Scientific Abstract Presentations: Research & Technology Session 6 Friday, September 27 Fri, Sep 27 5:00-6:00 p.m. PDT

Purpose

Lymphedema diagnosis relies on effective imaging of the lymphatic system. Indocyanine green (ICG) lymphography has become an essential diagnostic tool, but globally accepted protocols and objective analysis methods are lacking. In this study, we aim to investigate artificial intelligence (AI), specifically convolutional neural networks to categorize ICG lymphography images patterns into linear, reticular, splash, stardust and diffuse.

Methods

A data set comprised of 68 ICG lymphography images was compiled and labeled according to 5 recognized pattern types; linear, reticular, splash, stardust and diffuse. A convolutional neural network model, utilizing MobileNetV2 and TensorFlow, was developed and coded in Python for pattern classification.

Results

The AI model achieved 97.78% accuracy and 0.0678 loss in categorizing images into five ICG lymphography patterns, demonstrating high potential for enhancing ICG lymphography interpretation. The high level of accuracy with a low loss achieved by our model demonstrates its effectiveness in pattern recognition with a high degree of precision.

Conclusion

This is the very first study integrating AI technologies into ICG lymphography procedure. This study demonstrates that AI models can accurately classify ICG lymphography patterns and indicates potential for using AI to automate and standardize the interpretation of ICG lymphography imaging.

Background: The underrepresentation of women in leadership positions continues to be an issue within many surgical specialties, particularly within the field of plastic surgery. Understanding how the distribution of age and gender differs between ASPS and Academic leadership may shed light on patterns of plastic surgeons in leadership positions.

Methods: The total number of current ASPS members, ASPS leaders, Academic leaders, faculty, and their age and genders were collected. Descriptive statistics and statistical analysis was completed with the R program, comparing ASPS leaders and Academic leaders.

Results: Among ASPS leadership, 29.9% were female, and 70.1% males. Among Academic leadership, 25.3% were female and 74.7% male. Among ASPS members, women were more likely to be leaders (OR 1.68, p-value <0.001), and to be younger than men (48 vs. 51 years, p-value 0.043). Among Academic leadership, women were found to be younger than men (47 vs. 57 years, p-value <0.001). Among Academic leadership positions, men were more likely to be Chief/Chairs and women more likely to be Assistant Program Directors (OR 2.26, 3.5, p-value 0.01, <0.001 respectively). When comparing men versus women >50 and >60 years of age in Academic leadership, men were more likely to be found in leadership (OR 3.59, 3.86, p-value <0.001, 0.002, respectively).

Conclusions: Despite improvements in diversity among plastic surgery trainees and attendings, there remains a difference in age and gender among leaders in ASPS and in Academics. Increased support is needed for women in Academic leadership and to retain them in their positions like their male counterparts.

Introduction: Patient education is critical to optimizing surgical outcomes as it guides informed decision-making and peri-operative expectations. Traditionally, patient education has relied on face-to-face consultations with healthcare providers and the provision of written materials. The rise of artificial intelligence and natural language processing tools, however, has introduced new accessible avenues for patients to gather clinical information. This study examines the innovative use of ChatGPT-3, a free natural language processing tool, as an accessible alternative platform for patient education with a focus on deep inferior epigastric perforator (DIEP) flap reconstruction and carpal tunnel release (CTR), two commonly performed procedures within plastic and reconstructive surgery.

Methods: A set of six frequently asked questions (FAQs) on both DIEP flap reconstruction and CTR was derived from the American Society of Plastic Surgery (ASPS) website. Two plastic surgeons specializing in breast reconstruction and four plastic surgeons specializing in hand surgery systematically assessed ChatGPT's responses based on six criteria: clarity, relevance, reliability, validity, organization, and comprehensiveness using a 1-5 Likert scale (from poor to excellent). Inter-rater reliability was measured by Cohen's Kappa Score (for DIEP group) and Gwet's Agreeemt Coefficient Score (for CTR group).

Results: In both the DIEP and CTR groups, ChatGPT-3 generally received high scores in clarity and relevance (ranging from 4 to 5), demonstrating its ability to effectively convey understandable and pertinent information. However, the model scored lower in reliability (ranging from 3 to 4) and demonstrated significant range in validity (from 2 to 4), reflecting a variance in the alignment of ChatGPT-3's responses with the surgeons' expert knowledge. In addition, its organization and comprehensiveness of the responses received a mixed reception, with scores ranging from 1 to 5. The inter-rater analysis showed a Cohen's Kappa score of 0.5 (indicating moderate agreement beyond chance in the DIEP cohort) and a Gwet's Agreement Coefficient score of 0.507 (suggesting moderate to substantial agreement in the CTR cohort).

Conclusion: ChatGPT-3 demonstrates promise in delivering clear and relevant information for commonly performed reconstructive and hand surgeries such as DIEP flap reconstruction and CTR. Improvements in reliability, validity, organization, and comprehensiveness, however, will be critical for its adoption as a reliable patient education tool. The study also underscores the importance of precise and well-crafted prompts (i.e. Prompt Engineering) in order for patients to harness the full potential of ChatGPT-3.

Background: A medical student's experience during a core third-year clerkship serves two main roles: the first is to expose the student to a field to aid in specialty selection, and the second is to provide the student with foundational knowledge sufficient to pass USMLE examination, regardless of what field they ultimately pursue (1,2). Much has been written about educating students interested in plastic surgery about plastic surgery, but comparatively little has been written about educating students who are on a plastic surgery service during their surgery clerkship. While plastic surgery might be perceived as a niche subspecialty, its interaction with almost every anatomical region in the body can serve as a comprehensive setting for surgical education. We hypothesize that plastic surgeons are participating in primary surgical education more than they might realize, and propose a surgery clerkship curriculum in line with the American College of Surgeons (ACS) Core Competencies (3) to guide plastic surgeons serving as primary surgical educators.

Methods: A cross-sectional study of surgery clerkship curricula was performed in August 2023. All accredited US allopathic medical schools and plastic surgery residencies were identified using the LCME and ACAPS directories respectively. Information about surgery clerkship curricula was gathered on individual medical school websites. If information was unavailable or incomplete after an Internet search, the school's surgery clerkship director was contacted directly. The following information was collected: total length of surgery clerkship, number and length of surgical rotations, and availability of plastic surgery as a clerkship service.

Results: Clerkship information was obtained for 71 out of 158 US allopathic medical schools (45%). Students spent a median of 8 weeks on their surgical clerkship (range: 4-12) and rotated through a median of 2 (range: 1-5) different services. Plastic surgery was offered as a clerkship rotation at 46 surveyed schools (65%). At institutions where plastic surgery was offered, students spent an average of 40% of their total surgery clerkship on the plastics service. At 7 of the 46 schools offering plastic surgery (15%), students only rotated on one surgical service; meaning at these institutions, a student may obtain their entire primary surgical education on the plastics service. Additionally, 100 independent and integrated plastic surgery programs were identified via the ACAPS directory, and 95% were found to have direct affiliations with medical schools. Thus, almost all plastic surgery residents and academic faculty operate in a setting where they teach medical students.

Conclusions: While at first glance, plastic surgery might appear to be a highly specialized corner of medicine, its wide scope of practice and head-to-toe surgical breadth make it a hospitable learning environment for students with a wide array of interests. Moreover, many medical schools rely on plastic surgery services to host students on their surgery clerkships, with almost all academic plastic surgeons already participating in primary surgical education. We propose a curriculum of key skills obtainable on a plastic surgery rotation that fulfill the ACS Core Competencies, to ensure that plastic surgeons can provide students with a comprehensive undergraduate surgical education.

References 1. Chen, H., Hardacre, J. M., Martin, C., & Lillemoe, K. D. (2001). Do medical school surgical rotations influence subspecialty choice?. Journal of Surgical Research, 97(2), 172-178.

2. Lewis, B. D., Leisten, A., Arteaga, D., Treat, R., Brasel, K., & Redlich, P. N. (2009). Does the surgical clerkship meet the needs of practicing primary care physicians?. Wisconsin Medical Journal (WMJ), 108(8), 398.

3. de Moya, M. A., & Sudan, R. (2019). ACS/ASE Medical Student Core Curriculum. American College of Surgeons.

Background: Gender-related inequities in the field of plastic surgery are well described. However, the mechanisms that underlie gender-related gaps in professional achievement and advancement as well as the impact on personal well being are poorly understood. Therefore, the objective of this study was to collect the narrative experiences of women plastic surgeons in practice navigating the workforce to further investigate the systemic and societal barriers and inequities women plastic surgeons face over the course of their personal and professional lives.

Methods: To understand and describe experiences of women in plastic surgery we designed an interpretive description, a non-categorical qualitative method that allows for the understanding of a group's experiences. One-on-one semi-structured interviews were conducted in person and virtually with women plastic surgeons in various practice types across the United States between October 2023 and January 2024. Participants were asked to share narratives of their experiences across different domains. Interviews were analyzed inductively to generate themes across narratives.

Results: Twenty-four women plastic surgeons were interviewed, with purposive sampling by race and ethnicity, years in practice, subspecialty, practice setting, and geographic location. Participants reported experiencing inequities navigating the work environment, which were categorized across three main themes: 1) Participants reported a lack of transparency particularly pertaining to contract negotiation, salary, and promotion. In some instances, women described "inappropriate" comments about deservedness: "…the first time I ever negotiated for a raise…My chief [said] 'Well, you know…"he has five kids. So I could see why he would want a raise…I'm really surprised you want a raise.' Because I'm a single woman…". 2) Participants reported a lack of inclusivity in traditional surgical culture. For example, one participant describes feeling like an outsider among her colleagues, "…the chairman was a total bro [and] had known one of [my partners] since he was a medical student…[so he] made sure…that [his] practice got busier…he totally drove it for [him]. Whereas the same chairman said to me when I wasn't making my salary, 'You know, if you were my daughter, I'd have you go interview with [name]…down there at [the hospital].' It's like, I'm not your daughter. I'm a surgeon-that you hired!" 3) Participants reported frequently navigating social gender "norms", describing different societal expectations and overcoming common female stereotypes. One participant recounts, "I have been told that I have high expectations [and there have been complaints that] I was aggressive...[but] I am [much] easier to work with on a daily basis than some of my [male] partners who also have high expectations...they are demanding and at times inappropriate in the operating room, and I know that because I was their resident a few years ago… I was never to behave in any of those ways…but they've never had those complaints."

Conclusions: While the number of women entering plastic surgery is rising, gender-related barriers continue to create inequity in professional advancement and career satisfaction. A deeper understanding of these phenomena is critical to develop effective strategies that systematically address and enhance the diversity, inclusion, and equity in plastic surgery.

PURPOSE: Tissue engineering has made significant progress in generating skin constructs with diverse applications. However, inefficient vascularization and slow in-growth of new blood vessels are commonly cited issues contributing to construct failure (1). Incorporation of pre-fabricated blood vessels during the in vitro synthesis may circumvent this concern and improve construct integration and performance (2). We hypothesize that pre-vascularization will expedite the formation of anastomosis between host-recipient capillaries and enhance construct survival long-term in mice models. METHODS: Our research team has previously developed non-vascularized tissue-engineered skin constructs, incorporating human epidermal keratinocytes and dermal fibroblasts in a rat tail collagen type I matrix. We aimed to enhance these constructs with vascularization, introducing a collagen-fibrin layer beneath the dermis-like layer, embedding enhanced Green Fluorescent Protein-labeled Human Umbilical Vein Endothelial Cells (eGFP-HUVECs) and Dental Pulp Stem Cells (DPSCs). Evaluating these constructs, we treated full-thickness wounds (diameter = 8 mm) on athymic mice (aged 10-11 weeks) with them, assessing healing at 1-, 4-, 7-, and 14-day intervals. Post-application, mice were euthanized for assessing integration, with excised areas analyzed via Hematoxylin & Eosin (H&E) and immunohistochemical staining, identifying human vascular networks using anti-GFP antibody (primary antibody) and fast red chromogen (secondary antibody). RESULTS: Successful integration of the skin constructs into the wound bed was consistently observed across all examined time points via histological imaging. Initially, on Day 1, capillary networks were not visible; however, there was notable activity indicative of macrophages/cytokines migration from the wound site toward the construct's epidermis. By Day 4, the emergence of a few capillaries within the construct hinted at the onset of angiogenesis and anastomosis. By the 7th day, there was substantial healing and wound contraction, with an absence of scar tissue formation. The utilization of anti-GFP antibodies played a crucial role in demonstrating the successful integration of the tissue-engineered skin constructs with the mouse's wound bed. The observed lumens, filled with red blood cells, indicated effective blood flow within the construct, and subsequent anti-GFP analysis confirmed the connection of human blood vessels containing mouse blood to the host's circulatory system. By Day 14, a decrease in the density of blood vessels was noted. CONCLUSION: We have established the optimal culture conditions for the creation of pre-vascularized skin constructs that integrate seamlessly and form anastomoses with the recipient mice's blood vessels. Our ongoing research is dedicated to understanding the importance of pre-vascularization in improving the anastomotic process. Furthermore, we plan to evaluate similar parameters in decellularized versions of these pre-vascularized skin constructs, expanding our insights into their functionality and potential applications.

1. Shahin H, Elmasry M, Steinvall I, Söberg F, El-Serafi A. Vascularization is the next challenge for skin tissue engineering as a solution for burn management. Burns & trauma. 2020;8:tkaa022.
2. Son J, Mohamed HJ, Ha W, et al. Bioprinting of pre-vascularized constructs for enhanced in vivo neo-vascularization. Biofabrication. 2023;15(3).

Purpose: To evaluate the utility of training ChatGPT chatbots on ASPS and ERAS guidelines to help provide sufficient perioperative patient-centered recommendations for breast reconstruction.

Background and Rationale: Integrating artificial intelligence (AI)-powered large language models and application-based programs into multiple surgical care pathways, including breast reconstruction, has improved efficiency at different checkpoints. This technological paradigm shift has prompted the integration of open-source chatbots into breast reconstructive consultation, albeit without success [1].

Efficiency and accessibility are two critical aspects of post-mastectomy breast reconstruction. Guidelines released by the American Society of Plastic Surgeons (ASPS) and the Enhanced Recovery After Surgery (ERAS) Society outline best practices for optimizing surgical outcomes in autologous and implant-based reconstruction [2-4]. They provide strong recommendations for preoperative optimization, chronic disease management, and perioperative pain control. Critical appraisals reveal improved postoperative outcomes and patient satisfaction when following these guidelines.

Despite evidence-based guidelines, patients defer to the Internet or AI chatbots for medical advice [5]. The delivery, quality, and accessibility of recommendations via AI chatbots remain potential barriers to evidence-based practice. However, training chatbots to deliver evidence-based guidelines may benefit patient education. Therefore, our study aims to evaluate how well ChatGPT, trained on current ERAS and ASPS guidelines, can deliver sufficient preoperative recommendations for breast reconstruction candidates.

Methods: Using a Large Language Model application builder, we created three custom ChatBots using current ASPS guidelines for expander- and implant-based breast reconstruction, ASPS guidelines for autologous flaps in breast reconstruction, and ERAS guidelines to simulate a standardized patient interview. A fourth non-trained chatbot served as a control.

Trainees and faculty at our institution simulated a fictional patient encounter with each chatbot. They completed a short survey evaluating the chatbot's knowledge of surgery type, lifestyle management, comorbidity management, preoperative optimization, and knowledge of ASPS and ERAS guidelines.

Results: Participants nearly unanimously agreed that all four chatbots adequately educated patients on chronic disease and lifestyle risk factors. All bots provided information that was helpful and factually correct. There was more heterogeneity between the chatbots when counseling patients on type of surgery and perioperative optimization. Additionally, not all bots acknowledged gaps in their knowledge.

A chi-square regression analysis was conducted using SPSS. Although this analysis did not yield statistically significant results, it provided valuable insights into which bots best delivered information concordant with guidelines. Overall, there was disagreement on its ability to counsel patients on optimization and type of surgery.

Conclusion:

Our study indicates potential benefits and drawbacks of guideline-trained AI chatbots in educating patients on evidence-based breast reconstruction guidelines. As AI becomes more ubiquitous, further research, including expanding the respondent pool, is needed to comprehensively analyze its role in delivering patient recommendations. These data can refine the integration of guideline-driven chatbots into other plastic and reconstructive procedures. However, while AI may be helpful for some aspects of patient education regarding breast reconstruction, medical evaluation and recommendations from a plastic surgeon remain the gold standard.

References: 1. Saturno, M.P., et al., Generative artificial intelligence fails to provide sufficiently accurate recommendations when compared to established breast reconstruction surgery guidelines. J Plast Reconstr Aesthet Surg, 2023. 86: p. 248-250. 2. Lee, B.T., et al., Evidence-Based Clinical Practice Guideline: Autologous Breast Reconstruction with DIEP or Pedicled TRAM Abdominal Flaps. Plastic and Reconstructive Surgery, 2017. 140(5). 3. Alderman, A., et al., ASPS clinical practice guideline summary on breast reconstruction with expanders and implants. Plast Reconstr Surg, 2014. 134(4): p. 648e-655e. 4. Temple-Oberle, C., et al., Consensus Review of Optimal Perioperative Care in Breast Reconstruction: Enhanced Recovery after Surgery (ERAS) Society Recommendations. Plastic and Reconstructive Surgery, 2017. 139(5): p. 1056e-1071e. 5. Nguyen, J., P. Thompson, and A. Losken, Breast Reconstruction in the Social Media Age. Aesthet Surg J, 2021. 41(2): p. 200-205.

Background: Indocyanine green (ICG) helps preoperatively identify target organs and tumors. While research on ICG in trauma is still being explored, this project provides an opportunity to examine its utility in the reconstruction of traumatic defects. The most recent review on this topic was done in 2017, however, the rapid innovations in the field make an update necessary. This review is an update on recommendations and applications for ICG use in pre-, intra- and post- operative reconstructive plastic surgery(RPS) procedures.

Methods: A review was performed on Medline and PubMed for articles on recommendations for the use of ICG in RPS. Full text articles in English that analyzed ICG applications in PS procedures between Jan 2018 and Feb 2024 were included. This time period was selected following the most recent review published in 2017. Case reports, reviews, meta-analyses, and experimental studies were excluded during screening.

Results: 92 abstracts were reviewed and 10 full text articles that met criteria were included for full review. Updated applications of ICG use in RPS were identified in free flap perfusion assessment, treatment of secondary lymphedema, and sentinel lymph node biopsy in breast cancer. An additional clinical application microsurgery update was in the use of ICG with near infrared window II (NIR-II) compared to current standard NIR-I.

Conclusion: ICG angiography continues to be highlighted as a valuable intraoperative tool to assess intra- and post-operative flap viability, neovascularization perfusion, and guide operative decision making in reconstructive plastic surgery(1). In the case of thin and superthin flaps, ICG is helpful in assessing location and patterns of pure skin perforators within superthin flaps and minimizing risk of facial nerve injury in temporoparietal fascia flap harvesting(2,3). ICG angiography can also be used to detect flap congestion after anastomosis and inset. In breast cancer related lymphedema, ICG lymphangiography (ICG-L) may be useful in preoperative staging to better stratify and select patients for microsurgical treatment(4). This may be useful to improve microsurgical outcomes such as in lymphovenous anastomosis for lymphedema treatment. In microsurgery, use of ICG with NIR-II (1,000-1700 nm) imaging is a promising new tool with advantages over other fluorescence imaging. ICG with NIR-II has real-time capturability with higher resolution in cutaneous microvasculature compared to NIR-I(5). Overall, there is potential ICG application in traumatic defect reconstruction which our institution is continuing to explore.

References: 1. George RE, Elwood ET, Jones GE. Indocyanine Green Angiography Overpredicts Postoperative Necrosis Compared to Multispectral Reflectance Imaging. Plast Reconstr Surg. 2023;151(3):412e-419e. doi:10.1097/PRS.0000000000009917

2. Narushima M, Yamasoba T, Iida T, et al. Pure Skin Perforator Flaps: The Anatomical Vascularity of the Superthin Flap. Plast Reconstr Surg. 2018;142(3):351e-360e. doi:10.1097/PRS.0000000000004698

3. Patel V, Kulich M, Kochhar A, Gomez G. Endoscopic-Indocyanine Green Angiography Assisted Microtia Reconstruction. Laryngoscope. 2023;133(12):3615-3618. doi:10.1002/lary.30759

4. Jørgensen MG, Toyserkani NM, Hansen FCG, Thomsen JB, Sørensen JA. Prospective Validation of Indocyanine Green Lymphangiography Staging of Breast Cancer-Related Lymphedema. Cancers. 2021;13(7):1540. doi:10.3390/cancers13071540

5. Wu Y, Suo Y, Wang Z, et al. First clinical applications for the NIR-II imaging with ICG in microsurgery. Front Bioeng Biotechnol. 2022;10:1042546. doi:10.3389/fbioe.2022.1042546

Background: Securing a position in plastic and reconstructive surgery residency programs remains highly competitive. As applicants strive to enhance their profiles through various achievements such as Alpha Omega Alpha honor society membership, impactful letters of recommendation, and superior clerkship grades, research productivity has emerged as a critical differentiator. This importance has been magnified with the USMLE Step 1 exam's transition to a pass/fail system, leading to an increase in the average number of research publications cited in residency applications. However, less is known about publications that are listed as "submitted" to a journal on an application, rather than published or accepted. There is no way to truly verify the status or quality of publications that are listed as submitted on an applicant's record. Therefore, the purpose of this study was to investigate the fate of submitted publications and subsequently what value they possess for programs evaluating resident candidates.

Materials & Methods: This retrospective study included applications to an integrated plastic and reconstructive surgery residency between 2015-2023. Data on demographics and the number of submitted publications were collected. The publication status of each submitted paper was verified using PubMed and Web of Science. We compared the impact factor of the journals at submission against those of eventual publication using two-year impact factors corresponding to the submission year. Statistical analysis was conducted using univariate t-tests with an alpha threshold of 0.05.

Results: A total of 564 applications, encompassing 750 submitted papers, were evaluated. On average, applications featured 1.85 submitted papers (SD=3.03), with 67.7% (n=508) eventually published and 32.3% (n=242) remaining unverifiable. Papers were initially submitted to journals with an average impact factor of 3.501, which differed to a statistically significant extent from the 2.58 impact factor of journals where they were ultimately published (p=0.048). Notably, applicants were not a listed author on 2.13% (n=16) of submitted papers upon final publication.

Conclusions: This study found that the majority of submitted papers on applications are indeed published. However, given that the publication status of 32.3% of submitted papers was unverifiable, a significant amount of submitted papers never actually end up published. While the papers listed on the 2023 application cycle have had almost an entire year to be published, publication delay could partially explain these results. Furthermore, there was a statistically significant discrepancy between the original and final impact factors of the journals where papers were published. Additionally, authorship discrepancies in 2.13% of submitted papers call for cautious interpretation of submitted research's significance. Our paper concludes that while papers that are listed as submitted at the time of application do indicate scholastic endeavors, they should be distinguished from published works in application evaluations. Further discussion during interviews may provide a more accurate assessment of their contribution to an applicant's scholarly profile.