Pathologists have tended to take a fairly straightforward approach to understanding tumours – by simply cutting them into slices and examining them under the microscope. To some extent, they have succeeded in predicting how tumours might behave based on how they looked. But doctors have also frequently been baffled that patients with apparently indistinguishable tumours could receive the same treatment and yet respond so differently to it. It has become clear that there is a need for more sophisticated ways of analysing tumours, allowing doctors to more effectively match the treatment to the group of patients who will benefit most.
Over the past decade, we have learned that many cancers may not show much difference from each other under the microscope, but are likely to display significant variability at the molecular level. This variability poses a significant hurdle in treating patients, since it means that the best treatment for one cancer may not work for another.
But researchers are increasingly combining established indicators of prognosis – such as histology and treatment information – with state-of-the-art molecular profiling for each individual patient, in order to create therapeutic strategies precisely tailored to a subset of patients with similar profiles. This approach is known as ‘stratified medicine’, and is a key first step towards the ultimate goal of precision medicine, in which treatments are tailored for each individual patient.
Dr Anguraj Sadanandam is one of our newest faculty members here at The Institute of Cancer Research, London, applying his expertise to stratify tumours into molecular subtypes, with shared characteristics that may predict disease course and treatment response.
One of the cancers Dr Sandanandam has studied is pancreatic ductal adenocarcinoma – a particularly lethal form of cancer with an average survival of only about six months from diagnosis. Numerous phase III trials of agents effective in other cancers have failed to benefit unselected patients with the disease. Dr Sadanandam says: “My previous research found that pancreatic ductal adenocarcinoma has three molecular subtypes that may predict patient outcomes. For instance, one subtype is sensitive to erlotinib – a drug that targets epidermal growth factor receptors (EGFR) – whereas another subtype is sensitive to gemcitabine – a type of chemotherapy. We are now looking at the specific mechanisms that influence heterogeneity in this disease, such as genetic alterations and cellular origin. We are interested in developing biomarkers specific to each subtype and assays that can be used in the clinic. Armed with this knowledge, we could look at a patient’s molecular fingerprint of their tumour and match the treatment accordingly.”
Colorectal cancer is the third most common cancer type in the world. In the case of metastatic colorectal cancer, cetuximab – another EGFR-targeted therapy– has shown promise. But some patients seem to respond well to their therapy, and others do not. Dr Sadanandam explains: “Current treatments are being given to unselected patient populations, and do not account for the molecular variations among individual patient tumours, which could account for the differences in treatment response. We wanted to investigate why this was happening, and stratified patients’ colorectal cancer samples into six subtypes using transcriptome data and therapy response information for cetuximab.” Results from this study were intriguing. Dr Sadanandam reveals: “We were surprised to see that three subtypes had markedly better disease-free survival after surgery, suggesting these patients could be spared from having chemotherapy. Patients with this subtype would just require watchful surveillance until the tumour progresses to distant metastatic sites – then they can probably be prescribed appropriate therapies. One of these subtypes does not respond to cetuximab, and a MET tyrosine kinase receptor could be given to these patients who go on to have metastatic cancer. Two other subtypes with poor and intermediate disease-free survival respond well to chemotherapy in adjuvant or metastatic settings. We are now working to validate our findings using pre-clinical therapeutic trials. This information will not only match patients to specific therapies, but also will minimise the side-effects of traditional chemotherapies and suggest possible alternative therapies.” The study also showed that each subtype shared characteristics with different cells within the normal colon, giving researchers an indication of where the cancer started. One of the subtypes displayed stem cell-like characteristics, and this type had the worst prognosis, highlighting the need for early detection for this particular subtype.
Dr Sadanandam says: “Ultimately, our hope is that this information will be translated into the clinic so we can subtype cancer patients in order to identify the most appropriate therapy or combination of therapies to confer the greatest benefit to the patient. Putting the patient on the right therapy at the right time will no doubt prolong the lives of cancer patients. Unwanted side-effects, time-wasting and costs associated with ineffective treatment options will also be reduced. Overall, this will help to improve disease management and avoid over- and under-diagnosis and treatment.”
We have come a long way from trying to predict how tumours will behave by just looking at microscope slides. Delving deeper to the molecular level, we are now gaining new insights on how to match patients with treatments.
Image: Metastatic carcinoma from thyroid x 100 (c) Wellcome Photo Library Creative / Commons by-nc-nd 2.0 UK