METHOD: In the present study, three separate data cohorts containing 1288 breast lesions from three countries (Malaysia, Iran, and Turkey) were utilized for MLmodel development and external validation. The model was trained on ultrasound images of 725 breast lesions, and validation was done separately on the remaining data. An expert radiologist and a radiology resident classified the lesions based on the BI-RADS lexicon. Thirteen morphometric features were selected from a contour of the lesion and underwent a three-step feature selection process. Five features were chosen to be fed into the model separately and combined with the imaging signs mentioned in the BI-RADS reference guide. A support vector classifier was trained and optimized.
RESULTS: The diagnostic profile of the model with various input data was compared to the expert radiologist and radiology resident. The agreement of each approach with histopathologic specimens was also determined. Based on BI-RADS and morphometric features, the model achieved an area under the receiver operating characteristic (ROC) curve (AUC) of 0.885, which is higher than the expert radiologist and radiology resident performances with AUC of 0.814 and 0.632, respectively in all cohorts. DeLong's test also showed that the AUC of the ML protocol was significantly different from that of the expert radiologist (ΔAUCs = 0.071, 95%CI: (0.056, 0.086), P = 0.005).
CONCLUSIONS: These results support the possible role of morphometric features in enhancing the already well-excepted classification schemes.
METHODS AND FINDINGS: The web-based Joint Asia Diabetes Evaluation (JADE) platform provides a protocol to guide data collection for issuing a personalized JADE report including risk categories (1-4, low-high), 5-year probabilities of cardiovascular-renal events, and trends and targets of 4 risk factors with tailored decision support. The JADE program is a prospective cohort study implemented in a naturalistic environment where patients underwent nurse-led structured evaluation (blood/urine/eye/feet) in public and private outpatient clinics and diabetes centers in Hong Kong. We retrospectively analyzed the data of 16,624 Han Chinese patients with type 2 diabetes who were enrolled in 2007-2015. In the public setting, the non-JADE group (n = 3,587) underwent structured evaluation for risk factors and complications only, while the JADE (n = 9,601) group received a JADE report with group empowerment by nurses. In a community-based, nurse-led, university-affiliated diabetes center (UDC), the JADE-Personalized (JADE-P) group (n = 3,436) received a JADE report, personalized empowerment, and annual telephone reminder for reevaluation and engagement. The primary composite outcome was time to the first occurrence of cardiovascular-renal diseases, all-site cancer, and/or death, based on hospitalization data censored on 30 June 2017. During 94,311 person-years of follow-up in 2007-2017, 7,779 primary events occurred. Compared with the JADE group (136.22 cases per 1,000 patient-years [95% CI 132.35-140.18]), the non-JADE group had higher (145.32 [95% CI 138.68-152.20]; P = 0.020) while the JADE-P group had lower event rates (70.94 [95% CI 67.12-74.91]; P < 0.001). The adjusted hazard ratios (aHRs) for the primary composite outcome were 1.22 (95% CI 1.15-1.30) and 0.70 (95% CI 0.66-0.75), respectively, independent of risk profiles, education levels, drug usage, self-care, and comorbidities at baseline. We reported consistent results in propensity-score-matched analyses and after accounting for loss to follow-up. Potential limitations include its nonrandomized design that precludes causal inference, residual confounding, and participation bias.
CONCLUSIONS: ICT-assisted integrated care was associated with a reduction in clinical events, including death in type 2 diabetes in public and private healthcare settings.
MATERIALS AND METHODS: This is a retrospective study on patients with newly diagnosed hepatocellular carcinoma (HCC) at the University Malaya Medical Centre between 2011 and 2014. Survival times were analysed using the Kaplan- Meier procedure and comparison between groups was done using the log rank test.
RESULTS: The data of 190 patients was analysed. Chronic hepatitis B was the most common aetiology for HCC (43.7%), but a large proportion was cryptogenic or non-alcoholic steatohepatitis-related (41.6%). Only 11.1% were diagnosed early (BCLC Stage 0-A) while majority were diagnosed at an intermediate stage (BCLC Stage B, 53.7%). The median survival rate was significantly different between the different groups when either of the staging systems was used (p<0.05 for all comparisons). However, the two staging systems lacked agreement (weighted kappa 0.519, 95%CI: 0.449, 0.589) with significant difference in median survival rates between BCLC Stage A and HKLC Stage 2, and between BCLC Stage C and HKLC Stage 4.
CONCLUSION: Both staging systems were able to stratify patients according to survival, but they only had moderate agreement with significant differences observed in two groups of the staging systems.
AIM: To study factors associated with nonalcoholic steatohepatitis (NASH) and advanced fibrosis, and medical treatment of biopsy-proven nonalcoholic fatty liver disease (NAFLD) patients.
METHODS: Retrospective study of biopsy-proven NAFLD patients from centres in the GO ASIA Workgroup. Independent factors associated with NASH and with advanced fibrosis on binary logistic regression analyses in a training cohort were used for the development of their corresponding risk score, which were validated in a validation cohort.
RESULTS: We included 1008 patients from nine centres across eight countries (NASH 62.9%, advanced fibrosis 17.2%). Independent predictors of NASH were body mass index ≥30 kg/m2 , diabetes mellitus, dyslipidaemia, alanine aminotransferase ≥88 U/L and aspartate aminotransferase ≥38 U/L, constituting the Asia Pacific NASH risk score. A high score has a positive predictive value of 80%-83% for NASH. Independent predictors of advanced fibrosis were age ≥55 years, diabetes mellitus and platelet count <150 × 109 /L, constituting the Asia-Pacific NAFLD advanced fibrosis risk score. A low score has a negative predictive value of 95%-96% for advanced fibrosis. Only 1.7% of patients were referred for structured lifestyle program, 4.2% were on vitamin E, and 2.4% were on pioglitazone.
CONCLUSIONS: More severe liver disease can be suspected or ruled out based on factors identified in this study. Utilisation of structured lifestyle program, vitamin E and pioglitazone was limited despite this being a cohort of biopsy-proven NAFLD patients with majority of patients having NASH.
METHODS: This is a retrospective study of confirmed leptospirosis cases in Larut, Matang and Selama (LMS) district in Perak reported in 2016. The demographic, clinical presentation, laboratory result and clinical outcomes data were analysed and presented.
RESULTS: Forty-two patients with confirmed diagnosis of leptospirosis were included into the study. Majority of patients were males and Malays. The case fatality rate was 14.3%. Patients with leptospirosis present with variable clinical presentations and are commonly seen with coinfection. Patients 70-year-old and older, have clinical presentations suggestive of organ dysfunction and require intensive care are associated with higher mortality.
CONCLUSION: Leptospirosis is endemic in LMS district of Perak with high incidence and case fatality rate. The clinical presentation of leptospirosis is variable. Co-infection of leptospirosis with other acute febrile illness is common. Patients presenting with symptoms and signs of organ dysfunctions or require intensive care are associated with an increased odds of death.
METHODS: This is a cross-sectional study of all the children of HBsAg-positive mothers who delivered at the University of Malaya Medical Centre between 1993 and 2000.
RESULTS: A total of 60 HBsAg-positive mothers and their 154 children participated in the study. HBsAg was detected in four children (2.6%) while IgG antibody to the hepatitis B core antigen (anti-HBc IgG) was detected in seventeen children (11.0%). The mother's age at childbirth was significantly lower in the children with detectable HBsAg (22.5±6.1 years vs. 29.7±4.5 years, p=0.043) and anti-HBc IgG (26.6±6.1 years vs. 30.0±4.3 years, p=0.004). Children born in the 1980s were significantly more likely to have detectable HBsAg (18.8% vs. 0.7%, p=0.004) and anti-HBc IgG (37.5% vs. 8.0%, p=0.000) compared with those born later. All children with detectable HBsAg were born via spontaneous vaginal delivery, and hepatitis B immunoglobulin was either not given or the administration status was unknown. The majority of mothers with chronic HBV infection (70.4%) were not under any regular follow-up for their chronic HBV infection and the main reason was the lack of awareness of the need to do so (47.4%).
CONCLUSION: Transmission of HBV infection among children of HBsAg-positive mothers in Malaysia is low. However, attention needs to be given to the high rate of HBsAgpositive mothers who are not on any regular follow-up.
AIM: To determine how to use CAP in interpreting liver stiffness measurements.
METHODS: This is a secondary analysis of data from an individual patient data meta-analysis on CAP. The main exclusion criteria for the current analysis were unknown aetiology, unreliable elastography measurement and data already used for the same research question. Aetiology-specific liver stiffness measurement cut-offs were determined and used to estimate positive and negative predictive values (PPV/NPV) with logistic regression as functions of CAP.
RESULTS: Two thousand and fifty eight patients fulfilled the inclusion criteria (37% women, 18% NAFLD/NASH, 42% HBV, 40% HCV, 51% significant fibrosis ≥ F2). Youden optimised cut-offs were only sufficient for ruling out cirrhosis (NPV of 98%). With sensitivity and specificity-optimised cut-offs, NPV for ruling out significant fibrosis was moderate (70%) and could be improved slightly through consideration of CAP. PPV for significant fibrosis and cirrhosis were 68% and 55% respectively, despite specificity-optimised cut-offs for cirrhosis.
CONCLUSIONS: Liver stiffness measurement values below aetiology-specific cut-offs are very useful for ruling out cirrhosis, and to a lesser extent for ruling out significant fibrosis. In the case of the latter, Controlled Attenuation Parameter can improve interpretation slightly. Even if cut-offs are very high, liver stiffness measurements are not very reliable for ruling in fibrosis or cirrhosis.