Molecular imaging in early phase trials
Susan Galbraith
Bristol-Myers Squibb, Princeton, USA
Novel imaging technologies offer unprecedented opportunities to image tumour biology. Rather than merely documenting site, size and morphology of tumours we can now image microvascular function, hypoxia, tumour metabolism and proliferation, and apoptosis. New imaging technologies also offer exciting opportunities to learn about a drug’s characteristics at earlier stages of development. The 3 most frequent reasons for a drug to fail in development are - efficacy, safety/ toxicity and inadequate pharmacokinetics. Imaging technologies can provide information in all of these areas, which could enable earlier No Go decisions to be made. Currently anatomical imaging measurements are used in late phase trials - e.g. MRI appearance of multiple sclerosis lesions or definition of tumour response in oncology, used as a surrogate for the clinical endpoint of change in overall survival. However there is great potential to affect decision making in early phase I and II trials, where the focus is more likely to be imaging of function, molecular targets such as enzyme inhibition or receptor occupancy and pharmacokinetics. Demonstration that the drug does not hit its target or reach the target tissue would be a clear No Go for example. Imaging of drug uptake into and distribution within tumours may be more informative than analysis of the relation of plasma concentration to anti-tumour effect.
The development of ‘cytostatic’ drugs in oncology presents particular challenges. The maximum tolerated dose may not be the optimal dose for Phase II so measurement of the change in tumour microvasculature, metabolism or proliferation could be used in dose and schedule selection. Use of an imaging technique will require expansion of cohorts to ensure an effect of clinical significance can be measured. This will require a larger investment in Phase I, but would allow a ‘proof of confidence’ decision, and a No Go if no/limited effects are seen in the tolerable dose range.
Response rates in Phase II have been used with cytotoxic agents as an indicator of efficacy, but if lack of progression rather than tumour shrinkage is expected from the mechanism of action then imaging of changes in tumour biology could also provide an alternative efficacy indicator, assessed earlier and with fewer patients than required for a time to progression endpoint. Prolonged stable disease has also been suggested as a potential endpoint for trials with cytostatic compounds, but imaging of biology could better determine if stable disease is associated with true anti-tumour activity and improve the confidence around decisions to move to later phases of development.
In order for this potential to be realized several hurdles need to be overcome. Ideally the same techniques planned for early phase clinical trials should be used in pre-clinical models to compare dose response and time course of the imaging endpoint with dose response for anti-tumour efficacy. The more novel techniques are by their nature less standardized, with significant differences in methodology between centers even for such a widespread technique as FDG PET. There is frequently a lack of data on reproducibility between and within patients and sites and over the timepoints of interest. Image analysis methodology needs validation, with quality control of initial image acquisition. If data are to be shared across multiple sites there is a need for a centralized database, compatible with the different hardware and software at each site. Industry needs to work with academia to develop acceptable standards for these steps. The presentation will discuss updates on reproducibility in multi-center trials for dynamic MRI, FDG and FLT-PET, and illustrations of decision-driving data utilizing these techniques in preclinical and clinical experiments.
