scholarly journals Deep Learning for Discovering and Identifying Morphological Heterogeneity of Neutrophils in Primary Hematological Diseases Based on Bone Marrow Neutrophils Analysis

Blood ◽  
2020 ◽  
Vol 136 (Supplement 1) ◽  
pp. 18-18
Author(s):  
Lintao Bi ◽  
Wen Gao ◽  
Lingjun Meng ◽  
Guiying Gu ◽  
Zhangzhen Shi ◽  
...  
Keyword(s):  
Multiple Myeloma ◽  
Bone Marrow ◽  
Deep Learning ◽  
Roc Curve ◽  
Learning Model ◽  
Normal Bone Marrow ◽  
Chronic Leukemia ◽  
Hematological Diseases ◽  
Normal Bone ◽  
Deep Learning Model

It has been known that neutrophils play an important role in regulating homeostasis and disease. Tumor-associated neutrophils (TANs), as an important member of the tumor microenvironment, have gradually been proved their roles in a variety of solid tumors. It is generally believed that the changes in blood cell morphology (including neutrophils) are the phenotype of hematological diseases (such as in myelodysplastic syndromes) or tumor cells themselves. However, whether there is a possibility that the accumulation of abnormal neutrophils function leads to the change of hematopoietic stem cells and this is just the reason of hematological diseases? Do neutrophils play a key role in the pathogenesis and development of hematological tumors, especially acquired or age-related blood diseases, such as most acute and chronic leukemia, multiple myeloma and other diseases? TAN also has polarization, which is similar to tumor-associated macrophages (TAM), suggesting that the function and morphology of neutrophils are closely associated. Therefore, we assumed that there are function-related morphological differences in neutrophils in different hematological diseases. Finding these differences may provide clues for the functional research of neutrophils in hematological diseases. Artificial intelligence represented by deep learning can distinguish images efficiently and accurately (such as face recognition). Here we try to apply deep learning to discovery and recognize the morphological difference among neutrophils in different hematological diseases. We obtained whole slide images (WSI) from 4 types of malignant hematological diseases, which is chronic myelogenous leukemia (CML), multiple myeloma (MM), acute myeloblastic leukemia with maturation (AML-M2), acute monocytic leukemia (AML-M5) and normal bone marrow. Neutrophils were segmented from WSI by two diagnostic physicians (one with more than 40 years of diagnostic experience and the other with 13 years of diagnostic experience) There are 6115 neutrophils, and the number of cells in each disease and normal bone marrow is 593, 1404, 2509, 850, and 759, respectively. We trained these neutrophils using the transfer learning algorithm and the ratio of training and verification groups is 80:20. We established a convolutional neural network (CNN) model based on the morphological phenotype of neutrophils to judge their disease classification and used confusion matrix and receiver operator characteristic (ROC) curve for model evaluation. We found that neutrophils from different diseases can be classified into different categories, and the deep learning model has a high accuracy rate for judging the neutrophils from different diseases. Moreover, according to the obtained mixed matrix results, it is found that some M2 and M5 neutrophils are prone to misjudgment, while M2 and M5 is rarely confused with other diseases. The reason for this may be that M2 and M5 are both acute myeloid leukemia. Neutrophils from MM and normal bone marrow are prone to misjudge each other or judged as CML neutrophils, and MM often involves the plasma cell system, so some neutrophils of MM may be similar to normal bone marrow. Compared with acute leukemia, some chronic leukemia neutrophils are close to MM or normal bone marrow. Based on these results, we can further confirm that there are morphological and phenotypic differences between different types of hematological diseases. According to the ROC curve results, it is suggested that the deep learning model constructed based on the feature extraction of the CNN model can more accurately determine different hematological diseases according to morphological phenotypes of neutrophils. These findings suggest that neutrophils in different hematological diseases have their own features. These features may provide more evidence for the diagnosis of the disease and also provide clues for further research on the function of TAN in primary hematological diseases. Disclosures No relevant conflicts of interest to declare.

The Parthenolide Derivative LC-1 Is An Effective Single Agent and Is Highly Synergistic with Existing Therapies in Multiple Myeloma.

Blood ◽  
2008 ◽  
Vol 112 (11) ◽  
pp. 1708-1708
Author(s):  
Elisabeth J Walsby ◽  
Saman Hewamana ◽  
Alan Burnett ◽  
Steven Knapper ◽  
Chris Fegan ◽  
...  
Keyword(s):  
Multiple Myeloma ◽  
Bone Marrow ◽  
Cell Lines ◽  
Nuclear Localization ◽  
Plasma Cells ◽  
Bone Marrow Cells ◽  
Binding Capacity ◽  
Single Agent ◽  
Normal Bone Marrow ◽  
Normal Bone

Abstract Multiple myeloma (MM) remains incurable with conventional therapeutic agents and has a median survival of only 3–5 years. Therefore, there is clearly a need for novel treatment strategies that can change the natural pathology of this condition. The nuclear factor κB (NF-κB) family of transcription factors is constitutively activated in MM cell lines and the majority of MM patients. Since NF-κB has known oncogenic activity in a number of human malignancies, targeted inhibition of this family of proteins may be useful in the treatment of MM. We and others have recently shown that the parthenolide derivative LC-1 has activity in acute myeloid leukaemia (AML) and chronic lymphocytic leukaemia (CLL) cells. Unusually, it induces apoptosis via the activation of both the intrinsic and extrinsic pathways and apoptosis is preceded by marked inhibition of NF-κB. Importantly, LC-1 is more potent against primary AML blasts and CLL lymphocytes than normal bone marrow progenitors and normal B-cells and T-cells. In this study we set out to evaluate LC-1 in MM cell lines and plasma cells derived from MM patients. LC-1 was cytotoxic to MM cell lines H929, U266 and JJN3 and induced apoptosis in a dosedependent manner resulting in an overall LD50 of 3.6mM (±1.8) after 48 hours in culture. Primary myeloma plasma cells, identified by CD38 and CD138 positivity, had a mean LD50 for LC-1 of 5.4mM (±1.6) after 48 hours of in vitro culture. Normal bone marrow cells were significantly less sensitive to the effects of LC-1 under the same conditions (P = 0.0007). Treatment of MM cell lines with LC-1 resulted in a decrease in the nuclear localization of NF-κB, as evidenced by a dose-dependent decrease in the DNA binding capacity of the NF-κB subunit RelA after 4 hours of treatment. To demonstrate whether synergy exists between LC-1 and existing MM therapies, the H929 cell line was treated for 48 hours with LC-1 and doxorubicin (32:1), melphalan (1:1) or bortezimib (1:500) and the combination indices (CI) calculated using the median effect method. A combination index of less than 1 denotes synergy. LC-1 did not show synergy with doxorubicin (CI >1) but was synergistic with melphalan and bortezimib (CI values of 0.53 and 0.59 respectively). Taken together our data clearly demonstrate that LC-1 has activity in MM cell lines and primary MM cells. Its ability to inhibit the nuclear localization of NF-κB is important to its cytotoxic effects. Furthermore, it may also provide an explanation for the synergy demonstrated with melphalan and bortezimib. These results provide a rationale for exploring the potential of LC-1 in clinical studies.

scholarly journals Role and Mechanism of circ_0058063/miR-635 Axis in the Malignant Phenotype of Multiple Myeloma RPMI8226 Cells

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Xiaoya Li ◽  
Lingzhi Ding ◽  
Geyu Gu ◽  
Changjun Zheng ◽  
Chenshuai Pan ◽  
...  
Keyword(s):  
Multiple Myeloma ◽  
Bone Marrow ◽  
Protein Expression ◽  
Migration And Invasion ◽  
Normal Bone Marrow ◽  
Malignant Phenotype ◽  
Ki 67 ◽  
Myeloma Cells ◽  
Bone Marrow Tissue ◽  
Normal Bone

Objective. This study aims to explore circ_0058063 effect on multiple myeloma cells malignant phenotype and its feasible mechanism. Methods. We selected 47 cases of multiple myeloma tissues and 47 cases of normal bone marrow tissues and then used RT-qPCR method to test circ_0058063 and miR-635 expression in the tissues. Myeloma cells RPMI8226 were transfected with si-circ_0058063, miR-635 mimic, and si-circ_0058063 + anti-miR-635, respectively. Then, we adopt CCK-8 method, flow cytometry method, and Transwell and western blot methods to detect the influences of knockdown of circ_0058063 or miR-635 overexpression on RPMI8226 cell proliferation, apoptosis, migration, and invasion and also Ki-67, Bax, Bcl-2, MMP-2, and MMP-9 protein expression. The dual luciferase reporter gene assay experiment proved that it has regulatory relationship between circ_0058063 and miR-635. Results. circ_0058063 expression of multiple myeloma was higher than that in normal bone marrow tissue ( P < 0.05 ), while miR-635 expression was lower than that in normal bone marrow tissue ( P < 0.05 ). Knockdown of circ_0058063 or overexpression of miR-635 could reduce proliferation capacity, migration, invasion cell quantities, and Ki-67, MMP-2, MMP-9, and Bcl-2 protein expression ( P < 0.05 ), while increasing apoptosis rate together with Bax protein expression ( P < 0.05 ). circ_0058063 targets to negatively regulate miR-635, while knocking down miR-635 reverses the influences of knocking down circ_0058063 on RPMI8226 proliferation, apoptosis, migration, and invasion. Conclusion. circ_0058063 expression increased in multiple myeloma tissues. Knocking down its expression may inhibit myeloma proliferation, migration, and invasion by targeting and upregulating miR-635 and also promote cell apoptosis. As for multiple myeloma treatment, circ_0058063/miR-635 may provide new molecular targets.

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

10.2196/15963 ◽  
2020 ◽  
Vol 8 (4) ◽  
pp. e15963 ◽  
Author(s):  
Yi-Ying Wu ◽  
Tzu-Chuan Huang ◽  
Ren-Hua Ye ◽  
Wen-Hui Fang ◽  
Shiue-Wei Lai ◽  
...  
Keyword(s):  
Bone Marrow ◽  
Deep Learning ◽  
Validation Cohort ◽  
Learning Algorithm ◽  
Single Cells ◽  
Medical Image Analysis ◽  
Learning Model ◽  
Bounding Box ◽  
The Mean ◽  
Deep Learning Model

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.

Prognostication by miRNome-Profiling in Multiple Myeloma

Blood ◽  
2011 ◽  
Vol 118 (21) ◽  
pp. 1822-1822
Author(s):  
Anja Seckinger ◽  
Tobias Meißner ◽  
Vladimir Benes ◽  
Sabine Schmidt ◽  
Jonathan Blake ◽  
...  
Keyword(s):  
Gene Expression ◽  
Multiple Myeloma ◽  
Bone Marrow ◽  
Plasma Cells ◽  
Proliferation Index ◽  
Normal Bone Marrow ◽  
Global Test ◽  
Myeloma Cells ◽  
Normal Bone ◽  
Bone Marrow Plasma

Abstract Abstract 1822 INTRODUCTION. MicroRNAs are an abundant class of small non-protein-coding RNAs that function as negative gene regulators in diverse biological processes including cancer by affecting the stability and translation of mRNAs. METHODS. We determined expression of 559 miRNAs by miChip (Exiqon LNA Array probes V9.2) in CD138-purified myeloma cells from previously untreated patients (n=69), normal bone marrow plasma cells of healthy donors (pooled to n=3), and human myeloma cell lines (n=20). For normalization, an invariant-based method was applied. Gene expression profiling was performed using Affymetrix U133 2.0 DNA-microarrays. RESULTS. We found 29 miRNAs to be significantly up- and 35 down-regulated in myeloma cells vs. normal bone marrow plasma cells, respectively. Expression of 18 miRNAs was simultaneously significantly (P<.01) associated with event-free survival (EFS) and overall survival (OS), and allowed the delineation of prognostic groups. Of these, miRNAs miR-659 (located at 22q12.1) was significantly lower expressed in myeloma cells compared to normal bone marrow plasma cells. For this miRNA, low expression delineates a group with inferior EFS (median 19.7 months vs. not reached, P<.001) and OS (median 52.9 months vs. not reached, P<.001). This group shows a significantly higher gene expression based proliferation index. In contrast, high miR-590 expression (located at 7q11.23) delineates a group with inferior EFS (median 12.4 vs. 36 months, P<0.001) and OS (median 29.4 months vs. not reached, P<.001). By using Goeman's global test, a significant association of the predicted target gene signatures with survival could be found for both, EFS (miR-590-5p, P<.001; miR-659, P<.001) and OS (P=.009; P=.002). CONCLUSION. In conclusion, we demonstrate the prognostic relevance of miRNome profiling in multiple myeloma. Disclosures: No relevant conflicts of interest to declare.

Frequent Expression of Cancer/Testis Antigens CT7 and LAGE-1 in Advanced Stage Multiple Myeloma: Are They Possible Targets for Immunotherapy?.

Blood ◽  
2006 ◽  
Vol 108 (11) ◽  
pp. 5034-5034
Author(s):  
Valeria Cristina da Costa Andrade ◽  
Gisele Wally Braga Colleoni ◽  
Andre L. Vettore ◽  
Roberta Spetic Felix ◽  
Manuella S.S. Almeida ◽  
...  
Keyword(s):  
Multiple Myeloma ◽  
Bone Marrow ◽  
Cell Line ◽  
High Frequency ◽  
Positive Control ◽  
Proliferation Index ◽  
Advanced Stage ◽  
Normal Bone Marrow ◽  
Normal Tissues ◽  
Normal Bone

Abstract Introduction: Recently, Jungbluth and collaborators (Blood2005;106(1):167) demonstrated that CT7 and MAGE-A3/6 are frequently expressed in advanced stage MM patients and that higher levels of CT7 and MAGE-A3/6 proteins also correlate with elevated plasma-cell proliferation index. These findings suggest a possible pathogenic role of such proteins in MM and also show that they could be attractive targets for immunotherapy. Objectives: To analyze global expression of 13 CTs (MAGE-A1, MAGE-A2, MAGE-A3/6, MAGE-A4, MAGE-A10, MAGE-A12, BAGE-1, CT7, GAGE family, LAGE-1, PRAME, SP17 and SSX-1) in a panel of normal tissues and monoclonal gammopathies. Material and Method: We studied 13 normal tissues (skeletal muscle, bladder, lung, spleen, heart, brain, thymus, uterus, stomach, mammary gland, pancreas, prostate and colon, Clontech), six normal bone marrow aspirates (donors for allogeneic stem-cell transplants), one pool of 10 normal bone marrow samples (Clontech), three monoclonal gammophaties of undertermined significance (MGUS), five solitary plasmacytomas, 39 multiple myeloma samples (two Durie-Salmon stage II and 37 stage III) and MM cell line U266. Normal testis was used as positive control. Total RNA was extracted using Trizol reagent (Invitrogen). After cDNA synthesis, the expression of CTs was evaluated by RT-PCR and 2% agarose gel electrophoresis. Results: SP17 was positive in all seven normal bone marrow samples and in the 13 normal tissues tested. Thus, it was excluded from further analyses. CT7 was positive in one MGUS and in one plasmacytoma. U266 cell line was positive for all CTs, except SSX-1. The frequency of CTs expression in MM patients was: CT7 = 30/39 (77%); LAGE-1 = 18/39 (46%); MAGE-A3/6 = 16/39 (41%); MAGE-A2 = 14/39 (36%); GAGE family = 13/39 (33%); BAGE-1 = 11/39 (28%); MAGE-A1 = 10/39 (26%); PRAME = 9/39 (23%); SSX-1 = 10/39 (26%); MAGE-A12 = 8/39 (20,5%); MAGE-A4 and MAGE-A10 = 0%. It is important to note that from the 18 cases with less CT’s positivity (0, 1 or 2 positive-CTs), three were negative for all of them. Twelve of the remaining 15 cases (80%) were positive for CT7. We did not find association between International Scoring System and percentage of expressed CTs in 36 analyzed MM cases. Conclusion: Our results showed high frequency of expression of CT7 in advanced stage MM patients and support its importance as a target for immunotherapy, because it has a high incidence of positivity even in cases without expression of other CTs. As far we know, this is the second study showing high frequency (~ 50%) of LAGE-1 expression in MM (van Baren et al, 1999), also highlighting its importance as therapeutic target.

Over Expression of YY1 and RKIP in Multiple Myeloma (Cell Lines and MM Tissues).

Blood ◽  
2006 ◽  
Vol 108 (11) ◽  
pp. 5035-5035
Author(s):  
Stavroula Baritaki ◽  
Sara Huerta-Yepez ◽  
Alina Katsman ◽  
Kam C. Yeung ◽  
Devasis Chatterjee ◽  
...  
Keyword(s):  
Multiple Myeloma ◽  
Bone Marrow ◽  
Cell Lines ◽  
Bone Marrow Cells ◽  
Myeloma Cell ◽  
Gene Products ◽  
Normal Bone Marrow ◽  
Over Expression ◽  
Normal Bone ◽  
Marrow Cells

Abstract There is a need to identify underlying molecular mechanisms and gene products (targets for intervention and biomarkers) involved in MM chemo- or immuno-resistance. Recent findings from our laboratory have reported high expression of the transcription factor Yin Yang 1 (YY1) in several tumor types and have demonstrated its fundamental role in tumor cell chemo/immuno-resistance (Gordon, S., et al., Oncogene25:1125, 2006). In contrast, low levels of the Raf kinase inhibitor protein (RKIP) has been shown to play a pivotal role in the negative regulation of cell survival signaling pathways, and has been considered a metastasis tumor suppressor. The overall objective of our studies is to examine whether MM cell lines and tissues derived from MM patients present deregulated expression patterns of RKIP and YY1, and to demonstrate the roles of YY1 and RKIP in MM pathogenesis and response to various therapies. In the present study, the myeloma cell lines RPMI-8226 and MM-1S were screened for RKIP and YY1 expression at the protein and m-RNA levels. In addition, fresh tissues from MM patients were also examined for RKIP and YY1 expression by IHC and their patterns compared with those observed in normal bone marrow cells. The findings demonstrate that both MM cell lines express remarkably higher levels of RKIP protein and transcripts compared to other tumor cell lines examined (lymphomas and prostate cancer), as assessed by western, RT-PCR and IHC. There was significantly higher expression of RKIP in MM patient’s samples compared to normal bone marrow cells. YY1 protein and m-RNA levels were detected in the MM cell lines; however, they were lower compared to Ramos and PC-3 cells. YY1 expression in MM tissues was significantly elevated compared to normal bone marrow cells. RKIP was basically present in the cytoplasm while YY1 was mainly accumulated in the nucleus. In contrast, normal bone marrow cells presented primarily cytoplasmic YY1 and RKIP expression. The YY1 expressing population was the CD38+/CD138− cell subset, which has been reported to be more susceptible to apoptosis (Mitsiadis, S.C., et al., Blood 98:795, 2001). Overall, the present findings support the potential involvement of two new gene products, such as YY1 and RKIP, in MM pathogenesis and their implication in regulating MM resistance to conventional therapeutics. Moreover, these gene products suggest their potential use as new therapeutic targets and/or biomarkers in multiple myeloma. Future studies with large cohorts of tissues will elucidate the importance of the above findings and will give new insights in the unexpected RKIP over expression and activity in multiple myeloma.

scholarly journals Development of LSTM&CNN Based Hybrid Deep Learning Model to Classify Motor Imagery Tasks

2020 ◽  
Author(s):  
Caglar Uyulan
Keyword(s):  
Deep Learning ◽  
Motor Imagery ◽  
Language Processing ◽  
Roc Curve ◽  
Short Term Memory ◽  
Confusion Matrix ◽  
Learning Model ◽  
Classification Model ◽  
Performance Criteria ◽  
Deep Learning Model

AbstractRecent studies underline the contribution of brain-computer interface (BCI) applications to the enhancement process of the life quality of physically impaired subjects. In this context, to design an effective stroke rehabilitation or assistance system, the classification of motor imagery (MI) tasks are performed through deep learning (DL) algorithms. Although the utilization of DL in the BCI field remains relatively premature as compared to the fields related to natural language processing, object detection, etc., DL has proven its effectiveness in carrying out this task. In this paper, a hybrid method, which fuses the one-dimensional convolutional neural network (1D CNN) with the long short-term memory (LSTM), was performed for classifying four different MI tasks, i.e. left hand, right hand, tongue, and feet movements. The time representation of MI tasks is extracted through the hybrid deep learning model training after principal component analysis (PCA)-based artefact removal process. The performance criteria given in the BCI Competition IV dataset A are estimated. 10-folded Cross-validation (CV) results show that the proposed method outperforms in classifying electroencephalogram (EEG)-electrooculogram (EOG) combined motor imagery tasks compared to the state of art methods and is robust against data variations. The CNN-LSTM classification model reached 95.62 % (±1.2290742) accuracy and 0.9462 (±0.01216265) kappa value for datasets with four MI-based class validated using 10-fold CV. Also, the receiver operator characteristic (ROC) curve, the area under the ROC curve (AUC) score, and confusion matrix are evaluated for further interpretations.

scholarly journals Deep-Learning-Based Survival Prediction of Patients in Coronary Care Units

2021 ◽  
Vol 2021 ◽  
pp. 1-10
Author(s):  
Rui Yang ◽  
Tao Huang ◽  
Zichen Wang ◽  
Wei Huang ◽  
Aozi Feng ◽  
...  
Keyword(s):  
Deep Learning ◽  
Roc Curve ◽  
Learning Model ◽  
Survival Prediction ◽  
Training Cohort ◽  
Discrimination Ability ◽  
Model Based ◽  
Coronary Care ◽  
Deep Learning Model ◽  
Coronary Care Units

Background. A survival prediction model based on deep learning has higher accuracy than the CPH model in predicting the survival of CCU patients, and it also has a better discrimination ability. We collected information on patients with various diseases in coronary care units (CCUs) from the Medical Information Mart for Intensive Care III (MIMIC-III) database. The purpose of this study was to use this information to construct a neural-network model based on deep learning to predict the survival probabilities of patients with conditions that are common in CCUs. Method. We collected information on patients in the United States with five common diseases in CCUs from 2001 to 2012. We randomly divided the patients into a training cohort and a testing cohort at a ratio of 7 : 3 and applied a survival prediction method based on deep learning to predict their survival probability. We compared our model with the Cox proportional-hazards regression (CPH) model and used the concordance indexes (C-indexes), receiver operating characteristic (ROC) curve, and calibration plots to evaluate the predictive performance of the model. Results. The 3,388 CCU patients included in the study were randomly divided into 2,371 in the training cohort and 1,017 in the testing cohort. The stepwise regression results showed that the important factors affecting patient survival were the type of disease, age, race, anion gap, glucose, neutrophils, white blood cells, potassium, creatine kinase, and blood urea nitrogen ( P < 0.05 ). We used the training cohort to construct a deep-learning model, for which the C-index was 0.833, or about 5% higher than that for the CPH model (0.786). The C-index of the deep-learning model for the test cohort was 0.822, which was also higher than that for the CPH model (0.782). The areas under the ROC curve for the 28-day, 90-day, and 1-year survival probabilities were 0.875, 0.865, and 0.874, respectively, in the deep-learning model, respectively, and 0.830, 0.843, and 0.806 in the CPH model. These values indicate that the survival analysis model based on deep learning is better than the traditional CPH model in predicting the survival of CCU patients. Conclusion. A survival prediction model based on deep learning has higher accuracy than the CPH model in predicting the survival of CCU patients, and it also has a better discrimination ability.

Sign in / Sign up

Export Citation Format

Share Document