Assessment of Generative Artificial Intelligence (AI) Models in Creating Medical Illustrations for Various Corneal Transplant Procedures.

PURPOSE: This study aimed to task and assess generative artificial intelligence (AI) models in creating medical illustrations for corneal transplant procedures such as Descemet's stripping automated endothelial keratoplasty (DSAEK), Descemet's membrane endothelial keratoplasty (DMEK), deep anterior lamellar keratoplasty (DALK), and penetrating keratoplasty (PKP).  Methods: Six engineered prompts were provided to Decoder-Only Autoregressive Language and Image Synthesis 3 (DALL-E 3) and Medical Illustration Manager (MIM) to guide these generative AI models in creating a final medical illustration for each of the four corneal transplant procedures. Control illustrations were created by the authors for each transplant technique for comparison. A grading system with five categories with a maximum score of 3 points each (15 points total) was designed to objectively assess AI's performance. Four independent reviewers analyzed and scored the final images produced by DALL-E 3 and MIM as well as the control illustrations. All AI-generated images and control illustrations were then provided to Chat Generative Pre-Trained Transformer-4o (ChatGPT-4o), which was tasked with grading each image with the grading system described above. All results were then tabulated and graphically depicted. RESULTS: The control illustration images received significantly higher scores than produced images from DALL-E 3 and MIM in legibility, anatomical realism and accuracy, procedural step accuracy, and lack of fictitious anatomy (p<0.001). For detail and clarity, the control illustrations and images produced by DALL-E 3 and MIM received statistically similar scores of 2.75±0.29, 2.19±0.24, and 2.50±0.29, respectively (p=0.0504). With regard to mean cumulative scores for each transplant procedure image, the control illustrations received a significantly higher score than DALL-E 3 and MIM (p<0.001). Additionally, the overall mean cumulative score for the control illustrations was significantly higher than DALL-E 3 and MIM (14.56±0.51 (97.1%), 4.38±1.2 (29.2%), and 5.63±1.82 (37.5%), respectively (p<0.001)). When assessing AI's grading performance, ChatGPT-4o scored the images produced by DALL-E 3 and MIM significantly higher than the average scores of the independent reviewers (DALL-E 3: 10.0±0.0 (66.6%) vs. 4.38±1.20 (29.2%), p<0.001; MIM: 10.0±0.0 (66.6%) vs. 5.63±1.82 (37.5%), p<0.001). However, mean scores for the control illustrations between ChatGPT-4o and the independent reviewers were comparable (15.0±0.0 (100%) vs. 14.56±0.13 (97.1%); p>0.05). CONCLUSION: AI is an extremely powerful and efficient tool for many tasks, but it is currently limited in producing accurate medical illustrations for corneal transplant procedures. Further development is required for generative AI models to create medically sound and accurate illustrations for use in ophthalmology.

Evaluating the Accuracy of Artificial Intelligence (AI)-Generated Illustrations for Laser-Assisted In Situ Keratomileusis (LASIK), Photorefractive Keratectomy (PRK), and Small Incision Lenticule Extraction (SMILE).

PURPOSE: To utilize artificial intelligence (AI) platforms to generate medical illustrations for refractive surgeries, aiding patients in visualizing and comprehending procedures like laser-assisted in situ keratomileusis (LASIK), photorefractive keratectomy (PRK), and small incision lenticule extraction (SMILE). This study displays the current performance of two OpenAI programs in terms of their accuracy in common corneal refractive procedures. METHODS: We selected AI image generators based on their popularity, choosing Decoder-Only Autoregressive Language and Image Synthesis 3 (DALL-E 3) for its leading position and Medical Illustration Master (MiM) for its high engagement. We developed six non-AI-generated prompts targeting specific outcomes related to LASIK, PRK, and SMILE procedures to assess medical accuracy. We generated images using these prompts (18 total images per AI platform) and used the final images produced after the sixth prompt for this study (three final images per AI platform). Human-created procedural images were also gathered for comparison. Four experts independently graded the images, and their scores were averaged. Each image was evaluated with our grading system on "Legibility," "Detail & Clarity," "Anatomical Realism & Accuracy," "Procedural Step Accuracy," and "Lack of Fictitious Anatomy," with scores ranging from 0 to 3 per category allowing 15 points total. A score of 15 points signifies excellent performance, indicating a highly accurate medical illustration. Conversely, a low score suggests a poor-quality illustration. Additionally, we submitted the same AI-generated images back into Chat Generative Pre-Trained Transformer-4o (ChatGPT-4o) along with our grading system. This allowed ChatGPT-4o to use and evaluate both AI-generated and human-created images (HCIs). RESULTS: In individual category scoring, HCIs significantly outperformed AI images in legibility, anatomical realism, procedural step accuracy, and lack of fictitious anatomy. There were no significant differences between DALL-E 3 and MiM in these categories (p>0.05). In procedure-specific comparisons, HCIs consistently scored higher than AI-generated images for LASIK, PRK, and SMILE. For LASIK, HCIs scored 14 ± 0.82 (93.3%), while DALL-E 3 scored 4.5 ± 0.58 (30%) and MiM scored 4.5 ± 1.91 (30%) (p<0.001). For PRK, HCIs scored 14.5 ± 0.58 (96.7%), compared to DALL-E 3's 5.25 ± 1.26 (35%) and MiM's 7 ± 3.56 (46.7%) (p<0.001). For SMILE, HCIs scored 14.5 ± 0.68 (96.7%), while DALL-E 3 scored 5 ± 0.82 (33.3%) and MiM scored 6 ± 2.71 (40%) (p<0.001). HCIs significantly outperformed AI-generated images from DALL-E 3 and MiM in overall accuracy for medical illustrations, achieving scores of 14.33 ± 0.23 (95.6%), 4.93 ± 0.69 (32.8%), and 5.83 ± 0.23 (38.9%) respectively (p<0.001). ChatGPT-4o evaluations were consistent with human evaluations for HCIs (3 ± 0, 2.87 ± 0.23; p=0.121) but rated AI images higher than human evaluators (2 ± 0 vs 1.07 ± 0.73; p<0.001). CONCLUSION: This study highlights the inaccuracy of AI-generated images in illustrating corneal refractive procedures such as LASIK, PRK, and SMILE. Although the OpenAI platform can create images recognizable as eyes, they lack educational value. AI excels in quickly generating creative, vibrant images, but accurate medical illustration remains a significant challenge. While AI performs well with text-based actions, its capability to produce precise medical images needs substantial improvement.

The Utilization of Brillouin Microscopy in Corneal Diagnostics: A Systematic Review.

Corneal biomechanical data has been used since 2005 to screen for keratoconus and corneal ectasia by corneal specialists. Older technology uses force applanation techniques over a 3 mm area in the central cornea, making it highly dependent on extraneous variables and unable to calculate the elasticity of the tissue. Brillouin microscopy is a newer method that uses a natural shift in the frequency of light as it passes through a material. This frequency shift can be used to estimate the viscoelasticity of the tissue. The advantage of Brillouin microscopy is that it can create a full three-dimensional (3D) map of the entire cornea without direct contact. A literature search was conducted using the databases PubMed, Google Scholar, and Ovid regarding the applications of Brillouin microscopy in corneal diagnostics. A final total of 16 articles was included describing the various ex vivo and in vivo studies conducted using Brillouin microscopy. Applications of this technology spanned from keratoconus diagnosis to post-corneal refractive surgery evaluation. All studies evaluated corneal biomechanics and other corneal properties through the quantification of Brillouin frequency shifts. Many of the studies found that this diagnostic device is capable of detecting subtle changes in corneal thickness and biomechanics in keratoconic corneas at a high level of specificity and sensitivity. However, limitations of Brillouin microscopy may include the duration of time required for use and fluctuations in accuracy depending on the corneal hydration state. Future technology seems to be geared toward a combination of optical coherence tomography (OCT) and Brillouin microscopy, using OCT as a three-dimensional pupil-tracking modality. Further research and understanding of the technology involved will lead to better care of patients in the field of ophthalmology.

Visual Prognosis Following Cataract Surgery in Highly Myopic Patients with Prior History of Verisyse Phakic Intraocular Lens Implantation.

Background/Objectives: This study aimed to evaluate the visual outcomes and prognosis after cataract surgery in patients with prior history of Verisyse phakic intraocular lens (pIOL) implantation. Methods: A retrospective cohort study involving 215 Verisyse pIOL implantations and 17 explantations was conducted. The Verisyse pIOL was disenclaved and removed through a superior scleral tunnel incision. Cataract extraction with phacoemulsification was then performed through a temporal clear corneal incision. Results: An occurrence rate of 7.9% of eyes with cataract formation was found. Both uncorrected (UDVA) and corrected visual acuity (CDVA) three months after cataract surgery were significantly improved (0.24 ± 0.30 vs. 0.73 ± 0.48; p < 0.001 and 0.10 ± 0.14 vs. 0.30 ± 0.31; p = 0.004, respectively). The UDVA was 20/20 or better in 41% of eyes and 20/40 or better in 65% of eyes. The CDVA was 20/20 or better in 53% of eyes and 20/40 or better in 88% of eyes. The safety and efficacy indices were 1.96 ± 1.68 and 1.60 ± 1.36, respectively. Conclusions: Various complications including cataracts may develop in these patients. Verisyse pIOLs have a lower incidence of cataract formation and are more likely to lead to age-related cataracts rather than the anterior subcapsular cataracts commonly seen in implantable collamer lens (ICL) patients. Patients with a prior history of Verisyse pIOL can expect to have a good visual prognosis after cataract extraction.

Comparing the Utility of Landmark-Palpation Guided to Ultrasound-Guided Teaching Methodologies for Subclavian Central Venous Access Using a Formalin-Embalmed Cadaver Model.

Background: Low-risk and realistic simulation strategies are needed to train clinical learners to perform hands-on invasive procedures. This follow-up study compares the utility of landmark-guided palpation-based to ultrasound-guided teaching techniques for subclavian central venous access using formalin-embalmed cadavers. Methods: The subclavian veins of 3 cadavers were imaged with ultrasound to evaluate vein patency before palpation-based venous access was attempted. Twenty-three first-year medical students were trained to access the subclavian vein using palpation-based techniques. Training involved ten minutes of didactic orientation and ten minutes of hands-on practical instruction using cadavers. Participant confidence was measured using a 10-point Likert scale on pre- and post-training questionnaires. Objective skills testing for each participant included quantifying the number of skin punctures and recording the time elapsed from first skin puncture to fluid flashback into the syringe. Data was analyzed using a generalized linear model (GLM) approach. Results: Participant confidence significantly increased following training in both ultrasound and palpation training groups across all questionnaire items (P<0.001). The ultrasound group had fewer skin punctures (P<0.001) and fewer failures (1) than the palpation group (6). Participants in the ultrasound group were more confident than those in the palpation group in their ability to locate the vein and select the optimal site for needle access (P<0.05). Conclusion: Formalin-embalmed cadavers provide a safe, stress-free, and effective means by which to train students in subclavian vein access using both palpation and ultrasound-based techniques. Repeated practice accessing and aspirating fluid from a cadaveric subclavian vein significantly increases trainee confidence, an essential factor in physician performance that may lead to fewer complications. Introducing this type of low-risk and hands-on practice may be beneficial for trainees before they attempt subclavian vein access on live patients.

Advancing Point-of-Care Ultrasound Training in Medical Schools: Ultrasound-Guided Subclavian Vein Access Training Using Formalin-Embalmed Cadavers.

This brief report examines the feasibility of using formalin-embalmed cadavers in training medical students to use ultrasound guidance to access the subclavian. This novel educational approach is discussed in the context of the ongoing integration of point-of-care ultrasound training into medical education. Additionally, this report explores how cadavers can provide practical, effective, and hands-on skills training opportunities for medical students to learn to perform common clinical procedures under ultrasound guidance. DESIGN: This report presents subjective and objective data evaluating the utility of teaching medical students to perform ultrasound-guided subclavian vein access on formalin-embalmed cadavers. SETTING: Rocky Vista University College of Osteopathic Medicine in Ivins, UT. SUBJECTS: Twenty-five first-year medical students at Rocky Vista University. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Pre and posttraining questionnaires were administered to assess each participant's self-confidence in using ultrasonography to access the subclavian vein of a cadaver. A statistically significant increase in participant self-confidence was observed across all questionnaire items from pre to posttraining. Objective evaluation consisted of a supervised skills test. Participants were evaluated on their ability to visualize the subclavian vein with ultrasound and achieve flashback of blood/embalming fluid into a syringe. During skills testing, the number of needle sticks and the time taken to achieve flashback were recorded for each participant. Twenty-three of the 25 participants were able to successfully complete the skills testing assessment. CONCLUSIONS: The formalin-embalmed cadaver can be a readily available and effective learning tool for medical education programs seeking to provide training opportunities in ultrasound-guided clinical procedures. The use of cadavers allows learners to train in a low stress and anatomically authentic environment without risk of patient discomfort.

A Scoping Review of Emergency Department Discharge Risk Stratification.

INTRODUCTION: Although emergency department (ED) discharge presents patient-safety challenges and opportunities, the ways in which EDs address discharge risk in the general ED population remains disparate and largely uncharacterized. In this study our goal was to conduct a review of how EDs identify and target patients at increased risk at time of discharge. METHODS: We conducted a literature search to explore how EDs assess patient risk upon discharge, including a review of PubMed and gray literature. After independently screening articles for inclusion, we recorded study characteristics including outcome measures, patient risk factors, and tool descriptions. Based on this review and discussion among collaborators, major themes were identified. RESULTS: PubMed search yielded 384 potentially eligible articles. After title and abstract review, we screened 235 for potential inclusion. After full text and reference review, supplemented by Google Scholar and gray literature reviews, we included 30 articles for full review. Three major themes were elucidated: 1) Multiple studies include retrospective risk assessment, whereas the use of point-of-care risk assessment tools appears limited; 2) of the point-of-care tools that exist, inputs and outcome measures varied, and few were applicable to the general ED population; and 3) while many studies describe initiatives to improve the discharge process, few describe assessment of post-discharge resource needs. CONCLUSION: Numerous studies describe factors associated with an increased risk of readmission and adverse events after ED discharge, but few describe point-of-care tools used by physicians for the general ED population. Future work is needed to investigate standardized tools that assess ED discharge risk and patients' needs upon ED discharge.

High-intensity endurance training is associated with left atrial fibrosis.

INTRODUCTION: Endurance athletes are at higher risk for developing atrial fibrillation as compared to the general population. The exact mechanism to explain this observation is incompletely understood. Our study aimed to determine whether degree of left atrial fibrosis detected by late gadolinium-enhancement magnetic resonance imaging (LGE-MRI) differed between Masters athletes and non-athlete controls. METHODS: We recruited 20 endurance healthy Masters athletes and 20 healthy control subjects who underwent cardiac MRI. Healthy controls were recruited during screening colonoscopies and Masters athletes were recruited through word of mouth and at competitions. The two groups were age and gender matched. None of the participants were known to have an arrhythmia. Fibrosis, as measured by late gadolinium-enhancement, was measured in each participant by blinded readers. The degree of left atrial fibrosis was compared between the two groups. All participants were recruited from the Salt Lake City region and scanned at the University of Utah healthcare complex. RESULTS: Left ventricular function was normal in all study participants. Left atrial volumes were significantly larger in the athletes (74.2 ml ±â€¯14.4) as compared to the healthy control subjects (60.8 mL ±â€¯21.4) (P = .02). Mean left atrial fibrosis score, reported as a percentage of the LA, was 15.5% ±â€¯5.9 in the athlete cohort compared to 9.6% ±â€¯4.9 in the controls (P = .002). CONCLUSIONS: To our knowledge this is the first study that describes, characterizes and specifically quantifies fibrotic changes within the left atrium of highly trained endurance athletes. Increased atrial fibrosis seen in this population may be an early indicator for endurance athletes at risk of developing atrial arrhythmias.

Efficient Segmentation Pipeline Using Diffeomorphic Image Registration: A Validation Study.

Functional measurements of the left atrium (LA) in atrial fibrillation (AF) patients is limited to a single CINE slice midway through the LA. Nonetheless, a full 3D characterization of atrial functional measurements would provide more insights into LA function. But this improved modeling capacity comes at a price of requiring LA segmentation of each 3D time point,a time-consuming and expensive task that requires anatomy-specific expertise.We propose an efficient pipeline which requires ground truth segmentation of a single (or limited) CINE time point to accurately propagate it throughout the sequence. This method significantly saves human effort and enable better characterization of LA anatomy. From a gated cardiac CINE MRI sequence we select a single CINE time point with ground truth segmentation, and assuming cyclic motion, we register other images corresponding to all time points using diffeomorphic registration in ANTs. The diffeomorphic registration fields allow us to map a given anatomical shape (segmentation) to each CINE time point, facilitating the construction of a 4D shape model.