Scientific Abstract Presentations: Research & Technology Session 6 Sunday, October 12 Sun, Oct 12 11:00 a.m.-12:00 p.m. CDT

PURPOSE: Vascularized composite allografts (VCAs) face substantial challenges, particularly in immunosuppression and ex vivo preservation. Subnormothermic machine perfusion (SNMP) offers a promising alternative to conventional static cold storage for preserving and reconditioning VCAs. This study evaluates our extended SNMP protocol, optimized for non-human primate (NHP) VCAs, and its impact on graft immunogenicity, paving the way for future ex vivo preconditioning strategies.

METHODS: Partial facial grafts were procured from non-human primates (NHPs) that had undergone terminal organ donor procedures, with body weights ranging from 4 to 8 kg (n=6). The grafts were perfused for 18 hours using a non-recirculating Steen+ solution under a low-flow, high-oncotic protocol previously optimized for VCAs. Parameters including pressure, flow, weight gain, and metabolic markers (lactate, pH, O₂, ions) were monitored every hour for the first 6 hours and every 3 hours for the remaining 12 hours of perfusion. Pre- and post-perfusion tissue samples were collected to analyze the skin immune compartment (FACS), gene expression profiling, and tissue histology. Outflow perfusate samples were taken at every time point for cytokine and Lactate Dehydrogenase analysis (Multiplex).

RESULTS: The grafts had a mean warm ischemia time of 3 hours before perfusion initiation. Following 18 hours of continuous SNMP, mean weight gain remained below 10%. Perfusion parameters and histological analysis indicated stable ex vivo metabolic equilibrium and well-preserved tissues, consistent with findings from previous models. Leukocyte viability was maintained in the skin. The leukocyte count in the skin remained stable throughout the experiment with no significant perfusion washout effect observed, while the number of leukocytes in the outflow decreased at H18, though not reaching statistical significance. Endothelial cells constituted less than 20% of the cellular population in the perfusate throughout the experiment. The cellular stress response to perfusion was characterized, with IL-6 showing the most significant increase during perfusion.

CONCLUSIONS: We present the first application of continuous SNMP in the experimentally relevant NHP partial face transplant model, advancing the understanding of the model's response to extended ex vivo preservation and laying the foundation for future ex vivo preconditioning strategies in conjunction with tolerance induction protocols.

Purpose: Breast reconstruction after mastectomy entails a complex discussion of multiple options and a nuanced balance of risks, benefits, and alternatives. Social media, however, remains a common source of sharing and receiving information, medical or otherwise, for many patients and family members. This study aims to evaluate the TikTok breast reconstruction landscape and analyze the representations of implant versus free flap reconstructions.

Methods: A review of 500 TikTok posts discovered using #DIEP, #diepflap, #flap, #freeflap, #breastreconstruction, #breastimplant, #siliconeimplant, #salineimplant, on TikTok was performed on June 28th, 2024. Engagement (likes, comments, bookmarks, shares), post categories (reconstructive option, content, audience) and creator characteristics (followers, relation, main content) between platforms were collected. An ANOVA analyzed engagement metrics and followers while a chi-square test analyzed the post and creator characteristics between types of reconstructions.

Results: 263 mentioned free flaps (52.6%), 174 mentioned implants (34.8%), and 63 (12.6%) mentioned both. However, free flap reconstructions had substantially less engagement, with significantly fewer average views (free flap: 54333.2, implant: 230287.7, both: 25185.5, p < 0.001), likes (free flap: 54333.2, implant: 230287.7, both: 25185.5, p < 0.001), and bookmarks (flap: 83.3, implant: 266.2, both: 76.6, p = 0.025). However, comments (free flap: 177.0, implant: 111, both: 50.9) and shares (free flap: 160.0, implant: 110.8, both: 65.9) were not significantly different. Although free flap and implant posts did not have a significantly different audience, posts about both targeted patients (82.5%, p = 0.002). Free flap posts were more likely to be update videos (40.7%, p < 0.001) while implant (27.6%) and posts about both (38.1%) were more likely to be educational. Implant-based post creators had significantly more average followers (441069.6, free flap: 33718.5, both: 25185.4, p < 0.001). Free flap post creators were also more likely to be patients (73.4%, p < 0.001) compared to implant creators (59.8%). Posts about both reconstruction options were more likely to be a medical provider (57.1%). Main content categories were significantly different (p < 0.001) with most free flap creators making daily life (33.1%) and cancer journey (25.9%) videos compared to the mainly medical educational posts of others (implants: 31.0%, both: 46.0%).

Conclusion: Although free flap reconstructions were mentioned in the majority of breast reconstruction posts (65.2%), those posts had significantly less "influence" based on audience engagement and creator impact. Implant videos were often cross-tagged with aesthetic videos, potentially increasing their visibility within TikTok's algorithm and elevate post interaction. This study highlights the positive efforts of patients and providers to increase awareness of free flap breast reconstruction, while also emphasizing the need for greater engagement and higher quality education for both patients and the general public.

Introduction

Integrated plastic surgery is amongst one of the most competitive residencies to match. Previous studies have identified predictors of matching such as research productivity, Step 1 scores, and ranking in the standardized letter.1 The importance of narrative letters of recommendation (LOR) is well known; however, previous models to predict match have not incorporated data from them. The purpose of this study is to develop a model that incorporates applicant data and the content and authorship of LOR to accurately predict matching into integrated plastic surgery.

Methods

A retrospective review was performed on all integrated plastic surgery applicants in the 2023-2024 application cycle. Applicant demographics, scores, research productivity, narrative LOR, and standardized letters with scores were obtained. Letter author characteristics were also obtained from each letter of recommendation. Applicants with key demographic variables missing were removed from the final analysis.

Linguistic analysis was performed on all narrative LOR with Linguistic Inquiry and Word Count (LIWC-22) using a customized data dictionary. To explore the relationship between the key variables and matching, LASSO (Least Absolute Shrinkage and Selection Operator) was used to build the multivariable logistic model. Receiver Operating Curve was used to evaluate the predictive performance of the model with and without narrative letter analysis and letter author data using the area under the curve (AUC).

Results

A total of 189 applicants were included. The model revealed variables most predictive of matching were higher Step 2 scores (OR=1.06, p=0.003), greater number of podium presentations (OR=1.09, p=0.01). Furthermore, greater mentions of words describing an applicant as likable/communal within the narrative LOR was predictive of match. Applicants who are not MD seniors (OR=0.30, p=0.07) or who had an overall rank >5 in the standardized letter may have worse odds of matching (OR=0.23, p=0.04).

The AUC for the model without narrative letter analysis and author data was 0.8081. However, when letter content and author data was added to the final model, the AUC improved to 0.8875 demonstrating a successful prediction of match 88.75% of the time.

Discussion

This is the first study to utilize applicant letters of recommendation to predict matching into integrated plastic surgery. Step 2 scores, number of podium presentations, and likability were predictive of successful match, while lower overall rank and DO or IMG applicant status were predictive of not matching. Furthermore, addition of the content and authorship data improved the accuracy of the model demonstrating the importance of these variables in predicting match.

References

Thomas HC, Chen S, Narisetti L, Janis JE, Perdikis G, Drolet BC. The Utility of the Plastic Surgery Standardized Letter of Recommendation Form in Predicting Residency Match Outcomes. J Surg Educ. 2023 Jul;80(7):948-956. doi: 10.1016/j.jsurg.2023.04.012. Epub 2023 May 5. PMID: 37150702.

Abstract Purpose Accurate wound classification is essential for effective wound management, yet well-trained specialists are often unavailable in primary healthcare settings perfectly. This study presents a deep learning-based wound classification tool capable of identifying five key wound conditions-deep wounds, infected wounds, arterial wounds, venous wounds, and pressure wounds-using color images captured by readily available cameras. The model's performance was evaluated against medical personnel to assess its clinical applicability. Methods A convolutional neural network (CNN)-based multi-task deep learning model was developed to classify wounds based on images obtained from 1,429 patients (900 men, 529 women; mean age: 62.6 ± 18.5 years) at a single institution. A total of 2,149 wound images were reviewed by a wound specialist and annotated for five binary classification tasks. The dataset was divided into training (80%) and testing (20%) cohorts, ensuring patient-level partitioning. Model performance was assessed using accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC). A comparison with seven medical personnel (two attending surgeons, two residents, one nurse practitioner, and two registered nurses) was conducted using Cohen's kappa coefficient. Results The model demonstrated high classification performance across all five wound types, with AUC values of 0.804 for deep wounds, 0.751 for infected wounds, 0.897 for arterial wounds, 0.924 for venous wounds, and 0.940 for pressure wounds. When compared to human medical personnel, the model exhibited significantly superior or non-inferior performance across all classification tasks. Sensitivity analysis revealed that the model performed comparably to experienced wound specialists but outperformed junior personnel in key classification tasks. The model struggled with venous wound classification due to class imbalance, which was reflected in a lower sensitivity (63.2%) despite maintaining high specificity (97.5%). Receiver operating characteristic (ROC) analysis further confirmed that the model's performance was either comparable to or exceeded that of human raters, particularly in arterial and pressure wound classification. Conclusion The proposed deep learning model provides a reliable and efficient method for wound classification, with performance that is either comparable to or exceeds that of human medical personnel. Its compact architecture allows for potential deployment in telemedicine applications, facilitating automated wound assessment in primary care settings where wound specialists are not readily available. By integrating deep learning into wound care, this model could improve diagnostic accuracy, optimize resource allocation, and enhance patient outcomes. Further research is needed to improve model generalizability by incorporating additional imaging datasets and integrating non-visual patient data such as clinical history and laboratory findings.

References (1) Harding K, et al. Adv Wound Care. 2013;2(5):149-150. (2) Sen CK. Plast Reconstr Surg. 2009;124(4 Suppl):10S-17S. (3) Frykberg RG, Banks J. Adv Wound Care. 2015;4(9):560-582. (4) Chang M, et al. Int Wound J. 2021;18(3):276-286. (5) Yadav G, et al. IEEE Access. 2020;8:147479-147492.

Introduction Ultraviolet B (UVB) radiation is a significant environmental stressor that causes various skin disorders, including DNA damage, inflammation, hyperpigmentation, and epidermal structural changes. Chronic UVB exposure leads to cumulative skin damage, contributing to photoaging and carcinogenesis, while acute high-dose exposure results in immediate epidermal injury and inflammation. Although melanin production is protective, it is often dysregulated, causing hyperpigmentation disorders therefore identifying effective therapeutic agents to mitigate UVB-induced skin damage is critical. Metformin, an antidiabetic drug with antioxidant, anti-inflammatory, and DNA repair properties. This study investigates metformin's protective effects on human skin perfusion model exposed to acute and chronic UVB radiation.

Materials and Methods Full-thickness human skin flaps, obtained as surgical waste, were maintained on a perfusion system and exposed to UVB radiation. The study was comprised of three study groups, Control (no exposure), chronic UVB (received 200 mJ/cm² UVB daily for seven days), and acute UVB (received a single dose of 1400 mJ/cm²). Each group was subdivided into metformin-treated and without metformin groups, with a non-UVB-exposed as control. Topical metformin (0.5%) was applied immediately after each UVB exposure in the UVB exposed groups. Skin biopsies were collected at days 0, 1, 3, 5, and 8. Histological analysis was performed using Hematoxylin and Eosin (H&E) staining for epidermal and dermal integrity, Fontana-Masson staining for melanin deposition, and gamma-H2AX (γH2AX) staining to detect DNA double-strand breaks as a marker of DNA damage.

Results Both acute and chronic UVB exposure induced visible epidermal thinning, separation of the dermo-epidermal junction, and increased melanin production. Fontana-Masson staining confirmed an increase in melanin deposition in both UVB treated groups, with the chronic exposure showing cumulative pigmentation over time. However, both chronic and acute UVB treated groups with metformin-treatment showed reduced melanin levels in comparison. H&E staining showed widespread epidermal separation and structural disorganization in only UVB exposed groups, particularly at later time points (days 5-8). In contrast, metformin-treated UVB exposed groups exhibited improved epidermal integrity and reduced separation, suggesting enhanced tissue resilience. γH2AX staining revealed elevated DNA damage in the acute UVB group from 24 hr post-exposure and beyond, while chronic UVB exposure showed progressive DNA damage. Metformin-treated UVB exposed group showed fewer γH2AX positive cells, indicating protection against UVB-induced DNA strand breaks. Control biopsies remained intact with minimal pigmentation and no/baseline DNA damage.

Conclusions Metformin demonstrated a protective role in mitigating UVB-induced skin damage, hyperpigmentation, and genetic injury in an ex vivo skin perfusion model. Metformin offers a promising therapeutic strategy for preventing UV-induced skin disorders, and this model serves as a valuable platform for further research on UVB-induced damage and protective agents. Further investigation is underway to confirm the protective effects and will be repeated to check the reproducibility of the data.

Background: For many plastic surgery residents, early microsurgery training involves anastomoses of artificial vessels pinned to a flat platform. This practice, while valuable, provides limited application to microsurgery within a depressed cavity, as is the case for internal mammary (IM) artery and vein anastomosis during microsurgical breast reconstruction, commonly performed with a deep inferior epigastric perforator (DIEP) flap. With the severity of possible complications arising from access of the IM vessels for microsurgical anastomosis - including flap failure, pneumothorax, and hemorrhage - it is critical to maximize resident pre-operative preparedness. Given the rising access to 3D printers in the academic setting, our team developed an ultra-low-cost microsurgical platform, using a 3D-printed base to replicate the sternum and ribs, with a 3D-printed retractor and customizable layers of commonly-available dressing supplies to mimic the mastectomy skin flaps and typical retractor placement.

Methods and Materials: DICOM data from a de-identified non-contrast CT Chest was used to generate a 3D model of the sternum and ribs, then printed in poly-lactic acid (PLA) with a 3D-printer (Ender 3). A modified Richardson retractor was designed and 3D-printed in similar fashion. Foam tape and 4x4 gauze were used to imitate soft tissue to mimic the mastectomy flaps, with foam tape and a 500 mL bag of normal saline used to mimic the bulk of the DIEP flap. Practice vessels were generated from the cuffs of sterile gloves which were prepared into tubular segments (1). The components were then arranged to mimic IM vessel access as encountered in a DIEP flap. Residents from PGY1 to PGY6 practiced microsurgical anastomoses on the model and then completed a survey evaluating the model and their microsurgical competencies using a Likert scale (with 5 being highest competency and strongest agreement).

Results: The designed components were printed in PLA for $1.00 per rib model and $0.45 per Richardson ($0.015/g filament). Plastic surgery residents across several years of training utilized the model effectively, with the majority of participants noting increased competency in vessel access, anastomosis technique, hand-eye coordination, and overall confidence after model use.

Conclusions: Our ultra-low-cost model promotes anticipation of and preparation for the challenges of internal mammary vessel microsurgery, enabling training within the anatomic confines of the third rib access site. Our model can be replicated worldwide by anyone with access to a 3D-printer and common medical supplies.

1 - Almarghoub MA. A Simple and Cost-effective Method for Practicing Microsurgery. Plast Reconstr Surg Glob Open. 2019 Mar 14;7(3):e2146. doi: 10.1097/GOX.0000000000002146. PMID: 31044117; PMCID: PMC6467636.

Background: Standard two-dimensional (2D) computed tomography angiography (CTA) is the current gold standard for preoperative visualization of the abdominal vasculature prior to deep inferior epigastric artery perforator (DIEP) flap autologous breast reconstruction. Virtual reality (VR) technologies offer an immersive, three-dimensional (3D) view of abdominal CTA scans that may enhance the visuospatial understanding of the DIEP flap vessels critical to this procedure. In this proof-of-concept study, the authors explore the educational efficacy of 3D VR compared to 2D CTA visualization in the context of DIEP flap vessel mapping. Methods: A prospective, proof-of-concept case study using a female cadaveric model was conducted at a single academic institution. Post-mortem CTA images were obtained and then reconstructed and exported to a VR headset. The abdominal vasculature was then visualized using both VR and CTA modalities by two advanced plastic surgery trainees. Following visualization with the different modalities, participants drew 3D maps of the abdominal vasculature in a virtual environment. Participants also completed a comprehensive questionnaire evaluating their anatomical understanding of the DIEP vessels and individual experience using the technology. Participant drawings were evaluated in multiple domains by a blinded board-certified plastic surgeon using a five point Likert scale. Results: Plastic surgery trainees identified more perforators when using 3D VR compared to 2D CTA. Plastic surgery trainee drawings were also rated to have a greater degree of anatomical completeness, detail, and clarity compared to 2D CTA. Furthermore, participants survey results showed trainees felt that 3D VR outperformed 2D CTA in providing a better understanding of DIEP vessel anatomy and complex vascular relationships. Trainees also rated 3D VR higher in terms of ease of identifying perforators, and felt that 3D VR was more intuitive, efficient, and user-friendly compared to 2D CTA. Overall, it was found that trainees preferred 3D VR over 2D CTA and performed equally or better when using 3D VR, highlighting its potential as an alternative modality for preoperative DIEP flap planning. Conclusion: This preliminary analysis shows that VR visualization may prove useful in DIEP flap vessel mapping and plastic surgery trainee education. Future directions include obtaining additional data from plastic surgery trainees and performing cadaveric dissections to determine if enhanced anatomical understanding from preoperative 3D VR visualization translates to enhanced surgical precision and skill.

Background: Scleroderma, characterized by excessive collagen deposition and dysregulated immune responses, leads to debilitating skin fibrosis. Adipose-derived stem cells (ADSCs) exhibit antifibrotic and immunomodulatory properties, but their dynamic interactions within fibrotic microenvironments remain underexplored. This study investigates the temporal effects of ADSCs on skin fibrosis and elucidates their molecular mechanisms.

Methods: A bleomycin (BLM)-induced skin fibrosis model was established in C57BL/6J mice. GFP-labeled ADSCs were subcutaneously injected into fibrotic lesions, with follow-up analyses at days 1, 7, 14, and 21 post-injection. Histopathological evaluation (H&E, Masson's trichrome, α-SMA staining), qRT-PCR, flow cytometry, and RNA sequencing were performed to assess fibrosis progression, inflammatory cytokine levels, and ADSC behavior. In vitro co-culture experiments with human scleroderma-derived fibroblasts validated key pathways.

Results: ADSCs localized preferentially to hair follicles and perivascular regions, with survival declining by day 21. BLM-induced dermal thickening, collagen disorganization, and α-SMA expression were significantly reduced by ADSC treatment (p<0.05). Proinflammatory cytokines (TNF-α, IL-6, IL-1β) surged post-BLM induction but decreased sharply after ADSC administration. RNA sequencing revealed time-dependent transcriptomic shifts: immunomodulatory pathways (e.g., IL-6 production, VEGF signaling) dominated at day 7, while antifibrotic pathways (collagen matrix regulation, TGF-β/Smad inhibition) emerged by day 14. TSG-6, a critical antifibrotic mediator, was upregulated in ADSCs within fibrotic environments. Silencing TSG-6 in ADSCs abolished their antifibrotic effects, restoring COL1 and α-SMA expression in fibroblasts (p<0.05). In vivo, TSG-6-deficient ADSCs failed to suppress TGF-β/Smad2/3 activation, confirming its pivotal role.

Conclusion: ADSCs exert dual-phase therapeutic effects: early immunomodulation (days 1–7) followed by sustained antifibrotic action (days 14–21), mediated partly via TSG-6-dependent TGF-β/Smad pathway inhibition. These findings highlight ADSCs as a promising cell-based therapy for fibrotic skin disorders, offering insights into optimizing timing and mechanisms for clinical translation.

Purpose: Pain sketches can qualitatively depict neuropathic pain and predict surgical outcomes after targeted muscle reinnervation (TMR) in lower extremity (LE) amputees (1-5). This study aims to develop a machine learning (ML) tool to automate the analysis of these drawings.

Methods: Pain sketches were obtained from adult transtibial or transfemoral amputees who underwent unilateral primary or secondary TMR at a tertiary care center (2021-2024). Sketches were resized, grayscaled, and mirrored. Clinical experts labeled the images based on pain categories: focal (FP), radiating (RP), diffuse (DP), or no pain (NP). A convolutional neural network (CNN) was trained to classify sketches into the categories. Data was split into training, validation, and test sets (70%, 15%, 15%). Binary classification models were built and evaluated using precision, recall, F1-score, and area under the receiver operator characteristic (AUROC ).

Results: We included 588 sketches from 68 primary and 138 secondary LE TMR patients. The NP vs. DP classification model achieved the highest AUROC (0.799), while other models ranged from 0.587 to 0.760. Unsupervised clustering revealed distinct pain patterns not captured by human categorization.

Conclusions: Computer vision and ML techniques can be used to successfully classify pain sketches in LE amputees, automating the identification of common pain distributions. This tool has potential clinical utility in assisting surgeons with preoperative planning for neuroma-related surgical interventions, particularly for patients undergoing TMR.

1. Hwang CD, Hoftiezer YAJ, Raasveld FV et al. Biology and Pathophysiology of Symptomatic Neuromas. Pain. Published online 2023. doi:PAIN-D-22-01167R1
2. Eberlin KR, Ducic I. Surgical algorithm for neuroma management: A changing treatment paradigm. Plast Reconstr Surg Glob Open. 2018;6(10):1-8. doi:10.1097/GOX.0000000000001952
3. Gfrerer L, Hansdorfer MA, Ortiz R, et al. Patient Pain Sketches Can Predict Surgical Outcomes in Trigger-Site Deactivation Surgery for Headaches. Plast Reconstr Surg. 2020;146(4):863-871. doi:10.1097/PRS.0000000000007162
4. Gomez-Eslava B, Raasveld F V, Hoftiezer YA, et al. Pain Sketches Demonstrate Patterns of Pain Distribution and Pain Progression following Primary Targeted Muscle Reinnervation in Amputees. Plast Reconstr Surg. Published online May 26, 2023:online ahead of print.
5. Raasveld F V., Hao D, Gomez-Eslava B, Hwang C, Valerio IL, Eberlin KR. Predictive Value of Pre-operative Pain Sketches in Lower Extremity Amputees Undergoing Secondary Targeted Muscle Reinnervation for Treatment of Neuropathic Pain. J Am Coll Surg. under peer review.

BACKGROUND The American College of Surgeons' National Surgical Quality Improvement Program (NSQIP) database tracks 30-day postoperative outcomes across >150 perioperative variables. To analyze NSQIP data, researchers write code in languages like R or Python. However, most clinical journals do not require the publication of analytical code, making reproduction of NSQIP-based research challenging. Without the original code, researchers must infer methodological details that impact study findings. As experimental protocols are necessary for reproducing wet lab experiments, specific code is essential for replicating computational analyses. For example, in a study on 30-day readmission rates, researchers use NSQIP's readmission classifications of "Yes" (readmitted), "No" (not readmitted), or "NA" (missing data). In 2021, 95% of cases were labeled "NA", forcing researchers to exclude these cases or assume patients were not readmitted. Such choices go unreported yet influence study outcomes. This study assesses reproducibility in NSQIP-based research by replicating three plastic surgery studies, evaluating the impact of undisclosed methodological decisions on patient cohort selection and statistical outcomes.

METHODS A PubMed search identified studies using NSQIP. Articles using pediatric, geriatric, procedure-targeted, or additional datasets were excluded as they require additional access. Three plastic surgery articles were selected for reproduction using NSQIP data from 2005-2023. Each study's methodology was reviewed, and an independent analysis was performed following the described methods. Reproduced statistical results were compared with published findings.

RESULTS The literature search identified 2,699 studies using adult NSQIP data, including 252 plastic surgery articles. None of the studies published their code. Three articles were selected from the Journal of the American College of Surgeons, Aesthetic Surgery Journal, and Microsurgery examining immediate breast reconstruction (IBR), outpatient cosmetic outcomes, and free flap outcomes by frailty index, respectively. None of the patient cohorts or statistical analyses were fully replicated. In reproducing the IBR study, 14,260 cases were identified compared to 18,194 cases published. Both the original and reproduced analyses showed significant associations between race, cardiovascular comorbidities, and complication rates. However, the reproduced analysis also identified a significant association between pulmonary comorbidities and complication rates not reported in the original study. Similarly, reproduction of the cosmetic study identified 2,586 outpatient cosmetic cases compared to 2,879 published. The original analysis found operative time as a significant predictor of readmission, whereas the reproduced logistic regression instead identified obesity. For the free flap study, demographic trends across frailty scores were similar, but statistical significance varied. Both original and reproduced analyses found length of stay, mortality, and infection rates differed by frailty, but unplanned readmission was only significant in the original study.

CONCLUSIONS NSQIP is a widely used database in plastic surgery; however, our attempt to replicate three studies revealed that without access to the original methodology or code, patient cohorts and statistical findings could not be fully reproduced. As the original studies were well-designed and peer-reviewed, these discrepancies do not invalidate their findings, but instead highlight how unreported methodological choices influence results and hinder reproducibility. Journals should encourage clear methodological reporting, including code publication, to enhance verification and trust in NSQIP-based research.

Purpose Breast cancer related lymphedema (BCRL) is a chronic disease with a significant impact on patients lives. The disease pathology is driven by Th2 mediated inflammation, resulting in fibrosis, chronic inflammation, and skin changes. Previous therapies targeting the vascular endothelial growth factor (VEGF) pathway, particularly VEGF-C (the primary ligand for VEGFR3), have promoted lymphangiogenesis but also increased inflammation, vessel leakage, and tumor spread. Therefore, our lab aimed to investigate a VEGFC independent method to modulate lymphatics and used an intracellular modulation of receptor tyrosine kinase signaling to bypass the effect of VEGF-C and assess the effect on secondary lymphedema pathology. We used lymphatic gain and loss of function models, by employing PTEN-deficient and PI3K-deficient mice to test this hypothesis. Methods We utilized two mouse models, LEC specific PTEN knockout (LECPTEN) and LEC specific PI3K knockout (LECPI3K), with wild type (WT) mice as controls. We performed tail lymphatic excision to induce lymphedema and euthanized the mice at the 6-week time-point, after which we performed flow cytometry, H&E and immunofluorescence staining. We also performed a functional study using TritC conjugated to high molecular weight dextran, ensuring the molecule is only taken up by lymphatic vessels. For statistical analysis, we used the Kruskal Wallis test to identify differences between our three groups. Results Following tail lymphatic excision, LECPTEN mice demonstrated significantly increased lymphangiogenesis evidenced by an increase in both lymphatic vessel density distal to the wound, and bridging vessel formation across the wound, as compared to LECPI3K and WT mice (p<0.0001). Functional analysis showed that LECPTEN mice had increased amounts of TritC uptake, as compared to LECPI3K and WT mice, showing that the observed lymphangiogenesis resulted in vessels that were effective (p=0.002 overall). Epidermal thickness differed across the groups, with LECPTEN showing the thinnest epidermal thickness, and LECPI3K mice the thickest (LECPTEN vs. LECPI3K mice, P=0.0053). Interestingly, tail lymphedema and fibrosis following surgery did not differ between groups (overall P=0.2423, n=6/group), despite the above-mentioned differences. In addition, we did not find significant differences in tissue inflammation using flow cytometry to analyze T cells, macrophages, and neutrophils.
Conclusion This study suggests that increased lymphangiogenesis alone is not a viable method of decreasing the pathologic changes of lymphedema thus highlighting the importance of inflammation in this disease. These findings may provide a rationale for the inconclusive findings of prior clinical trials using lymphangiogenic approaches. Thus, modulation of inflammatory responses (with or without changes in lymphangiogenesis) may be a more fruitful approach.