skin necrosis with topical nitroglycerin (TN). This study compares the efficacy of TT versus TN.

Methods: A cranially based dorsal skin flap measuring 3×10cm was raised and reinset on 25 Sprague-Dawley rats. Rats were randomized to receive either TT or TN. 0.2g of selected ointment was administered to each rat both immediately post-operatively and then twice daily. Rats were sacrificed 7 days postoperatively. Fiji software was used to measure surface area of healthy tissue, reversible ischemia and full-thickness necrosis. Biopsies of tissue samples were taken from three randomly chosen animals in each group. Hematoxylin and eosin-stained slides were prepared. A blinded pathologist described the conditions of the dermal layer.

Results: The average area of the dorsal flaps in the TT and TN groups was 26.4 and 24.5 cm, respectively. In the TT cohort, the average healthy, ischemic and necrotic areas were 47.9%, 30.0%, and 22.1%, respectively. In the TN cohort, the average healthy, reversible ischemic and necrotic areas were 31.4%, 36.3%, and 32.3%, respectively. Total healthy tissue was significantly greater for rats receiving TT when compared to TN (p=0.035). Under light microscopy, reversibly ischemic tissue of rats treated with TT demonstrated increased granulation tissue compared to rats treated with TN. Similarly, necrotic tissue of TT treated rats demonstrated granulation tissue, fibroblasts, myofibroblasts and inflammatory cells while rats treated with TN demonstrated only inflammatory cells, without associated granulation tissue.

Conclusions: TT treatment was associated with significantly more healthy tissue when compared to TN. The presence of granulation tissue in reversibly ischemic and necrotic flaps treated with TT and the absence of granulation tissue in flaps treated with TN correlates with our gross findings of increased healthy tissue in our TT treated rodent model compared to TN. This study has identified that topical tacrolimus may be a better treatment for postoperative skin ischemia than topical nitroglycerin. Topical tacrolimus warrants clinical investigation for this purpose.

(1) Robertson S, Jeevaratnam J, Agrawal A, Cutress R. Mastectomy skin flap necrosis: challenges and solutions. BCTT. 2017;Volume 9:141-152. doi:10.2147/BCTT.S81712 (2) Jung W, Van Y vu R, Huang H, Wald G, Otterburn DM. The Effect of Pre-and Postsurgical Topical Tacrolimus on Pedicled Flap Survival in Rats: A Pilot Study. Plastic and Reconstructive Surgery Global Open. 2021;9(10 Suppl). (3) Wald G, Van YV, Towne W, Otterburn DM. The Effect of Topical Tacrolimus on Pedicled Flap Survival: A Histological Analysis. Ann Plast Surg. 2021;87(1s):S57-S59. doi:10.1097/SAP.0000000000002799

Purpose: Computed tomography angiography (CTA) remains the gold standard for evaluation of vascular perforators for pre-operative planning of flaps. However, CTA requires radiation exposure, contrast administration, is unable to resolve small caliber vessels, and is done in advance of surgery. Here, we present a novel method utilizing doppler ultrasound and power doppler ultrasound for the rapid generation of high quality, 3D flap angiograms. Materials and Methods: A total of 48 flaps were analyzed, as part of pre-operative evaluation and planning, and included DIEP flaps (n=10), anterolateral thigh flaps (n=8), anteromedial thigh flaps(n=10), nasolabial flaps(n=4), forehead flaps(n=4), latissimus dorsi myocutaneous flaps(n=6), radial forearm flaps(n=4), and dorsal scapular artery perforator flaps(n=2). Clarius handheld ultrasound was utilized for this study. A standardized ultrasound motion sequence was optimized for each type of flap, to ensure consistency between flaps. A method was designed and optimized to perform rapid 3D reconstruction of the ultrasound data using Slicer software. Results: The mean time from completion of ultrasound imaging to 3D reconstruction was 3.77 minutes (95% confidence interval 3.58-3.95 minutes). Importantly, our method also 3D reconstructs the tissue architecture enabling delineation of soft tissue and fibrous structures. Our method demonstrated efficacy in reconstruction of angiograms of large vessels, such as radial artery but also microvasculature at cutaneous, subcutaneous, and fascial levels, representing an advance beyond CTA. The use of power doppler enables highly sensitive and specific detection of low flow microvasculature and submillimeter vessels. The 3D reconstructions can be translated, rotated, zoomed in and out to facilitate visualization of structures. Moreover, measurements and annotations can be performed on the 3D reconstructions to facilitate, inform, and revise surgical planning. Conclusions: Our method enables the rapid generation of 3D flap angiograms in a variety of flap types. The highly sensitive detection and 3D reconstruction of microvascular flow also represents a major advance. The data demonstrates that the procedure can be performed in under 5 minutes, making the method feasible as part of the pre-operative or post-operative setting, as well as intraoperatively to monitor flap perfusion.

Purpose: A common postoperative complication associated with skin flaps is wound dehiscence (1). One factor that leads to dehiscence is tissue ischemia, which we treat clinically with topical nitroglycerin. We designed a study comparing topical nitroglycerin and topical tacrolimus in an ischemic rat model with flap dehiscence as one of our clinical endpoints.

Methods: A cranially based dorsal skin flap measuring 3×10cm was raised and reinset on 24 Sprague-Dawley rats by a single surgeon using a running suture technique with a 4-0 monofilament suture. Rats were randomized to receive either topical nitroglycerin or topical tacrolimus. 0.2g of ointment was administered to each rat immediately post-operatively and twice daily afterward. For seven days, rats were evaluated twice daily for flap dehiscence. Fiji software was used to measure surface area of viable and necrotic tissue. Bivariate analysis comparing rates of dehiscence was conducted.

Results: Seven rats from the nitroglycerin group and 1 rat from the tacrolimus group experienced dehiscence (p=0.013). In all rats, suture knots were intact at the time of dehiscence. There were increased rates of necrosis in rats treated with topical nitroglycerin relative to topical tacrolimus, 32.4% vs. 22.1%, respectively. However, this did not reach a level of significance (p=0.068). There was one incidence of post-operative death.

Conclusions: Significantly more skin flap dehiscence events were reported in rats treated with topical nitroglycerin than topical tacrolimus. Our findings suggest that the increased incidence of dehiscence in the topical nitroglycerin group relative to the topical tacrolimus group was secondary to increased necrosis along flap margins where tissue is most vulnerable to ischemia. The efficacy of topical tacrolimus in decreasing skin flap ischemia warrants further investigation.

1. Muller-Sloof E, de Laat E, Kenç O, et al. Closed-Incision Negative-Pressure Therapy Reduces Donor-Site Surgical Wound Dehiscence in DIEP Flap Breast Reconstructions: A Randomized Clinical Trial. Plastic and Reconstructive Surgery. 2022;150. https://journals.lww.com/plasreconsurg/fulltext/2022/10001/closedincisionnegativepressuretherapy_reduces.6.aspx

Introduction Surgeons apply autologous fat grafting (AFG) to correct soft tissue defects and for tissue regeneration. However, outcomes remain inconsistent and unpredictable. Vitamin D3 (VD3) offers anti-inflammatory and pro-angiogenic factors known to improve wound healing. Our small animal proof of concept studies demonstrated promising results with VD3 for graft retention and quality. The aim of this study was to evaluate oral VD3 for AFG retention in a porcine model.

Methods Three Yucatan pigs underwent AFG of inguinal adipose grafted into 16 defined dorsal zones, 5cc for each graft. Pig 1 received no medical intervention, Pig 2 received post-surgical oral VD3 thrice weekly (100K IU), and Pig 3 received oral VD3 (100K IU) thrice weekly beginning two weeks prior to surgery and continued post-operatively. The selected VD3 dose was chosen from agricultural literature showing inefficient porcine metabolism of nutritional VD3 to its metabolically active form. Pigs underwent monthly ultrasounds to estimate graft retention and collect biopsies until the study endpoint at 3 months. At sacrifice, all grafts were excised, weighed, and processed for histology. Statistical analyses (ANOVA, post-hoc tests) were performed via GraphPad.

Results Mean graft retention for Control Pig 1 was 18.99% (SEM=4.57) versus 40.19% (SEM=5.65) for Pig 2 and 41.98% (SEM=6.57) for Pig 3 (p=0.016, p=0.008). Histologically, graft architecture was similar across all study groups and resembled intact adipose with some areas of necrosis, adipocyte loss, small accumulations of free oil, and fibrosis. Quantification of longitudinally obtained biopsies for macrophages, vasculature, and perilipin-positive adipocytes is underway.

Conclusion Oral VD3 significantly increased fat graft retention compared to no treatment, with no demonstrated benefit for pre-treating graft recipients compared to post-surgical supplementation. Given the demonstrated safety profile of VD3 supplementation, with widely available serum quantification capabilities, we believe this is a feasible approach to improve fat grafting outcomes. Based on the known human conversion of VD3 from inactive precursors, we predict lower doses will be necessary in future clinical studies to achieve therapeutic benefit, as pigs are inefficient in VD3 metabolism.

Introduction

In the era of digital information, patients often rely on Google to find answers to their questions on plastic surgery.1 However, due to the recent surge in popularity of the large language model (LLM) ChatGPT, patients may turn to ChatGPT for their plastic surgery questions.2,3 Thus, this study aims to assess the quality of ChatGPT as a potential source of patient education by comparing the answers to frequently asked questions on various plastic surgery topics (free flaps, breast reconstruction, breast implant-associated anaplastic large cell lymphoma, breast implant illness, and body contouring) provided by ChatGPT and Google.

Methods

The Google and ChatGPT answers to the 10 most frequently asked questions on the search terms "free flap surgery," "breast reconstruction," "breast implant associated anaplastic large cell lymphoma," "breast implant illness," and "body contouring surgery" were recorded. Five blinded plastic surgeons were then asked to grade the quality of the answers according to the Global Quality Score (GQS), a scale to assess the quality of online sources. Wilcoxon paired t-tests were performed to evaluate the difference in GQS ratings for Google and ChatGPT answers.

Results

In a comparison of answers provided by Google and ChatGPT on various plastic surgery topics, ChatGPT significantly outperformed Google. For free flap surgery, Google scored an average of 2.49±1.25, whereas ChatGPT scored an average of 4.47±0.84 (p<0.001). For breast reconstruction, Google scored an average of 3.18±1.14, whereas ChatGPT scored an average of 4.34±0.89 (p=0.002). For breast implant-associated anaplastic large cell lymphoma, Google scored an average of 2.72±1.44, whereas ChatGPT scored an average of 4.18±1.04 (p=0.002). For breast implant illness, Google scored an average of 2.66±1.24, while ChatGPT scored an average of 4.28±0.97 (p=0.002). For body contouring, Google scored an average of 2.55±1.29, while ChatGPT scored an average of 4.38±0.67 (p=0.001).

These GQS scores indicate that for Google's responses, some information was present but important topics were missing. ChatGPT's responses were good quality, with most important topics covered. The superiority of ChatGPT's responses was attributed to their comprehensive nature and recognition of existing knowledge gaps. When asked which source they would recommend overall for patient education among ChatGPT, Google, or none, most surgeons chose ChatGPT.

Conclusion

ChatGPT outperforms Google in providing high-quality answers to commonly asked questions in plastic surgery. Although ChatGPT is limited by the accessibility of its references, we foresee the potential of this technology to revolutionize patient education when used in adjunct to clinical counseling.

References

1. Rohrich RJ, Weinstein A. Paging Dr. Google: The Changing Face of Plastic Surgery. Plast Reconstr Surg. 2021;148(5S):104S-107S. doi:10.1097/01.prs.0000794912.60516.94
2. Alessandri-Bonetti M, Liu HY, Giorgino R, Nguyen VT, Egro FM. The First Months of Life of ChatGPT and Its Impact in Healthcare: A Bibliometric Analysis of the Current Literature. Ann Biomed Eng. Published online July 23, 2023. doi:10.1007/s10439-023-03325-8
3. Liu HY, Alessandri-Bonetti M, Arellano JA, Egro FM. Can ChatGPT be the Plastic Surgeon's New Digital Assistant? A Bibliometric Analysis and Scoping Review of ChatGPT in Plastic Surgery Literature. Aesthetic Plast Surg. Published online October 18, 2023. doi:10.1007/s00266-023-03709-0

PURPOSE

Patient-facing materials in gender-affirming surgery are often written at a level higher than the NIH-recommended 8th grade reading level for patient education materials.(1) However, ChatGPT has been shown to successfully optimize certain patient-facing resources in plastic surgery.(2) Therefore, the purpose of this study is to evaluate and compare the ability of large language models to analyze readability and simplify online patient-facing resources for gender affirming procedures.

METHODS

Google Incognito searches were performed on 15 terms relating to gender affirming surgery ("Facial feminization surgery, "FTM top surgery," "MTF vaginoplasty," etc.). The first 20 text results for 11 out of 15 key terms were analyzed for reading level difficulty by a traditional online readability calculator, Readability Scoring System v2.0. Four keywords resulted in fewer than 20 texts and were analyzed in the same manner: "MTF Adams apple reduction" (N=13); "Gender affirming vocal surgery" (N=20); "Gender affirming rhinoplasty" (N=10); "Gender affirming forehead surgery" (N=5). Academic articles, non-English texts, and sponsored websites were excluded from this study. Seven different, easily accessible large language models were then used to assess website texts for reading grade level and simplify each text to an 8th grade reading level. The re-written texts were then re-evaluated for reading grade level with the Readability Scoring System. Descriptive statistics and t-tests were used for statistical analyses.

RESULTS

Online patient-facing resources on gender affirming surgery were written with a mean reading grade level of 12.7 +/- 2.7. Texts written by academic hospitals had the easiest readability at a 11.6 +/-2.8 grade level compared to private practice (13.2 +/- 2.6) and online blog sources (12.0 +/- 2.4). "Adams apple reduction" texts were the easiest to understand with a mean grade level of 10.8 +/- 2.4, and "Facial masculinization surgery" texts were the most difficult to understand with a mean grade level of 15.4 +/- 3.7. When asked to simplify texts to an 8th grade reading level, Google Gemini was most successful ( 8.5 +/- 1.5), followed by Anthropic Claude (9.6 +/- 1.9) and ChatGPT 4 (10.0 +/- 1.9). Artificial intelligence tools produced better reading level estimates (p <0.0001) and simplifications (p<0.001) for online texts describing feminizing procedures compared to masculinizing procedures.

CONCLUSIONS

Online patient-facing texts on gender affirming procedures are written with a readability more challenging than the NIH-recommended 8th grade reading level for patient-centered resources. Certain artificial intelligence tools were better at simplifying texts than others. Text readability and simplification were more difficult for texts on masculinization procedures. Providers should take caution when using large language models for patient education in plastic surgery, and improvements in artificial intelligence tools are required to provide more understandable patient-facing materials to transgender men.

REFERENCES

1. Maisner, R.S., Keenan, E., et al. (2023). A multimetric health literacy analysis of online gender affirmation surgery materials: From facial to genital surgery. JPRAS, 87, 449–460. https://doi.org/10.1016/j.bjps.2023.10.067

2. Vallurupalli, M., Shah, N.D., & Vyas, R.M. (2024). Validation of ChatGPT 3.5 as a Tool to Optimize Readability of Patient-facing Craniofacial Education Materials. PRS GO, 12(2), e5575. https://doi.org/10.1097/GOX.0000000000005575

Background : Genital lymphedema is an uncommon and disabling disease that manifests as enlargement of the genital region resulting from the disturbance of lymphatic drainage. Lymphaticovenular anastomosis (LVA) has been reported to improve genital lymphedema. Indocyanine green (ICG) lymphangiography has been a gold standard of preoperative mapping of lymphatic vessels during LVA surgery. However, there are several limitations to this technique. Previously, contrast-enhanced US (CEUS) was reported to be effective identifying lymphatic vessels in lymphedema of the upper and lower extremities. Therefore, we aimed to examine if CEUS with intradermal injections of microbubbles successfully identifies target lymphatic vessels in the preoperative management for Lymphovenous Anastomosis surgery in patients with non-extremity lymphedema.

Methods : IRB-approved, single-center retrospective chart review was performed for patients with non-extremity lymphedema undergoing LVA surgery between 2020-2024. Intraoperative CEUS with intradermal injection of microbubble suspension (Lumason, Bracco Suisse) was used in patients before LVA surgery. We will retrieve data retrospectively from chart review regarding sociodemographic variables, comorbidities, clinical characteristics, possible complications, and outcomes.

Results : A total of three patients with non-extremity lymphedema underwent a Lymphaticovenular anastomosis (LVA) surgery. The first patient was a 34-year-old female who was diagnosed with lymphangioma circumscriptum involving the lower abdomen, mons and vulva. The second patient was a 16-year-old male who was diagnosed with congenital crotal and bilateral lower extremity end-stage lymphedema. The third patient was a 31-year-old female who was diagnosed with primary lymphedema of right lower extremity and right lower abdomen and genital region. Preoperative planning for the three patients with CEUS with intradermal injections of microbubbles successfully detected lymphatic vessels in all patients, including those where ICG imaging was unsuccessful. The follow-up varied between patients. The first patient follow-up consisted in 4 monthly evaluations, even though the recovery was uneventful, 8 months later had recurring lymphedema requiring debulking surgery in the vulva. The second patient follow-up consisted of one week, edema improved, and recovery was uneventful. The third patient follow-up consisted in 4 visits in the first nine months, lower extremity edema improved substantially, but one year later required a LVA surgery for the lymphedema on the mons, which reported an improvement on the edema.

Conclusion : Contrast-enhanced US (CEUS) with intradermal injections of microbubbles was able to successfully identify target lymphatic vessels in the preoperative management for Lymphovenous Anastomosis surgery in patients with non-extremity lymphedema. To our knowledge, this is the first report on the use of CEUS with intradermal injections of microbubbles as a preoperative technique to perform a LVA in patients with non-extremity lymphedema. Our case series, despite limitations like small sample size and short follow-up time, the patients showed a decrease in lymphedema without adverse effects and post-operative complications. More research on this technique for this patient population is essential. This study highlights CEUS lymphatic mapping as an efficient technique and alternative to current methods for genital lymphedema LVA surgery.

INTRODUCTION Scarring, and in some cases hypertrophic scarring, is an unavoidable consequence of injuries and surgical procedures that represents a significant clinical challenge. Prominent scars, characterized by excessive collagen production, develop during the natural wound healing process, leaving patients with functional and psychosocial concerns. As such, managing the wound healing process, in an attempt to minimize abnormal scarring, proves to be an important facet of improving surgical treatment and postoperative recovery. Previous studies have shown that one of the core circadian clock genes, Neuronal PAS domain 2 (Npas2), regulates wound healing process through collagen synthesis and fibroblast biological activity. Npas2 appears to be a safe target for downregulation, as Npas2 knockout (KO) mice exhibited accelerated wound closure, less scarring, and did not result in embryonic or developmental pathology. The purpose of this study utilizes single-cell RNA sequencing to identify Npas2 gene expression in various identified cell classes over the wound healing process in order to begin understanding the mechanism of action of Npas2 for wound healing.

METHODS A previously established mouse excisional wound healing model was used to examine wound healing. The dorsal skin of the 12-week-old wild type C57BL/6J mice was shaved and two 5 mm diameter full skin thickness excisional wounds were created using skin biopsy punch. The open wound margin was tied with silicon splint by nylon stitch to prevent the skin contraction. On days 1, 3, 7, and 12, dermal tissues were harvested. Subcutaneous fat tissue was carefully removed, and the tissues were incubated with enzymes to isolate single cells. Cells were counted and the equal number of cells (4,000~8,000 cells/µL) from each animal were combined in the group for single-cell RNA sequencing (10X Genomics). The Cell Ranger output of scRNA-seq data was analyzed using R software.

RESULTS Npas2 expression shows using ViolinPlots, and identified cell classes including epithelial cells and fibroblasts. Healthy, naïve skin showed low Npas2 expression across all cell types. In situ hybridization demonstrated Npas2 expression in the epidermis and dermis. Wound-induced Npas2 was expressed from day 3 onwards, primarily in all types of fibroblasts (myofibroblasts, dermal fibroblasts, and papillary fibroblasts).

CONCLUSION Npas2 has the potential to serve as a genetic target for therapeutic strategies aimed at improving the outcome of surgical or injury wounds in the skin. Single-cell RNA sequencing has yielded significant results, revealing the expression of Npas2 in fibroblasts and epithelial cells. This discovery represents a major advance and provides a solid foundation for future studies focusing on the mechanisms by which Npas2 functions in the context of wound healing. By elucidating the precise role of Npas2 in these cell populations, we will aim to gain deeper insights into the underlying processes that govern wound healing, ultimately paving the way for novel therapeutic interventions and treatment strategies.