Scientific Abstract Presentations: Aesthetic Session 1 Thursday, September 26 Thu, Sep 26 1:00-2:30 p.m. PDT

Background: Connective tissue diseases (CTD), such as rheumatoid arthritis, lupus, and scleroderma, may influence wound healing and postoperative outcomes due to autoimmune inflammatory processes. However, the association between CTDs and complications following cosmetic surgery is understudied and lacks consensus. The aim of this study is to compare complications after cosmetic surgery in CTD patients versus healthy controls.

Methods: An IRB-approved retrospective review (IRB#22-764) of all patients diagnosed with CTD who underwent cosmetic procedures at our institution from 2003-2022 was conducted. Demographic data, comorbidities, medications, procedures, and postoperative complications were collected. Patients were excluded for missing data or lack of follow-up. Healthy controls were identified for each procedure and matched 1:1 based on propensity scores derived from race, gender, BMI, smoking status, and comorbidities. We estimated a sample size of 229 patients per study group needed to achieve statistical power (α = 0.05, β = 0.20, p1 = 43%, p2 = 30%). Statistical analysis included t-tests and chi-squared tests, where appropriate. All data are presented as mean and standard deviation (mean ± SD).

Results: Six hundred and thirty-eight patients were included, comprising 319 (50%) patients diagnosed with CTD, and 319 (50%) healthy controls. 597 (93.6%) were female and 41 (6.4%) were male, with an average age at surgery of 56.3 years. CTD patients were significantly older than their healthy counterparts (59.5 ± 13.6 versus 53.1 ± 15.3, p<0.0001). Within the CTD group, connective tissue pathologies included rheumatoid arthritis (73.4%), lupus (27.6%), and scleroderma (3.45%). Between the CTD and control groups, there were no differences with respect to sex (p=0.420), race (p=0.671), BMI (p=0.634), smoking status (p=0.797), or comorbidities (diabetes mellitus, p=0.715; chronic kidney disease, p=0.541; heart failure, p=0.613; peripheral vascular disease, p=0.677; coronary artery disease, p=0.501). The predominant procedures were blepharoplasty (340, 53.3%), breast reduction (137, 21.5%), and abdominoplasty (80, 12.5%). The overall complication rate for the study population was 20.2% (129 complications). Predominant complications included wound dehiscence (35, 5.49%), infection (32, 5.02%), and soft tissue necrosis (17, 2.67%). There was no difference between the CTD group and control group in overall complication rate (21.6% versus 18.8%, p=0.375), or the rate of complications requiring reoperation (33.3% versus 26.7%, p=0.411). Subgroup analysis of individual CTDs showed no difference in complications for patients with rheumatoid arthritis (p=0.386), lupus (p=0.618), or scleroderma (p=0.958) versus controls. 65.2% of CTD patients were prescribed disease modifying therapy including methotrexate, corticosteroids, hydroxychloroquine, and other immunosuppressant biologics at the time of surgery. These patients had no difference in complication rate compared to the control group (24.0% versus 18.8%, p=0.148). Subgroup analysis of individual procedures showed no difference in complications between CTD patients and controls who underwent blepharoplasty (p=0.384), breast reduction (p=0.909), abdominoplasty (p=0.459), rhytidectomy (p=0.499), or rhinoplasty (p=1.00). However, CTD patients who underwent breast augmentation had a significantly higher rate of complications compared to breast augmentation controls (35% versus 5%, p=0.0177*).

Conclusion: Most cosmetic procedures are safe in patients with connective tissue diseases including rheumatoid arthritis, lupus, and scleroderma, and do not carry an increased burden of post-operative complications.

Background: The nasal tip is a complex three-dimensional structure whose projection is supported by several anatomic elements including but not limited to the lateral crura, medial crura, nasal spine, septal angle, alar base, and anterior nasal spine. Modification of tip projection thus requires judicious consideration of its various structural contributors. However, given that multiple structures may be concomitantly altered during rhinoplasty, it is often difficult to disentangle the contributions of specific maneuvers to postoperative tip projection. This study aims to clarify the distinct effects of intraoperative maneuvers affecting tip projection.

Materials & Methods: This is a retrospective study of patients who underwent primary rhinoplasty at a single private practice between 2001 and 2022 with standardized preoperative and postoperative photography. Preoperative and postoperative standardized right profile images were compared. Tip projection was assessed using the Goode ratio. Patients were stratified into one of three categories according to the preoperative Goode ratio: underprojected (<0.55), ideally projected (>0.55, <0.60), or overprojected (>0.60). Matched pairs analyses were used to determine significant differences between preoperative and postoperative values. Linear regression and MANOVA were employed to detect intraoperative maneuvers with significant effects on tip projection.

Results: A total of 386 patients (310 female, 76 male; mean age: 31.15±13.99 years) underwent primary rhinoplasty with standardized preoperative and postoperative photography. The mean time between surgery and photographic follow-up was 897.09±888.79 days. Preoperatively, 301 patients were overprojected, 13 patients were underprojected, and 72 patients had ideal tip projection. The mean preoperative Goode ratio for overprojected patients was 0.67±0.08, while the mean postoperative Goode ratio was 0.62±0.04. For underprojected patients, the preoperative mean was 0.51±0.04 and the postoperative mean was 0.53±0.04. For both overprojected (p<0.0001) and underprojected patients (p=0.0330), there was a significant projection change in the desired direction. In comparison, the mean preoperative ratio for ideally projected patients was 0.58±0.01 and the postoperative mean was 0.58±0.04-postoperative projection was not found to be significantly different from preoperative projection (p=0.4891). For overprojected patients, ratio reductions were associated with lateral crural repositioning (difference: -0.04, p=0.0007), increased ratios were associated with transdomal sutures (difference: +0.02, p=0.0479), membranous septum resections (difference: 0.03, p=0.0320), columellar strut grafts (difference: +0.02, p=0.0286), footplate approximation (difference: +0.03, p=0.0021), and subdomal grafts (difference: +0.03, p=0.0046). For ideally projected patients, lateral crural onlay grafts were associated with increased projection (difference: 0.1. p=0.0068). Linear regression revealed subdomal grafts (p=0.0083), tip rotation sutures (p=0.0139), soft triangle grafts (p=0.0376), and lateral crural struts (p=0.0433) as the most significant drivers of tip projection, with subdomal grafts having the largest observed effect.

Conclusions: These data highlight several techniques for their significant long-term influence on nasal tip projection and validate the utility of factors contentiously postulated to drive changes in tip projection, including the columellar strut graft and footplate approximation. Additionally, regression modeling identified the subdomal graft as the most significant driver of increased tip projection.

Background Emerging do-it-yourself (DIY) trends in hyaluronic acid lip fillers have gained traction on most social media platforms. One popular trend involves using a "needle free" device, called the Hyaluron Pen, distributed mostly by non-medical professionals, which employs compressed air pressure to administer hyaluronic acid into the lips. However, plastic surgeons and medical agencies have raised concerns over this product due to the inability to control the depth of penetration and nonsterile methods of refilling the device. Currently, there is a lack of research investigating its usage and potential enduring complications. In this study, we aimed to explore both patient and injector perspectives and outcomes associated with the Hyaluron Pen.

Methods Anonymous surveys were administered via Amazon Mturk in February 2024. Eligible participants were 18 years or older who had used the Hyaluron Pen in the United States. Perspectives on use of Hyaluron Pen were assessed through questions regarding Hyaluron Pen administration, distribution, costs, experiences, complications, and training. Data was analyzed using descriptive statistics.

Results A total of 161 participants indicated using the Hyaluron Pen. A majority of the respondents (53.4%) learned about the Hyaluron Pen from social media. About 43% of participants indicated that licensed health professionals administered their Hyaluron Pen treatments, followed by physicians (36%) and self-administration (18.6%). Participants paid on average $536.80 per injection session and had about 47 treatments. Almost all participants (98.8%) rated their experience with the Hyaluron Pen as positive and would overwhelmingly (97.5%) recommend it to others. Most started using the Hyaluron Pen because of its ease of use (70.2%), to have fuller lips (53.4%), and its affordability over conventional injectable filler (47.8%). Common side effects included infection (39.8%), bleeding/bruising/swelling (34.8%), gross asymmetry (31.7%), and scarring (31.7%). Concerningly, 79% developed complications from the Hyaluron Pen that resulted in either hyaluronidase injections (59.1%), corrective surgery (48.8%), emergency room visit (37.8%), or hyperbaric oxygen treatment (24.4%). About 78% of participants with complications reported permanent disfigurement from the Hyaluron Pen and paid an average of $2493.90 out-of-pocket for corrective treatments. Individuals who administer Hyaluron Pen indicated training from Hyaluron Pen courses (81.9%) and videos on social media (17.2%). Alarmingly, 15% of injectors do not counsel on any potential complications.

Discussion While the allure of the Hyaluron Pen may be enticing to consumers seeking a quick fix for achieving plump, youthful lips, self reported complications reported in this study justify growing concerns from the medical community. This study highlights the need for further comprehensive research, strict regulatory oversight, and plastic surgeon-led initiatives against its usage and distribution to safeguard the well-being of patients seeking cosmetic enhancements.

Purpose: The use of before-and-after images in body contouring is routine, and patients often refer to these photos as a reference for expectations of results. However, recent developments in image-generating artificial intelligence (AI) now allow lifelike 'photos' of nonexistent individuals. Models are directly accessible to patients, who already report using these tools in medical decision-making. Various AI models have been trained on divergent sets of images, and it is unclear if generated images accurately represent human figures, cosmetic defects, or expected surgical results. The authors therefore sought to examine the quality of these generated images.

Methods: AI Models, Image Generation, and Criteria for Textual Prompts: Text prompts for abdominoplasty and buttock augmentation were written by the lead author, including desired pre- and postoperative features but excluding patient-specific information. Prompts were fed into ChatGPT-4, which was iteratively used to develop complex prompts use in image-generation models. After final prompt generation, the same prompt was fed into GetIMG, Leonardo, and Perchance to produce one image each for abdominoplasty and for buttock augmentation. Surgeon Involvement and Validation: Grading systems from various other literature were adopted to create an 11-item assessment (scale 1-5) of image quality relating to surgically relevant factors, including consistent image presentation, anatomical veracity, appropriate presentation of the aesthetic deformity, and appropriate postoperative result. A panel of two board-certified plastic surgeons and two plastic surgery residents evaluated the images using the 11-item assessment, and provided free text feedback. Results were analyzed by Excel for means and standard deviations. Statistical testing was performed with ANOVA, Kruskal-Wallis H-testing, and Mann-Whitney U-testing where appropriate.

Results: When comparing models, GetIMG had the highest mean score (3.66 ± 0.95), followed by Leonardo (3.07 ± 0.84) and Perchance (2.45 ± 0.75). The highest scoring metrics were Correct Simulation of Age (3.36 ± 0.85) and Symmetry Analysis (3.45 ± 0.94), while Healing and Scarring Prediction was rated lowest (2.13 ± 1.14), due in part to the fact that no models were able to generate scar patterns. Metrics for image realism showed significant variation between panelists (F=12.51, p=0.0034). Conversely, metrics for clinical value demonstrated no significant differences between panelists (F=0.77, p=0.55). There was consistent performance across the eleven metrics (H=16.03, p=0.099). When comparing AI models, there were no significant differences in realism metrics (F=0.437, p=0.663), but significant disparities in clinical value metrics (F=10.317, p=0.017). However, the paired tests as followed revealed no significant differences between any models: GetIMG vs. Leonardo (U=5.5, p=0.236), GetIMG vs. Perchance (U=6.0, p=0.2), and Leonardo vs. Perchance (U=9.0, p=0.1).

Conclusion: While some images demonstrate accuracy in depicting appropriate surgical candidates and expected postoperative results, some were anatomically unrealistic or failed to show appropriate postoperative results. Patients should wait to use AI-generated images in the surgical decision-making process until systems significantly improve, as they may be led to expect impossible results at this time. In addition, surgeons should continue to use photos of actual patients in their surgical consultations rather than generated images. Further planned research includes comparison to 'real' before-and-after imagery and improved image generation.

Background

Hybrid fractional lasers (HFL) have redefined the landscape of facial laser resurfacing treatments, owing to their dual ablative and non-ablative wavelength technology. While HFL therapy has already shown strong outcomes, the safety of HFL at skin depths beyond 150μ is debated, particularly for its non-ablative wavelength. Some suggest that HFL delivery at deeper levels has the potential to further enhance skin remodeling and rejuvenation. As such, we conducted a clinical safety assessment of deep HFL therapy among patients concurrently undergoing rhytidectomy for facial rejuvenation.

Methods

A retrospective review was performed at a single site from 2017 to 2022. All patients who underwent rhytidectomy with or without concurrent intraoperative HFL therapy were included in the study. Data was collected on patients' demographics, medical history, the details of HFL therapy, and postoperative complications. Differences between the two groups were determined using descriptive statistics.

Results

A total of 169 patients were included, most of whom were female (94.7%) and white (97.0%) with an average age of 63.3 ± 7.5 years and Fitzpatrick score of 2.7 ± 0.7. All patient characteristics, such as demographics, comorbidities, and relevant surgical history, were similar between the HFL group and the control group. Among those who received HFL therapy at the time of rhytidectomy, the average depth of delivery for the non-ablative laser was 355μ ± 25μ across all areas of the face. The maximum depth observed in the HFL cohort was 425μ. In the 12 months following index rhytidectomy, 7.6% of the HFL group experienced postoperative complications, compared to 3.0% in the control group. This difference was not significant (p = 0.13). Known complications of HFL, such as burns, skin breakdown, and hypo- and hyperpigmentation were not observed among any of the patients in the HFL group.

Conclusions

HFL therapy at depths of up to 425μ has been proven to be a safe option for facial skin photorejuvenation. Furthermore, patients with Fitzpatrick skin types I-VI, who are especially susceptible to epidermal injury, did not experience any adverse effects secondary to HFL therapy. Compared to rhytidectomy alone, combining HFL with rhytidectomy is not associated with a greater rate of complications, offering a promising, multifaceted approach for comprehensive facial rejuvenation. Our findings serve to expand the clinical safety parameters of safe HFL use, which is an instrumental step in optimizing the clinical outcomes and long-term success of laser skin resurfacing.

Background: Modern lipoabdominoplasty techniques optimize aesthetic results while minimizing risk of wound healing complications related to malperfusion or ischemia. This descriptive study seeks to characterize the volume of adipose tissue above and below Scarpa's fascial plane spanning the anterior abdominal wall to better understand the potential impact of direct sub-Scarpal fat excision and flank liposuction in lipoabdominoplasty.

Methods: Retrospective chart and CT imaging review was performed on 41 Deep Inferior Epigastric Perforator (DIEP) breast reconstruction and 24 lipoabdominoplasty patients between November 2010 to June 2022. Variables collected included demographics, surgical history, BMI at multiple time points and abdominal fat measurements (in millimeters) above and below Scarpa's fascia. Individual measurements were taken at 20 unique sites across the anterior abdominal wall. Statistical analysis was performed to assess potential relationships between supra- and sub- Scarpa's Fascia adipose volume with BMI and age via multivariate analysis. Ratios of adipose volume above and below Scarpa's fascia were calculated using a Mean of Ratios equation. These ratios were compared along the same vertical axis in pre- and post-operative DIEP flap breast reconstruction patients in an attempt to assess how tension at closure may modify adipose anatomy and aesthetic outcomes.

Results: No significant statistical difference was found between the two patient cohorts for both age (p=0.6461) and BMI (p=0.6434). In the DIEP cohort, the average time from preoperative CT imaging to surgery was 1.1 years [2 days-3.8 years], and 2.5 years from surgery to postoperative CT [15 days-10 years]. Age and BMI were found to be unrelated to fat volumes above or below Scarpa's layer in the preoperative DIEP CT measurements. In all three groups (lipoabdominoplasty, pre- and post- operative DIEP) there is a trend toward higher above/below ratios at the lateral-most sites. This trend in ratios is driven by a combination of higher supra-Scarpa's adipose volume laterally and higher sub-Scarpal volume at midline. The inferiormost horizontal axis was analyzed in pre- and post-operative DIEP patients to assess tension at closure site. RL3 and LL3 (lateral-most sites) ratios decrease from pre-op to post-op, while the locations nearing midline (R3, M3 and L3) increase. These differing ratios at the most midline sites are driven by a decrease in supra-Scarpa's adipose volume following lipoabdominoplasty.

Conclusion: The results of this study support direct excision of sub-Scarpal fat with flank liposuction in abdominoplasty following our observation that sub-scarpal fat volumes were highest near midline and the supra-scarpal fat volumes were highest laterally. Further, this data indicates that the majority of fat remains in the supra-Scarpal plane following prior abdominoplasty or DIEP, which would be most amenable to liposuction should a patient present seeking further abdominal contouring after one of these procedures. Lastly, the pre- versus post-operative DIEP CT values in the upper abdomen changed insignificantly, suggesting what seems intuitive, that inferiorly directed tension applied to the lower abdomen at the time of closure only minimally affects the upper abdomen.

Background: Marijuana is the most prevalent recreational drug used in the United States and is available for medicinal prescription in 37 states. Shifts in public acceptance and legal reform coincide with historic highs in reported marijuana use. Health impact of marijuana has been examined in anesthetic and psychiatric disciplines but has not been investigated in association with surgical outcomes. This investigation examines outcomes for insurance-covered body contouring in the setting of Perioperative marijuana use.

Methods: A retrospective chart review was performed for patients who received an insurance-covered panniculectomy from 2018-2022. Data collection included demographics, comorbidities, perioperative data, outcomes, and chart documentation of regular marijuana use. Patients were considered to have Perioperative marijuana use if regular inhalation use was clinically documented within a year of surgery. Exclusion criteria included tobacco use, diabetes, edible THC consumption, or lack of insurance approval for panniculectomy. Statistical analysis included descriptive statistics, t-tests, Chi-squared analysis and Odds Ratios (OR) with 95% Confidence Intervals (CI).

Results: 192 patients were identified who underwent insurance-based panniculectomy from 2018-2022. 3 had never been screened for marijuana use (1.6%). 25 (13.0%) had regular marijuana use clinically documented within a year of surgery. Patients with documented regular marijuana use were 5.07 times as likely (p-value<0.001, OR 95% CI 2.04-12.58) to develop a complication compared to patients without documented marijuana use. Patients with documented regular marijuana use were 5.02 times as likely (p-value=0.010, OR 95% CI 1.31-19.24) to require a return to the OR to address a complication, and 4.27 times (p-value=0.019, OR 95% CI 1.153-15.83) as likely to develop a complex wound persisting past 30 days when compared to patients without clinically documented marijuana use. Finally, the marijuana use cohort was found to be 3.17 times (p-value=0.038, OR 95% CI 1.02-9.928) as likely to have an infection requiring treatment with antibiotics compared to the non-marijuana use cohort. There was no difference in average age, BMI, or resection weight between the two cohorts. However, there was also no difference between the marijuana use and non-marijuana use cohorts for wound dehiscence or seromas/hematomas requiring drainage.

Conclusions: Patients with regular marijuana use documented within a year of insurance-based panniculectomy had significantly increased risk of developing a complication, requiring a return to the OR, developing a complex wound, as well as increased likelihood of developing an infection that necessitated treatment with antibiotics. With increasingly widespread use in the United States, patients with regular marijuana use will present for plastic surgery at a higher rate. Counseling about inhaled marijuana use in the Perioperative period may mitigate complications. Continued research is needed to further elucidate the relationship between perioperative marijuana use and surgical outcomes.

Background: Perioperative bleeding poses significant challenges to rhinoplasty/septoplasty procedures; tranexamic acid (TXA), an antifibrinolytic, may help reduce this bleeding, though its effectiveness in these specific surgeries is not well-established. This systematic review and meta-analysis aims to evaluate the impact of TXA on perioperative bleeding in rhinoplasty and septoplasty. This review addresses the methodological insufficiencies of similar previous reviews which were assessed using the AMSTAR 2.0 tool

Methods: A comprehensive search was made across multiple databases, including PubMed, Embase, Google Scholar, and Web of Science, up to October 2023. Eligible studies were randomized or non-randomized comparative studies focusing on adult patients undergoing rhinoplasty or septoplasty. The primary outcomes assessed were intraoperative blood loss, surgery duration, and surgeon satisfaction, analyzed using RevMan 5.4.1.

Results: The search yielded 154 initial results; 12 studies were deemed eligible, covering 1296 participants, 54.7% of whom received TXA. A quality and bias assessment was undertaken using the GRADE, Cochrane RoB 2.0, and ROBINS-I tools which let to the exclusion of a further 2 studies due to poor methodological quality / high risk of bias. The meta-analysis showed a significant reduction in intraoperative blood loss with TXA (Mean difference -55.50; 95% CI: -87.84 to -23.26, P < 0.0007), but no significant changes in surgery duration or surgeon satisfaction, albeit with high heterogeneity across studies (I^2 > 90%).

Discussion: The findings support TXA's efficacy in reducing blood loss during rhinoplasty and septoplasty without affecting surgery duration or surgeon satisfaction. The high heterogeneity and concentration of studies in specific regions suggest the need for standardized research methodologies. Future studies should optimize TXA's dosing and administration while exploring its long-term effects and patient-centered outcomes. This review confirms TXA's potential as a valuable surgical adjunct while underscoring the need for further standardized research, which is necessary to elucidate TXA's benefits and inform clinical practice fully.

Introduction: Excess skin following massive weight loss poses both a cosmetic and functional challenge for patients. As the demand for post-bariatric brachioplasty procedures rises, we must understand the incidences of wound complications and best practices for treatment options. This study presents data from a single institution's experience with brachioplasty body contouring surgery and its associated wound complications.

Methods: A single-center review of cases was performed for patients undergoing brachioplasty over a 15-year period from July 2008 to October 2023. Patient information was acquired through Electronic Medical Records or paper source documents. Any paper records with missing or unreliable data points were excluded from the final analyses. All patients were analyzed for preoperative clinical data and timelines of post-operative wound complications. Descriptive statistics were conducted, and Fisher Exact tests were used for further statistical analyses; a P value of 0.05 or less was considered statistically significant.

Results: In this study, 205 patients undergoing brachioplasty procedures were analyzed. Of these patients, 10 (5%) patients were male, and 195 (95%) patients were female. The average age of brachioplasty patients was 47, with an average BMI of 30.34 kg/m2 after weight loss and an average maximum BMI of 52.19 kg/m2 prior to massive weight loss. Of the three most common wound complications, 57 (27.9%) patients had ≥1 event of minor wound dehiscence, 33 (16.2%) patients had ≥1 event of seroma, and 20 (9.8%) had ≥1 event of suture extrusion. These complications were, in all cases, treated with office-based wound management. Interestingly, there was a positive association between minor wound dehiscence and receiving liposuction within the zone of resection at the same time as brachioplasty, and suture extrusion incidence and liposuction (p=0.01 and p=0.01, respectively). Additionally, we found that there was no significant association between concurrent liposuction of the posterior arm, away from the site of resection, and wound complications, which mirrors the conclusion of our earlier published work. The average total follow-up time for our cohort was 369.1 days (about 1 year).

Conclusion: Minor wound dehiscence and small seromas are commonly observed after brachioplasty, but amenable to office-based care. Our analysis shows concurrent liposuction within the zone of tissue resection may be associated with a higher rate of wound complications.

Background: A commonly stated goal of aesthetic rhinoplasty is to improve nasal aesthetics while preserving the patient's individual features for a personalized result. To the contrary, we hypothesized that primary rhinoplasty has a homogenizing effect on a patient's appearance and, furthermore, in a distinct manner for a given rhinoplasty surgeon.

Purpose: This study aims to evaluate and measure the homogenizing effects of rhinoplasty.

Methods: Standardized and unedited preoperative and postoperative photos, including frontal, lateral, and oblique views, of 100 female-presenting, Caucasian-presenting primary rhinoplasty from the public websites of two independent experienced rhinoplasty surgeons from separate practices were included in this study. Using the Amazon Rekognition API and all 100 patients' photos, similarity between photos were calculated by comparing: all the preoperative photos to each other, all postoperative photos to each other, each patient to their own postoperative photo. These scores were averaged to provide a similarity score for each comparison type. Next, two cohort subgroups were created, with Surgeon A's 50 patients comprising one and Surgeon B's the other. The comparisons were evaluated, and similarity scores were calculated within and between cohorts. The change in similarity after surgery was calculated as the difference of similarity scores preoperatively and postoperatively. Paired t-tests were used to evaluate the changes in similarities between the before photos of patients and the after photos of patients. A simple t-test was used to evaluate the absolute value of change of the patients' preoperative state with others' postoperative states and between cohorts.

Results: As hypothesized, patients were significantly more similar after surgery (1.02%, p < 0.001) with the greatest effect in the lateral (1.28%, p < 0.001), oblique (0.91%, p < 0.001), and then the frontal (0.64%, p < 0.001) views. In the cohort subgroup analysis, Surgeon A had a 1.13% (front: 0.78%, lateral: 1.66%, oblique: 1.05%), whereas Surgeon B had a 0.89% increase (front: 0.46%, lateral: 0.90%, oblique: 0.78%), indicating a greater homogenizing effect by Surgeon A than B. Moreover, the inter-cohort similarity increase of 0.61% was significantly less than the intra-cohort similarity increase (1.02%, p = 0.003), indicating that the cohorts were more similar to each other in appearance than the other cohort on average. This was supported on the frontal (0.64%, p < 0.001), lateral (0.49%, 1.28%, p < 0.001), and oblique (0.49%, 0.91%, p = 0.002) views.

Conclusion: This study demonstrates that primary rhinoplasty patients have an increased similarity of appearance, or homogeneity, postoperatively. Rhinoplasties by both surgeons had this effect, greater for Surgeon A than Surgeon B, and patients within a cohort were more similar to each other than to the other cohort. We believe that this is the first study that measures and quantifies the homogenizing effect of rhinoplasties in general through the use of neural networks, and more surprisingly, the quantifiable "signature" effect of a specific surgeon on the appearance of their patients as a population. This study may impact how surgeons set patients' expectations, affect how patients select surgeons, and ultimately increase patient satisfaction.

Background: Liposuction is a common surgical procedure used to remove excess adipose tissue in targeted regions of the body. According to the 2022 ASPS annual report, liposuction is the most performed cosmetic procedure. Despite its increasing popularity, current literature lacks a unified understanding of the risks associated with this aesthetic procedure. This study represents the first meta-analysis of its kind to explore the risks associated with liposuction and aims to improve current methods by identifying and analyzing complications associated with the procedure.

Methods: A systematic review and meta-analysis was reported according to PRISMA guidelines and registered on the PROSPERO database (CRD42023471626). The primary outcome of the study was overall complication rate. Absolute risk for individual complications was also assessed. Ultrasound-assisted, power-assisted, laser-assisted, and radio-assisted liposuction were included in this review, while high-definition liposuction was excluded. Methodological Index for Non-Randomized Studies (MINORS) was utilized for study quality assessment.

Results: From 2,957 articles, 39 studies were selected for analysis. In total, 29,368 patients were included, with a mean age of 40.62 years and mean BMI of 26.36 kg/m2. Overall, the rate of any complication was 2.62 [95% CI: 1.78; 3.84]. The most common complication was contour deformity, with a prevalence of 2.35% [95% CI: 1.05%; 5.16%]. The prevalence of hyperpigmentation was 1.49% [95% CI: 1.12%; 1.99%], seroma 0.65% [95% CI: 0.33%; 1.24%], hematoma 0.27% [95% CI: 0.12%; 0.60%], superficial burn 0.25% [95% CI: 0.17%; 0.36%], infection 0.20% [95% CI: 0.082%; 0.50%], allergic reaction 0.16% [95% CI: 0.050%; 0.52%], skin necrosis 0.046% [95% CI: 0.013%; 0.16%], generalized edema 0.041% [95% CI: 0.0051%; 0.32%], deep vein thrombosis and thromboembolism 0.017% [95% CI: 0.0055%; 0.053%], and local anesthesia toxicity 0.016% [95% CI: 0.0040%; 0.064%].

Conclusion: Liposuction is a safe procedure with an overall low complication rate. When they do occur, contour deformity and hyperpigmentation are among the most prevalent complications. Enhancing awareness of the specific risks associated with individual complications can contribute to improved physician training and potentially reduce the incidence of costs for patients. Furthermore, these findings help define expectations for liposuction procedures, improving surgical outcomes and patient-physician alliance.

INTRODUCTION: Accurate risk assessment is paramount in plastic surgery practice, particularly for the abdominoplasty, which is known for increased risks of complications like surgical site infection (SSI). Therefore, abdominoplasty risk calculators can provide insights into patient outcomes and complications. Vu et al. (1) developed an internally validated 30-day risk calculator for abdominoplasty, incorporating factors like age, body mass index, and comorbidities. With the rise of artificial intelligence like OpenAI's ChatGPT (2) in medicine, one may question how these models can be compared to and integrated with existing risk calculators such as that of Vu et al. This study aims to evaluate how accurately ChatGPT-3.5 and ChatGPT-4 can predict complication risks for abdominoplasty when compared to a validated calculator.

METHODS: Twenty-five abdominoplasty patient scenarios were presented to both ChatGPT-3.5 and ChatGPT-4, prompting the models to compute the 30-day risk of medical complications, superficial SSIs, deep or organ-space SSIs, dehiscence, and unplanned operations. Each scenario was provided twice to ensure consistency. The Abdominoplasty Risk Calculator developed by Vu et al. served as the benchmark for calculating gold standard values, against which the predictions of ChatGPT were measured. The deviations between the predictions of ChatGPT and the gold standard values were analyzed using descriptive statistics. Differences between models were assessed using a paired, two-tailed t-test.

RESULTS: Overall, ChatGPT-3.5 overestimated the risk of abdominoplasty complications by 2.2%, and ChatGPT-4 underestimated by 1.6% (p-value <0.001). When considering the deviation from the actual risks, ChatGPT-3.5 had an average deviation of 4.7% (SD 3.4), whereas ChatGPT-4 had an average deviation of 3.5% (SD 4.5). Both models exhibited their weakest performance in estimating medical complication risks, with ChatGPT-3.5 and ChatGPT-4 averaging deviations of 6.3% and 6.1%, respectively. Notably, ChatGPT-4 performed best in predicting the risk of dehiscence, with an average deviation of only 1.4%, while ChatGPT-3.5 excelled in predicting deep or organ-space SSI, with an average deviation of 3.5%. ChatGPT-3.5 performed better than ChatGPT-4 only in predicting unplanned reoperation (4.2% vs. 4.5% deviation, p-value 0.015).

CONCLUSION: Accurate risk assessment tools are important for preoperative planning, especially for the abdominoplasty with its heightened risk profile compared to other cosmetic surgeries. The integration of LLMs, like ChatGPT, presents an intriguing avenue for potentially enhancing risk prediction accuracy. The evaluation of ChatGPT-3.5 and ChatGPT-4 in this study revealed both models, on average, provided a risk calculation within 5% of the gold standard. However, even minor deviations can have notable implications for patient safety, emphasizing the importance of precise risk estimation tools in medical practice and further refinement of LLMs.

REFERENCES: 1. Vu MM, Ellis MF, Blough JT, Gutowski KA, Kim JYS. Development and Internal Validation of the Abdominoplasty Risk Calculator. Plast Reconstr Surg. Jan 2018;141(1):34e-45e. doi:10.1097/prs.0000000000003922 2. OpenAI. ChatGPT. https://chat.openai.com/chat

Purpose: Rates of lumbar spine surgery are increasing due to advancements in surgical techniques and an aging population [1]. Due to the potential for large incisions in complex patients, plastic surgeons are often consulted to close lumbar spinal surgeries. The purpose of the present study is to examine if plastic surgeon closures of lumbar spine surgeries as a standard practice result in superior outcomes compared to closures performed by the primary surgeon as documented in the National Surgical Quality Improvement Program (NSQIP) Database.

Methods: Patients undergoing lumbar spine surgery at a large academic surgical center who were closed by the senior author (PJT) between January 2016 and June 2023 were included. Patient charts were reviewed for demographics, preoperative risk factors, and operative course including total operation time, incision length, estimated blood loss, and length of stay. The following wound-related complications were collected: superficial surgical site infection (SSI), deep SSI, organ/space SSI, wound dehiscence, seroma, hematoma, and CSF leak. The following medical complications were also collected: acute renal failure, transfusion, pneumonia, unplanned reintubation, failure to wean off ventilator, pulmonary embolism, urinary tract infection, stroke, cardiac arrest, myocardial infarction, deep venous thrombosis (DVT), sepsis, septic shock, and C. difficile infection. Rates of unplanned readmission and revision surgeries related to the primary procedure were noted. All included outcomes fell within a 30-day postoperative timeframe. Univariate and multivariate analyses were performed (statistical significance set at p<0.05), and outcomes were compared to those reported in 47,426 NSQIP cases after matching by CPT and ICD-10 codes.

Results: A total of 888 consecutive lumbar surgery patients whose closure was performed by a plastic surgeon were reviewed. Patients in the plastics closure group had lower rates of hypertension (52.7% vs 61.8%, p<0.001) but higher rates of diabetes (19.8% vs 6.7%, p<0.001) than those undergoing lumbar spine surgeries nationally. However, both wound-related revisional surgeries (1.2% vs 2.6%, p=0.01) and wound-related readmissions (1.5% vs 4.2%, p<0.001) were significantly lower for the plastics closure group than the NSQIP group. Rates of transfusion (2.7% vs 4.7%, p=0.005) and urinary tract infection (0.2% vs 1.3%, p=0.004) were also lower for the plastics closure group. Within the plastics closure group, ASA was the only operative characteristic that was significantly higher in patients with wound-related complications than those without (2.8 vs 2.5, p<0.001). There were no demographic or operative characteristics that were predictive of wound-related complications following plastic surgery closure.

Conclusion: Plastic surgeon closure of lumbar spine surgeries may decrease the risk of wound-related revisional surgeries and readmissions, even within a population predisposed to higher wound-related complication rates [2]. Therefore, the findings of this study provide support for increased plastic surgery consultation in lumbar spine cases.

References:

1. Martin BI, Mirza SK, Spina N, Spiker WR, Lawrence B, Brodke DS. Trends in Lumbar Fusion Procedure Rates and Associated Hospital Costs for Degenerative Spinal Diseases in the United States, 2004 to 2015. Spine (Phila Pa 1976). 2019;44(5):369-376. doi:10.1097/BRS.0000000000002822
2. Bajaj A, Shah RM, Brecount H, Kurapaty S, Patel AA, Divi SN. The Effect of Diabetes and Metabolic Syndrome on Spine Surgery Outcomes. Curr Rev Musculoskelet Med. 2023;16(2):39-47. doi:10.1007/s12178-022-09814-y