Application of EfficientNet-B0 and GRU-based deep learning on classifying the colposcopy diagnosis of precancerous cervical lesions.

BACKGROUND: Colposcopy is indispensable for the diagnosis of cervical lesions. However, its diagnosis accuracy for high-grade squamous intraepithelial lesion (HSIL) is at about 50%, and the accuracy is largely dependent on the skill and experience of colposcopists. The advancement in computational power made it possible for the application of artificial intelligence (AI) to clinical problems. Here, we explored the feasibility and accuracy of the application of AI on precancerous and cancerous cervical colposcopic image recognition and classification. METHODS: The images were collected from 6002 colposcopy examinations of normal control, low-grade squamous intraepithelial lesion (LSIL), and HSIL. For each patient, the original, Schiller test, and acetic-acid images were all collected. We built a new neural network classification model based on the hybrid algorithm. EfficientNet-b0 was used as the backbone network for the image feature extraction, and GRU(Gate Recurrent Unit)was applied for feature fusion of the three modes examinations (original, acetic acid, and Schiller test). RESULTS: The connected network classifier achieved an accuracy of 90.61% in distinguishing HSIL from normal and LSIL. Furthermore, the model was applied to "Trichotomy", which reached an accuracy of 91.18% in distinguishing the HSIL, LSIL and normal control at the same time. CONCLUSION: Our results revealed that as shown by the high accuracy of AI in the classification of colposcopic images, AI exhibited great potential to be an effective tool for the accurate diagnosis of cervical disease and for early therapeutic intervention in cervical precancer.

Thyroid Function Test Abnormalities in Twin Pregnancies.

Background: Compared with singletons, a twin pregnancy is associated with a larger thyroid hormone demand and an increased stimulation of gestational thyroid function due to higher concentrations of human chorionic gonadotropin. However, such effects have been sparsely quantified. The aim of this study was to evaluate thyroid function and thyroid function test abnormalities in twin pregnancies during early and late pregnancy compared with singletons. Methods: We included 1208 twin pregnancies and 46,834 singleton pregnancies with thyroid function tests available. Thyroid function test abnormalities were defined using population-based reference ranges. The analyses were adjusted for potential confounders including maternal age and body mass index. Results: Compared with singletons, a twin pregnancy was associated with a lower thyrotropin (TSH) (ß = -0.46 [95% confidence interval, CI -0.49 to -0.44], p < 0.001) and a higher free thyroxine (fT4) (ß = 0.91 [CI 0.69-1.16], p < 0.001) during early pregnancy. During late pregnancy, a twin pregnancy was associated with a higher TSH (ß = 0.35 [CI 0.29-0.42], p < 0.001) while fT4 did not differ (ß = -0.11 [CI -0.22 to 0.01], p = 0.065). During early pregnancy, a twin pregnancy was associated with a higher risk of overt hyperthyroidism (odds ratio, OR = 7.49 [CI 6.02-9.33], p < 0.001), subclinical hyperthyroidism (OR = 5.26 [CI 4.17-6.64], p < 0.001), and isolated hypothyroxinemia (OR = 1.89 [CI 1.43-2.49], p < 0.001), but with a lower risk of subclinical hypothyroidism (OR = 0.27 [CI 0.13-0.54], p < 0.001). In late pregnancy, a twin pregnancy was associated with a higher risk of subclinical hypothyroidism (OR = 4.05 [CI 3.21-5.11], p < 0.001), isolated hypothyroxinemia (OR = 1.48 [CI 1.04-2.10], p = 0.028), and subclinical hyperthyroidism (OR = 1.76 [CI 1.27-2.43], p < 0.001). Conclusions: During early pregnancy, a twin pregnancy was associated with a higher thyroid function and a higher risk of (subclinical) hyperthyroidism, as well as a higher risk of isolated hypothyroxinemia. During late pregnancy, a twin pregnancy was associated with a higher TSH concentration and a higher risk of subclinical hypothyroidism, as well as a persistently higher risk of isolated hypothyroxinemia and subclinical hyperthyroidism. The study was approved by Chinese Clinical Trial Registry (registration no. ChiCTR1800014394).

Study protocol for the Sino-Canadian Healthy Life Trajectories Initiative (SCHeLTI): a multicentre, cluster-randomised, parallel-group, superiority trial of a multifaceted community-family-mother-child intervention to prevent childhood overweight and obesity.

INTRODUCTION: Childhood overweight and obesity (OWO) is a primary global health challenge. Childhood OWO prevention is now a public health priority in China. The Sino-Canadian Healthy Life Trajectories Initiative (SCHeLTI), one of four trials being undertaken by the international HeLTI consortium, aims to evaluate the effectiveness of a multifaceted, community-family-mother-child intervention on childhood OWO and non-communicable diseases risk. METHODS AND ANALYSIS: This is a multicentre, cluster-randomised, controlled trial conducted in Shanghai, China. The unit of randomisation is the service area of Maternal Child Health Units (N=36). We will recruit 4500 women/partners/families in maternity and district level hospitals. Participants in the intervention group will receive a multifaceted, integrated package of health promotion interventions beginning in preconception or in the first trimester of pregnancy, continuing into infancy and early childhood. The intervention, which is centred on a modified motivational interviewing approach, will target early-life maternal and child risk factors for adiposity. Through the development of a biological specimen bank, we will study potential mechanisms underlying the effects of the intervention. The primary outcome for the trial is childhood OWO (body mass index for age ≥85th percentile) at 5 years of age, based on WHO sex-specific standards. The study has a power of 0.8 (α=0.05) to detect a 30% risk reduction in the proportion of children with OWO at 5 years of age, from 24.4% in the control group to 17% in the intervention group. Recruitment was launched on 30 August 2018 for the pilot study and 10 January 2019 for the formal study. ETHICS AND DISSEMINATION: The study has been approved by the Medical Research Ethics Committee of the International Peace Maternity and Child Health Hospital in Shanghai, China, and the Research Ethics Board of the Centre Intégré Universitaire de Santé et Services Sociaux de l'Estrie-CHUS in Sherbrooke, Canada. Data sharing policies are consistent with the governance policy of the HeLTI consortium and government legislation. TRIAL REGISTRATION NUMBER: ChiCTR1800017773. PROTOCOL VERSION: November 11, 2020 (Version #5).

The Association Between Isolated Maternal Hypothyroxinemia in Early Pregnancy and Preterm Birth.

Background: The association between isolated maternal hypothyroxinemia (IMH) during pregnancy and preterm birth (PTB), especially for subtypes of PTB, is unclear. This study aimed at determining the association between IMH diagnosed in early pregnancy and PTB, with further investigation into various subtypes of PTB. Methods: This study included 41,911 pregnant women (963 with IMH and 40,948 euthyroid women) who underwent first-trimester prenatal screening at the International Peace Maternity and Child Health Hospital (IPMCH) in Shanghai, China between January 2013 and December 2016. PTB was defined as birth before 37 weeks of gestation. PTB was further classified into three clinically relevant groups to investigate the clinical heterogeneity of PTB: (a) preterm birth with premature rupture of membranes (PROM-PTB); (b) spontaneous preterm birth with intact membranes (S-PTB); and (c) medically-induced preterm birth (MI-PTB). The overall and sex-specific effect of IMH on PTB and various subtypes of PTB were estimated by using logistic regression in crude and adjusted models. Results: Pregnant women with IMH had an increased risk of PTB (odds ratio [OR]: 1.32 [95% confidence interval; CI: 1.02-1.70], p = 0.03) compared with women with euthyroid function. The increased risk of PTB is mainly driven by S-PTB (OR: 1.57 [CI: 1.11-2.24], p = 0.01), while women with early pregnancy IMH had no statistically significant increased risk of PROM-PTB and MI-PTB. The effect of IMH on PTB was modified by fetal sex (p-values for interaction = 0.04). More prominent effects were observed in women carrying a female fetus, while no statistically significant effects were found in women carrying a male fetus. Conclusions: This study revealed that pregnant women with IMH in early pregnancy have a higher risk of PTB compared with euthyroid women. The effect of IMH on PTB is mainly driven by S-PTB and is modified by fetal sex.