Week 7: Personalised medicine I
Principles in laboratory diagnostics
Step 1: Biomarker identification
● Done via genomics, proteomics, metabolomics, lipidomics, and glycomics.
Step 2: Test development
● What is a good test? → should be easy, cheap, easy sample collection
● Validation of a laboratory test: precision and trueness = accuracy
○ Precision: how close is the observed
sample to the mean of samples? = random
error
○ Trueness: how close is the observed
sample to the real sample value? =
systematic error.
■ Reference to measure should be
available,
■ Clear what is analysed.
■ reference material should be
identical to the analyst in patients.
Step 3: Clinical validation
● Confidence interval: reference interval of “normal”.
Determined by:
○ Selection bias for reference interval: within and between subject variation.
○ Preanalytical consideration analytical considerations
● Sensitivity = true positives in an ill population
● Specificity = true negatives in a healthy population
disease
yes no
test result = positive true positive (A) false positive (B)
negative false negative (C) true negative (D)
how to calculate? sensitivity = A/A+C specificity = D/B+D
● ROC curves = sensitivity vs
1-specificity for different cut-off
values.
Principles in laboratory diagnostics
Step 1: Biomarker identification
● Done via genomics, proteomics, metabolomics, lipidomics, and glycomics.
Step 2: Test development
● What is a good test? → should be easy, cheap, easy sample collection
● Validation of a laboratory test: precision and trueness = accuracy
○ Precision: how close is the observed
sample to the mean of samples? = random
error
○ Trueness: how close is the observed
sample to the real sample value? =
systematic error.
■ Reference to measure should be
available,
■ Clear what is analysed.
■ reference material should be
identical to the analyst in patients.
Step 3: Clinical validation
● Confidence interval: reference interval of “normal”.
Determined by:
○ Selection bias for reference interval: within and between subject variation.
○ Preanalytical consideration analytical considerations
● Sensitivity = true positives in an ill population
● Specificity = true negatives in a healthy population
disease
yes no
test result = positive true positive (A) false positive (B)
negative false negative (C) true negative (D)
how to calculate? sensitivity = A/A+C specificity = D/B+D
● ROC curves = sensitivity vs
1-specificity for different cut-off
values.
