ABSTRACT
Segmentation of organs and lesions could be employed for the express purpose of dosimetry in nuclear medicine, assisted image interpretations, and mass image processing studies. Deep leaning created liver and liver lesion segmentation on clinical 3D MRI data has not been fully addressed in previous experiments. To this end, the required data were collected from 128 patients, including their T1w and T2w MRI images, and ground truth labels of the liver and liver lesions were generated. The collection of 110 T1w-T2w MRI image sets was divided, with 94 designated for training and 16 for validation. Furthermore, 18 more datasets were separately allocated for use as hold-out test datasets. The T1w and T2w MRI images were preprocessed into a two-channel format so that they were used as inputs to the deep learning model based on the Isensee 2017 network. To calculate the final Dice coefficient of the network performance on test datasets, the binary average of T1w and T2w predicted images was used. The deep learning model could segment all 18 test cases, with an average Dice coefficient of 88% for the liver and 53% for the liver tumor. Liver segmentation was carried out with rather a high accuracy; this could be achieved for liver dosimetry during systemic or selective radiation therapies as well as for attenuation correction in PET/MRI scanners. Nevertheless, the delineation of liver lesions was not optimal; therefore, tumor detection was not practical by the proposed method on clinical data.
Subject(s)
Deep Learning , Image Processing, Computer-Assisted , Liver Neoplasms , Liver , Magnetic Resonance Imaging , Humans , Liver/diagnostic imaging , Liver Neoplasms/diagnostic imaging , Liver Neoplasms/pathology , Female , Male , Imaging, Three-Dimensional , Middle AgedABSTRACT
Aims: Growing evidence suggests that poor sleep health is associated with cardiovascular risk. However, research in this area often relies upon recollection dependent questionnaires or diaries. Accelerometers provide an alternative tool for measuring sleep parameters objectively. This study examines the association between wrist-worn accelerometer-derived sleep onset timing and cardiovascular disease (CVD). Methods and results: We derived sleep onset and waking up time from accelerometer data collected from 103â712 UK Biobank participants over a period of 7 days. From this, we examined the association between sleep onset timing and CVD incidence using a series of Cox proportional hazards models. A total of 3172 cases of CVD were reported during a mean follow-up period of 5.7 (±0.49) years. An age- and sex-controlled base analysis found that sleep onset time of 10:00 p.m.-10:59 p.m. was associated with the lowest CVD incidence. An additional model, controlling for sleep duration, sleep irregularity, and established CVD risk factors, did not attenuate this association, producing hazard ratios of 1.24 (95% confidence interval, 1.10-1.39; P < 0.005), 1.12 (1.01-1.25; P = 0.04), and 1.25 (1.02-1.52; P = 0.03) for sleep onset <10:00 p.m., 11:00 p.m.-11:59 p.m., and ≥12:00 a.m., respectively, compared to 10:00 p.m.-10:59 p.m. Importantly, sensitivity analyses revealed this association with increased CVD risk was stronger in females, with only sleep onset <10:00 p.m. significant for males. Conclusions: Our findings suggest the possibility of a relationship between sleep onset timing and risk of developing CVD, particularly for women. We also demonstrate the potential utility of collecting information about sleep parameters via accelerometry-capable wearable devices, which may serve as novel cardiovascular risk indicators.