Ophthalmology Operation Note Encoding with Open-Source Machine Learning and Natural Language Processing.

INTRODUCTION: Accurate assignment of procedural codes has important medico-legal, academic, and economic purposes for healthcare providers. Procedural coding requires accurate documentation and exhaustive manual labour to interpret complex operation notes. Ophthalmology operation notes are highly specialised making the process time-consuming and challenging to implement. This study aimed to develop natural language processing (NLP) models trained by medical professionals to assign procedural codes based on the surgical report. The automation and accuracy of these models can reduce burden on healthcare providers and generate reimbursements that reflect the operation performed. METHODS: A retrospective analysis of ophthalmological operation notes from two metropolitan hospitals over a 12-month period was conducted. Procedural codes according to the Medicare Benefits Schedule (MBS) were applied. XGBoost, decision tree, Bidirectional Encoder Representations from Transformers (BERT) and logistic regression models were developed for classification experiments. Experiments involved both multi-label and binary classification, and the best performing model was used on the holdout test dataset. RESULTS: There were 1,000 operation notes included in the study. Following manual review, the five most common procedures were cataract surgery (374 cases), vitrectomy (298 cases), laser therapy (149 cases), trabeculectomy (56 cases), and intravitreal injections (49 cases). Across the entire dataset, current coding was correct in 53.9% of cases. The BERT model had the highest classification accuracy (88.0%) in the multi-label classification on these five procedures. The total reimbursement achieved by the machine learning algorithm was $184,689.45 ($923.45 per case) compared with the gold standard of $214,527.50 ($1,072.64 per case). CONCLUSION: Our study demonstrates accurate classification of ophthalmic operation notes into MBS coding categories with NLP technology. Combining human and machine-led approaches involves using NLP to screen operation notes to code procedures, with human review for further scrutiny. This technology can allow the assignment of correct MBS codes with greater accuracy. Further research and application in this area can facilitate accurate logging of unit activity, leading to reimbursements for healthcare providers. Increased accuracy of procedural coding can play an important role in training and education, study of disease epidemiology and improve research ways to optimise patient outcomes.

The Role of Tocilizumab in Glucocorticoid Resistant Giant Cell Arteritis: A Case Series and Literature Review.

BACKGROUND: Glucocorticoid-resistant giant cell arteritis (GCA) describes a subgroup of patients whose visual acuity further declines despite immediate administration of high-dose intravenous glucocorticoids. Tocilizumab, a recombinant humanized anti-interleukin-6 receptor monoclonal antibody may be used as "rescue" therapy in these cases but requires more research. METHODS: We present a literature review on tocilizumab as rescue therapy and a retrospective case series of 5 consecutive glucocorticoid (GC) resistant, temporal artery biopsy (TAB)-positive [corrected] patients. RESULTS: The use of tocilizumab as rescue therapy for GC-resistant GCA is limited to 3 case reports. Two cases saw visual acuity improvement in the newly affected contralateral eye after 8 mg/kg of intravenous tocilizumab (from 6/60 to 6/15 in one case and hand motion to 6/6 in another). The third described stabilization and prevention of further best-corrected visual acuity (BCVA) decline. All 5 of our patients presented with acute monocular vision loss. BCVA ranged from 6/12 to light perception (LP). All patients were promptly commenced on 1 g intravenous methylprednisolone daily. Weekly 162 mg of subcutaneous tocilizumab was commenced once contralateral eye involvement was noted. Tocilizumab resulted in bilateral BCVA gains in 2 cases, recovery of the contralateral eye in one, and no effect in the remaining 2 cases. BCVA recovery was no light perception to 6/6 after 6 weeks of tocilizumab. Tocilizumab had no effect in cases with severe vision loss and high C-reactive protein on presentation. CONCLUSIONS: We agree with existing hypothesis that tocilizumab likely prevents a "pending" central retinal artery occlusion by maintaining retinal vasculature perfusion. Our case series suggests that there is a role for tocilizumab as "rescue" therapy for GC-resistant GCA, where vision loss would otherwise be imminent.

Machine learning in the prediction of medical inpatient length of stay.

Length of stay (LOS) estimates are important for patients, doctors and hospital administrators. However, making accurate estimates of LOS can be difficult for medical patients. This review was conducted with the aim of identifying and assessing previous studies on the application of machine learning to the prediction of total hospital inpatient LOS for medical patients. A review of machine learning in the prediction of total hospital LOS for medical inpatients was conducted using the databases PubMed, EMBASE and Web of Science. Of the 673 publications returned by the initial search, 21 articles met inclusion criteria. Of these articles the most commonly represented medical specialty was cardiology. Studies were also identified that had specifically evaluated machine learning LOS prediction in patients with diabetes and tuberculosis. The performance of the machine learning models in the identified studies varied significantly depending on factors including differing input datasets and different LOS thresholds and outcome metrics. Common methodological shortcomings included a lack of reporting of patient demographics and lack of reporting of clinical details of included patients. The variable performance reported by the studies identified in this review supports the need for further research of the utility of machine learning in the prediction of total inpatient LOS in medical patients. Future studies should follow and report a more standardised methodology to better assess performance and to allow replication and validation. In particular, prospective validation studies and studies assessing the clinical impact of such machine learning models would be beneficial.

Professionalism and clinical short answer question marking with machine learning.

Machine learning may assist in medical student evaluation. This study involved scoring short answer questions administered at three centres. Bidirectional encoder representations from transformers were particularly effective for professionalism question scoring (accuracy ranging from 41.6% to 92.5%). In the scoring of 3-mark professionalism questions, as compared with clinical questions, machine learning had a lower classification accuracy (P < 0.05). The role of machine learning in medical professionalism evaluation warrants further investigation.

The use of frontalis sling in the management of variable ptosis secondary to congenital myasthenic syndrome.

Congenital myasthenic syndrome (CMS) describes a group of rare inherited disorders caused by impaired neuromuscular transmission at the motor endplate. Common ophthalmic manifestations associated with CMS include ptosis and ophthalmoplegia. A 19-year-old female presented with variable day-to-day ptosis secondary to CMS that was refractory to medical therapy. Bilateral silicone frontalis slings were used to stabilise the upper lid height and reduce fluctuation in severity of ptosis. Blepharoptosis surgery has been performed in patients with chronic myasthenia gravis (MG), but rarely in the setting of CMS. Blepharoptosis surgery in CMS patients with variable ptosis is difficult due to the risk of upsetting the original lid position and developing post-operative exposure keratopathy. Our case demonstrates that the frontalis sling procedure may be considered as an option in the management of variable blepharoptosis secondary to CMS.

Medical student knowledge and critical appraisal of machine learning: a multicentre international cross-sectional study.

To utilise effectively tools that employ machine learning (ML) in clinical practice medical students and doctors will require a degree of understanding of ML models. To evaluate current levels of understanding, a formative examination and survey was conducted across three centres in Australia, New Zealand and the United States. Of the 245 individuals who participated in the study (response rate = 45.4%), the majority had difficulty with identifying weaknesses in model performance analysis. Further studies examining educational interventions addressing such ML topics are warranted.

Triaging ophthalmology outpatient referrals with machine learning: A pilot study.

IMPORTANCE: Triaging of outpatient referrals to ophthalmology services is required for the maintenance of patient care and appropriate resource allocation. Machine learning (ML), in particular natural language processing, may be able to assist with the triaging process. BACKGROUND: To determine whether ML can accurately predict triage category based on ophthalmology outpatient referrals. DESIGN: Retrospective cohort study. PARTICIPANTS: The data of 208 participants was included in the project. METHODS: The synopses of consecutive ophthalmology outpatient referrals at a tertiary hospital were extracted along with their triage categorizations. Following pre-processing, ML models were applied to determine how accurately they could predict the likely triage categorization allocated. Data was split into training and testing sets (75%/25% split). ML models were tested on an unseen test set, after development on the training dataset. MAIN OUTCOME MEASURE: Area under the receiver operator curve (AUC) for category one vs non-category one classification. RESULTS: For the main outcome measure, convolutional neural network (CNN) provided the best AUC (0.83) and accuracy on the test set (0.81), with the artificial neural network (AUC 0.81 and accuracy 0.77) being the next best performing model. When the CNN was applied to the classification task of identifying which referrals should be allocated a category one vs category two vs category three priority, a lower accuracy was achieved (0.65). CONCLUSIONS AND RELEVANCE: ML may be able to accurately assist with the triaging of ophthalmology referrals. Future studies with data from multiple centres and larger sample sizes may be beneficial.

Clinical Application of Temporary External Fixator for Immediate Mandibular Reconstruction.

Immediate mandibular reconstruction is always necessary for the patients receiving segmental mandibulectomy to recover the facial contour and function of occlusion. When 3D modeling is unavailable, temporary external fixator is necessary to maintain the occlusion relationship and facial contour. In this study, we introduce the clinical application of temporary external fixator for immediate mandibular reconstruction in patients receiving segmental mandibulectomy, which consists of 2 anchor claws, 2 all-round retentive arms, and 1 central locking structure. From August 2016 to September 2017, temporary external fixator was applied in 13 patients. Clinical data of gender, age, surgical procedure, duration of operation, and clinical outcomes were recorded and analyzed. Among the 13 patients, there were 4 men and nine women whose ages ranged from 21 to 64 (mean 47.7) years old. There were 9 benign and 4 malignant lesions. All lesions expended at the buccal side of mandible. 12 fibular flaps and 1 vascularized iliac bone graft were used. The mandibular defect ranged from 6 to 14 (mean 10) cm. The operation duration of surgery ranged from 5 to 10 (mean 7) hours. All flaps survived with primary healing. The occlusion and facial contour were good, without significant changes of the length of mandibular body and width of mandible before and after surgery. No functional sequelae were noted at the donor sites. From these results, the temporary external fixator is easy to operate; the surgical procedure is simple and time-saving for surgeon when 3D modeling is unavailable. The indication for temporary external fixator usage is the mandibular lesion growing outward to cheek soft tissue.

Mutation allele frequency threshold does not affect prognostic analysis using next-generation sequencing in oral squamous cell carcinoma.

BACKGROUND: With the development of sequencing technologies, there may be some disputes on sequencing analysis. The aim of this study was to investigate different allele frequency thresholds of mutations in targeted genes on prognostic analyses using a panel of cancer associated gene exons (CAGE) in oral squamous cell carcinoma (OSCC). METHODS: Forty-six patients were included in this study. Twelve genes were sequenced and analyzed using next-generation sequencing from formalin-fixed paraffin-embedded tissues. Allele frequency thresholds of 10, 5, and 3% were used for prognostic analyses. RESULTS: With a mean sequence depth of 3199-fold, 99% of CAGE were represented by at least 10 reads. Ninety-four non-synonymous (missense [70.2%], nonsense [11.7%], splice site [10.6%], and insertion/deletion [7.5%]) mutations were detected in 40 OSCC patients with an allele frequency threshold of 10%. TP53 (78.3%), NOTCH1 (30.4%), CASP8 (13.0%), CDKN2A (10.9%), and CDH1 (6.5%) were the most frequently mutated genes. Using allele frequency thresholds of 10, 5, and 3%, there were no significant differences in clinical outcomes between patients with non-synonymous mutations and wild type genotypes. CONCLUSIONS: TP53, NOTCH1, CASP8, CDKN2A, and CDH1 are the most frequently mutated genes in OSCC patients. The allele frequency threshold used in this study does not affect the results of clinical outcome analysis.

Decreased expression of glutathione S-transferase pi correlates with poorly differentiated grade in patients with oral squamous cell carcinoma.

BACKGROUND: Glutathione S transferase pi (GSTP1) is a member of phase II detoxification enzymes as a major regulator of cell signaling in response to stress, hypoxia, growth factors, and other stimuli. The clinical role of GSTP1 in cancer is still unclear. The aim of this study was to investigate the serum GSTP1 level in patients with oral squamous cell carcinoma (OSCC) and the GSTP1 expression in tissue samples from patients with OSCC and OSCC lines. METHODS: One hundred and sixty-six patients with OSCC and 120 normal persons were used to screen potential serum peptide biomarkers using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Serum GSTP1 concentration was detected in 18 patients with OSCC and 18 normal persons using ELISA. Immunohistochemistry was used to detect GSTP1 expression in tissue samples from twenty-eight OSCC patients. Western blot and real-time PCR were used to detect GSTP1 expression in nine OSCC lines. RESULTS: Decreased GSTP1 concentration was found in the patients with OSCC compared with the normal persons by MALDI-TOF-MS, which was then confirmed by ELISA (P = 0.019). Decreased GSTP1 mRNA level and protein expression were also found in the OSCC lines. Decreased GSTP1 expression was found correlating with pathological differentiation grade in the tissue samples from OSCC patients, a lower GSTP1 expression indicating a poorer pathological differentiation grade (P = 0.041). CONCLUSIONS: These results suggest that decreased GSTP1 expression in patients with OSCC and a lower GSTP1 expression indicating a poorer pathological differentiation grade in OSCC tissue samples.

Knowledge-Driven Experimental Discovery of Ce-Based Metal Oxide Composites for Selective Catalytic Reduction of NOx with NH3 through Interpretable Machine Learning.

Mining the scientific literature, combined with data-driven methods, may assist in the identification of optimized catalysts. In this paper, we employed interpretable machine learning to discover ternary metal oxides capable of selective catalytic reduction of nitrogen oxides with ammonia (NH3-SCR). Specifically, we devised a machine learning framework utilizing extreme gradient boosting (XGB), identified for its optimal performance, and SHapley Additive exPlanations (SHAP) to evaluate a curated database of 5654 distinct metal oxide composite catalytic systems containing cerium (Ce) element, with records of catalyst composition and preparation and reaction conditions. By virtual screening, this framework precisely pinpointed a CeO2-MoO3-Fe2O3 catalyst with superior NOx conversion, N2 selectivity, and resistance to H2O and SO2, as confirmed by empirical evaluations. Subsequent characterization affirmed its favorable structural, chemical bulk properties and reaction mechanism. Demonstrating the efficacy of combining knowledge-driven techniques with experimental validation and analysis, our strategy charts a course for analogous catalyst discoveries.

Vital sign measurements demonstrate terminal digit bias and boundary effects.

OBJECTIVE: The measurement and recording of vital signs may be impacted by biases, including preferences for even and round numbers. However, other biases, such as variation due to defined numerical boundaries (also known as boundary effects), may be present in vital signs data and have not yet been investigated in a medical setting. We aimed to assess vital signs data for such biases. These parameters are clinically significant as they influence care escalation. METHODS: Vital signs data (heart rate, respiratory rate, oxygen saturation and systolic blood pressure) were collected from a tertiary hospital electronic medical record over a 2-year period. These data were analysed using polynomial regression with additional terms to assess for underreporting of out-of-range observations and overreporting numbers with terminal digits of 0 (round numbers), 2 (even numbers) and 5. RESULTS: It was found that heart rate, oxygen saturation and systolic blood pressure demonstrated 'boundary effects', with values inside the 'normal' range disproportionately more likely to be recorded. Even number bias was observed in systolic heart rate, respiratory rate and blood pressure. Preference for multiples of 5 was observed for heart rate and blood pressure. Independent overrepresentation of multiples of 10 was demonstrated in heart rate data. CONCLUSION: Although often considered objective, vital signs data are affected by bias. These biases may impact the care patients receive. Additionally, it may have implications for creating and training machine learning models that utilise vital signs data.

Prevalence and causes of visual impairment and blindness in Lao People's Democratic Republic: the Vientiane Eye Study.

AIMS: To determine the prevalence and causes of visual impairment and blindness in Vientiane Province, the Lao People's Democratic Republic (Lao PDR). METHODS: We conducted a population-based, cross-sectional study of 1264 participants aged 40 years and older from urban and rural areas of Vientiane Province. The ophthalmic examination included presenting and pinhole Snellen visual acuity (VA) with an illiterate E chart, slit-lamp examination of the anterior segment and dilated stereoscopic fundus examination. Visual impairment and blindness were defined by both presenting and pinhole VA based on the better eye according to WHO criteria: VA worse than 6/12 to 6/18 for mild impairment, VA worse than 6/18 to 6/60 for moderate impairment, VA worse than 6/60 to 3/60 for severe impairment and VA worse than 3/60 for blindness. RESULTS: Comprehensive ophthalmic examinations were performed on 1264 participants (77.8% participation rate). Population-weighted prevalence of presenting bilateral blindness was 1.4% (95% CI 0.8 to 2.0) and bilateral visual impairment was 22.4% (95% CI 14.7 to 30.1). After pinhole correction, the corresponding prevalence of blindness was 1.3% (95% CI 0.8 to 2.0) and that of visual impairment was 12.6% (95% CI 8.2 to 16.9). Cataract was the leading cause of presenting bilateral blindness (52.9%), whereas uncorrected refractive error was the predominant cause of presenting visual impairment (40.3%). CONCLUSIONS: Visual impairment and blindness remain major public health problems in Lao PDR. There is an ongoing need to fund ophthalmic care resources and community education programmes to improve access to healthcare in this region.

Prevalence of Refractive Error in Vientiane Province, Lao People's Democratic Republic.

PURPOSE: To determine the prevalence of adult refractive error and associated risk factors in the Vientiane Province of the Lao People's Democratic Republic. METHODS: Population-based, cross-sectional ophthalmic survey of individuals ≥ 40 years of age in Vientiane Province. Suitable refractive data was available in 1058 individuals. Demographic data, which included age and gender, was obtained from all participants. Smoking status, presence of diabetes and hypertension was also recorded. The ophthalmic examination included autorefraction, grading of cataract, applanation tonometry and ultrasound pachymetry for ocular biometry, including axial length. RESULTS: Mean refractive error measured -0.36 diopters (D) (standard deviation [SD], 1.41) and mean cylindrical error measured -0.33 D (SD 0.95). Myopia ≤ -0.5 D and ≤ -5.0 D occurred in 53.2% (95% confidence interval [CI]: 43.7 to 62.6) and 2.0% (95% CI: 0.4 to 3.6) of participants, respectively. There was a correlation between myopia and both age and higher grading of nuclear cataract (p < .001). Hyperopia ≥ +0.5 D was present in 26.4% of participants and was associated with increasing age (p < .001). Astigmatism was present in 55.8% (95% CI: 51.5 to 60.2) of the population and was associated with increased nuclear cataract (p < .001). Urban participants had a reduced prevalence of myopia compared with rural participants. CONCLUSION: Myopia was associated with younger age and a higher grade of nuclear cataract. The prevalence of myopia in the adult population of Vientiane Province was higher than that reported in neighbouring Asian regions and contributed to low vision.

Immune Checkpoint Inhibitors and Optic Neuropathy: A Systematic Review.

INTRODUCTION: Immune checkpoint inhibitors are a class of monoclonal antibodies that are used as a mainstay of immunotherapy for multiple solid organ malignancies. With the recent increase in popularity of these agents, immune-related adverse events including optic neuropathy are becoming more frequently reported. This review aims to explore the association between immune checkpoint inhibitors and optic neuropathy through analysis of incidence, clinical features, investigations, treatment, and patient outcomes. METHOD: A systematic search of the databases PubMed/MEDLINE, Embase, and CENTRAL was performed from inception to September 2022. Data collection and risk of bias analysis was subsequently conducted in accordance with the PRISMA guidelines. RESULTS: Eleven articles fulfilled the inclusion criteria. The results showed an increased incidence of optic neuropathy among patients receiving immune checkpoint inhibitor therapy compared to the general population. Presentation with painless reduced visual acuity and optic disc swelling was most common. Investigation findings were poorly documented. The only two patients who achieved full resolution of symptoms were treated with oral prednisolone. CONCLUSION: There is a strong association between immune checkpoint inhibitor therapy and development of optic neuropathy. Although it remains uncommon, the incidence of optic neuropathy in this population exceeds that of the general population. Future research is needed to further characterise the risk profiles of patients who are most likely to develop ICI-associated optic neuropathy, and treatment pathways for these patients.