Search

Purpose: To evaluate the capabilities of Chat Generative Pre-Trained Transformer (ChatGPT), as a large language model (LLM), for diagnosing glaucoma using the Ocular Hypertension Treatment Study (OHTS) dataset, and comparing the diagnostic capability of ChatGPT 3.5 and ChatGPT 4.0. Design: Prospective data collection study. Participants: A total of 3170 eyes of 1585 subjects from the OHTS were included in this study. Methods: We selected demographic, clinical, ocular, visual field, optic nerve head photo, and history of disease parameters of each participant and developed case reports by converting tabular data into textual format based on information from both eyes of all subjects. We then developed a procedure using the application programming interface of ChatGPT, a LLM-based chatbot, to automatically input prompts into a chat box. This was followed by querying 2 different generations of ChatGPT (versions 3.5 and 4.0) regarding the underlying diagnosis of each subject. We then evaluated the output responses based on several objective metrics. Main Outcome Measures: Area under the receiver operating characteristic curve (AUC), accuracy, specificity, sensitivity, and F1 score. Results: Chat Generative Pre-Trained Transformer 3.5 achieved AUC of 0.74, accuracy of 66%, specificity of 64%, sensitivity of 85%, and F1 score of 0.72. Chat Generative Pre-Trained Transformer 4.0 obtained AUC of 0.76, accuracy of 87%, specificity of 90%, sensitivity of 61%, and F1 score of 0.92. Conclusions: The accuracy of ChatGPT 4.0 in diagnosing glaucoma based on input data from OHTS was promising. The overall accuracy of ChatGPT 4.0 was higher than ChatGPT 3.5. However, ChatGPT 3.5 was found to be more sensitive than ChatGPT 4.0. In its current forms, ChatGPT may serve as a useful tool in exploring disease status of ocular hypertensive eyes when specific data are available for analysis. In the future, leveraging LLMs with multimodal capabilities, allowing for integration of imaging and diagnostic testing as part of the analyses, could further enhance diagnostic capabilities and enhance diagnostic accuracy. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

BackgroundIn this paper, we present an automated method for article classification, leveraging the power of large language models (LLMs). ObjectiveThe aim of this study is to evaluate the applicability of various LLMs based on textual content of scientific ophthalmology papers. MethodsWe developed a model based on natural language processing techniques, including advanced LLMs, to process and analyze the textual content of scientific papers. Specifically, we used zero-shot learning LLMs and compared Bidirectional and Auto-Regressive Transformers (BART) and its variants with Bidirectional Encoder Representations from Transformers (BERT) and its variants, such as distilBERT, SciBERT, PubmedBERT, and BioBERT. To evaluate the LLMs, we compiled a data set (retinal diseases [RenD] ) of 1000 ocular disease–related articles, which were expertly annotated by a panel of 6 specialists into 19 distinct categories. In addition to the classification of articles, we also performed analysis on different classified groups to find the patterns and trends in the field. ResultsThe classification results demonstrate the effectiveness of LLMs in categorizing a large number of ophthalmology papers without human intervention. The model achieved a mean accuracy of 0.86 and a mean F1-score of 0.85 based on the RenD data set. ConclusionsThe proposed framework achieves notable improvements in both accuracy and efficiency. Its application in the domain of ophthalmology showcases its potential for knowledge organization and retrieval. We performed a trend analysis that enables researchers and clinicians to easily categorize and retrieve relevant papers, saving time and effort in literature review and information gathering as well as identification of emerging scientific trends within different disciplines. Moreover, the extendibility of the model to other scientific fields broadens its impact in facilitating research and trend analysis across diverse disciplines.

Abstract Introduction The purpose of this study was to evaluate the capabilities of large language models such as Chat Generative Pretrained Transformer (ChatGPT) to diagnose glaucoma based on specific clinical case descriptions with comparison to the performance of senior ophthalmology resident trainees. Methods We selected 11 cases with primary and secondary glaucoma from a publicly accessible online database of case reports. A total of four cases had primary glaucoma including open-angle, juvenile, normal-tension, and angle-closure glaucoma, while seven cases had secondary glaucoma including pseudo-exfoliation, pigment dispersion glaucoma, glaucomatocyclitic crisis, aphakic, neovascular, aqueous misdirection, and inflammatory glaucoma. We input the text of each case detail into ChatGPT and asked for provisional and differential diagnoses. We then presented the details of 11 cases to three senior ophthalmology residents and recorded their provisional and differential diagnoses. We finally evaluated the responses based on the correct diagnoses and evaluated agreements. Results The provisional diagnosis based on ChatGPT was correct in eight out of 11 (72.7%) cases and three ophthalmology residents were correct in six (54.5%), eight (72.7%), and eight (72.7%) cases, respectively. The agreement between ChatGPT and the first, second, and third ophthalmology residents were 9, 7, and 7, respectively. Conclusions The accuracy of ChatGPT in diagnosing patients with primary and secondary glaucoma, using specific case examples, was similar or better than senior ophthalmology residents. With further development, ChatGPT may have the potential to be used in clinical care settings, such as primary care offices, for triaging and in eye care clinical practices to provide objective and quick diagnoses of patients with glaucoma.

This paper presents the data set of Optic coherence tomography (OCT) and fundus Images of human eye. The OCT machine TOPCON'S 3D OCT-1000 camera is employed to acquire the images. The dataset is comprised of 50 images which includes control and glaucomatous images. For each OCT Image there is a corresponding fundus Image with annotation. Cup to disc ratio (CDR) values annotated by glaucoma specialists through fundus Images are provided in excel file. OCT images are optic nerve head (ONH) centred. Manually annotation is performed for the delineation of the Inner Limiting Membrane (ILM) Layer and Retinal pigmented epithelium (RPE) layer with the help of ophthalmologist. The data is valuable for the development of automated algorithm for glaucoma diagnosis. Keywords: Fundus images, OCT images, Optic nerve head (ONH), Cup to disc ratio (CDR) and glaucoma

Glaucoma is an asymptomatic neurological disease. It causes damage to optic nerve due to increased fluid pressure within eyes. In the proposed system, cup-to-disc-ratio has been computed considering internal layers of the retina using spectral-domain optical coherence tomography images. In the cupdiameter-calculation process, cup contour has been extracted from inner-limiting-membrane (ILM) layer. The paper introduces a new method to improve the precision of the ILM-layer extraction. It also employs a novel technique to refine contour of an ILM layer. The novel method has outperformed interpolation and Bezier curve fitting in term of outliers' removal and surface refinement. In the disc-diameter-calculation process, the retinal-pigment-epithelium (RPE) layer end points have been used to define disc margin. Prior to RPE-layer extraction, ILM-Layer removal has been done by an innovative strategy to locate and remove ILM-layer. Finally, precise RPE-layer extraction has been done based on the novel thickness-value (TV) estimation method. Furthermore, a new criterion for cup edges determination, based on the mean value of RPE-layer end points, is proposed. The proposed system has shown a clear precedence over its contemporary systems in terms of accuracy and handling of acute cases. Satisfactory results have been obtained when compared with the clinical results.

Background: The prevalence of genital tract malignancies is rising worldwide making it one of the most common malignancies in females after breast cancer. Despite cervical cancer being the most commonly occurring cancer globally, a multitude of studies in Pakistan have identified ovarian cancer as the most commonly occurring malignant tumor. The morbidity and mortality rates from genital tract tumors are considerably higher in developing countries. In developed countries, screening tests for early diagnosis of a malignancy followed by effective treatment are readily available. The objective of this study is to study the pattern and relative frequency of the malignancies involving the genital tract among women admitted to Ziauddin Hospital in 2015. Methods: The study is a retrospective, observational analysis of the case files of female patients admitted into Ziauddin University Hospital, North Nazimabad. The patients coming to the OPD have been followed for the last 3 years to check for the suspected presence of a genital tract malignancy. Results: A total of 55 patients over the last three years were confirmed to have some form of genital tract malignancy. Ovarian cancer accounted for 49% of the outcome, with around 23.6% of the cases diagnosed at stages 3 and 4. Endometrial cancer was the second most common cancer, with 27.2% being diagnosed at stage 1. Cancers of the cervix and vulva accounted for 3.6% and 1.8% respectively. No cancer of the vagina was documented. Conclusion: The most common genital tract malignancy in female patients at Ziauddin University Hospital is ovarian cancer, in contrast to the vast majority of developing countries where cervical cancer is the leading cause of genitourinary malignancies.

Background: Cord blood relative telomere length(RTL), the biological chronometer which can determine life span, has been found highly variable at birth due to fetal programming under the influence of different risk factors and markers of biological aging. This study was designed to compare RTL of maternal and new born and its association with blood groups of mothers as well as other related maternal and new born parameters. Methods: Maternal and cord RTL’s were measured in 250 samples using a quantitative real-time PCR method. Correlation was used to check the association of maternal and cord blood RTL. Kruskal Wallis was used to examine the comparison of cord blood RTL with maternal demographics. Results: Maternal mean age (Mean±S.D) was 27.17± 5.11 and paternal mean age was 34.03 ± 4.90. Maternal and cord RTL showed 0.01 level of significance at 95% confidence level. Cord blood RLT kilo base pair (kbp) (6.76± 1.35) was higher than maternal RTL (6.43± 1.35).Maternal blood groups were distributed highest (n=84, 33%) as B+ve and the lowest (n=1, 0.4%) as B-ve. Regarding maternal telomere length the longest was (6.68 ±1.34) in A+ve group and A-ve had the smallest length 5.32 ±0.95, however the results were not statistically significant (0.134). On contrary cord blood RTL was longest in O-ve blood group, 6.95±0.56and smallest in B-ve 5.41 RTL with p-value 0.159. Conclusion: Cord Blood RLT was higher than in maternal blood in target population of Karachi, Pakistan. The longest maternal RTL was in A+ve and smallest in O+ve cord blood. Majority of mothers were B+ve followed by O+ve.

Background: The foremost cause of maternal morbidity and mortality in third-world countries is Post-Partum Hemorrhage (PPH). To prevent PPH, often Misoprostol is used. This study aimed to find out the efficacy and safety profile of Misoprostol when administered rectally in comparison to oxytocin, to avoid the incidence of PPH. Methods: This quasi-experimental study on n=72 patients were performed at Ziauddin University and Hospital Karachi for 6 months (2021). The patients were divided into two groups. Oxytocin (10 IU) was administered intramuscularly to Group A(n=36) while 400ug of Misoprostol was given to group B(n=36) rectally, at the time of delivery. The level of hemoglobin was checked at the time of admission and 24 hours after delivery and side effects were also monitored. Chi-square and t-test were applied between both groups and a p-value

BACKGROUND: Nicotine use is reported more in Schizophrenic patients as compared to general population, which results is decrease efficacy of psychotropic medicines, thus increased severity of illness. OBJECTIVE: The objective of this study is to evaluate the severity of schizophrenia following tobacco use. METHODOLOGY: This cross-sectional study was conducted at department of Psychiatry Civil Hospital Karachi, Pakistan. Patients were recruited using convenience sampling technique. Clinical symptoms in patients were measured by the 18-item Brief Psychiatric Rating Scale (BPRS), scored 1–7 on each item. RESULTS: Out of the 276 patients included in the study, more than half of the participants were tobacco addicts (n=150, 54%). The statistical analysis showed significant relation (p value=0.04) between the Brief Psychiatric Rating Scale (BPRS) and tobacco addiction. CONCLUSION: This study found that schizophrenia patients with tobacco addiction had worse psychiatric symptoms as compared to those who are not addicted to tobacco.