Diagnosing acute promyelocytic leukemia by using convolutional neural network.

PURPOSE: To evaluate the efficacy of diagnosis systems based upon instance segmentation with convolutional neural networks (CNNs) for diagnosing acute promyelocytic leukemia (APL) in bone marrow smear images. MATERIALS AND METHODS: A self-established dataset was used in this study that was exempted from review by the institution review board, which consisted of 13,504 bone marrow smear images. One subset of the dataset with 12,215 labeled images was split into training (80%) and validation (20%), another with 1289 labeled images was used to test, in which each test entry consists of about 130 images. An instance segmentation method named Mask R-CNN was used to detect and classify the nucleated cells. Here, we train a trained neural network from scratch; for comparison, we also use a network pre-trained on MS COCO (common objects in context, a data set provided by Microsoft which can be used for image recognition, the images in MS coco dataset are divided into training, validation and test sets) and fine-tuned with our dataset and both were trained with same data augmentation scheme. Diagnosis systems based on trained models and "FAB Classification" (French-American-British classification systems, a series of diagnostic criteria for acute leukemia, which was first proposed in 1976) were developed for diagnosing the test entry as APL or as not. Average precision (AP) and average recall (AR) were used to evaluate model performance. RESULTS: The best-performing model had an average precision of 62.5%, which was the augmented pre-trained Mask R-CNN with average recall 84.1%. The average precision of the pre-trained model was greater than that of the model trained from scratch (P < 0.05). Augmenting the dataset further increased accuracy (P < 0 0.03). CONCLUSION: Deep learning technology such as instance segmentation with Mask R-CNN may accurately diagnose APL in bone marrow smear images with an average precision of 62.5% when 0.5 as IoU thresholds. A data augmentation and pre-trained approach further improved accuracy.

Analytical performances and heart failure research of the BNP and NT-proBNP assays on the Cobas E601 and ADVIA Centaur.

BACKGROUND: The analytical performances of the NT-proBNP and BNP assays on the Cobas E601 and ADVIA Centaur were thoroughly evaluated. In addition, the values of BNP and NT-proBNP, which are heart failure markers, were compared in the diagnosis of HF patients with or without acute cerebral infarction since they could also be elevated in ischemic stroke. METHODS: Clinical and Laboratory Standards Institute (CLSI) documents were employed in the analytical evaluation of NT-proBNP and BNP assays on the Cobas E601 and ADVIA Centaur. Then 100 heart failure patients and 103 cerebral infarction complicated with heart failure patients, who had been diagnosed by clinical doctors blinded to NT-proBNP and BNP concentrations, were chosen to compare their values in the diagnosis of heart failure with or without acute cerebral infarction. RESULTS: The NT-proBNP and BNP methods are precise and accurate (total CV < 2.9%, deviation < 3.6%), have wide dynamic measuring ranges (8 pg/mL to 35 126 pg/mL and 2.0 pg/mL to 5094 pg/mL, respectively) with maximum dilutability of 1:2, and are free of common interferences. The most suitable sample types for NT-proBNP and BNP are serum and EDTA plasma, respectively, and both methods correlate well in simple-HF patients. Unlike BNP, the level of NT-proBNP is much higher in HF patients with acute cerebral infarction (p < 0.001). The Cobas E601 and ADVIA Centaur systems have good analytical performances. CONCLUSIONS: In HF patients with acute cerebral infarction, the NT-proBNP and BNP levels did not correlate and thus had implications for clinical diagnosis.

[Amino-terminal pro-brain natriuretic peptide and brain natriuretic peptide measurements under various detection conditions in patients with chronic heart failure].

OBJECTIVE: To find the potential interference factors for the detection of NT-proBNP and BNP in patients with chronic heart failure. METHODS: EP15-A2 issued by Clinical and Laboratory Standards Institute (CLSI) was employed to compare the precision and accuracy of commercial NT-proBNP and BNP analyzer electrochemiluminescence immunoassay system Cobas E601 and chemiluminescence system ADVIA Centaur. Moreover, NT-proBNP and BNP were detected in different time interval and in different interfered sampling conditions (haematolysis, choloplania, lipemia). NT-proBNP and BNP of 203 patients with heart failure or heart failure complicated with acute cerebral infarction were analyzed to find the deviation caused by patients' endogenous factors. RESULTS: The precision and accuracy were comparable for NT-proBNP and BNP detection using Cobas E601 and ADVIA Centaur (total-CV below 2.9% and 3.5%, the deviation from definite value below 2.38% and 3.91%). The most suitable sample type for NT-proBNP and BNP detection was serum and EDTA-anticoagulant plasma. The detection results of NT-proBNP and BNP were comparable for at least 120 min post sampling and not affected by Hb (2 g/L), DB (428 µmol/L) and chyle (2000 FIU). NT-proBNP was significantly higher in heart failure patients complicated with cerebral infarction (P = 0.003) than in heart failure patients. BNP was significantly higher in heart failure grade III patients complicated with cerebral infarction (P < 0.01). CONCLUSIONS: Cobas E601 and ADVIA Centaur supplied satisfactory detection of NT-proBNP and BNP in patients with chronic heart failure with strong anti-interference capacity. The diagnostic value of NT-proBNP and BNP for chronic heart failure should be analyzed objectively in the presence of complicating diseases.