Scientific Abstract Presentations: Residents Craniomaxillofacial and Research & Technology Abstracts Session 4 Friday, October 16 Fri, Oct 16 10:30 a.m.-12:30 p.m. CDT

Abstract

Background: The optimal timing for surgical intervention in reconstructive procedures remains a subject of significant debate. This study evaluated whether the timing of intervention for Le Fort fractures -early versus delayed-affects functional and quality-of-life outcomes across multiple clinical domains. Traditionally, proponents of early intervention (within 24 hours) argue that immediate stabilization prevents the early onset of scar tissue and fracture malunion, thereby simplifying the reduction process and potentially reducing the risk of surgical site infections [1]. Furthermore, early repair has been associated with shorter intensive care unit stays and reduced total hospital costs [2]. Conversely, many clinicians advocate for a delayed approach, typically between five and seven days, to allow for the resolution of significant facial edema and the stabilization of concomitant medical conditions [3]. This delay is thought to facilitate more accurate soft tissue closure and provide a window for more meticulous preoperative planning using three-dimensional computed tomography.

Methods: A comparative analysis was performed on 112 data points categorized into five domains: occlusion, vision, breathing, pain, and quality of life (QOL). Subjects were divided into early intervention and delayed intervention cohorts. Early intervention was defined as primary surgical repair occurring within 24 hours of the injury while delayed intervention was defined as primary repair occurring within 5-7 days after injury. Continuous variables were compared using independent sample t-tests, with Levene's test utilized to assess equality of variance. Statistical significance was set at p<0.05.

Results: No statistically significant differences were observed between the early and delayed groups across any of the assessed domains. Mean scores for occlusion (Early: 1.33 vs. Delayed: 1.31, p=0.953), breathing (Early: 1.11 vs. Delayed: 1.50, p = 0.237), pain (Early: 2.00 vs. Delayed: 2.17, p = 0.766), and QOL (Early: 1.56 vs. Delayed: 1.42, p = 0.728) demonstrated clinical equivalence. While vision scores showed the greatest numerical variance (Early: 2.22 vs. Delayed: 1.25), this also failed to reach significance (p = 0.125).

Conclusions: Our findings suggest that delaying surgical intervention by 5-7 days does not result in inferior functional or psychosocial outcomes. In the absence of acute surgical emergencies, a delayed approach appears to be a safe and viable management strategy, allowing for the resolution of perioperative factors without compromising the final reconstructive result.

References 1. Hurrell MJ, Batstone MD. The effect of treatment timing on the management of facial fractures: a systematic review. Int J Oral Maxillofac Surg. 2014 Aug;43(8):944-50.

2. Barker DA, et al. Timing for repair of mandible fractures. Laryngoscope. 2011 Jun;121(6):1160-3.

3. Janus SC, et al. Analysis of results in early versus late midface fracture repair. Otolaryngol Head Neck Surg. 2008 Apr;138(4):464-7.

Background/Purpose: Early discrimination between superficial partial-thickness (IIA) and deep partial-thickness (IIB) burns remains clinically challenging. This study evaluated the diagnostic performance, calibration, and clinical utility of BURNS-PBIA, a smartphone-based artificial intelligence system designed to support early burn-depth prediction and TBSA estimation. Methods: A retrospective diagnostic accuracy study was performed using 320 standardized smartphone images from 80 adult patients treated between January 2024 to January 2025. Images were acquired within 24 hours using a standardized protocol. Burn depth was determined by consensus of three certified burn surgeons (κ = 0.86). A Vision Transformer model (ViT-B/16) was fine-tuned using a patient-level 70/15/15 split. Performance was assessed using AUC, sensitivity, specificity, calibration metrics, TBSA mean absolute error, and decision-curve analysis. Results: The study included 80 patients and 320 images. The model achieved an AUC of 0.93 (95% CI, 0.88–0.98), sensitivity of 90%, specificity of 85%, and accuracy of 87% for distinguishing IIA from IIB burns. Calibration was strong, with a Brier score of 0.07. TBSA estimation showed a mean absolute error of 0.54% compared with Lund–Browder assessment. The system achieved an accuracy of 0.88 for grafting requirement. Decision-curve analysis demonstrated net clinical benefit across relevant threshold probabilities. Conclusions: BURNS-PBIA demonstrated high diagnostic accuracy, reliable calibration, and clinically meaningful utility for early burn-depth prediction using standardized smartphone photography. Its low-cost, workflow-compatible design may support objective risk stratification where specialized imaging is unavailable. External multicenter validation is warranted before broader clinical implementation.

Background: The secondary (mature) cleft nasal deformity poses a significant challenge for surgeons. Secondary cleft rhinoplasty is often performed with the goal of restoring nasal cosmesis and function. However, persistent deforming forces including cicatricial forces and cartilage memory can predispose patients to recurrent asymmetry, incomplete correction and relapse. We aimed to describe a comprehensive structural approach to secondary cleft rhinoplasty that prioritizes integrated stabilization of the central and lateral nasal framework and robust fixation of cartilage constructs. Key elements include a caudal septal extension fixated to the maxilla in the true facial midline, circumferential graft support of the cleft ala, dome and soft triangle to achieve symmetric stabilization, and an autologous rib cartilage preparation technique designed to minimize long-term warping. We present a quantitative assessment of resultant nasal symmetry and tip projection. Methods: Thirty five patients with unilateral cleft lip and palate who underwent secondary cleft rhinoplasty using a caudal septal extension graft and circumferential support of the cleft sided ala will be included. Standardized photographs taken pre-operatively and an average of 20.54 months (range 11-60 months) post-operatively were used to compare tip projection and ratios of cleft side to non-cleft side nostril height, width, and area. Wilcoxon signed-rank test evaluated statistically significant differences between the measurements. Intra-rater and inter-rater reliability were measured using two-way mixed-effects and random-effects intraclass correlation coefficient (ICC) models for absolute agreement, respectively. Results: To date results from 13 of 35 patients with a mean age of 23.31 ± 3.47 years have been analyzed. Six patients (46.15%) were male and 7 patients (53.85%) were female. 8 patients (61.54%) had left-sided cleft lip and palate. Intra-rater and inter-rater reliability were good to excellent across variables with intra-rater ICCs ranging from 0.86 to 0.99 (mean ICC 0.95 (rater 1) and 0.97 (rater 2)) and inter-rater ICCs ranging from 0.81 to 0.99 (mean ICC 0.92). From the preoperative to postoperative period, tip projection, represented as a ratio of subnasale-pronasale over inter-alar width significantly improved by 14.2% (p<0.01). Nostril width symmetry (ratio of cleft side width over non-cleft side width) significantly improved by 30.1% (p=0.03). Nostril area symmetry (ratio of cleft side nostril area over non-cleft side nostril area) demonstrated a non-significant trend towards improvement by 29.1% (p=0.09). Nostril height symmetry (ratio of cleft side height over non-cleft side height) showed no significant change (p=0.38). 4 patients experienced complications requiring revision surgery, including nasal synechiae (n=1), migration of caudal septum (n=1) and migration/loss of septal extension graft (n=2). Conclusions: Using a boney anchored caudal septal extension graft harvested from the 6th 7th intercostal bridge and circumferential support of the cleft ala appears a promising approach to stable secondary cleft rhinoplasty that can improve nostril width symmetry and tip projection with possible optimization of nostril area symmetry. Updated results will be presented.

Abstract

Background: Robin Sequence (RS) is characterized by micrognathia, glossoptosis, and airway obstruction, often necessitating airway interventions in infancy, including mandibular distraction osteogenesis (MDO)1-2. The mechanism by which MDO improves airway patency is directly related to the forward advancement of the mandible, which in turn pulls the tongue base anteriorly due to its muscular attachments along the floor of the mouth. This forward movement increases the pharyngeal airway space (PAS) and affects critical anatomical structures such as the hyoid bone3-5. The effects of MDO on hyoid position in RS is poorly characterized. This study aims characterize the linear and rotational changes of the hyoid bone after MDO.

Methods: A 10-year retrospective review of patients less than one year of age with RS who underwent MDO. Thirty patients met inclusion criteria with standardized pre/post-distraction imaging and were compared with age/sex matched controls. Cephalometric landmarks (retrognathion (Rgn), hyoid (Hy), C3 vertebra (C3), sella (S)) were used to generate linear and angular measurements (Figure 1). Pre/post-distraction anesthesia Mallampati grades were reviewed. Paired t-tests compared pre/post-distraction changes and outcomes relative to matched controls.

Results: The mean age at distraction was 74.2 days, with an average advancement of 18.6 mm. Post-distraction, the Rgn–Hy distance increased significantly (1.73 cm vs 2.04 cm, p=0.01), as did the Hy–C3 distance (2.36 cm vs 2.93 cm, p<0.001). Inferior displacement was demonstrated by an increase in S–Hy distance (5.00 cm vs 6.31 cm, p<0.001), which exceeded control values (p<0.001). Angular measurements showed an increase in the Rgn–Hy–S angle (94.6° vs 106.0°, p=0.02) and a decrease in the S–Hy–C3 angle (83.2° vs 66.6°, p<0.001), consistent with anterior–inferior hyoid linear and rotational repositioning. Composite ratios corroborated these shifts, demonstrating significant post-distraction realignment compared to controls. Mallampati scored decreased from a mean of 2.9 pre-distraction to 1.3 post-distraction.

Conclusions: This study provides compelling evidence that mandibular distraction osteogenesis in infants with Robin Sequence results in significant anterior, inferior displacement of the hyoid bone, accompanied by angular rotation of the suprahyoid complex. Significant changes in post-distraction relative realignment are observed when comparing pre-and post-distraction and to matched controls. These changes likely reflect the anatomic basis for improved airway patency observed clinically following mandibular distraction osteogenesis.

1. Abramson, Z.R., et. al. (2013). Effects of Mandibular Distraction Osteogenesis on Three-Dimensional Airway Anatomy in Children With Congenital Micrognathia. J Oral Maxillofac Surg, 71,90-97.

2. Rachmiel, A., et. al. (2005). Bilateral mandibular distraction for patients with compromised airway analyzed by three-dimensional CT. Int. J. Oral Maxillofac. Surg., 34,9–18.

3. Rachmiel, A., et. al. (2012). Management of obstructive sleep apnea in pediatric craniofacial anomalies. Ann Maxillofac Surg, 2,111-115.

4. Mohamed, A.M., et. al. (2011). Distraction osteogenesis as followed by CT scan in Pierre Robin sequence. Journal of Cranio-Maxillo-Facial Surgery, 39,412–419.

5. Ramieri, V., et. al. (2017). Three-dimensional airways volumetric analysis before and after fast and early mandibular osteodistraction. Journal of Cranio-Maxillo-Facial Surgery, 45,377–380.

Background: The multi-staged nasolabial interpolated flap is often preferred over the single-staged nasolabial transposition flap for reconstructing large nasal alar defects due to its robust vascularity and nasal contour preservation. We challenge this paradigm by comparing outcomes of both flaps, hypothesizing that the simpler nasolabial transposition flap achieves comparable results. Methods: A retrospective cohort study of patients undergoing nasal alar reconstruction with either an interpolated nasolabial or nasolabial transposition flap was performed. Demographics, comorbidities, surgical data, complications, revision procedures, and number of follow-up visits were obtained from the medical record. Pre- and post-operative photos were analyzed in Adobe Photoshop. Nasal axis deviation (NAD), alar-facial angle (AFA), and nostril show right/left (NS-R/L) were measured. The change in pre- and post-operative measurements was calculated and comparative analyses were conducted using RStudio. Results: The cohort had a mean age of 70.8 years and was 81% female and 19% male. A total of 12 interpolated and 19 transposition flaps were performed. Pre-operative to post-operative mean change in NAD, AFA, and NS-R was comparable between groups, while NS-L showed a greater mean change in the interpolated group (p = 0.03). Interpolated flaps showed higher rates of nasal valve incompetence (16.7% vs. 0%, p = 0.142) and trapdoor deformity (50% vs. 15.8%, p = 0.056). There was one case of flap loss (interpolated), one donor site wound dehiscence (transposition), and one infection treated with oral antibiotics (transposition). The interpolated group required more follow-up visits (median 7.0 [IQR 1.25] vs. 4.0 [IQR 0.50]) and revision surgeries (2 vs. 0). Conclusion: The interpolated nasolabial and nasolabial transposition flap techniques achieve comparable aesthetic and functional outcomes for reconstruction of nasal alar defects, with interpolated flaps showing higher complication rates and greater follow-up burden. These results suggest that the single-stage nasolabial transposition flap may be superior, offering equivalent outcomes with fewer office visits, procedures, and complications.

Background: Laser imaging detection and ranging (LiDAR) is a three-dimensional (3D) technology that measures the round trip of an infrared laser beam to accurately detect the presence and features of objects. Notably, iPhone with built-in LiDAR sensors have existed since 2020. Our team developed a software application for 3D head scanning and made therapeutic helmet using 3D printer based on LiDAR scan. Therefore, this study evaluated the therapeutic effect of this new helmet. Methods: Plagiocephaly and brachycephaly patients were scanned with iPhone application. Based on this 3D data, therapeutic helmet was made using 3D printer. Longer diagonal distance, shorter diagonal distance, mediolateral dimension, anteroposterior dimension, CVAI (cranial vault asymmetry index) and CI (cephalic index) were measured. We compared CVAI and CI before and after helmet therapy. Results: A total 66 patients of plagiocephaly were participated on this study. CVAI was improved from 7.47 ± 4.54 to 2.05 ± 6.31 with significant difference. Mean treatment period was 165 ± 45.50 days. A total 60 patients of brachycephaly were participated on this study. CI was improved from 96.94 ± 3.32 to 90.82 ± 6.93. Mean treatment period was 181.47 ± 60.88 days. Conclusion: The proposed software application installed on iPhone with a built-in LiDAR sensor provides convenient 3D scanning. This new 3D scan provides good accuracy, uses smart mobile devices that people already own, and is more affordable than contemporary tapeline measurement and 3D scanning systems. Additionally, therapeutic helmet made by 3D printer based on this 3D scan for plagiocephaly and brachycephaly showed reliable treatment result.

Background: Postoperative oral feeding practices following head and neck mucosal free-flap reconstruction vary widely, largely due to concerns regarding wound integrity and fistula formation. Although delayed oral intake of 6-12 days has traditionally been favoured, emerging evidence spearheaded by Enhanced Recovery After Surgery programmes suggests that earlier feeding may optimise physiological function, reduce stress response and facilitate recovery and healing. Nevertheless, the ideal timing for oral nutrition is uncertain and its effects on post-operative complications and hospital length-of-stay (LOS) remains contentious.

Methods: A systematic review and meta-analysis was performed following PRISMA guidelines. Medline, EMBASE, Cochrane Central, and Scopus were searched for studies comparing early (≤5 postoperative days) and delayed (>5 days) oral feeding after head and neck mucosal free-flap reconstruction. Random- and fixed-effects models were used where appropriate. Primary outcomes included fistula formation, with secondary outcomes comprising pneumonia, flap failure, wound complications, and length of hospital stay (LOS).

Results: Thirteen studies met inclusion criteria, representing 1,657 patients. Early oral feeding was associated with a significantly lower incidence of fistula formation compared with delayed feeding (3.66% vs 11.35%; RR 0.37, 95% CI 0.22-0.64; p = 0.0004). Rates of postoperative pneumonia were also reduced in the early feeding group (6.31% vs 12.38%; RR 0.53, 95% CI 0.33-0.87; p = 0.011). Mean LOS was significantly shorter among patients who commenced oral intake early (9.85 vs 13.11 days; MD −4.10 days, 95% CI −7.07 to −1.14; p = 0.0067). No statistically significant differences were observed between groups for flap failure, haematoma, or wound dehiscence. Available data did not demonstrate an increased risk of fistula formation associated with early feeding in patients who had received preoperative radiotherapy.

Conclusion: Early initiation of oral feeding following head and neck mucosal free-flap reconstruction is not associated with increased reconstructive complications and is linked to lower rates of fistula formation, reduced pneumonia, and shorter hospital stays. These findings support consideration of earlier oral intake in appropriately selected patients and highlight the need for evidence-based postoperative feeding pathways.

Purpose: To evaluate long-term surgical outcomes and patient-reported satisfaction following excision of head and neck dermoid cysts at a tertiary care center.

Methods: We conducted a retrospective cohort study of patients who underwent surgical excision of head and neck dermoid cysts between 1992 and 2024 at a single tertiary institution. Demographic, clinical, operative, and postoperative data were collected from the medical record. To assess long-term outcomes, a standardized survey was distributed to patients or caregivers, capturing recurrence, complications, aesthetic outcomes, and overall satisfaction. Patients were stratified by lesion location to explore potential differences in outcomes. Descriptive statistics were used to analyze both clinical and survey data.

Results: Of 267 eligible patients, 58 completed the survey (response rate 21.7%), with a median follow-up of 7.4 years. Recurrence was rare, occurring in 3.4% of patients, and none required reoperation. Postoperative complications were similarly uncommon (3.4%) and were minor in nature. Structural abnormalities, including bony contour irregularities, were infrequently reported and did not necessitate additional intervention. Patient-reported outcomes were highly favorable, with all respondents indicating satisfaction with both functional and aesthetic results. No significant differences in outcomes were observed based on lesion location.

Conclusions: Surgical excision of head and neck dermoid cysts is associated with low rates of recurrence and complications, along with excellent long-term patient satisfaction. These findings support early surgical management as a safe, effective, and durable treatment strategy with meaningful long-term benefits.

Objective: Robin Sequence is defined by micrognathia, glossoptosis, and airway obstruction. In infancy, patients with severe airway obstruction can be managed with mandibular distraction. However, long-term outcomes are not well known. A main concern with significant overcorrection is that patients may require mandibular setback later in life. The purpose of this study was to describe the long-term cephalometric features and incidence of revision orthognathic surgery in patients with Robin sequence treated with mandibular distraction early in life.

Methods: A retrospective chart review was performed using the University of Michigan EMERSE database. Patients diagnosed with Robin Sequence and managed with mandibular distraction in infancy or early childhood were identified. To assess long-term outcomes, patients younger than age 10 years at most recent follow up were excluded. Age, sex, timing of distraction, amount of distraction (mm), need for revision orthognathic surgery and most recent follow up date were recorded. Profile facial photographs were imported and traced in Dolphin Imaging software version 12.0.64 (Patterson Dental, Chatsworth, CA). Cephalometric analysis was performed on soft tissue measures and averages were calculated.

Results: Six patients met final inclusion criteria. One third of patients were male. Average total distraction measured 30.3+/-6.65 mm. Length of follow-up averaged 14.7+/-3 years. Cephalometric analysis of soft tissue measures identified facial angle convexity 17+/-4.7 degrees (Norm 12+/-4 degrees), nasolabial angle 113.3+/-10.5 degrees(Norm 102+/-8 degrees), mentolabial angle 130.2+/-12.65 degrees (Norm 120+/-1 degrees), and cervicofacial angle 124.7+/-10.5 degrees (Norm 121+/-1 degrees). At most recent follow up, all patients were tolerating oral feeds and none had obstructive sleep apnea. No patients required orthognathic surgery for mandibular setback. One patient underwent bilateral sagittal split osteotomy for additional mandibular advancement, and two patients were considering advancement.

Conclusion: Patients with Robin Sequence managed with mandibular distraction can expect reasonable long-term appearance and function of their jaw. In our experience, significant overcorrection was rewarded with excellent results without risking the need for mandibular setback. In fact, families should be counseled that despite significant overcorrection patients often experience relapse or sustained growth impairment and might require additional advancement in young adulthood.

Purpose: Persistent calvarial defects after craniosynostosis surgery remain a challenge, with wide variation in reossification rates and no clear guidelines on when to intervene. Given the potential for cosmetic, developmental, and neurological sequelae, timely and appropriate management is critical. The objective of this study is to review existing literature and our institutional data to identify factors that guide the decision between cranioplasty and observation in pediatric patients with craniosynostosis, with the goal of developing a clinical algorithm to support surgical decision-making.

Methods: A comprehensive literature review was performed to identify studies evaluating outcomes of persistent calvarial defects after craniosynostosis surgery, including risk factors and decision-making regarding observation versus cranioplasty. A retrospective review of our institutional data was also performed on pediatric patients with craniosynostosis who underwent endoscopic, open strip craniectomy, or open cranial vault remodeling at a tertiary care center.

Results: The primary predictive factor in the decision between cranioplasty versus observation was defect size. Defects greater than five square centimeters were unlikely to reossify, favoring cranioplasty. Symptom development including headaches, developmental delay, or visible cosmetic deformities encouraged consideration of a CT scan and if a defect was confirmed, cranioplasty was preferred. Additional factors favoring surgical repair included defect location over critical structures and patient age greater than three to five years old. Routine surveillance through physical examination proved sufficient for monitoring, while CT imaging should be reserved for concerning findings and preoperative planning.

Our institutional data demonstrates that incomplete ossification after endoscopic, open strip craniectomy, and open cranial vault remodeling can be reliably identified during routine postoperative physical exam, and in most cases, managed conservatively without the need for intervention. This correlates with what was found in our literature review.

Conclusions: Based on our data and review of the literature, a decision-making algorithm was developed from the following criteria: physical exam findings, patient age, defect characteristics, and symptoms. This algorithm uses a patient-centric model to manage persistent calvarial defects without the need for routine CT monitoring. Our institutional experience supports a conservative management strategy in most cases, with CT imaging and cranioplasty reserved for clinically indicated scenarios. Applying this decision-making framework will improve postoperative management of patients with craniosynostosis by minimizing radiation exposure and over-treatment, while ensuring those who would benefit from surgical intervention receive the appropriate care in an adequate time frame.

Introduction: Leiomyosarcoma (LMS) of the head and neck is a rare malignant tumor derived from smooth muscle cells. LMS of the scalp is an especially uncommon site of origin that is locally aggressive with potential for cranial invasion, particularly in cases of delayed diagnosis due to misidentification as a benign lesion. Reports describing LMS of the scalp are scant and only identified in adults. In contrast, LMS in the pediatric population is extremely rare and disproportionately associated with immunocompromised individuals. To our knowledge, we present the only case of cutaneous/subcutaneous LMS of the scalp in an immunocompetent pediatric patient highlighting the unique oncologic and reconstructive challenges.

Case Presentation: Our patient is a 12-year-old female who presented from an outside facility for biopsy-proven occipital scalp leiomyosarcoma. Wide local excision was planned with pediatric general surgery and resulted in a 10 × 9 cm scalp defect. Over 18 months, she underwent staged scalp reconstruction. At time of wide local excision, plastic surgery performed burring of the outer table of the calvaria with placement of a dermal regeneration template. Once margins returned negative, placement of a split-thickness skin graft (11 × 7 cm) was performed. A single, rectangular tissue expander was placed eight months later and progressively expanded to a total volume of 475cc. Her final stage included tissue expander removal and rotation advancement flap. She experienced no complications, and achieved her desired hair length by three months postoperatively.

Discussion: LMS is broadly categorized into superficial, somatic, and vascular subtypes, with somatic lesions being the most common. Superficial LMS accounts for 2-3% of soft tissue sarcomas and is categorized into dermal and subcutaneous variants.(1) Prognosis in superficial LMS is largely determined by depth of invasion, with dermal tumors demonstrating excellent outcomes (5-year disease-specific survival ~96.8%, 10-year metastasis ~2.7%) compared to subcutaneous lesions (~62.9%, 25%).(2) After multidisciplinary evaluation, PET/CT was performed to rule out metastasis and patient was planned for excision of margins. Both imaging and margins were negative deeming the patient appropriate to forego radiation. By delaying definitive reconstruction with temporizing STSG and confirming negative margins our patient was able to undergo successful scalp reconstruction.

Conclusion: LMS of the scalp in the pediatric population is exceedingly rare and presents a unique reconstructive challenge requiring progression up the reconstructive ladder. Goals for reconstruction include restoration of a hair-bearing scalp, attention to psychosocial outcomes, and aggressive surgical resection to achieve negative margins and minimize the need for adjuvant radiation. Clinicians should maintain a broad differential for slow-growing scalp masses, as early diagnosis may reduce reconstructive complexity and improve overall outcomes.

References: 1. Soares Queirós C, Filipe P, Soares de Almeida L. Cutaneous leiomyosarcoma: a 20-year retrospective study and review of the literature. An Bras Dermatol. 2021;96(3):278-283. doi:10.1016/j.abd.2020.10.003 2. Abebe, K., Munch, A.V., Wagenblast, A.L., Schmidt, G., Jensen, D.H., Petersen, M.M., Loya, A., Daugaard, S., Mentzel, T., Herly, M., Vester-Glowinski, P., & Ørholt, M. (2025). Classification of High- and Low-Risk Groups in Patients with Dermal Leiomyosarcoma: An Exploratory Register-Based Nationwide Cohort Study. EMJ Dermatology.

Purpose Self-directed learning has become increasingly important in surgical education. Large language models using artificial intelligence (AI) may support surgical education; however, their accuracy and reliability remains unclear. This study evaluated the accuracy of procedural descriptions generated by ChatGPT's newest model (GPT-5.2) across a range of peripheral nerve procedures for surgical residents.

Methods Ten peripheral nerve procedures were randomly selected from the Washington University in St. Louis Surgical Education Website as reference standards. Videos were transcribed and coded into discrete procedural steps. Prompt-engineered GPT-5.2 outputs were evaluated against the reference standards for procedural accuracy and qualitative content gaps with regression analysis by procedural complexity. Accuracy was calculated as the proportion of GPT-5.2 content that matches the reference standard. Each procedure was queried six times to evaluate reliability.

Results The mean steps per procedure was 10.8 (SD 3.4). Mean accuracy ranged from 69.4% (SD 12.3%) to 91.0% (SD 12.3%) across six independent runs per procedure. The overall mean accuracy was 81.9% (SD 12.9%). There was no association between procedural complexity and accuracy on regression analysis. Across all 60 runs, there were 122 missed steps, most involving judgement-based maneuvers likely to vary by surgeon preference (68/122), neurovascular structure identification (25/122) or major steps (24/122).

Conclusion GPT-5.2 generated largely accurate descriptions in peripheral nerve surgery for junior residents, but had limited reliability and inconsistently captured surgeon-dependent, judgement-based maneuvers. While useful for foundational procedural learning, AI outputs should be utilized as adjunctive rather than definitive educational resources.

Purpose Call schedule creation in surgical training programs is a time-intensive administrative task that often results in inequitable distribution of workload among trainees. Disparities in call burden can lead to trainee dissatisfaction, burnout, and perceptions of unfairness. Automated scheduling has been implemented in other specialties with improvements in equitable distribution and satisfaction (1,2); however, no such systems have been described for plastic surgery, where multi-service coverage (plastic, hand, and facial trauma) creates distinct complexity. We sought to implement an algorithmic scheduling solution to improve equity in call distribution while reducing administrative time burden.

Methods and Materials
This study was conducted in the Division of Plastic Surgery at Henry Ford Health, Detroit, MI. A weighted point system was developed to quantify call difficulty based on service coverage requirements. Base difficulty points were assigned to each service according to institutional consensus on relative workload: Plastic Only Call (1 point), Facial Trauma Call (2 points), and Hand Surgery Call (3 points), reflecting the recognized hierarchy of call intensity. Daily call difficulty was calculated by summing the base points for all services covered on a given shift, yielding the following weekday values:

Weekday Plastic Only: 1 point Weekday Plastic + Face: 3 points Weekday Plastic + Hand: 4 points Weekday Plastic + Face + Hand: 6 points

Weekend call was assigned to a single resident covering Friday through Sunday. Weekend difficulty was calculated by summing the daily service points across all three days plus a 10-point weekend premium to account for extended duration and disruption. At our institution, weekend call follows one of two coverage patterns:

Full Coverage Weekend: Friday (Plastic + Hand) + Saturday (Plastic + Face + Hand) + Sunday (Plastic + Hand) = 14 + 10 = 24 points Limited Coverage Weekend: Friday (Plastic + Face) + Saturday (Plastic Only) + Sunday (Plastic + Face) = 7 + 10 = 17 points

Holidays falling on weekdays were scored using the standard daily service calculation plus a 7.5-point holiday premium.

A mobile scheduling application utilizing a fairness-optimization algorithm (FairShift) was implemented to distribute call assignments among three plastic surgery residents. The algorithm iteratively assigns shifts to the resident with the lowest cumulative point total, ensuring balanced workload distribution. Seven months of pre-implementation schedules (July 2025 – January 2026) were retrospectively scored using the same weighted system and compared to schedules prospectively generated by the algorithm for 4 months (February – May 2026). Primary outcomes included standard deviation of total points per resident (equity measure), coefficient of variation, and time required for schedule creation.

Results Over 7 months of manual scheduling, mean monthly difficulty points per resident were 51.7 ± 9.7 with a coefficient of variation (CV) of 14.6%, indicating substantial inequity in call distribution. The highest-burden resident accumulated 32% more total points than the lowest (397 vs 301 cumulative points). Schedule creation averaged 4.7 hours per month. Following implementation of algorithm-assisted scheduling, mean monthly points per resident were 47.3 ± 3.1 (CV 3.0%) over 4 months. Point distribution variance decreased by 98.8%, maximum inter-resident disparity fell from 32% to 6%, and the average monthly standard deviation across residents decreased from 9.9 to 2.0. Schedule creation time was reduced to 15 minutes per month, representing a 94.7% reduction in administrative time.

Conclusions Implementation of an algorithmic scheduling tool significantly improved equity in call distribution among plastic surgery residents while reducing administrative time by over 90%. The coefficient of variation in call burden decreased from 14.6% to 3.0%, and inter-resident point variance was reduced by 98.8%. The weighted point system provided a transparent, objective framework for quantifying call burden that can be adapted to other surgical subspecialties. This study demonstrates that technology-assisted scheduling can enhance fairness, efficiency, and transparency in graduate medical education.

Reference: 1. Porche K, Mohan A, Dow J, Melnick K, Laurent D, Hoh B, Murad G. Automated and Optimized Neurosurgery Scheduling System Improves Resident Satisfaction. Neurosurgery. 2024 Jan 8. doi: 10.1227/neu.0000000000002821. 2. Howard FM, Gao CA, Sankey C. Implementation of an automated scheduling tool improves schedule quality and resident satisfaction. PLoS One. 2020 Aug 11;15(8):e0236952. doi: 10.1371/journal.pone.0236952.

Title Artificial Intelligence–Assisted Simplification of Plastic Surgery Patient Education Materials While Preserving Clinical Accuracy: A Pilot Study

Background Online patient education materials in plastic surgery are frequently written above recommended reading levels, limiting patient comprehension and contributing to disparities in health literacy. National guidelines from organizations such as the American Medical Association and National Institutes of Health recommend materials be written at approximately a sixth-grade level. This study evaluates whether artificial intelligence (AI) can improve readability of plastic surgery educational content across aesthetic, reconstructive, and hand procedures while maintaining clinical accuracy.

Methods Ten patient-facing webpages describing common plastic surgery procedures across aesthetic, reconstructive, and hand domains were selected from large academic medical center websites. Original text was extracted and Flesch-Kincaid Grade Level (FKGL) was calculated. Each passage was rewritten using AI targeting an approximately sixth-grade reading level while preserving key clinical information. FKGL was recalculated for AI-generated content. Two independent physician reviewers assessed accuracy and content preservation using a 5-point Likert scale (1 = inaccurate/unsafe, 5 = fully accurate and clinically appropriate). Pre- and post-rewrite readability scores were compared using the Wilcoxon signed-rank test.

Results Mean FKGL decreased significantly from 7.35 in original materials to 5.33 following AI-assisted rewriting, representing a mean reduction of 2.02 grade levels (p < 0.01). AI-generated materials approached recommended readability targets. Content accuracy was largely preserved, with a mean Likert score of 3.95 across reviewers. Seventy percent of ratings were ≥4, indicating acceptable to high clinical fidelity. Inter-rater agreement was moderate.

Conclusions AI-assisted rewriting significantly improves readability of plastic surgery patient education materials across aesthetic, reconstructive, and hand procedures, bringing content closer to recommended literacy levels while maintaining acceptable clinical accuracy. These findings suggest AI may serve as a scalable tool to enhance accessibility of patient-facing surgical information, though careful review remains necessary to ensure preservation of critical clinical details.

Introduction: There is increasing evidence supporting the clinical effectiveness of negative pressure wound therapy with instillation and dwell (NPWTi-d), but limited economic data regarding its impact on care costs (1, 2).

Objective: To examine differences in overall and wound-related inpatient costs for patients with acute traumatic wounds who received NPWTi-d versus costs for patients who received layered gauze dressing or traditional negative pressure wound therapy (NPWT).

Methods: Overall hospitalization costs and inputs to model wound-related costs were obtained from a recently published single-center randomized controlled trial of patients treated with NPWTi-d, NPWT, or gauze layer dressings for acute traumatic wounds and hospital and country-level cost data. The total per patient wound-related cost was calculated as the sum of the cost of mean hospital length of stay (LOS) for the duration of wound-related therapy, wound-related surgical procedures, and cost of wound therapy.

Results: Total overall cost of hospitalization was lower for patients who received NPWTi-d ($7,814) compared to patients who received NPWT ($7,890) or gauze layer dressing ($9,735). When analyzing modeled wound-related costs only, NPWTi-d also showed a cost advantage. Wound-related cost per patient for NPWTi-d was $3,390, representing an 18.3% reduction compared with NPWT ($4,149) and a 20.7% reduction compared with gauze dressing ($4,275). Differences were due to reductions in wound-related LOS and surgeries.

Conclusion: The results from this economic study demonstrate a potential cost savings with use of NPWTi-d compared to NPWT and gauze layer dressing in patients with acute traumatic wounds due to reductions in wound-related surgical procedures and wound-related LOS. These findings corroborate existing evidence suggesting potential cost savings of NPWTi-d in complex wound management (3, 4, 5).

References:

1. Gabriel A, et al. Cost-effectiveness of negative pressure wound therapy with instillation versus conventional negative pressure wound therapy in the treatment of complex acute wounds. Wound Repair Regen. 2018;26(4):313-321.

2. Kim PJ, et al. Economic impact of negative pressure wound therapy with instillation in the treatment of complex surgical wounds. Wound Repair Regen. 2019;27(2):161-168.

3. Leavitt T, et al. Cost-effectiveness of negative pressure wound therapy with instillation and dwell time in the management of complex surgical wounds. Plast Reconstr Surg. 2020;145(1):241-250.

4. Scales K, et al. The economic impact of negative pressure wound therapy with instillation and dwell time on acute traumatic wounds: a systematic review. Int Wound J. 2021;18(3):345-356.

5. Krug E, et al. The role of negative pressure wound therapy with instillation and dwell time in the management of acute traumatic wounds: a systematic review and meta-analysis. J Trauma Acute Care Surg. 2022;92(1):191-200.

Aim: To compare the effectiveness of dialkylcarbamoyl chloride (DACC) dressings versus polyamide with silver dressings in critically colonized or infected stage III–IV pressure ulcers requiring surgical management.

Method: This prospective randomized clinical trial included patients with stage III–IV pressure injuries indicated for surgical treatment. Participants were randomized to receive either a DACC-based dressing (n = 11) or a silver-impregnated dressing (n = 11) after initial debridement. On postoperative day 7, patients underwent repeat debridement, deep tissue culture collection, and assessment for flap closure. Clinical evolution, microbiology, laboratory parameters, and complications were monitored for 28 days (1,2).

Results: Mean age was 43.5 years (SD 14.2; 95% CI 38.7–49.0). Wounds were longstanding (mean 60.5 months; median 72). Most lesions were sacral (65.2%), and osteomyelitis was present in 60.9% of cases. Groups were comparable except for age, with the DACC group being older (p = 0.046). Initial cultures were positive in 65.2% of wounds, increasing to 76.2% on day 7, predominantly polymicrobial. No significant microbiological or laboratory differences were observed between groups. Minor dehiscence occurred in two different patients in the DACC group (18.2%) and in none of the silver-treated wounds. All were minor and managed conservatively. No major dehiscence, flap loss, or reoperation occurred.

Discussion: In this interim analysis, DACC and silver dressings showed equivalent early outcomes in critically colonized or infected stage III–IV pressure injuries undergoing surgical reconstruction. Fisher's exact test showed no significant differences, consistent with the small sample size. Prior studies similarly reported comparable outcomes between DACC and silver dressings (3,4). Both were safe and clinically effective. A larger sample is needed to clarify potential differences in complication rates and long-term efficacy (5).

Conclusion: This interim analysis suggests that DACC and silver dressings offer comparable early safety and effectiveness in the management of critically colonized or infected stage III–IV pressure ulcers requiring surgical intervention. Further research with a larger cohort is warranted to confirm these findings and explore subtle differences in outcomes.

References: 1. Schwarzer S, et al. Does the use of DACC-coated dressings improve clinical outcomes for hard-to-heal wounds? Int Wound J. 2024;21(5):e1449527. 2. Jeyaraman M, et al. Efficacy of Dialkylcarbamoylchloride (DACC)-Impregnated Dressings in Chronic Wounds: A Systematic Review. Cureus. 2025;17(2):e11843900. 3. Yi Q, et al. Impact of Silver Dressings on Wound Healing Rate in Patients with Chronic Wounds: A Meta-Analysis. J Clin Med. 2024;13(10):2876. 4. Idensohn PJ. Dialkylcarbamoyl Chloride-coated wound dressing: An Evidence Review and Position Document. Wounds International. 2025. 5. Gijón MM, et al. The efficacy of negative-pressure wound therapy (NPWT) in the prevention of surgical site occurrences in open abdominal surgery: A randomized clinical trial. Surgery. 2025;177(2):412-419.

Background: Free tissue transfer remains a cornerstone of reconstruction, with venous thrombosis as the leading cause of flap failure. The gold standard of flap monitoring relies on clinical examination, which is subjective, intermittent, and less reliable in patients with darker skin tones. Adjunct perfusion-based technologies provide additional information but remain limited by skin pigmentation and signal variability. Transcutaneous porphyrin spectrophotometry (TPS) is being developed as a continuous, objective monitoring platform to detect changes in heme-derived products during flap compromise, independent of melanin. This study describes iterative in vitro testing of the first two TPS prototype versions developed. Methods: Extinction coefficient ratios were calculated for hemoglobin and melanin using spectrophotometric measurements. Version 1 was tested using samples of known hemoglobin concentrations representing physiologic and pathologic tissue states, including normal skin, subdermal plexus, venous congestion, and hematoma level concentrations. Reflected light intensity was measured near the Soret peak (400 to 430 nm) to evaluate concentration dependent heme detection. Version 2 incorporated discrete LED wavelengths at 410nm, corresponding to the Soret band; 604nm (orange), selected as a candidate reference wavelength; and 660nm (red), selected to evaluate oximetry-based monitoring. Reflection values were recorded across known hemoglobin and melanin concentrations. Linear, logarithmic, exponential, power, and quadratic models were evaluated using R². Results: Extinction coefficient analysis demonstrated that hemoglobin-to-melanin absorption contrast was greatest near 405 nm (ratio of 142:1) supporting the TPS focus on the Soret band. Version 1 demonstrated successful concentration dependent detection of heme. Linear regression showed a negative relationship between reflected light intensity and hemoglobin concentration (R² = 0.63, p = 0.01). However, an exponential decay model provided the strongest fit (R² = 0.98) consistent with Beer-Lambert law. In Version 2 testing, all three wavelengths demonstrated significant relationships with hemoglobin concentration using their best fit models. The 410 nm and 604 nm channels showed the strongest fits using logarithmic models, with R² = 0.8653, p = 0.0219, and R² = 0.9535, p = 0.0043, respectively. The 660 nm channel demonstrated a nonmonotonic relationship with hemoglobin concentration, with the strongest fit using a quadratic model, R² = 0.9308, p = 0.0182. Version 2 also demonstrated melanin associated signal variation. The 410 nm and 604 nm channels demonstrated significant linear relationships with melanin concentration, with R² = 0.9181, p = 0.0418, and R² = 0.9274, p = 0.0370, respectively.
Conclusion: Iterative prototype testing supports the feasibility of TPS as a novel optical approach to free flap monitoring. Version 1 demonstrated proof-of-concept heme detection with an exponential relationship between reflected light intensity and hemoglobin concentration. Version 2 expanded testing across multiple wavelengths and demonstrated that raw reflected signal varies with both hemoglobin and melanin concentration. The 660 nm channel showed a nonmonotonic hemoglobin response and variable melanin-associated signal, highlighting limitations of red-light and oximetry-based monitoring. The 604 nm channel demonstrated significant relationships with both hemoglobin and melanin concentration, supporting continued evaluation as a reference or correction wavelength. Future work will focus on reference-wavelength algorithms and in vivo animal studies for reliable monitoring across diverse skin tones.

Titile: Evaluation of Artificial Intelligence in Translating Patient Aesthetic Demands into Rhinoplasty Outcomes: An Expert Panel Evaluation Abstract

Background: Artificial intelligence (AI) is increasingly being explored in aesthetic surgery for facial analysis, surgical planning, and outcome simulation. However, most existing approaches rely on image-based prediction rather than interpretation of patient-expressed aesthetic goals. This study evaluates whether a generative AI system can translate structured rhinoplasty patient demands into realistic postoperative simulations using fully synthetic data.

Methods: This observational cross-sectional study used AI-generated rhinoplasty cases representing diverse nasal deformities, facial types, and ethnic backgrounds. Preoperative and postoperative images were entirely generated using a generative AI model under a standardized imaging protocol. Structured patient aesthetic demands, formulated to reflect real clinical consultations, were applied to generate postoperative outcomes in a controlled synthetic workflow. An expert panel of consultant plastic surgeons evaluated AI performance using a structured questionnaire assessing accuracy of demand translation, aesthetic outcome, surgical plausibility, and preservation of facial identity.

Results: AI-generated simulations demonstrated consistent ability to reproduce major structural nasal modifications, particularly dorsal hump correction and overall facial harmony, while maintaining ethnic and gender-related facial characteristics. Performance was more variable in fine tip-definition adjustments. Evaluators generally rated outcomes as clinically plausible and suitable for educational and counseling purposes.

Conclusion: This study presents a novel framework for evaluating AI-driven rhinoplasty simulation based on structured patient demands rather than image-only prediction. The findings suggest that generative AI may serve as a supportive tool in preoperative counseling and surgical planning by translating patient aesthetic goals into visually coherent outcomes within a standardized and reproducible synthetic environment.

Mesenchymal stem cell–derived exosomes have emerged as promising cell-free therapeutics in regenerative medicine, particularly in wound healing.(1,2) However, the optimal dosing strategy and the extent of their paracrine effects remain poorly defined. This study aimed to evaluate the dose-dependent and systemic (paracrine) effects of human umbilical cord mesenchymal stem cell–derived exosomes (hUC-MSC-Exos) in an acute wound model.

  Twenty-four Wistar rats were randomly assigned to four groups (n=6): sham (saline), and three treatment groups receiving 1×10⁹, 2.5×10⁹, or 5×10⁹ exosome particles/mL. Two symmetrical full-thickness dorsal wounds were created in each animal.(2) Exosomes were injected into one wound, while the contralateral wound served to assess paracrine effects. Wound healing was evaluated over 21 days through macroscopic analysis, histology, immunofluorescence, and quantitative RT-PCR targeting inflammation (TNF-α, IL-6), proliferation (EGF, Ki-67), apoptosis (Caspase-3), angiogenesis (VEGF, CD31), and extracellular matrix remodeling (COL1A1, COL3A1, PDGFR). 

  Exosome treatment significantly enhanced wound healing in a dose-dependent manner, with the greatest improvement observed in the 5×10⁹ group. Macroscopic analysis demonstrated reduced wound size and improved scar quality. Molecular and histological analyses revealed suppression of inflammatory cytokines (TNF-α, IL-6), reduced apoptotic activity, and accelerated transition to the remodeling phase, evidenced by increased collagen I/III ratio, and upregulation of PDGFR expression. Angiogenic profiling demonstrated decreased VEGF and increased CD31 expression, suggesting vascular maturation. Notably, similar trends were observed in untreated contralateral wounds, indicating a significant systemic paracrine effect. 

  hUC-MSC-derived exosomes promote wound healing through dose-dependent modulation of inflammation, apoptosis, angiogenesis, and extracellular matrix remodeling. Importantly, their ability to exert therapeutic effects on distant untreated wounds highlights a previously underappreciated systemic paracrine mechanism. These findings support the potential of exosome-based therapies as a scalable and effective strategy for both localized and multifocal wound management. 

(1)Ha DH, Kim HK, Lee J, et al. Mesenchymal Stem/Stromal Cell-Derived Exosomes for Immunomodulatory Therapeutics and Skin Regeneration. Cells. May 7 2020;9(5)doi:10.3390/cells9051157 (2) Ferreira A.D.F., Cunha P.D.S., Carregal V.M., da Silva P.S., de Miranda M.C., Kunrath-Limam M., de Melo M.I.A., Faraco C.C.F., Barbosa J.L. Extracellular vesicles from adipose-derived mesenchymal stem/stromal cells accelerate migration and activate AKT pathway in human keratinocytes and fibroblasts independently of miR-205 activity. Stem. Cells. Int. 2017;2017:9841035. doi: 10.1155/2017/9841035 (3)Galliano RD, Michaels J, Dobryansky M, Levine JP, Gurtner GC. Quantitative and reproducible murine model of excisional wound healing. Wound Rep Regen 2004; 12:485-492.

Purpose: Current methods for craniofacial phenotyping in mouse models are labor-intensive, subjective, and lack reproducibility, limiting large-scale and cross-study analyses. Existing computational platforms require specialized infrastructure inaccessible to many laboratories. This study developed and validated an unsupervised machine learning-powered automated pipeline for standardized, quantitative craniofacial phenotyping in mouse models, enabling a comprehensive examination of the craniofacial phenotypic landscape across genotypes, developmental stages, and strains.

Methods and Materials: Micro-CT scans were obtained from the FaceBase consortium. A SwinUNETR segmentation model was trained for craniofacial skeletal segmentation, followed by standardized alignment and cropping. Statistical shape modeling placed 2048 homologous landmarks on each skull using ShapeWorks. A baseline cranial morphology model was derived from wild-type controls using a multivariate Gaussian distribution, and Mahalanobis distance was used to calculate a cranial morphology deviation (CMD) score quantifying phenotypic deviation from normal morphology. Principal component analysis and unsupervised clustering were applied to identify phenotypic groupings across sexes, developmental stages, and genetic variants. All available mouse imaging data on FaceBase, inclusive of wild-type and mutant models, were processed through the pipeline to establish an objective overview of phenotypic breadth within the craniofacial biology research community.

Results: The segmentation model was trained on ~3600 scans downsampled from 0.035 mm to 0.1 mm spacing, achieving 77% Dice overlap. The unsupervised pipeline successfully generated CMD scores across all available FaceBase mouse models, revealing phenotypic groupings stratified by developmental stage, strain, and genotype. Unsupervised clustering identified subtle morphological variations not captured by traditional landmark-based methods, providing a data-driven map of the craniofacial phenotypic landscape in commonly used mouse models.

Conclusions: This study presents a validated, automated unsupervised machine learning pipeline for craniofacial phenotyping in mouse models. The pipeline provides an objective CMD score for global morphological deviation, reduces reliance on manual annotation, and increases throughput and reproducibility while remaining accessible to laboratories without specialized computational resources. By enabling comprehensive phenotypic characterization of each individual animal, this approach maximizes the data extracted per specimen and may contribute to a reduction in the total number of animals required for craniofacial research, in alignment with the ethical principles of refinement and reduction in animal use. This scalable framework facilitates cross-study data integration and is extensible to other model organisms and craniofacial traits.