Hazlina Hamdan

1
Modelling survival prediction in medical data

• By
• Hazlina Hamdan
• Dr. Jon Garibaldi

2
Presentation Content

• Introduction
• Cancer and Breast Cancer
• Medical Prognosis
• Survival Analysis
• Research Background
• Aims Objectives
• Understanding Previous Approach
• Artificial Neural Network
• PLANN
• Analysis and Results
• Conclusion
• Future Work

3
Cancer

• Cancer is basically a disease which occurs when
  cells behave abnormally and divide out of control
  form visible mass or tumour.
• There are two general types of tumours namely
• benign
• malignant
• Breast cancer is the most common cancer causing
  fatality amongst women.
• Breast cancer is a malignant tumour that develops
  from the uncontrolled growth of cells in the
  breast.

4
Medical Prognosis

• The principal factor in estimating of cure,
  complication, disease recurrence or survival for
  a patient or group of patients after treatment.
• Prognosis is important because the type and
  intensity of the medications are based on it.
• Prognosis is only a prediction.

5
Survival Analysis

• The analysis of data that corresponds to the time
  from when an individual enter a study until the
  occurrence of some particular event or end-point.
• Concerned with the comparison of survival curves
  for different combinations of risk factors.
• Data contains uncensored (reach until end point)
  and censored (lost to follow-up or die from
  unrelated cause) observations.

6
Survival Analysis Survival Function

• Probability an individual survive at least up to
  a certain time t.
• S(tl)P(Tt)
• Kaplan-Meier survival curve.

7
Survival Analysis Hazard Function

• Probability an individual will die at a certain
  time, conditioned on survival up to that time,
  and denotes the instantaneous death rate. (Collet
  D., 1994)
• hl P(T ? AlTgttl-1) fl /S(tl-1)
• known also as conditional failure probability
• Survival and Hazard function are related to each
  other
• S(t)?(1-hl)
• ltlt

8
Research Objectives

• Understand previous approaches.
• Apply previous approaches to our data.
• Develop novel approaches based on Artificial
  Neural Network (ANN) and Fuzzy method.
• In clinical perspective is to assist doctor in
  predicting survival of individual patients and
  planning future treatments.

9
Previous ApproachArtificial Neural Network (ANN)

• Artificial Neural Network (ANN) is defined as an
  information processing system inspired by the
  structure of the human brain.
• ANN gathers its knowledge by detecting a common
  pattern and relationships in raw data, then
  learning from such relationships and adapting the
  results as required.
• The knowledge is then used to predict the outcome
  for new combinations of data.

10
Previous Approach - ANN

• In the field of medicine, ANN have been used
  since the late 1980s, initially as an aid to
  diagnosis and treatment, and recently as a tool
  for the analysis of survival data in the presence
  of censorship.
• The ability of neural networks to generalise to
  new cases based on existing patterns is used as a
  basis to compute and predict the survival of
  individual patient or group of patients.

11
Previous Approach-Feed-Forward ANN
Patients
Variables 1x Inputs Hidden units
Outputs
A A1 Ax . . . . N N1 Nx
bias
Transfer Function
12
Previous Approaches-PLANN

• Partial Logistic Artificial Neural Network
• Proposed by Biganzoli et. al (1998)

13
PLANN Model
14
PLANN Model-Pre-processing

• Categorical variables ?indicator variables
• 
  
• Continuous variables ? range(-1,1) or (0,1)

Treatment Type Indicator variables Indicator variables Indicator variables
Radiotherapy 1 0 0
Hormone therapy 0 1 0
Chemotherapy 0 0 1
15
PLANN Model-Pre-processing

• Training data - each subjects are replicated for
  all the intervals in which the subjects is
  observed and coupled with the event indicator

Time Size Treat1 Treat2 Treat3 Event
Subject1 3 1.0 1 0 0 1
Subject2 5 0.5 0 0 1 0
Time Size Treat1 Treat2 Treat3 Event
Subject1 1 1.0 1 0 0 0
2 1.0 1 0 0 0
3 1.0 1 0 0 1
Subject2 1 0.5 0 0 1 0
2 0.5 0 0 1 0
3 0.5 0 0 1 0
4 0.5 0 0 1 0
5 0.5 0 0 1 0
16
PLANN Model-Pre-processing

• Testing each subjects are replicated into full
  number of time interval of observed with all
  event indicator as zero.

Time Size Treat1 Treat2 Treat3 Event
Subject1 3 1.0 1 0 0 1
Subject2 5 0.5 0 0 1 0
Time Size Treat1 Treat2 Treat3 Event
Subject1 1 1.0 1 0 0 0
2 1.0 1 0 0 0
3 1.0 1 0 0 0
4 1.0 1 0 0 0
5 1.0 1 0 0 0
Subject2 1 0.5 0 0 1 0
2 0.5 0 0 1 0
3 0.5 0 0 1 0
4 0.5 0 0 1 0
5 0.5 0 0 1 0
17
PLANN Model-Post-processing

• Predicted hazard is the mean calculated from the
  distribution of the activation.

18
Analysis and Result

• Head and Neck Cancer disease recurrence

Radiation therapy (Arm A) Radiation Chemotherapy (Arm B)
Total patients 51 45
End of time interval (in month) 47 76
Total patient recur until end interval 42 31
Total patient lost to follow up 9 14
Total training replication 628 967
Total testing 47 76
19
Analysis and Result
20
Analysis and Result
21
Analysis and Result
22
Conclusion

• ANN have been considered as alternative methods
  for analysis of survival for individual patient
  or group of patients.
• A smooth discrete hazard possible be model by
  treating the time interval and the covariates as
  an input variable with standard feed forward
  network and logistic activation function.

23
Future Work

• Implementing PLANN model to our data (breast
  cancer data from QMC).
• Develop fuzzy set rules in producing the survival
  rate prediction for breast cancer patient.

24
References

• Bishop, C. M. (1995). Neural Networks for Pattern
  Recognition, Oxford University Press Inc., New
  York,.
• Burke, H.B., Goodman, P.H., Rosen, D.B., Henson,
  D.E., Weinstein, J.N., Harrell, F.E., Marks,
  J.R., Winchester, D.P. Bostwick, D.G. (1997).
  Artificial neural network improve the accuracy of
  cancer survival prediction. Cancer, vol. 79,
  pp.857-862
• Collett, D. (1994). Modelling Survival Data In
  Medical Research. Chapman and Hall, London.
• Elia Biganzoli, P. B. L. M. E. M. (1998). "Feed
  forward neural networks for the analysis of
  censored survival data a partial logistic
  regression approach." Statistics in Medicine
  17(10) 1169-1186.
• Lisboa, P. J. G., H. Wong, et al. (2003). "A
  Bayesian neural network approach for modelling
  censored data with an application to prognosis
  after surgery for breast cancer." Artificial
  Intelligence in Medicine 28(1) 1-25.
• Ohno-Machado, L. (2001). "Modeling Medical
  Prognosis Survival Analysis Techniques." Journal
  of Biomedical Informatics 34(6) 428-439.
• Ripley, R. M., A. L. Harris, et al. (1998).
  "Neural network models for breast cancer
  prognosis." Neural Computing Applications 7(4)
  367-375.
• Ravdin, P. and G. Clark (1992). "A practical
  application of neural network analysis for
  predicting outcome of individual breast cancer
  patients." Breast Cancer Research and Treatment
  22(3) 285-293.