How do you know a treatment works? Flashcards
Current cancer treatments
Surgery Chemotherapy Radiotherapy Hormone therapy Immunotherapy Biological therapies
Objectives of cancer therapies
Cure the patient - kill or remove ALL cancer cells
Prolong patient survival - kill or remove MOST cancer cells
Palliative symptoms - kill SOME cancer cells
Cure therapy complete removal
You can think you have cured but not completely so the patient goes into remission
Prolong patient survival the patient will go into remission.
Tumour efficacy
Whether the tumour responds to the treatment or not
- complete response: disappearance of all the signs of disease
- partial response: a reduction of tumour volume by at least 30%
- Stable disease: no significant change (not 30%)
- disease progression: an increase in tumour volume by at least 20% or new metastasises
Assessing tumour efficacy - survival time
Overall survival: survival time from the start of the treatment
Disease free survival: survival time prior to relapse after radical treatment - how long no tumour lasts for
Progression free survival: survival time prior to tumour progression - how long before growth or metastasis
How much evidence is needed to persuade you that a treatment works?
Large group of patients from a variety of centres
Anecdotal/1 patient evidence is not enough
Small groups from one centres not enough
What is a clinical trial
Any form of planned experiment which involves patients and is designed to elucidate the most appropriate method of treatment of future patients with a medical condition
A single arm clinical trial for assessment of new treatment. New treatment new patients, show 5 year survival rate. Does this mean the treatment works?
No, there no control arm of the study to compare results too
Single arm clinical trial compared against historical controls with the old standard treatment - show 5 year 60% survival is this enough?
Historical patients are not comparable the patient population may be from different groups so cannot be compared. Age difference the whole package of care may have changed since the controls were used which may be confounding.
Clinical trial with comparison against concurrent controls there is a systematic assignment of patient to the two groups e.g. Birthday before or after July 1st.
This is not enough as it is not blind or randomised. Therefore patients who are unfavourable may be left out because of the criteria patients should be selected randomly and randomised to groups
Eligible patients identified are then randomised to either the standard or new treatment - 5 year survival rate = 60%
This is a randomised controlled trial. Bias is removed by randomisation. 2 arms to compare currently in real time.
Remember this still isn’t enough to prove. It needs good statistical analysis to evaluate any uncertainty.
How do we assess whether the new treatment works better than the old?
Estimate the treatment effect from the sample data. Calculate the hazard ratio
Hazard of death with new treatment/compared to hazard of death with old treatment.
If a number less than 1 then the new treatment is has a reduced risk of death than the old treatment
The need to estimate the confidence interval, whether the results fit into the confidence interval need to make sure the confidence interval doesn’t over lap either the null hypothesis otherwise it is not statistically significant
Test the research hypothesis that the new treatment works better than the standard.
Kaplan Meier survival curve show?
Each step is a death
Shows amount of deaths over time, so compared the survival rate of new drug compared to the old drug
Symbols on the curve are censored data. So the patient dropped out and that is the last time they were alive, this data point can’t be used.
What is the hazard ratio?
Measuring the relative difference in survival between new and standard treatments
risk of death on NEW/ risk of death on STANDARD
Hazard ratio less than 1 means there is a reduction. Reduction of 0.7 means the risk of death on the NEW is 70% of that on the standard so a reduction of 30%
HR=1 no difference between the treatment
HR<1 NEW reduces risk
HR> NEW increases risk
When do single case studies/ clinical experience provide compelling evidence. No RCT conducted.
Case study 1 patient PET scan before treatment cancer is everywhere then after treatment vast reduction in all cancers
Single are clinical trial
Baseline and then the change from baseline is such a reduction in all patients that the change can effect clinical studies.
What is the research pathway before the RCT?
Phase 1= first time on man - they are cancer patients not healthy patients
However this is not necessarily - the first time the drug has been used it may have been use in other circumstances it may be the first time it is being used this way
Phase 1 you asses for doses and toxicity - if. It not first time used on man dosed for this particular target
Aim of phase 1 is to find safe dose of a NEW treatment and understand toxicities.
