Hofmann, Dr. Carsten Eickhoff, Dr. Alexander Meyer
- Главная
- Разное
- Дизайн
- Бизнес и предпринимательство
- Аналитика
- Образование
- Развлечения
- Красота и здоровье
- Финансы
- Государство
- Путешествия
- Спорт
- Недвижимость
- Армия
- Графика
- Культурология
- Еда и кулинария
- Лингвистика
- Английский язык
- Астрономия
- Алгебра
- Биология
- География
- Детские презентации
- Информатика
- История
- Литература
- Маркетинг
- Математика
- Медицина
- Менеджмент
- Музыка
- МХК
- Немецкий язык
- ОБЖ
- Обществознание
- Окружающий мир
- Педагогика
- Русский язык
- Технология
- Физика
- Философия
- Химия
- Шаблоны, картинки для презентаций
- Экология
- Экономика
- Юриспруденция
Prediction of Postoperative Complications in Cardiac Surgery презентация
Содержание
- 1. Prediction of Postoperative Complications in Cardiac Surgery
- 2. Outline Problem Definition Data Set Methods Results & Discussion Conclusion
- 3. Problem Definition
- 4. Motivation Huge amounts of data collected at
- 5. Postoperative Bleeding Coagulation Problems: Bleeding due to
- 6. Problem Statement Predicting the need for surgical
- 7. Related Work Electronic Health Records (EHRs) for
- 8. Data Set
- 9. Patients Bleeding patient: surgical re-exploration within 25
- 10. Features Continuous or categorical Static features: e.g.
- 11. Time Slices Time window: end of initial
- 12. Representation
- 13. Methods
- 14. Clinical Baseline Decision in favor of a
- 15. Machine Learning Approaches Naive Bayes AdaBoost (Decision
- 16. Recurrent Neural Network (RNN) x: input
- 17. Results & Discussion
- 18. Evaluation Metrics P: number of positive time
- 19. Results Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
- 20. Problem Definition - Data Set - Methods - Results & Discussion - Conclusion Accuracy
- 21. Problem Definition - Data Set -
- 22. Problem Definition - Data Set -
- 23. Problem Complexity and Limitations Ground truth unknown
- 24. Conclusion
- 25. Conclusion All approaches perform significantly better than
- 26. Thank you!
- 27. Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
- 28. Problem Definition - Data Set -
- 29. Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
- 30. Problem Definition - Data Set -
- 31. Problem Definition - Data Set - Methods - Results & Discussion - Conclusion Likelihood
- 32. Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
- 33. Problem Definition - Data Set -
- 34. Problem Definition - Data Set - Methods - Results & Discussion - Conclusion Likelihood
- 35. Feedforward Neural Network (FNN) Final model: Hidden
- 36. Recurrent Neural Network (RNN) Final model:
Outline Problem Definition Data Set Methods Results & Discussion Conclusion
Слайд 1
Prediction of Postoperative Complications in Cardiac Surgery
Master Thesis
Dina Zverinski
Supervisors: Prof. Thomas
Слайд 4Motivation
Huge amounts of data collected at the intensive care unit (ICU)
High
workload for the ICU staff
Harder to recognize postsurgical complications
Early recognition can lower the risk of late complications
No clinical real-time decision support system
Harder to recognize postsurgical complications
Early recognition can lower the risk of late complications
No clinical real-time decision support system
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 5Postoperative Bleeding
Coagulation Problems:
Bleeding due to non-clotting
Treatment: transfusion (blood products)
Problem Definition -
Data Set - Methods - Results & Discussion - Conclusion
Surgical Bleeding:
Unstaunched bleeding
Treatment: transfusion at first, if no improvement, surgical re-exploration
Early recognition can be crucial
Hard to distinguish at the beginning!
Слайд 6Problem Statement
Predicting the need for surgical re-exploration due to postoperative bleeding
in real-time.
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 7Related Work
Electronic Health Records (EHRs) for prediction
Mortality prediction in real-time at
the ICU
Methods: e.g. logistic regression, deep learning
Risk factor analysis of surgical bleeding
Methods: e.g. logistic regression, deep learning
Risk factor analysis of surgical bleeding
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 9Patients
Bleeding patient: surgical re-exploration within 25 hours after initial surgery
Control group:
no surgical re-exploration after initial surgery
All initial surgeries are open heart surgeries
Adult patients only (18+)
3650 patients in total (50% bleeding patients)
All initial surgeries are open heart surgeries
Adult patients only (18+)
3650 patients in total (50% bleeding patients)
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 10Features
Continuous or categorical
Static features: e.g. age, gender, initial surgery type, …
Dynamic features: e.g. bleeding rate, blood pressure, laboratory results, …
72 features in total
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 11Time Slices
Time window: end of initial surgery until start of surgical
re-exploration
Time slice: feature vector (one per half an hour) labelled with its patient’s class
69996 time slices in total
Missing values imputed with:
last measured value
default value
Time slice: feature vector (one per half an hour) labelled with its patient’s class
69996 time slices in total
Missing values imputed with:
last measured value
default value
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 12Representation
a
Time Slice Representation
a
Sequence Representation
Problem Definition - Data Set - Methods -
Results & Discussion - Conclusion
Слайд 14Clinical Baseline
Decision in favor of a surgical re-exploration, if the bleeding
rate is
> 400 mL/h for 1 hour
> 300 mL/h for 3 hours
> 200 mL/h for 4 hours
Otherwise, no surgical re-exploration needed
from: Robert M. Bojar. Manual of Perioperative Care in Adult Cardiac Surgery. John Wiley & Sons, 2005.
> 400 mL/h for 1 hour
> 300 mL/h for 3 hours
> 200 mL/h for 4 hours
Otherwise, no surgical re-exploration needed
from: Robert M. Bojar. Manual of Perioperative Care in Adult Cardiac Surgery. John Wiley & Sons, 2005.
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 15Machine Learning Approaches
Naive Bayes
AdaBoost (Decision Trees)
Logistic Regression
Support Vector Machines (SVM)
K-Nearest Neighbors
(KNN)
Feedforward Neural Network (FNN)
Feedforward Neural Network (FNN)
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 16Recurrent Neural Network (RNN)
x: input
s: hidden state
o: output
U, V, W:
weight matrices
Figure from Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. Deep learning. Nature, 521(7553):436–444, 2015.
Figure from Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. Deep learning. Nature, 521(7553):436–444, 2015.
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 18Evaluation Metrics
P: number of positive time slices
N: number of negative time
slices
TP: number of true positive time slices
TN: number of true negative time slices
FP: number of false positive time slices
FN: number of false negative time slices
TP: number of true positive time slices
TN: number of true negative time slices
FP: number of false positive time slices
FN: number of false negative time slices
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Accuracy:
ROC AUC: area under the true positive vs. false positive rate curve
Precision:
Recall:
F1 score:
TP + TN
P + N
TP
TP + FN
TP
P
precision * recall
precision + recall
2 *
Слайд 21
Problem Definition - Data Set - Methods - Results & Discussion
- Conclusion
Different Feature Sets
Слайд 22
Problem Definition - Data Set - Methods - Results & Discussion
- Conclusion
Given actual time s until re-exploration and the first time f RNN predicts re-exploration, the relative saved time d is defined as:
Per-Patient-Specificity:
Possible Time Savings
number of true negative patients
number of negative patients
<
Слайд 23Problem Complexity and Limitations
Ground truth unknown
Real-time prediction
Missing or incorrect data
Coarse temporal
resolution
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 25Conclusion
All approaches perform significantly better than the clinical baseline
RNN performs with
accuracy of 0.818
ROC AUC of 0.889
F1 score of 0.802
RNN could help decrease the time until re-exploration by up to 65%
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 30
Problem Definition - Data Set - Methods - Results & Discussion
- Conclusion
Distribution of Patients
Слайд 33
Problem Definition - Data Set - Methods - Results & Discussion
- Conclusion
RNN Classification Options
Слайд 35Feedforward Neural Network (FNN)
Final model:
Hidden layers: 1
Hidden nodes: 20
Activation function: sigmoid
Regularization:
L2-norm
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
Слайд 36Recurrent Neural Network (RNN)
Final model:
Hidden layers: 1 (GRU)
Hidden nodes: 40
Activation
function: sigmoid
Problem Definition - Data Set - Methods - Results & Discussion - Conclusion
