Exploring the potential utility of AI large language models for medical ethics: an expert panel evaluation of GPT-4.

Integrating large language models (LLMs) like GPT-4 into medical ethics is a novel concept, and understanding the effectiveness of these models in aiding ethicists with decision-making can have significant implications for the healthcare sector. Thus, the objective of this study was to evaluate the performance of GPT-4 in responding to complex medical ethical vignettes and to gauge its utility and limitations for aiding medical ethicists. Using a mixed-methods, cross-sectional survey approach, a panel of six ethicists assessed LLM-generated responses to eight ethical vignettes.The main outcomes measured were relevance, reasoning, depth, technical and non-technical clarity, as well as acceptability of GPT-4's responses. The readability of the responses was also assessed. Of the six metrics evaluating the effectiveness of GPT-4's responses, the overall mean score was 4.1/5. GPT-4 was rated highest in providing technical (4.7/5) and non-technical clarity (4.4/5), whereas the lowest rated metrics were depth (3.8/5) and acceptability (3.8/5). There was poor-to-moderate inter-rater reliability characterised by an intraclass coefficient of 0.54 (95% CI: 0.30 to 0.71). Based on panellist feedback, GPT-4 was able to identify and articulate key ethical issues but struggled to appreciate the nuanced aspects of ethical dilemmas and misapplied certain moral principles.This study reveals limitations in the ability of GPT-4 to appreciate the depth and nuanced acceptability of real-world ethical dilemmas, particularly those that require a thorough understanding of relational complexities and context-specific values. Ongoing evaluation of LLM capabilities within medical ethics remains paramount, and further refinement is needed before it can be used effectively in clinical settings.

A Nationwide Network of Health AI Assurance Laboratories.

Importance: Given the importance of rigorous development and evaluation standards needed of artificial intelligence (AI) models used in health care, nationwide accepted procedures to provide assurance that the use of AI is fair, appropriate, valid, effective, and safe are urgently needed. Observations: While there are several efforts to develop standards and best practices to evaluate AI, there is a gap between having such guidance and the application of such guidance to both existing and new AI models being developed. As of now, there is no publicly available, nationwide mechanism that enables objective evaluation and ongoing assessment of the consequences of using health AI models in clinical care settings. Conclusion and Relevance: The need to create a public-private partnership to support a nationwide health AI assurance labs network is outlined here. In this network, community best practices could be applied for testing health AI models to produce reports on their performance that can be widely shared for managing the lifecycle of AI models over time and across populations and sites where these models are deployed.

Testing and Evaluation of Health Care Applications of Large Language Models: A Systematic Review.

Importance: Large language models (LLMs) can assist in various health care activities, but current evaluation approaches may not adequately identify the most useful application areas. Objective: To summarize existing evaluations of LLMs in health care in terms of 5 components: (1) evaluation data type, (2) health care task, (3) natural language processing (NLP) and natural language understanding (NLU) tasks, (4) dimension of evaluation, and (5) medical specialty. Data Sources: A systematic search of PubMed and Web of Science was performed for studies published between January 1, 2022, and February 19, 2024. Study Selection: Studies evaluating 1 or more LLMs in health care. Data Extraction and Synthesis: Three independent reviewers categorized studies via keyword searches based on the data used, the health care tasks, the NLP and NLU tasks, the dimensions of evaluation, and the medical specialty. Results: Of 519 studies reviewed, published between January 1, 2022, and February 19, 2024, only 5% used real patient care data for LLM evaluation. The most common health care tasks were assessing medical knowledge such as answering medical licensing examination questions (44.5%) and making diagnoses (19.5%). Administrative tasks such as assigning billing codes (0.2%) and writing prescriptions (0.2%) were less studied. For NLP and NLU tasks, most studies focused on question answering (84.2%), while tasks such as summarization (8.9%) and conversational dialogue (3.3%) were infrequent. Almost all studies (95.4%) used accuracy as the primary dimension of evaluation; fairness, bias, and toxicity (15.8%), deployment considerations (4.6%), and calibration and uncertainty (1.2%) were infrequently measured. Finally, in terms of medical specialty area, most studies were in generic health care applications (25.6%), internal medicine (16.4%), surgery (11.4%), and ophthalmology (6.9%), with nuclear medicine (0.6%), physical medicine (0.4%), and medical genetics (0.2%) being the least represented. Conclusions and Relevance: Existing evaluations of LLMs mostly focus on accuracy of question answering for medical examinations, without consideration of real patient care data. Dimensions such as fairness, bias, and toxicity and deployment considerations received limited attention. Future evaluations should adopt standardized applications and metrics, use clinical data, and broaden focus to include a wider range of tasks and specialties.

Principled estimation and evaluation of treatment effect heterogeneity: A case study application to dabigatran for patients with atrial fibrillation.

OBJECTIVE: To apply the latest guidance for estimating and evaluating heterogeneous treatment effects (HTEs) in an end-to-end case study of the Long-term Anticoagulation Therapy (RE-LY) trial, and summarize the main takeaways from applying state-of-the-art metalearners and novel evaluation metrics in-depth to inform their applications to personalized care in biomedical research. METHODS: Based on the characteristics of the RE-LY data, we selected four metalearners (S-learner with Lasso, X-learner with Lasso, R-learner with random survival forest and Lasso, and causal survival forest) to estimate the HTEs of dabigatran. For the outcomes of (1) stroke or systemic embolism and (2) major bleeding, we compared dabigatran 150 mg, dabigatran 110 mg, and warfarin. We assessed the overestimation of treatment heterogeneity by the metalearners via a global null analysis and their discrimination and calibration ability using two novel metrics: rank-weighted average treatment effects (RATE) and estimated calibration error for treatment heterogeneity. Finally, we visualized the relationships between estimated treatment effects and baseline covariates using partial dependence plots. RESULTS: The RATE metric suggested that either the applied metalearners had poor performance of estimating HTEs or there was no treatment heterogeneity for either the stroke/SE or major bleeding outcome of any treatment comparison. Partial dependence plots revealed that several covariates had consistent relationships with the treatment effects estimated by multiple metalearners. The applied metalearners showed differential performance across outcomes and treatment comparisons, and the X- and R-learners yielded smaller calibration errors than the others. CONCLUSIONS: HTE estimation is difficult, and a principled estimation and evaluation process is necessary to provide reliable evidence and prevent false discoveries. We have demonstrated how to choose appropriate metalearners based on specific data properties, applied them using the off-the-shelf implementation tool survlearners, and evaluated their performance using recently defined formal metrics. We suggest that clinical implications should be drawn based on the common trends across the applied metalearners.

APLUS: A Python library for usefulness simulations of machine learning models in healthcare.

Despite the creation of thousands of machine learning (ML) models, the promise of improving patient care with ML remains largely unrealized. Adoption into clinical practice is lagging, in large part due to disconnects between how ML practitioners evaluate models and what is required for their successful integration into care delivery. Models are just one component of care delivery workflows whose constraints determine clinicians' abilities to act on models' outputs. However, methods to evaluate the usefulness of models in the context of their corresponding workflows are currently limited. To bridge this gap we developed APLUS, a reusable framework for quantitatively assessing via simulation the utility gained from integrating a model into a clinical workflow. We describe the APLUS simulation engine and workflow specification language, and apply it to evaluate a novel ML-based screening pathway for detecting peripheral artery disease at Stanford Health Care.

Selumetinib in the Treatment of Orbital Plexiform Neurofibroma: A Case Report.

A 4-year-old boy with a known diagnosis of neurofibromatosis 1 (NF1) and a diffusely infiltrative plexiform neurofibroma (PN) of the left orbit was started on selumetinib treatment for progressively worsening amblyopia. The patient first presented with new-onset left ptosis at 11 months old. He subsequently developed refractory anisometropic amblyopia of the left eye, in addition to clinically significant left proptosis and hypoglobus that interfered with glasses wear for his amblyopia treatment. The plexiform neurofibroma was not amenable to surgical resection and selumetinib treatment was initiated 3 years after the initial diagnosis. The patient showed remarkable clinical and radiographic improvement in tumor burden after treatment. Best corrected visual acuity improved from 20/50 to 20/20- in his amblyopic eye. Relative proptosis of the affected eye also improved from 4mm to 2mm on Hertel measurements, which allowed for consistent glasses wear. Adverse effects from the treatment were limited to an acneiform rash, which resolved following dose reduction according to the FDA dosing guidelines.

Teprotumumab and Hearing Loss: Case Series and Proposal for Audiologic Monitoring.

PURPOSE: To present a protocol for audiologic monitoring in the setting of teprotumumab treatment of thyroid eye disease, motivated by 4 cases of significant hearing loss, and review the relevant literature. METHODS: Cases of hearing loss in the setting of teprotumumab were retrospectively elicited as part of a multi-institutional focus group, including oculoplastic surgeons, a neurotologist and an endocrinologist. A literature review was performed. RESULTS: An aggregate of 4 cases of teprotumumab-associated hearing loss documented by formal audiologic testing were identified among 3 clinicians who had treated 28 patients. CONCLUSIONS: Teprotumumab may cause a spectrum of potentially irreversible hearing loss ranging from mild to severe, likely resulting from the inhibition of the insulin-like growth factor-1 and the insulin-like growth factor-1 receptor pathway. Due to the novelty of teprotumumab and the lack of a comprehensive understanding of its effect on hearing, the authors endorse prospective investigations of hearing loss in the setting of teprotumumab treatment. Until the results of such studies are available, the authors think it prudent to adopt a surveillance protocol to include an audiogram and tympanometry before, during and after infusion, and when prompted by new symptoms of hearing dysfunction.

Research and Reporting Considerations for Observational Studies Using Electronic Health Record Data.

Electronic health records (EHRs) are an increasingly important source of real-world health care data for observational research. Analyses of data collected for purposes other than research require careful consideration of data quality as well as the general research and reporting principles relevant to observational studies. The core principles for observational research in general also apply to observational research using EHR data, and these are well addressed in prior literature and guidelines. This article provides additional recommendations for EHR-based research. Considerations unique to EHR-based studies include assessment of the accuracy of computer-executable cohort definitions that can incorporate unstructured data from clinical notes and management of data challenges, such as irregular sampling, missingness, and variation across time and place. Principled application of existing research and reporting guidelines alongside these additional considerations will improve the quality of EHR-based observational studies.

Compound Coherent Plane-Wave Ultrasound Imaging of Vascular Malformations of the Orbit.

PURPOSE: Prior color-flow Doppler ultrasound studies of the eye have been performed with systems that exceed US Food and Drug Administration permissible ophthalmic ultrasonic energy limits. The authors report a study of orbital vascular malformations using a novel, Food and Drug Administration compliant, ultrafast compound coherent plane-wave ultrasound device to produce power Doppler images. METHODS: Using a Verasonics Vantage 128 ultrasound engine and a user-developed MATLAB program with a 5-MHz linear-array probe, compound coherent plane-wave ultrasound data were collected on patients with orbital vascular malformations. Real-time color-flow Doppler visualized orbital blood flow. Power Doppler images were produced by post-processing compound coherent plane-wave ultrasound data acquired continuously for 2 seconds. RESULTS: Compound coherent plane-wave ultrasound was performed on 3 orbital vascular malformations (1 venolymphatic malformation, 1 infantile hemangioma, and 1 arteriovenous malformation). Compound coherent plane-wave ultrasound produced a high-resolution depiction of orbital blood flow for orbital vascular malformations with high sensitivity to slow flow. CONCLUSIONS: Analysis of blood flow within orbital lesions informs treatment planning. Compound coherent plane-wave ultrasound is an emerging ultrasound modality that falls within the Food and Drug Administration guidelines for use in the orbit and provides information to characterize orbital vascular malformations.

Carcinoid Tumors of the Orbit and Ocular Adnexa.

PURPOSE: Well-differentiated neuroendocrine or carcinoid tumors are found most commonly in the gastrointestinal tract. When metastatic to the orbit, they tend to have a propensity for the extraocular muscles. The purpose of this study was to better understand the diversity in presentation of orbital carcinoid disease and to determine predictors for survival. METHODS: In this observational cross-sectional cohort study, data from 8 tertiary orbital practices were compiled. Demographic, clinical, pathologic, American Joint Committee on Cancer stage and grade, imaging, and management data were extracted for all the patients. Descriptive statistics were calculated. Subgroups were compared utilizing analysis of variance analyses and Kaplan-Meier curves. Time to progression and disease-specific and overall mortality were calculated. Comparisons were performed for the following a priori pairs: unknown versus known primary tumor, single versus multiple extraocular muscle involvement, unilateral versus bilateral orbital disease, extraocular muscle versus other orbital involvement, and excisional versus incisional surgery. RESULTS: A total of 28 patients with carcinoid tumors of the orbit were identified. Of these, 57.1% of patients were female, the mean age at diagnosis of the primary tumor was 58.8 years and the mean age at diagnosis of orbital disease was 62.6 years. At primary presentation, all patients were American Joint Committee on Cancer stage III or IV and 21.4% demonstrated carcinoid syndrome. Muscle involvement was noted in 78.6% of patients, and of these, 72% were noted to have single muscle disease. Eight patients had no primary tumor identified; 3 of these 8 demonstrated disseminated disease at the time of diagnosis. The overall 5-year survival rate was 81.8% from diagnosis of primary tumor and 50% from diagnosis of orbital disease. Subgroup analysis revealed that patients with unilateral orbital disease when compared with bilateral orbital disease had a longer progression-free survival and time to death from all causes (p = 0.025). Patients with disease localized to the orbit at presentation had longer time to death than those with disseminated disease. Treatment with surgery, radiation, or octreotide did not appear to affect survival. Patients managed with systemic chemotherapy had a shorter time of survival than the rest of the group. All other subgroup comparisons were not found to be statistically significant. CONCLUSIONS: Neuroendocrine tumors of the orbit represent a wide spectrum of disease, with some cases being part of disseminated disease, while others being localized presentations. This heterogeneity may be responsible for the slightly higher overall survival in these patients than others with metastatic carcinoid tumors in other locations.

Feasibility and evaluation of a large-scale external validation approach for patient-level prediction in an international data network: validation of models predicting stroke in female patients newly diagnosed with atrial fibrillation.

BACKGROUND: To demonstrate how the Observational Healthcare Data Science and Informatics (OHDSI) collaborative network and standardization can be utilized to scale-up external validation of patient-level prediction models by enabling validation across a large number of heterogeneous observational healthcare datasets. METHODS: Five previously published prognostic models (ATRIA, CHADS2, CHADS2VASC, Q-Stroke and Framingham) that predict future risk of stroke in patients with atrial fibrillation were replicated using the OHDSI frameworks. A network study was run that enabled the five models to be externally validated across nine observational healthcare datasets spanning three countries and five independent sites. RESULTS: The five existing models were able to be integrated into the OHDSI framework for patient-level prediction and they obtained mean c-statistics ranging between 0.57-0.63 across the 6 databases with sufficient data to predict stroke within 1 year of initial atrial fibrillation diagnosis for females with atrial fibrillation. This was comparable with existing validation studies. The validation network study was run across nine datasets within 60 days once the models were replicated. An R package for the study was published at https://github.com/OHDSI/StudyProtocolSandbox/tree/master/ExistingStrokeRiskExternalValidation. CONCLUSION: This study demonstrates the ability to scale up external validation of patient-level prediction models using a collaboration of researchers and a data standardization that enable models to be readily shared across data sites. External validation is necessary to understand the transportability or reproducibility of a prediction model, but without collaborative approaches it can take three or more years for a model to be validated by one independent researcher. In this paper we show it is possible to both scale-up and speed-up external validation by showing how validation can be done across multiple databases in less than 2 months. We recommend that researchers developing new prediction models use the OHDSI network to externally validate their models.

Predominant Contribution of Superior Rectus-Levator Complex Enlargement to Optic Neuropathy and Inferior Visual Field Defects in Thyroid Eye Disease.

PURPOSE: To compare extraocular muscle volumes in thyroid eye disease patients with and without compressive optic neuropathy. METHODS: A retrospective review of 44 orbital CT scans (28 orbits without compressive disease and 16 orbits with compressive optic neuropathy) was conducted. The extraocular muscle volumes, summated soft tissue volumes, and optic nerve volumes were calculated at a section in the posterior 1/3 of the orbit. The visual fields of the orbits with compressive optic neuropathy were analyzed. RESULTS: The mean combined extraocular muscle/summated soft tissue volume ratio and the mean superior rectus-levator complex/summated soft tissue volume ratio were greater in those with compressive optic neuropathy than in those without compressive optic neuropathy (p = 0.02, 0.008, respectively). The ratio of the mean inferior, medial, or lateral rectus/summated soft tissue volume did not differ significantly between patients with or without compressive optic neuropathy (p values of 0.315, 0.615, and 0.254, respectively). Visual field analysis of the compressive optic neuropathy group demonstrated that 58% of the orbits with visual field defects had inferior field defects. CONCLUSIONS: When measured at a section near the orbital apex, the mean combined muscle/summated soft tissue volume ratio and the mean superior rectus-levator complex/summated soft tissue volume ratio are greater in those with compressive disease than those without. This suggests that the specific enlargement of the superior rectus-levator complex makes a significant contribution to thyroid eye disease-compressive optic neuropathy and may explain the inferior visual field deficits classically found in this group of patients.

Doxycycline Injection for Sclerotherapy of Lower Eyelid Festoons and Malar Edema: Preliminary Results.

PURPOSE: To investigate the safety and efficacy of direct, intralesional doxycycline hyclate injection for improving the appearance of cosmetically significant lower eyelid festoons and malar edema. METHODS: An Institutional Review Board approved, retrospective review was performed of 15 consecutive patients with malar edema and/or festoons injected with doxycycline hyclate at a concentration of 10 mg/ml. Pre- and postinjection photographs were reviewed and graded on a scale of 0 to 3 (0: no festoon; 1: small festoon; 2: medium festoon; 3: large festoon) by 2 masked physician observers. Patients were excluded from the final analysis if they received an alternate dose concentration, had incomplete photographic records, or did not follow up. Student t test was used for statistical analysis. RESULTS: Twenty consecutive treatment areas of 11 patients were included in the analysis. Final follow up ranged from 3 to 104 weeks, with a mean follow up of 22.5 weeks. The average (standard deviation) initial festoon grade of 2.5 (0.58) decreased to 0.9 (0.82) with a p value of <0.001. The average number of injections performed per side was 1.4 (range: 1-2). The mean volume per injection was 0.72 ml (range: 0.15-2.0 ml). Commonly documented subjective complaints were burning sensation with injection, pain, bruising, and erythema. There were no other dermatologic or visual complications following treatment. CONCLUSIONS: These preliminary results suggest that intralesional injections of doxycycline hyclate at a concentration of 10 mg/ml may be an effective treatment option for cosmetically significant lower eyelid festoons and malar edema. Future prospective studies with increased patient numbers, increasing concentrations, combination therapies with local anesthetic or regional nerve blocks, and longer follow up are needed to validate these results and determine optimal injection technique.

Nasal Glioma: A Rare Cause of Congenital Inner Canthal Swelling.

Nasal glioma, encephalocele, and ectopic brain are rare congenital anomalies. The terminology applied to these entities has been historically confusing. In many cases, the terms overlap and may be employed synonymously although some authors emphasize their differences. The authors describe herein a child with an inner canthal mass of brain-like tissue that they interpret as nasal glioma, a variety of encephalocele that has lost its connection to the intracranial contents. This research was conducted in conformity with the Helsinki Declaration and Health Insurance Portability and Accountability Act regulations.

The Columbia Thyroid Eye Disease-Compressive Optic Neuropathy Diagnostic Formula.

PURPOSE: Diagnosing thyroid eye disease-compressive optic neuropathy (TED-CON) is challenging, particularly in cases lacking a relative afferent pupillary defect. Large case series of TED-CON patients and accessible diagnostic tools are lacking in the current literature. This study aims to create a mathematical formula that accurately predicts the presence or absence of CON based on the most salient clinical measures of optic neuropathy. METHODS: A retrospective case series compares 108 patients (216 orbits) with either unilateral or bilateral TED-CON and 41 age-matched patients (82 orbits) with noncompressive TED. Utilizing clinical variables assessing optic nerve function and/or risk of compressive disease, and with the aid of generalized linear regression modeling, the authors create a mathematical formula that weighs the relative contribution of each clinical variable in the overall prediction of CON. RESULTS: Data from 213 orbits in 110 patients derived the formula: y = -0.69 + 2.58 × (afferent pupillary defect) - 0.31 × (summed limitation of ductions) - 0.2 × (mean deviation on Humphrey visual field testing) - 0.02 × (% color plates). This accurately predicted the presence of CON (y > 0) versus non-CON (y < 0) in 82% of cases with 83% sensitivity and 81% specificity. When there was no relative afferent pupillary defect, which was the case in 63% of CON orbits, the formula correctly predicted CON in 78% of orbits with 73% sensitivity and 83% specificity. CONCLUSIONS: The authors developed a mathematical formula, the Columbia TED-CON Diagnostic Formula (CTD Formula), that can help guide clinicians in accurately diagnosing TED-CON, particularly in the presence of bilateral disease and when no relative afferent pupillary defect is present.

IgG4-related orbital disease masquerading as thyroid eye disease, vice versa, or both?

A 40 year-old male presented after one year of unilateral, progressive, steroid-responsive, orbital inflammatory disease causing proptosis, extraocular muscle (EOM) restriction, and compressive optic neuropathy. The development of anti-thyroidal antibodies prompted the diagnosis of thyroid eye disease (TED); however, the prolonged active phase, remarkable reversibility of ophthalmic features with high-dose corticosteroids, unilaterally of disease, uncharacteristic EOM involvement (including both obliques), and the absence of autoimmune thyroid disease provoked consideration of alternative diagnoses. Inferior oblique biopsy stained positive for IgG4 with histologic features atypical of TED. The patient received rituximab for presumed IgG4-related orbital disease (IgG4-ROD) with subsequent reversal of compressive optic neuropathy, near complete resolution of EOM restriction, and improved proptosis, the latter two of which are not routinely anticipated in advanced TED. The possible role for B-cell depletion in both TED and IgG4-ROD suggests a degree of overlap in the underlying immune-related pathophysiology that is yet to be defined.

Low-Cost 3D Printing Orbital Implant Templates in Secondary Orbital Reconstructions.

PURPOSE: Despite its increasing use in craniofacial reconstructions, three-dimensional (3D) printing of customized orbital implants has not been widely adopted. Limitations include the cost of 3D printers able to print in a biocompatible material suitable for implantation in the orbit and the breadth of available implant materials. The authors report the technique of low-cost 3D printing of orbital implant templates used in complex, often secondary, orbital reconstructions. METHODS: A retrospective case series of 5 orbital reconstructions utilizing a technique of 3D printed orbital implant templates is presented. Each patient's Digital Imaging and Communications in Medicine data were uploaded and processed to create 3D renderings upon which a customized implant was designed and sent electronically to printers open for student use at our affiliated institutions. The mock implants were sterilized and used intraoperatively as a stencil and mold. The final implant material was chosen by the surgeons based on the requirements of the case. RESULTS: Five orbital reconstructions were performed with this technique: 3 tumor reconstructions and 2 orbital fractures. Four of the 5 cases were secondary reconstructions. Molded Medpor Titan (Stryker, Kalamazoo, MI) implants were used in 4 cases and titanium mesh in 1 case. The stenciled and molded implants were adjusted no more than 2 times before anchored in place (mean 1). No case underwent further revision. CONCLUSIONS: The technique and cases presented demonstrate 1) the feasibility and accessibility of low-cost, independent use of 3D printing technology to fashion patient-specific implants in orbital reconstructions, 2) the ability to apply this technology to the surgeon's preference of any routinely implantable material, and 3) the utility of this technique in complex, secondary reconstructions.

Enhanced Fasanella-Servat Procedure for the Graded Repair of Blepharoptosis.

PURPOSE: To describe in detail a technique for a modified Fasanella-Servat procedure that provides a simplified method for graded Mueller muscle excision with minimal and controlled tarsal excision. METHODS: A retrospective study of 71 patients (102 eyelids) who underwent the modified Fasanella-Servat procedure is reported. Measurements include the preoperative, post-phenylephrine, and postoperative margin-to-reflex distance-1, and postoperative upper eyelid height symmetry as determined by the absolute difference between right- and left-sided margin-to-reflex distance-1. Postoperative complications are reported. RESULTS: The average increase in margin-to-reflex distance-1 was 2.4 mm with an average postoperative upper eyelid height symmetry of 0.4 mm. Postoperative asymmetry was 1.5 mm or less in 68 patients, a success rate of 96%. Four patients (6%) exhibited overcorrection, 2 of which required additional surgery. No lagophthalmos or notable eyelid contour abnormalities were seen. CONCLUSIONS: The modified Fasanella-Servat technique offers a simple method to isolate and resect Mueller muscle and a minimal segment of tarsus in a quantitative fashion, allowing for a graded repair of blepharoptosis and thereby decreasing the risk of postoperative overcorrection, lagophthalmos, and eyelid contour asymmetry whilst preserving the bulk of tarsus.

Atypical Rapidly Enlarging Orbital Schwannoma.

Orbital schwannomas are typically slow-growing benign tumors that can cause gradual loss of vision, proptosis, and limitation of ocular motility. The authors present an atypical case of a rapidly growing orbital apex schwannoma in a patient with preexisting vision loss secondary to presumed sarcoidal optic neuritis. Contrary to the slowly progressive nature of a typical orbital schwannoma, the lesion was observed to enlarge from radiologically undiscernible to 3.5 cm over 4 years.

Low-Flow Arterialized Venous Malformations of the Orbit.

PURPOSE: Orbital vascular malformations are classified by their hemodynamic properties, either high or low flow. Low-flow lesions may be simple venous, lymphatic, or combined lymphaticovenous malformations. The authors report a series of cases in which predominantly low flow, venous lesions were unexpectedly noted to have arterial feeders. METHODS: A retrospective chart review of patients identified by the authors as having orbital varices with arterial components was conducted. The authors identified 7 such cases. After careful review, 2 cases were excluded due to inconclusive neuroradiographic findings. The authors review the clinical, radiologic, histopathologic, and surgical information from the remaining 5 cases and discuss their clinical significance. RESULTS: All 5 cases were most consistent with variceal lesions: 3 as clinically distensible lesions and 2 as thrombosed lesions. Additional arterial feeder vessels were noted by angiography (3) or intraoperative visualization (2). The arterial contribution varied from faint vessels to distinct branches of the ophthalmic artery. Ages ranged from 13 to 61 years without predilection for gender. Treatments consisted of excision, embolization, and observation. Two poignant cases are highlighted: the first illustrating that an angiogram in isolation of its clinical picture can be misleading and result in treatment intervention with undue risk, and the second illustrating that inadequate treatment of unrecognized arterial components may contribute to recurrences. CONCLUSIONS: Low-flow orbital variceal lesions may have less prominent, arterial components. This type of combined arterialized venous malformation is largely unrecognized in the ophthalmic literature. Correct identification of these lesions is critical in providing safe, effective, and durable treatment.