Annals of the Academy of Medicine, Singapore最新文献

Introduction: Automated machine learning (autoML) removes technical and technological barriers to building artificial intelligence models. We aimed to summarise the clinical applications of autoML, assess the capabilities of utilised platforms, evaluate the quality of the evidence trialling autoML, and gauge the performance of autoML platforms relative to conventionally developed models, as well as each other.

Method: This review adhered to a prospectively registered protocol (PROSPERO identifier CRD42022344427). The Cochrane Library, Embase, MEDLINE and Scopus were searched from inception to 11 July 2022. Two researchers screened abstracts and full texts, extracted data and conducted quality assessment. Disagreement was resolved through discussion and if required, arbitration by a third researcher.

Results: There were 26 distinct autoML platforms featured in 82 studies. Brain and lung disease were the most common fields of study of 22 specialties. AutoML exhibited variable performance: area under the receiver operator characteristic curve (AUCROC) 0.35-1.00, F1-score 0.16-0.99, area under the precision-recall curve (AUPRC) 0.51-1.00. AutoML exhibited the highest AUCROC in 75.6% trials; the highest F1-score in 42.3% trials; and the highest AUPRC in 83.3% trials. In autoML platform comparisons, AutoPrognosis and Amazon Rekognition performed strongest with unstructured and structured data, respectively. Quality of reporting was poor, with a median DECIDE-AI score of 14 of 27.

Conclusion: A myriad of autoML platforms have been applied in a variety of clinical contexts. The performance of autoML compares well to bespoke computational and clinical benchmarks. Further work is required to improve the quality of validation studies. AutoML may facilitate a transition to data-centric development, and integration with large language models may enable AI to build itself to fulfil user-defined goals.

Introduction: Tuberculosis (TB) remains endemic in Singapore. Singapore's clinical practice guidelines for the management of tuberculosis were first published in 2016. Since then, there have been major new advances in the clinical management of TB, ranging from diagnostics to new drugs and treatment regimens. The National TB Programme convened a multidisciplinary panel to update guidelines for the clinical management of drug-susceptible TB infection and disease in Singapore, contextualising current evidence for local practice.

Method: Following the ADAPTE framework, the panel systematically reviewed, scored and synthesised English-language national and international TB clinical guidelines published from 2016, adapting recommendations for a prioritised list of clinical decisions. For questions related to more recent advances, an additional primary literature review was conducted via a targeted search approach. A 2-round modified Delphi process was implemented to achieve consensus for each recommendation, with a final round of edits after consultation with external stakeholders.

Results: Recommendations for 25 clinical questions spanning screening, diagnosis, selection of drug regimen, monitoring and follow-up of TB infection and disease were formulated. The availability of results from recent clinical trials led to the inclusion of shorter treatment regimens for TB infection and disease, as well as consensus positions on the role of newer technologies, such as computer-aided detection-artificial intelligence products for radiological screening of TB disease, next-generation sequencing for drug-susceptibility testing, and video observation of treatment.

Conclusion: The panel updated recommendations on the management of drug-susceptible TB infection and disease in Singapore.

Introduction: The most prevalent type of fragility fractures is osteoporotic vertebral fractures (OVFs). However, only a few studies have examined the relationship between anti-osteoporosis treatments and malignancy-related mortality following an OVF. The goal of this study is to determine the effect of anti-osteoporosis therapy on mortality in OVF patients with and without cancer.

Method: Data from older people over the age of 65 who were hospitalised for OVFs between 1 January 2003 and 31 December 2018 were analysed retrospectively. A total of 6139 persons getting osteoporosis treatment and 28,950 who did not receive treatment were analysed, together with 2 sets of patients, comprising cancer patients (794) and cancer-free patients (5342), using anti-osteoporosis medication or not, in 1:1 propensity score-matched analyses. The hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated.

Results: In all, 35,089 patients with OVFs were included in the population; 29,931 people (85.3%) were women, and the mean (standard deviation) age was 78.13 (9.27) years. Overall survival was considerably higher in those undergoing osteoporosis therapy. This was true both for those without cancer (adjusted HR 0.55; 95% CI 0.51-0.59; P<.0001) as well as those with cancer (adjusted HR 0.72; 95% CI 0.62-0.84; P<.0001). Even among cancer patients, those who received anti-osteoporotic drugs had a lower mortality rate than those who did not.

Conclusion: Our findings suggest that anti-osteoporosis therapy should be initiated regardless of the presence of cancer in the elderly, as it increases survival following OVFs.

Introduction: Two decades after the Rotterdam 2003 consensus workshop, there have been considerable advances in elucidating the pathophysiology and epidemiology of polycystic ovary syndrome (PCOS). This has prompted the re-examination of the features that characterise this common condition. Current definitions have led to great heterogeneity in the prevalence of PCOS and have contributed to inconsistent treatment protocols and assessment of therapeutic outcomes. Diagnosis is further complicated by the lack of universal agreement on threshold cut-offs for ovarian dysfunction and ethnic differences in hirsutism; both of which are key features in the definitions that are commonly used currently. These challenges often result in dissatisfaction with medical care among PCOS patients and their physicians.

Method: Our factor analysis mathematically identified anti-Mullerian hormone (AMH), associated polycystic ovarian morphology (PCOM) and serum testosterone as the only significant cluster associated with menstrual cycle length variability.

Results and conclusion: As such, we propose a simplified criteria wherein the presence of at least 2 of the 3 features below would be sufficient to define PCOS: (1) chronic oligo-ovulation or anovulation as indicated by oligomenorrhea (cycle lengths >35 days) or amenorrhea; (2) PCOM: raised AMH ≥37.0 pmol/L instead of transvaginal ultrasound assessment of ovaries; and (3) Androgen excess, or raised serum androgens above the laboratory reference for women. Further studies are required to examine whether the proposed criteria would reduce diagnostic confusion and improve care and outcomes, especially among patients of East Asian ethnicities.

Introduction: Determining the exact weight of children is a challenging task during emergency situations. Current guidelines recommend the use of length-based weight-estimating tapes. However, healthcare providers must either always carry the tapes or take time to locate them. Moreover, they may not know how to use them. To address these issues, we developed an augmented reality smartphone application for length-based weight estimation called the Paediatric Augmented Reality Scale (PARS). We evaluated its performance and compared it to that of the Broselow tape (BT) and Paediatric Advanced Weight Prediction in the Emergency Room extra-long and extra-large (PAWPER-XL) tape methods.

Method: A prospective, single-blinded cross-sectional study was conducted with children aged 1 month to 12 years who visited the emergency department of the tertiary university hospital in Bucheon, South Korea between July 2021 and February 2022. This study aimed to evaluate the measurement agreement and performance of 3 methods: BT, PAWPER-XL and PARS.

Results: In all, 1090 participants were enrolled, and 639 (58.6%) were male. The mean age of the participants was 4.1 ± 2.8 years, with a mean height of 102.7 ± 21.7 cm and mean weight of 18.8 ± 9.5 kg. Compared to BT and PAWPER-XL, PARS exhibited lower mean absolute percentage error (9.60%) and root mean square percentage error (3.02%). PARS achieved a higher proportion of weights estimated within 10% of the actual weight (63.21%), outperform-ing BT (57.25%) and PAWPER-XL (62.47%). The intraclass correlation coefficients for the actual and estimated weights of BT, PAWPER-XL and PARS were 0.952, 0.969 and 0.973, respectively (P<0.001).

Conclusion: PARS exhibited a modestly better performance than BT and PAWPER-XL in estimating body weight. PARS-estimated body weights correlated fairly accurately with the actual body weights. PARS holds potential utility in paediatric emergencies.