A Hematologist-Level Deep Learning Algorithm (BMSNet) for Assessing the Morphologies of Single Nuclear Balls in Bone Marrow Smears: Algorithm Development.

BACKGROUND: Bone marrow aspiration and biopsy remain the gold standard for the diagnosis of hematological diseases despite the development of flow cytometry (FCM) and molecular and gene analyses. However, the interpretation of the results is laborious and operator dependent. Furthermore, the obtained results exhibit inter- and intravariations among specialists. Therefore, it is important to develop a more objective and automated analysis system. Several deep learning models have been developed and applied in medical image analysis but not in the field of hematological histology, especially for bone marrow smear applications. OBJECTIVE: The aim of this study was to develop a deep learning model (BMSNet) for assisting hematologists in the interpretation of bone marrow smears for faster diagnosis and disease monitoring. METHODS: From January 1, 2016, to December 31, 2018, 122 bone marrow smears were photographed and divided into a development cohort (N=42), a validation cohort (N=70), and a competition cohort (N=10). The development cohort included 17,319 annotated cells from 291 high-resolution photos. In total, 20 photos were taken for each patient in the validation cohort and the competition cohort. This study included eight annotation categories: erythroid, blasts, myeloid, lymphoid, plasma cells, monocyte, megakaryocyte, and unable to identify. BMSNet is a convolutional neural network with the YOLO v3 architecture, which detects and classifies single cells in a single model. Six visiting staff members participated in a human-machine competition, and the results from the FCM were regarded as the ground truth. RESULTS: In the development cohort, according to 6-fold cross-validation, the average precision of the bounding box prediction without consideration of the classification is 67.4%. After removing the bounding box prediction error, the precision and recall of BMSNet were similar to those of the hematologists in most categories. In detecting more than 5% of blasts in the validation cohort, the area under the curve (AUC) of BMSNet (0.948) was higher than the AUC of the hematologists (0.929) but lower than the AUC of the pathologists (0.985). In detecting more than 20% of blasts, the AUCs of the hematologists (0.981) and pathologists (0.980) were similar and were higher than the AUC of BMSNet (0.942). Further analysis showed that the performance difference could be attributed to the myelodysplastic syndrome cases. In the competition cohort, the mean value of the correlations between BMSNet and FCM was 0.960, and the mean values of the correlations between the visiting staff and FCM ranged between 0.952 and 0.990. CONCLUSIONS: Our deep learning model can assist hematologists in interpreting bone marrow smears by facilitating and accelerating the detection of hematopoietic cells. However, a detailed morphological interpretation still requires trained hematologists.

Dominance of the inactive Asian variant over activity and protein contents of mitochondrial aldehyde dehydrogenase 2 in human liver.

BACKGROUND: It has been well documented that a variant allele of mitochondrial aldehyde dehydrogenase 2 (ALDH2), ALDH2*2, commonly occurs in East Asians but rarely in other ethnic populations. This unique allelic variation significantly influences drinking behavior and susceptibility to development of alcoholism. Previous structural, functional, and cellular studies indicate that the resulting variant polypeptide subunit K (Lys-487) exerts dominance of null activity and shorter half-life over the tetrameric enzyme molecules in distinct manners. However, the in vivo evidence for the proposed dominance mechanisms remains lacking. METHODS: To address this question, we investigated 33 surgical liver samples identified to be normal homozygous ALDH2*1/*1 (n = 17), heterozygous ALDH2*1/*2 (n = 13), and variant homozygous ALDH2*2/*2 (n = 3). The ALDH2 activity was determined at a sufficient low acetaldehyde concentration (3 µM) and the isozyme protein amount by immunotitration using purified class-specific antibodies. RESULTS: The tissue ALDH2 activity in heterozygotes was 17% that of the ALDH2*1/*1 genotype (p < 0.001), whereas the activity of ALDH2*2/*2 was too low to be precisely determined. The protein amounts of tissue ALDH2 in variant homozygotes and heterozygotes were similar but only 30 to 40% that of normal homozygotes (p < 0.01). Linear regression analyses show that ALDH2 activities were significantly correlated with the protein contents in normal homozygotes and heterozygotes, respectively (p < 0.005). The specific activity of ALDH2 per enzyme protein in ALDH2*1/*2 was 38% that of ALDH2*1/*1 (p < 0.001). CONCLUSIONS: These results are in good agreement with those predicted by the model studies, thus providing in vivo evidence for differential impairments of hepatic acetaldehyde oxidation with alcohol metabolism in individuals carrying ALDH2*1/*2 and ALDH2*2/*2 genotypes.

Combining normobaric hypoxia with short-term resistance training has no additive beneficial effect on muscular performance and body composition.

The aim of this study was to determine the effects of short-term resistance training combined with systemic hypoxia on muscular performance and body composition. Eighteen resistance-untrained men (21.3 ± 2.0 years, 172.7 ± 5.5 cm, 67.3 ± 9.7 kg) were matched and assigned to 2 experimental groups: performing 6 weeks of squat exercise training under normobaric hypoxia (H, FiO2 = 15%) or normoxia (N). In both groups, subjects performed 3 weekly sessions (a total of 18 sessions) of 3 sets of back squat at 10-repetition maximum with 2 minutes of rest between sets. Dynamic, isometric, and isokinetic leg strength and body composition were measured under normoxia before and after resistance training. Squat 1 repetition maximum (1RM) improved significantly (p ≤ 0.05) after resistance training in both H and N groups (88.9 ± 16.9 to 109.4 ± 17.0 kg and 90.0 ± 12.2 to 105.6 ± 13.3 kg, respectively). However, there were no changes in maximal isometric and isokinetic leg strength, lean body mass, and fat mass after the resistance training in both groups. In addition, no significant differences were observed between H and N groups in squat 1RM, maximal isometric and isokinetic leg strength, and body composition. The major findings of this study suggest that short-term resistance training performed under normobaric hypoxia has no additive beneficial effect on muscular performance and body composition. In practical terms, our data suggest that the use of systemic hypoxia during short-term resistance training is not a viable method to further enhance muscular performance and body composition in previously resistance-untrained men.