n this interview, ISI Essential Science Indicators correspondent Simon Mitton and Professor Christopher Elston of the University of Nottingham discuss Prof. Elston’s career in breast cancer research. Prof. Elston, along with his colleague Ian Ellis of the City Hospital of Nottingham, wrote the highly cited paper, "Pathological prognostic factors in breast-cancer 1. The value of histological grade in breast-cancer—Experience from a large study with long-term follow-up," (Histopathology 19[5]: 403-10, November 1991). In a Special Topics survey of breast cancer research, this paper was cited 458 times, making it one of the 10 most-cited papers in breast cancer research of the past decade. Current data in ISI Essential Science Indicators Web product indicate that this paper has been cited a total of 621 times.

I received my initial training in general pathology and turned to specialization in histopathology on appointment in 1965 to Fulham Hospital, London. I went to King’s College Hospital Medical School, London, in 1969 as a Senior Lecturer in Morbid Anatomy and by pure chance my appointment coincided with the start-up of a trial on breast cancer screening. This trial compared simple mastectomy with mastectomy followed by axillary radiation. We found that the addition of the radiation did not affect survival. In 1970, as part of that investigation, I was appointed Trial Pathologist to the UK Cancer Research Campaign (CRC), which meant looking at and grading the tumors. We applied a method of grading first described in 1957 by H. J. G. Bloom and W. W. Richardson in a paper published in the British Journal of Cancer.

From King’s College Hospital I moved to Nottingham in 1974 to find, coincidentally, that Professor Roger W. Blamey had been appointed as surgeon a couple of months before my arrival. His intention was to set up a breast cancer unit. Pathology was of great importance to his thinking, and he had an interest in individualizing breast cancer treatment. During the CRC trial I had modified the original Bloom and Richardson method (which was vague, not objective). We improved on Bloom and Richardson by adding more objectivity to the counting of metastases.

 In your 1991 paper in Histopathology (19[5]: 403-10, November 1991), what was your logic in thinking that morphological grading might provide important prognostic information?

Ten years ago, it was already very clear that with the widening of options for the treatment of breast cancer patients, clinicians needed accurate prognostic information on which to base therapeutic decisions. A fundamental aspect of histopathology has been the recognition that the morphological appearance of tumors can be correlated with the degree of malignancy. Publications on grading malignancy go all the way back to 1925 when R. B. Greenberg, a pathologist in Boston, published a crude study on varying degrees of malignancy in cancer of the breast in the Journal of Cancer Research. In 1928 this was modified by D. H. Patey and R. W. Scarfe working at the Middlesex Hospital, but then it was promptly forgotten about. It took another 30 years for Bloom and Richardson to modify the Patey-Scarfe method: they made it quantitative by adding a score chart, but they did not provide cut-off points. That was the background of publications when I joined the field.

I had shown in the CRC trial that even with poorly prepared specimens you got a good correlation of grading with clinical outcome. That confirmed other studies. We decided once we started the Nottingham trial, which had the aim of assessing the relative importance of a range of potential prognostic factors, to make grading the central focus.

While we were doing our preliminary data collection and beginning to present some of our data showing correlation, there was coincidentally an increase in the papers reporting on more biological markers of differentiation, namely hormone receptors. All these publications showed that grading was a factor but they invariably cited papers showing poor reproducibility; we felt these were poor papers. Consequently, my colleagues and I decided to make grading as objective as possible. We adopted Scarf and Richardson as important pioneers, added my CRC work, and then strengthened the objectivity. Our 1991 paper was the first publication of our grading scheme, showing correlation between lymph node size and grade.

 That paper has become one of your most highly cited. What was the significance of this paper and why was it so highly cited?

The high impact is quite remarkable. Let me give you this background: pure scientists did not like subjective morphological studies. The serious researchers like numbers and we gave them numbers!

The first breakthrough was that our numerical grading was shown to be reproducible in three papers in peer-review journals. This knocked the idea on the head that quantitative grading would never become acceptable. My colleague, Dr. Ian O. Ellis, and I published several follow-up papers showing that in histological grading we had a robust method for the assessment of differentiation in breast carcinoma. Another very important factor in establishing the method was that I decided to go to the US and lecture on the method. This was at a time when clinical researchers from the UK seldom attended conferences in the US, as a result of which our American counterparts often ignored research done outside the US. In this ambassadorial role I went to big meetings, banging the drum. As a result of this campaign, David Page (Nashville, Tennessee) became a convert and a powerful advocate, as did James Connolly (Boston, Massachusetts). By now our method was taking off. I knew that the US was the key to becoming accepted. That’s where I made a big effort for two years, promoting the protocols.

The wide adoption of our method then followed. In 1988 The American Journal of Surgical Pathology (the official journal of the Association of Directors of Anatomical and Surgical Pathology) published a protocol which included the grade concept. Parallel to that, in 1988 UK screening started in a national pathology group under Professor John Sloane; I was in that group and we produced reporting guidelines for breast screening pathology. We argued for group grading and advocated that the Nottingham method must be included. The situation now is that the UK guidelines we advocated have been accepted by both the European Community and by the US. Other countries have taken it up as well. For example, I have been a frequent visitor to Australia and they have adopted our method.

Very rarely do you see the fruits of labors in your own lifetime. I have been fortunate in that our evidence-based work has persuaded oncologists that treatment should be individualized. The Nottingham Prognostic Index is being used increasingly to stratify patients for treatments. When I was in Australia recently, an oncologist told me he would not now treat a patient without knowing the histological grade, whereas five years ago he would not have bothered with the procedure.

 When you moved from King’s to Nottingham, what role did practical support or considerations of existing faculty play in your decision?

It was a gamble for me to move from a purely academic post to a position in the UK National Health Service. And I was moving to a small department. The attraction for me was that Nottingham had a new medical school and I coincided with the first cohort of medical students. The buzzwords were: expansion, young and enthusiastic staff and students, a great atmosphere. It was a teaching hospital where I could be an independent consultant And Roger Blamey was our inspiration: pushing individualization. Facilities were limited and I did all this research in my own time.

 Let’s turn to a more recent paper, published in Critical Reviews in Oncology/Hematology ("Pathological prognostic factors in breast cancer," 31[3]: 209-23, August 1999). You introduce this paper with the statement that "The management of patients has changed enormously in the last 20 years, for a variety of reasons." I’d like to invite you to comment on how you see your own work as having contributed to that.

We have the largest data sets in the world on histological grading for these tumors. Our work is characterized by excellent follow-up by physicians, obsessive attention to the data, and rigorous analysis of the data. We used multivariate analysis to prove conclusively that the method would lead to good stratification for treatment methods. The major advances have been our ability to demonstrate the benefits of grading, together with improvements in treatments, new drugs, and our ability to predict how patients would respond.

 Do any serendipitous events touch on your research?

Certainly: after embedding the excised tumor in paraffin, we take standard sections (4-6 mm) and then conventionally stain the specimens with hemotoxylin and eosin. These are extracts from a tree which grows in South America. When I started all of this I found that by looking down the microscope we could do the diagnostics. The method is low tech, and I am still surprised it works so well. Its success demonstrates you can do great research without having the best labs in the world.

 What were the greatest challenges in performing and presenting your work?

The hill we had to climb was persuading oncologists in North America to adopt the method, which was done on the conference circuit and by giving lectures. We also had to win over the skeptics by showing the method had a satisfactory level of consistency and reproducibility.

 These days, how do you decide where to submit or publish your papers?

I support several journals, and the choice naturally depends on the precise nature of the research being reported. Recently my colleagues and I have been publishing basic science papers in Histopathology, Human Pathology, the American Journal of Surgical Pathology, and the Journal of Pathology. Recent clinical papers are published in The Breast and the European Journal of Cancer.

 What message would you like to convey to the public about your work?

This has to be a very encouraging one: by careful study of the basic structure of tumors we can provide useful information to the surgeons so the treatment is tailor-made to the needs of the patient. We can place individual women into separate prognostic groups, each with an appropriate therapy and counselling.

 How do you see the current state of affairs in breast cancer research and its prospects for the future?

Right now the prognosis depends on basic pathology. The future is in accepting we have prognosis and then developing factors to measure therapy outcomes from new classes of drugs. We need to develop better predictive factors, through molecular markers such as the estrogen receptors.

Dr. Christopher Elston
Professor of Tumour Pathology
University of Nottingham Medical School
Nottingham, UK