Interest in finding biological features of a tumor that predict clinical behavior began in earnest about 15 years and has gained momentum ever since. Today, the type of treatment given to breast cancer patients is heavily dependent on knowing their tumors’ status for certain biomarkers, such as the estrogen receptor. Unfortunately, we only know about a few useful prognostic factors in any type of cancer.

Cancers arising in specific organs such as the breast can look very similar under the microscope but their clinical behavior can vary tremendously. There are underlying biological abnormalities responsible for these differences in behavior and we have learned that individual tumors may contain a large number of potentially important alterations. Until recently, due mainly to limitations in technology, investigators in this field were only able to evaluate one factor at a time by labor-intensive methods such as Western blotting, or immunohistochemistry, or manual sequencing of DNA. Progress has been slow and frustrating. Technical advances made in the past few years—such—as automated DNA sequencers, cDNA microarrays, tissue arrays, and so on—have the potential to dramatically increase the rate at which important factors can be identified, evaluated, and translated to the bedside. Using these tools, I think it is very likely that we will learn more in the next 5 or 10 years than we have up to now.

The treatment of patients with breast and other types of cancers is becoming increasingly dependent on the biological phenotype of their tumors. Specific alterations call for specific types of therapy. For example, breast cancers expressing the estrogen receptor are highly responsive to tamoxifen, and those over-expressing the erbB2 oncoprotein are responsive to an exciting new drug call Herceptin.

Our study of p53, along with others, confirmed that alterations of this gene were associated with very poor clinical outcome in untreated patients, and this information is now being used by some oncologists to justify using adjuvant chemotherapy in certain patients. Because p53 plays such important roles in DNA repair and programmed cell death, mutations of this gene may also influence tumor response to cytotoxic drugs or radiation - but the jury is still out on this issue, at least in breast cancer. Because most cancers contain several biological abnormalities, it is very unlikely that any single biomarker, including p53, will be able to tell us everything we want to know about a tumor’s potential behavior. In the future we will probably be assessing a large number of genes and pathways simultaneously to obtain a comprehensive biological fingerprint.

I have been very lucky to work in a time and in places where I had the freedom and resources to pursue my research interests. I guess, like anyone else, I get disappointed when a paper gets rejected or a grant fails to get funded, but in the long run even these minor setbacks have a way of working out for the best in the sense of improving the quality of the science.

I have gotten a lot of satisfaction by being able to bring my perspective as a pathologist to a multidisciplinary team of breast cancer researchers that includes oncologists, surgeons, biostatisticians, and molecular biologists.

For example, our bank of 120,000 frozen breast cancers consists largely of tissue remaining after hormone receptor assays performed at various clinical laboratories. Until recently, these assays were biochemical in nature and required the tissue to be "pulverized" to prepare an extract for the test. This was done in liquid nitrogen using an automated instrument called a tissue press, which is really just an expensive mortar-and-pestle. We believed that the tissue was smashed to paste and only useful for biochemical analyses. Because of the ultra-cold temperatures involved, however, it occurred to me that perhaps the tissue was shattered like glass into small but otherwise intact fragments. We developed a method to rehydrate, concentrate, and process the samples like routine clinical tissue in a pathology laboratory so they could be evaluated under the microscope. As suspected, the fragments were intact histologically. This trivial method made it possible for us to use our tumor bank in a variety of new ways we never thought possible before, such as immunohistochemistry and laser-capture microdissection. It has been very useful.

Become a better fly fisherman?

D. Craig Allred, M.D.
Baylor College of Medicine
Department of Pathology
Houston, TX, USA