| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, Room H-1203, New York, NY10021, USA
Correspondence: Address correspondence and reprint requests to: Murray F. Brennan, MD; E-mail: brennanm{at}mskcc.org
Key Words: Retrospection • Career • Clinician • Investigator • Surgeon
Mr. President, ladies and gentlemen, it is a privilege to deliver the John Wayne Lecture in Clinical Research. When asked to deliver this lecture, it gave me time to reflect on 30 years of clinical research. Of great concern, however, was how to make it interesting and how to make it sequential. I thought I would begin with lessons that I have learned about giving a talk. They are rather simple lessons and they involve the why, the when, the who, the how, and the clear awareness that boredom can kill the finest of talks. I have learned that the occasional frivolity will reawaken the audience and succeeds in delivering a message that is lost in repeatable and uninterpretive slides. I hope my presentation will not be too pixelated!
The question of when to give a talk is fairly simple—it is either by invitation, because you have new information, or less attractively, self-promotion for you or your institution. It is very important to emphasize the high risk of failure in an invited talk. The inviter often thinks you are better than you are, and almost always wants you to talk about that which you know least. This is particularly true as you reach my level of evolution or devolution. No longer am I invited to talk about the clinical issues that I know most about, but more often I am invited to speak of the future. Frustratingly, those who are invited to talk about the future are those almost certainly destined to not take part in it! If you think anyone is capable of predicting the health care future, you need only look at the failures in predicting surgical manpower needs over the last 20 years.1
I begin to plan a talk with a simple table of contents, and for this talk it consists of: why this title, why is it important, what is the lesson? Then the body of the talk is: how can I tell a story and not have you all asleep within three to four minutes? My decision was to describe how it all started and with whom; I could then weave learning into time, and include the heroes and the heroines.
Why this title? It was an invitation by President Eberlein, who I have admired since we worked together at the National Cancer Institutes (NCI) of the National Institutes of Health (NIH) more than 25 years ago. If the talk is invited, there is poetic license, so you can change the title! An invited title that began as "Lessons learned from 30 years of clinical research," and appears in your program as such, I have converted to the present title, "Lessons learned..." You can decide about the solatium. The question of why it is important and what a lesson is can be simply summarized: a lesson is something that needs to be learned!
In an effort to bring some thread to this presentation, I utilized the concept of Tusitala, who was the "Teller of Tales." For those who are unaware of who Tusitala was (living as he did at Vailima in Apia, Samoa), you can ask your colleagues beside you or you can wait until the end and I will remind you.
The story starts long before 30 years ago and it begins with an understanding of the accident of your birth. You could have been born Theodore Roosevelt or Winston Churchill or Mohandas Ghandi or Albert Einstein, or Adolf Hitler, or perhaps even Attila the Hun, the scourge of God (some of my fellows see me as his reincarnation). Being born is a true accident, but the rest is truly up to you. You might be born as John Kennedy, Jr. in an environment of privilege and expectation, but gain notoriety on the front page of Time magazine only because of an untimely and tragic death.
Throughout this talk, I will emphasize the importance of the lesson of needing "tools to grow." They are the initial tools that you are given by your parents, by family, by school and by university, but they will last a lifetime. As I have emphasized so often to my own fellows, we are not a finishing school, we are a school in which we endeavor to provide them with the tools to allow their continued growth throughout their professional lives. My initial tools were obtained in New Zealand, somewhat south of from whence Tusitala discoursed. I was privileged to go to a wonderful university, the University of Otago, and, like many, I did receive a lucky break—Dr. Francis Moore, who appeared on the cover of Time magazine, May 5, 1963. If you look at the cover of that Time magazine, the sidebar is: "If they can operate you’re lucky!" I was fortunate in that he came as a visiting professor when I was the equivalent of a senior resident. For me, Dr. Moore emphasized the importance of a mentor. Mentor was the advisor of Telemachus, the son of Odysseus and Penelope in Homer’s Odyssey. Even Homer’s Odyssey tells us a little about academic surgery—it was written almost three hundred years after the events that it describes! We like to think of a mentor as "an experienced trusted advisor, counselor, and guide." That does not say anything about his or her intelligence, political appropriateness, color, race, ethnicity, or indeed his or her interests. A genuine mentor is truly politically color-blind.
I learned a great deal from Dr. Moore. One of the many lessons was that you can learn a great deal from a single patient carefully studied. His book, The Metabolic Care of the Surgical Patient, published in 1959, contains 32 cases, 27 clinical procedures, and 1,000 pages. It is dedicated to "three surgeons whose work was prematurely cut short by death in their prime." In this, through careful attention to detail, he dissected the body composition of normal man and the surgical patient under varying conditions of stress and injury. This is a remarkable text, which deserves rereading today. He is most famous for his description of the surgical investigator: "The surgical investigator must be a bridge tender, channeling knowledge from biological science to the patient’s bedside and back again. He traces his origin from both ends of the bridge. He is thus a bastard, and is called this by everybody. Those at one end of the bridge say he is not a very good scientist, and those at the other end say he does not spend enough time in the operating room."2 This emphasizes the lesson that there is nothing new under the sun, and only the content, not the principle, has changed. During my time in Boston, I also learned what a Harvard answer is. A Harvard answer is always correct, but often has little to do with the question asked! Examples of this lesson can be enumerated; a hypothetical one would be the question: "What is the patient’s serum potassium?" Answer: "In the New England Journal of Medicine this week there is an article on the relative merits of ADH suppression in fulminant sepsis." The question is clear and the answer is probably true, there was an article on some erudite subject that had nothing to do with the question. Certainly, I have utilized this awareness to focus some of my trainees when they stray from the specifics of the question. The other component of such an answer is the "I would probably..." There is no "probably" in surgery. If you want "probably," become a dermatologist.
One cannot let the concept of mentoring go without being reminded of Menelaus’ comment to Telemachus in Homer’s Odyssey: "Treat a man well while he is with you, but let him go when he wishes." The lesson from this is that not all mentors realize that it is supposed to be about the future of the mentee, not the future of the mentor. For the mentees in the audience, look to your mentors. For the mentors, look to your role and be sure that it is about the mentee, not about yourself. Others have identified this concept of letting people go, letting the bird fly free, and if they return, that is a gift. If they do not return, you never really knew them anyway.
My next set of lessons was at the NIH, the most important one of which was that you should take advantage of the environment in which you work. The NIH was a wonderful environment in which to work, and I first began my obsession with the concept, or the lesson, that to know something you need to record the observations, gather the data, and look at the findings—nothing unique about that. Lord William Kelvin (of degrees Celsius fame), said, "To measure is to know." Once you begin to gather the data, record the observation, and define the question, you realize that collaborators matter. It seems ironic to me that we seem to be rediscovering today that the only way a clinician-investigator can survive is with a collaborator. That is not new, collaborators were always essential. This absurd misrepresentation comes about, I believe, because of a gross misunderstanding. When the study of man was man—nutrition, shock, hemorrhage, metabolism, technical issues of transplantation, then, indeed, the independent surgeon could be the all-embracing scientist. Francis Moore epitomized that. But times change; the biological revolution accompanying the human genome project changed all that. Science became molecular science, and the surgeon was slow to recognize that he or she could not compete if they wished to be both surgeon and molecular biologist. The latter field was moving too fast for part-time devotion. The realization that this could only occur as translational collaboration was natural and inevitable, but only the nature of the collaboration has changed. My collaborators were clinicians or clinician-scientists, pathologists, epidemiologists, radiologists, statisticians. Now they are basic scientists. There were many, but the most prescient was my collaboration with John Doppman, one of the very first interventional radiologists.3 He was a remarkable man, and his contributions were highly focused on interventional radiology designed for the localization of functional endocrine tumors, whether that be a pituitary adenoma, an aldosteronoma, or a patient who had persistent hyperparathyroidism following one or more prior operations. I worked with John Doppman in a way that truly taught me about the essence of collaboration. A totally unselfish man, he was absolutely uninterested in who got the credit for the work. His most famous surgical quote was, "The only localization that a patient needs who has primary hyperparathyroidism is the localization of an experienced surgeon!" In contradistinction were his remarkable efforts at localizing the offending parathyroid in those with persistent or recurrent disease.
I learned some other lessons at the NIH, one of which was that some old lions turn out to be real pussycats. I wrote a manuscript on the role of surgery in patients with Zollinger–Ellison syndrome, which was presented at the American Surgical Society in 1982.4 A young man presenting at that meeting can often be intimidated, but I was naïve and courageous, and so I sent the abstract to Dr. Zollinger asking him to discuss the paper. Neither I nor any collaborators were members. I did not hear anything in return, so a week or two before the presentation, I called him to say that I was sending him the manuscript. You can imagine my terror when he replied, "I sure as hell hope that it is better than the abstract." This fine-tuned me for the presentation, and as he got up to speak I am sure if beta blockers had been known then, I would have preferred to be on them. He began his opening dissertation with some comments stating that if he had realized how many problems he was going to create he would have "never invented the disease." He then proceeded to be quite charming and most complimentary. I still recall the feeling of one’s pulse returning to normality.
Another important lesson learned at the NIH was what I might call the "Charles Huggins Law," the concept of being willing to ask the simple question. Dr. Huggins received his Nobel Prize in December of 1966, and his talk was on endocrine-induced regression of cancer. If you think of the simplicity of the question: a Nobel Prize for cutting off the testes of a rabbit!
Another lesson that I learned was that if you wait until you have better agents, better operations, better treatments before defining a question or defining a trial, very little gets done. If I had a regret, it is that from the time I came to Memorial Hospital, I did not pursue all of the questions that I thought should be answered, often being dissuaded by others that somehow some treatment would be better next week or next year and we would have to ask the question again. At the NIH, Steve Rosenberg designed a trial comparing amputation and chemotherapy with limb-sparing surgery and radiation therapy for soft tissue sarcoma—a simple question with enormous consequences for the patient. The patients at that time were randomized to receive or not receive an amputation. You can imagine the extraordinary emotional difficulties of obtaining informed consent. While it was a two to one randomization, and the total number of patients was 44, there was no significant difference in local recurrence, disease-free survival, or overall survival.5 Such a trial today would be considered underpowered, underfocused, but in fact gave strength to our practice today. When we think prior to that trial, greater than 50% of all patients with soft tissue sarcoma of the extremities underwent an amputation, and now less than 5% of those we see come to amputation. Combine that with the belief at that time that all amputations for sarcoma had to be at least one full joint above. Think of all the hemipelvectomies and forequarter amputations we no longer do.
Another lesson I learned at the NIH was the concept that changing from mentee to mentor is tough. It is difficult and involves a number of concepts that have become lessons for me. I have recognized that greatness is not genetic, faculty development must be deliberate and strategic, admiration is definitely not mentoring, and that each mentee is unique. Most mentees develop far better if recruited into a well-defined position for which the mentee has the defined skills and the mentor and mentee both have appropriate expectations. It is valuable to have some form of measurable outcome, but that is absolutely not crucial. Jeff Norton began with me as a clinical associate at the NIH and subsequently succeeded me as the Chief of the Metabolism Section. We were fortunate enough to put together our experience with reoperation for persistent recurrent parathyroidism in 1985.6 A paper examining patients with severe persistent and recurrent parathyroidism was, I thought, an important contribution from mentor to mentee.
As Tusitala would say, I need to "get on with the tale." I came to New York in 1981. As some of you know, I have an interest in John Steinbeck and his description of New York in 1953: 7 "I don’t think New York City is like other cities. It does not have character like Los Angeles or New Orleans. It is all characters—in fact, it is everything. It can destroy a man, but if his eyes are open it cannot bore him. New York is an ugly city, a dirty city. Its climate is a scandal, its politics are used to frighten children, its traffic is madness, its competition is murderous. But there is one thing about it—once you have lived in New York and it has become your home, no place else is good enough." While most of you would not agree with that description, for me it has a large degree of truth. While I see New York City about as far removed from where I began as is possible to conceive, it has a wonderful challenge and is wonderfully character-building. The important lesson here is that where we work is important, but the resources to do the job are way more important than the personal rewards. It is terribly important as you begin to negotiate for your first job, or even your leadership job, that your own personal rewards are deferred. If the resources do not exist to achieve your goals, you will remain unhappy regardless of the personal and (particularly) the financial rewards.
It matters not only who you work with, but for whom you work. That is more important early in your career rather than later. I was recruited by my predecessor as Chairman, Jerome DeCosse, and fortunate to work for the President, Paul Marks, and the Physician-in-Chief, Samuel Hellman, at Memorial Sloan-Kettering—a wonderful example of what clinical leadership can achieve. Memorial Sloan-Kettering was a well-established institute with many achievements when I joined it. It taught me about the importance of recognizing what has been done in an institution before arriving. The description of the Stewart–Treves syndrome by Norman Treves and Fred Stewart in Cancer in 1943 was an important contribution in the recognition that lymphedema can result in the development of angiosarcoma after major breast operations and extensive radiation. This led me to learn a further lesson, that the observations of the past may well influence the future. We now see an increasing incidence of angiosarcoma after treatment for primary breast cancer, with the odds ratio of developing angiosarcoma of the chest or upper limb approximating 60 times what would be expected if the patient receives breast irradiation.8 While angiosarcoma is still a rare disease, this should give us pause for thought as we liberally utilize radiation following lesser surgical procedures for minimal breast cancer.
The most important lesson that I learned at Memorial was that the environment is the people. You can certainly succeed with good colleagues and limited facilities, but you cannot succeed with poor colleagues and great facilities. In that I have been extraordinarily fortunate. I began in 1985 as chairman with an extraordinary leadership group, and now, over 20 years later, I have a similarly distinguished leadership group; none, however, being retained from the first group. If nothing else, that suggests that it is also time for me to move on.
Another lesson that I have learned is the concept that if you throw a brick out of the window and it goes up, you do not need a randomized trial. I am appalled by our preoccupation with diseases in which we have changed nothing in 30 years, and yet we continue with the same tired and hard-worn approaches. If we look at the survival of adenocarcinoma of the pancreas following resection with 11% absolute survival at five years and half of those patients going on to die in the next five years, the decay curve for survival is relentless. We do not need a randomized trial; we need something that moves that curve far to the right. In similar fashion, while statistical significance is OK, the eyeball test is key. We have shown that the overall survival by year of surgery for resected pancreatic cancer has statistically improved since 1984 when we began our database (P = 0.001). But if you look at the curve, the initial decay is so dramatic that it is unlikely that we have made any impact that is clinically important, and statistical significance is only achieved by the large numbers. This reminds me to rant some more about the preoccupation of statistical significance in huge randomized trials when the clinical benefit is measured in single digits.
One of the great and most fortunate lessons that I have learned is the importance of youth. Without the commitment of youth nothing is achieved. An alternative way to describe this is that "fertilizer works better than insecticide." So, encouragement and support are hallmarks of development, not the cynicism and criticism that I experienced on occasions as a young trainee. One thing that I have had to learn is that it does take time for an old dog to learn new tricks. Our fellowship program was dominated by the Y chromosome until only recently. Things have changed and I now find myself, on occasions, in the operating room and I am the only Y chromosome in a sea of Xs, all extraordinarily capable of doing their job and much more attractive!
What I did consolidate at Memorial was the importance of databases, not only the development of prospective databases. Our sarcoma database, now with over 7,000 cases followed continuously for more than 20 years, provides a rich source of information and knowledge, also emphasizing that data teaches you not only about outcome but also about biology. A simple example is that if you examine local recurrence based on site, one will see that visceral sarcomas have a local recurrence rate of less than 20%, similar to that of the extremity. Retroperitoneal lesions, however, have a local recurrence rate of at least 60%. When you look at disease-specific survival (death from disease), then the extremity lesions have a mortality of approximately 40% and visceral and retroperitoneal lesions have a death rate that exceeds 60%. This teaches us that patients dying of extremity lesions die from systemic disease, not from local recurrence. Not too surprisingly, it also teaches us that visceral neoplasms also die from systemic disease with a very low local recurrence rate. Conversely, retroperitoneal lesions with a high local recurrence rate die of local recurrence. Without a randomized trial, biology is clearly defined and it makes limited sense to do systemic trials when death is from local recurrence, and little sense to focus on local control as an important factor in survival if change in local control does not translate into decrease in systemic disease. It has been shown that in soft tissue sarcoma, lymph node metastasis is extraordinarily rare and clearly dependent on the underlying histopathology. The presence of lymph node metastasis in a patient with liposarcoma makes one even question the diagnosis.9 It also says that we can rescue the occasional patient with a lymph node metastasis by lymph node dissection, and that they do not have the absolute mortality of those who have systemic disease.
A lesson that has been difficult to accept is the lesson that residents and fellows do not realize that you wrote the paper that they are quoting, and it is OK if they think anything done more than five years ago is either wrong or needs repeating. Early on, I thought I knew a great deal about total parenteral nutrition, and indeed wrote a lead article in the New England Journal of Medicine,10 but no one recognizes that and gives me lectures about the relevance of nutritional support. But then again, surgeons have given up, in the main, the management of nutritional support, pain control, antibiotic usage, intensive care, palliative care, and many other aspects inherent in care of the cancer patient—the subject of another dissertation! Dr. Moore would be appalled!
A lesson we all learn over and over again is that if you are going to do a study, try to do it correctly. We did a study looking at expression of the retinoblastoma gene product in sarcoma tissue derived from a heterogeneous group of patients, primary and meta-static, young and old, that was published in the New England Journal of Medicine11 in 1990, showing a strong association with the loss of the expression in patients with poor outcome, but then when we sat down and dissected that problem, examining only large, deep, high-grade extremity lesions in a sequential fashion, we found that the single gene product expression was not discriminatory when it mattered.12
One lesson that continues to bother me is the focus on the concept that a single molecular marker is sufficient to discriminate such a complex disease as sarcoma or, indeed, cancer. We have tried to help differentiate the weakness of conventional staging systems by use of nomograms.13 The wonderful potential for nomograms is that they may help us to dissect the significance of molecular markers. With enough information, it is possible that we can add molecular markers to current nomograms to add individual validity for both marker and patient outcome.
There are innumerable other lessons that I have learned, and one of the most important is that if at first you do not succeed, then you must continue. My first attempt at randomized controlled trials at our institution was accompanied by comments that "why would we possibly want to do such a trial when we already know the answer?" We do not know the answers, and surgeons can do randomized trials. We performed a randomized trial that involved an intraoperative randomization to receive or not receive adjuvant brachytherapy for soft tissue sarcoma; a well-controlled trial14 relying on our knowledge of the importance of appropriate stratification. My initial role beyond writing the trial was to beg other attendings for permission to randomize their patients! That trial confirmed the importance of radiation and its ability to control local recurrence of soft tissue sarcoma. Most dramatically, however, the impact was primarily on the patients with high-grade lesions, i.e., those most likely to have systemic disease. What better test of the concept that if local control can make a difference to survival, then a major impact on local control in high-grade tumors will translate into survival benefit. Unfortunately, this is not true, and there is no survival benefit now beyond ten years.14 In our initial randomized trial of radiation, a radiation therapist challenged us that we were depriving patients of something that they should receive. Only when they realized that 50% of the patients would get the therapy if they entered the trial and as many were not being treated because of clinical judgment by the attending surgeons, did they willingly participate.
This important observation was paralleled by other similar observations that we investigated by randomized trials. It is very important to challenge your own biases. It was my belief that total parenteral nutrition (TPN) was an important adjunct in the management of the cancer patient and particularly the postoperative cancer patient undergoing extensive resection. We therefore performed a prospective randomized trial looking at the routine application of TPN to patients undergoing pancreaticoduodenectomy. Despite randomizing 117 patients, we were unable to show a benefit, and indeed there was a greater infection and complication rate in those receiving parental nutrition.15 I encourage you to challenge your own prejudices and biases; sometimes you will be right, but often you will be wrong.
Such randomized trials do not, however, define the absolute, and I now find myself not routinely applying radiation therapy despite our randomized trial, but becoming progressively more selective as I make the judgment between risk and benefit. Time for another randomized trial! But randomized trials are difficult, and institutional review boards ever more rigorous. I do realize the need for patient protection, and full disclosure, but some of the actions generalized by slavish adherence to HIPAA (Health Insurance Portability and Accountability Act) are ludicrous. This tempts me to digress to lessons learned in patient safety, where we have learned that one way to make clinical care more effective is to minimize the steps from order to action. When will we realize that minimizing the steps from concept to entry in a trial will be the most efficient way to benefit patients? The relentless imposition of bureaucratic oversight has created barriers that, while intentionally sound, will so strangulate new information that while we may protect this generation in a politically correct way, we will injure the next generation because of ignorance and lack of valuable outcome analyses. Knowledge unknown is never missed.
Another lesson that is extraordinarily important is that while opinions matter and ideas matter, common sense matters. I have been enormously frustrated by the discussion of perceived benefits of adjuvant chemotherapy in soft tissue sarcoma. We tried desperately to write a prospective randomized trial and we failed, not because of people who believed it worked or did not believe it worked, but because different individuals held one opinion to the exclusion of another. The ideal scenario for a randomized trial you might think, but we have failed. When we look at the meta-analysis of chemotherapy in soft tissue sarcoma, one can say that ten-year overall survival without chemotherapy is 50%, and with chemotherapy it is 54%, not a significant difference.16 I ask you to apply the common sense rule: with treatment 46% recurred anyway and received no benefit, and without treatment 50% never recurred and so did not need it. So at best, one patient in 25 has any chance of potential benefit. The fact that this data was not significant makes it even more concerning that there is no benefit. I find this a frightening concept when I look at improvements in survival in other diseases from 90% to 92% by some form of adjuvant therapy. Emphasizing that only one in 50 patients can possibly benefit, presuming that there are no side-effects whatsoever, and believing that we can actually deliver anything in biology that is 2% is frustrating. No treatment is ever risk-free.
Finally, I leave you with a sobering thought. Treasure the people whom you meet, because some will indeed leave sooner than others. A distinguished member of this society, Michael Burt, won the laboratory research prize and the clinical research prize of the society in 1980 and 1981, but his ever-ascending career was tragically shortened at the age of 49. So, turn to those that you care for and remind them that you do care, because there may be an occasion when they will not be there to be with you, and you will regret not telling them.
And finally, who was Tusitala? It was Robert Louis Stevenson, as he sat on the verandah at Vailima in Apia, Samoa in 1892. A more fitting conclusion to this presentation would be the final verse of his Requiem:17 "This be the verse you grave for me: Here he lies where he longed to be; Home is the sailor, home from sea, And the hunter home from the hill."
 |
FOOTNOTES |
|---|
Received for publication March 4, 2005. Accepted for publication January 18, 2006.
|
REFERENCES |
|---|
 TOP REFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
