AI in RT

Question 1

machine learning + different types

Answer 1

Statistical methods of learning from data

• Supervised learning
• Unsupervised Learning
• Semi-supervised learning
• Reinforcement learning
• Transfer learning

Question 2

supervised learning - 2 categories

Answer 2

• classification
  
  • predicts a class outcome –> dosimetry goal (met or not met)
• regression
  
  • predicts a numerical outcome (OAR dose volume)

Question 3

How do we ensure safety of our patients is maintained?

Answer 3

• Relies on an understanding of how AI systems are designed and applied
• Education of undergraduate and qualified staff on AI basics
• Ability to understand the assumptions and limitations of an AI application to
  ensure accountability in decision making
• Communication with patients about how these tools are used

Question 4

In practice, MRPs need to…

Answer 4

• have an understanding of the technology, data science, including privacy and security
• Should use and deploy these systems, maintain and overview them, direct their application in their practice
• Should have control of AI supported systems to ensure that patient safety is not compromised.

Question 5

Both undergraduate & qualified medical radiation practitioners are expected to engage in:

Answer 5

• Understanding the language of AI
• Understanding the principles of AI, the development of data algorithms and the evaluation and validation
  of AI supported healthcare
• Awareness of the benefits and risks, including supervised and unsupervised machine learning,
• Informing and explaining to patients about the role of AI in their healthcare
• Understanding and applying health informatics to ensure safe AI-supported healthcare
• Data curation, governance and stewardship that assures patient safety, privacy and data protection
• Integrating research driven, evidence-informed approaches to safe AI-supported healthcare, and
• Partnering and leading AI-supported point-of-care image analysis and diagnosis, treatment planning and
  adaptive technologies

Question 6

causes for inappropriate quality in data sampling

Answer 6

• selection bias
• class imbalance
• insufficient quantity
• prejudice bias

Question 7

selection bias

Answer 7

collecting data that is not representative of the target population

Question 8

class imbalance

Answer 8

Underrepresentation of a class
in the outcome variable

Question 9

insufficient quantity

Answer 9

Low sample size insufficient to train model relative to the complexity of the outcome

Question 10

prejudice bias

Answer 10

training data includes (human) biases containing implicit racial, gender or ideological prejudices

Question 11

causes for inappropriate quality in data management

Answer 11

• noise
• uninformative or less informative
• exclusion bias
• measurement bias
• survey bias
• observer bias (confirmation bias)

Question 12

data quality issue - noise

Answer 12

Particularly related to medical
images where artefacts or blurry
images may reduce image quality

Question 13

uninformative or less informative

Answer 13

Records do not convey all the
necessary information to solve
the problem at hand

Question 14

exclusion bias

Answer 14

Deleting valuable data that was
thought to be unimportant

Question 15

measurement bias

Answer 15

systematically favoring a
particular result when observing
or measuring variables in data

Question 16

survey bias

Answer 16

Substantial missing, incomplete,
and inconsistent responses to
surveys, questionnaires, or
interviews used to collect data

Question 17

observer bias

Answer 17

Favoring information that does
not contradict the researcher’s
desire or previous beliefs

Question 18

causes for inappropriate quality in data annotation

Answer 18

• recall bias
• annotation bias
• lateral variability (interobserver variability)
• longitudinal variability (clinical drift)

Question 19

recall bias

Answer 19

Labelling similar types of data
inconsistently

Question 20

annotation bias

Answer 20

Inaccuracies induced by
human errors when labelling

Question 21

lateral variability (interobserver variability)

Answer 21

Difference in practice
between physicians or
institutions

Question 22

longitudinal variability (clinical drift)

Answer 22

Difference in practice over
time due to change in
techniques or guidelines

Question 23

missing data + different types

Answer 23

Missing data can be a common issue in healthcare research, particularly when relying on electronic health records

Missing data can be classified as:
- Missing completely at random (MCAR)
- Missing at random (MAR)
- Missing not at random (MNAR)

Question 24

how to fix missing data

Answer 24

• One option is to remove the rows with missing data
• Another option is to use multiple imputation, a statistical method for estimating the missing values
• Data augmentation is an additional method for increasing the dataset size when limited data is available

Question 25

Aleatoric uncertainty

Answer 25

• Aleator is the Latin word for someone who rolls the dice
• this uncertainty is based on randomness in the data
• it cannot be reduced by collecting more data
• It can only be measured and quantified
• It is seen when an experiment is repeated multiple times and there is variation in the outcome due to
  randomness

Question 26

Epistemic uncertainty

Answer 26

• From the Greek word relating to knowledge
• Often described as model uncertainty
• Can be reduced by gaining greater “knowledge” of the question being asked of the model.
• Gaining greater “knowledge” may be in the form of more data, or more input variables (higher dimensionality) to give the model a greater ability to fit to the data

Question 27

AI ethics principle

Answer 27

• human, societal, and environmental wellbeing
• human-centred values
• fairness
• privacy protection and security
• reliability and safety
• transparency and explainability
• contestability
• accountability

Question 28

supervised vs unsupervised learning

Answer 28

In supervised learning, the algorithm learns from labeled data, meaning the input data is paired with corresponding correct outputs, allowing the model to learn patterns and make predictions.

Unsupervised learning, on the other hand, deals with unlabelled data, aiming to discover underlying structures and relationships within the data without explicit guidance

Question 29

What QA should be completed in the implementation stage of AI?

Answer 29

1. identify workflow position
2. identify test data set
3. end to end testing of AI tool
4. set clinical limits for use
5. risk analysis

Question 30

What QA should be completed in the case-specific stage of AI?

Answer 30

1. manual checks of outcomes
2. recording instances of manual adjustment
3. send feedback to model developers

Question 31

What routine QA should be completed for AI?

Answer 31

1. use standard datasets from implementation to check any model updates
2. review logbook for trends
3. review patient cohort for shifts in background
4. send feedback to model developers

Question 32

What is involved in an AI contingency plan?

Answer 32

Set a threshold for action, identify the logistics of changing to a manual process, determine what the AI tool is failing because of (1 pt, group, or all)

Question 33

What risks need to be analysed before the use of AI?

Answer 33

• how much autonomy is the system given
• do we trust the results
• what are the potential consequences if something goes wrong
• what is the downstream impact
• what areas could the error impact

Question 34

What is distributed learning?

Answer 34

Data collected across multiple hospitals where data is sent to a central terminal

Question 35

Why is model interpretation important?

Answer 35

1- ethical considerations
2 - if something goes wrong understand how AI works
3- clinical reasoning

Question 36

model agonistic interpretation

Answer 36

can be applied to any models and are more flexible than model-specific interpretations

Question 37

global interpretations

Answer 37

provide information on what variables were most influential from the training dataset

Question 38

local interpretation

Answer 38

provide information on what variables from an individual’s data were most influential on model output ]