SABCS 2012

Diagnosis of blood cancer after a breast cancer diagnosis

Professor Antonio Wolff – The Johns Hopkins Hospital, Baltimore, USA

http://ecancer.org/tv/conference/161/1795

Today I had the opportunity of presenting data from the National Comprehensive Cancer Centre Network database on outcomes after breast cancer. This is a project that was started by Jane Weeks a number of years ago prospectively collecting a lot of data on patients in large cancer centres but receiving clinical care not necessarily in a clinical trial. So this gives us a lot of good information on what really happens in real life with the caveats that these are usually larger centres.

So today I had the opportunity of presenting some data on the frequency of blood cancers or leukaemias after diagnosis of breast cancer. We have known for a number of years that patients who receive adjuvant chemotherapy have a small increase in the risk of developing leukaemia and data have been described on the frequency of myelodysplastic syndromes as well as acute myelogenous leukaemia. What has been less recognised is actually that it appears, perhaps, to also be an increase in the risk of lymphoid leukaemias such as chronic lymphocytic leukaemia and acute lymphoid leukaemia. What we did see in the NCCN data is that there are some patients presenting with ALL, for instance, after diagnosis of leukaemia. Often they present with a translocation of the MLL gene. That is usually seen in kids’ ALL not in adults’ ALL and for chronic lymphocytic leukaemia there might be a correlation between the presence of genetic abnormalities in the ATM and BRCA2 genes which are also associated with an increased risk of breast cancer.

So the most important piece of information we showed today is that the observed risk of leukaemia after standard chemotherapy with drugs such as anthracyclines and alkylating agents like cyclophosphamide, it appears that the risk may be a little bit higher and potentially twice as higher as previously described from clinical trial datasets such as the NSABP. We looked at a little over 20,000 patients that were part of the database collection in eight institutions throughout the US that are part of the NCCN. We chose on purpose to look into centres that had joined the NCCN effort before 2000 but the collection of patients, or the patients that were part of this analysis, were those who had been diagnosed between 1997 and 2008, so essentially receiving contemporary treatments. We had a median follow-up of about 5.1 years so it’s a pretty robust dataset.

What should we do with these findings?

One of the most important pieces of information that came out of this analysis is that for those of us medical oncologists who take care of patients, we had the belief that the risk of leukaemia after breast cancer treatment with chemotherapy was essentially in the first five years. Actually it appears not to be the case. There are two types of leukaemias that we observed: one are the ones associated with anthracyclines, usually happening around years two and three and often presenting right away with leukaemia with no [?? 4:03]. But there’s a second type associated with alkylating agents such as cyclophosphamide that actually happens a little later towards the years four to six and perhaps as late as ten to twenty years and they often present with myelodysplastic syndromes. So many patients after diagnosis of breast cancer and treatment may present as subtle abnormalities in their blood counts that may not immediately be recognised as being MDS. So the key message was actually that we need to be attentative beyond five years because there is a risk that these cases could be happening later and this is something that was not recognised.

The most important message for the practising clinician is that number one, we should not run away from using, or offering, adjuvant therapy, adjuvant chemotherapy, for someone who has a higher risk early stage breast cancer. When we look at the overall survival data resulting from greater use of adjuvant systemic chemotherapy, there is no question that that overall survival already takes into account any negative survival impact by the development of leukaemia in a small number of patients; the number of patients benefitting is much greater. And that is the key message for anybody considering chemotherapy. We are now beginning to understand that not every patient with a diagnosis of early stage breast cancer has a high risk enough to justify chemotherapy or has a phenotype of cancer that would benefit from using adjuvant systemic chemotherapy. And sometimes we are in situations where there are concerns from the physician or concerns from the patient and the decision sometimes is made of doing and using adjuvant chemotherapy just in case. Well, the challenge is that for patients like that you are potentially giving them none of the potential benefits from adjuvant chemotherapy and putting them at risk for complications, including the risk of AML which approaches 0.5% after ten years. So that’s no question that it is small but for patients with a low likelihood of being helped by chemotherapy that risk, all of a sudden, could become more meaningful.

How can we identify patients at risk of blood cancer complications?

One of the big questions is how can we optimally identify who these patients are. Unfortunately at this point we don’t have perfect predictors because in the ideal world we would be able to have a set of metrics from the tumour phenotype from gene expression profiling or routine markers of prediction and prognosis identify which are the patients which are most likely to benefit from therapy. We then would like to also be able to identify which are the patients that could be at greater risk from developing complications such as leukaemia after diagnosis of breast cancer. There are some indirect measures and they are not perfect but there’s a sense that some of these complications are more likely to occur in older individuals and MDS and leukaemia, by default, is a disease of older people. In the dataset that we observed from the NCCN, the median age of all the patients was around 53 years or so but the median age of the patients diagnosed with AML MDS was about 60 so suggesting that age may play a role. Those were the patients where potential intensification of chemotherapy could be associated with a greater risk.

There is one possible risk factor that we don’t know yet whether it will be meaningful or not, which is the family history. Part of this project was started based on the clinical observation by my co-author, Judy Karp, who’s the Director of Leukaemia Programme at Johns Hopkins, observing in the clinical practice at Hopkins a large enough number of patients diagnosed with MDS and AML who happen either to have a personal history of breast cancer or a family history of breast cancer. So if familial breast cancer is a risk factor, patients who belong to a family with a strong family history of breast cancer or who are found to have mutations in known genes associated with breast cancer risk, such as BRCA1 or BRCA2 specifically, could potentially be at greater risk. I hope that the exercise that we had the opportunity to present in San Antonio can now be used as a stimulus for large groups that have access to familial breast cancer registry where the primary goal up until now has been to look into the frequency or the incidence of breast and ovarian cancer but also to look into the possibility of some other cancers or even other cancers in relatives and individuals who are not affected. Because I think that will begin to allow us to gradually be even more personalised and individualised when we make decisions about treatment. We must look not only at the benefits offered by treatment in terms of who’s likely to be helped by it or not, but also in terms of toxicities such as cardiotoxicity and, the topic of this presentation today, marrow toxicity with the risk of leukaemia, we need to better understand who those individuals are.