Chapter 03: Work at NCI/NIH, the Role of Statistics in Medical Research, and its Application in Sequential and Combination Cancer Treatment Evaluation


Chapter 03: Work at NCI/NIH, the Role of Statistics in Medical Research, and its Application in Sequential and Combination Cancer Treatment Evaluation



Media is loading


Dr. Gehan talks about working for, and remembrances of, Nathan Mantel (known for survival rate analysis) at the NCI/NIH. Also at that time, he received direction from Jerome “Jerry” Cornfield, senior statistician (best known for the causal relation between cigarette smoking and lung cancer). He talks about the importance of statistics in medical research and publication, gives an example of a “6-Mercaptopurine (6MP) and Methotrexate (MTX)” treatment response rate study, and how he provided “analytical support” for Drs. Emil “Tom” Frei III and Emil J. Freireich’s sequential and combination treatment “ideas” for Leukemia treatment.

At an MD Anderson dinner event the prior evening, Mr. Gehan mentions several prominent researchers (Dr’s Emil J Freireich, Emil “Tom” Frei III, Michael Keating, Jeffrey Gotlieb, Levy[?], and 6th Annual Emil J. Freireich Award for outstanding achievement of a young researcher in clinical cancer therapeutic award recipient Eric K. Rowinsky), their passion for research work, and being “turned on” serving as the statistical “member of the team.” The beginning of work with Drs. Frei and Freireich in January 1958, Dr. Gehan lists his roles over time as the replacement for Marvin A. Schneiderman, his career advancement to Acting Head, then Head of the Biometrics Section, Cancer Chemotherapy National Service Center and NCI Southwest and Eastern Clinical Trials Cooperative Groups. As a member of the Acute (now Cancer) Leukemia Group B, he reminisces about the face-to-face meeting style where Dr’s Frei III, Freireich, and James “Jim” F. Holland developed ideas for clinical studies. Additionally, he mentions his marriage to wife Brenda (nee McKeon) before his time working for Sir David Roxbee Cox (known for Regression Models and Life Tables) from 1962-1964 as a special Fellow at Birkbeck College of London, England.



Publication Date



The Historical Resources Center, Research Medical Library, The University of Texas Cancer Center


Houston, Texas

Topics Covered

The Interview Subject's Story - Professional Path; The Researcher; On Research and Researchers; Understanding Cancer, the History of Science, Cancer Research; MD Anderson Culture; Personal Background; Professional Path; Funny Stories; On Research and Researchers

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.


History of Science, Technology, and Medicine | Oncology | Oral History


Edmund A. Gehan, PhD:

Anyway, I started in 1958 [NCI] working with Nathan Mantel, who died last summer, actually. He comes into the story later. I worked for him for about a year. I think of all of these different stories… Nathan was a tremendously intuitive statistician. He was the one who first taught me about consulting because when I first got to NIH [National Institutes of Health], Jerry [Jerome] Cornfield, who also will come in later, was the head statistician. NIH has the Cancer Institute, the Heart Institute—about eight institutes. Jerry was the senior leader. He would tell me about certain investigators, “Go see so and so.” Well, so and so would transfer all the data on his desk to me to me [and say], “You take this away and try to make some sense out of it.” Nathan was my boss. Nathan was always walking around—he never seemed to have anything to do, and I couldn’t understand this. Then I went with him on one of his (?) and he taught the investigators to do their own statistics. He didn’t take anything away. I can tell a number of stories about shortcut methods that he used where he taught them to use their own data. I have come in that direction. Anyway, we are still not to M. D. Anderson.

James S. Olson, PhD:

May I ask you a question?

Edmund A. Gehan, PhD:

Sure, any time.

James S. Olson, PhD:

Was there any resentment among the physicians at NCI to having a mathematician or statistician come in and start looking at the data, or did they welcome it?

Edmund A. Gehan, PhD:

They welcomed it. I guess one of the things I had to do back at Chapel Hill is to teach statistics to the medical students, the dental students, and the nursing students. There must have been a hundred or so medical students in the class, but the ones interested in research are the ones that you could talk to. I mean, you don’t get a paper in the New England Journal of Medicine or the good journals unless it has solid analytical support. Some of the students would read the paper. The ones at NIH, particularly Tom Frei and J Freireich, and there are specific examples of this. I guess one particular example is not really on this list, but one of the first studies I worked on was the 6MP and methotrexate study. They studied 6MP treatment and methotrexate treatment, and the combination. Well, 6MP had a certain response rate and so did methotrexate have a certain response rate. You can find a write-up in the paper about what is the predicted response rate of the combination. The analogy I drew was in the paper, if you shoot an arrow at a target, the patient, with a certain probability of a hit with 6MP and you shoot another arrow at it with a certain probability of a hit with methotrexate, you can calculate from that. You don’t need both of them to hit to be a success. You only need one. That is a successful treatment. It is one minus the probability that neither one hits, is the probability that one or the other hits. What we were able to show in this particular paper is that the actual response rate of 6MP plus methotrexate was very close to what you would predict by shooting arrows at the target. You’d have to ask Tom Frei and J Freireich where the idea would come from to study multiple drugs given at the same time that didn’t have additive toxicity. If you are adding the toxicity, you are in trouble. That is one example within a paper that is in my curriculum vitae. Frei was the first author, “Studies of sequential and combination treatment in leukemia.” I helped to provide the analytical support to some of the ideas—both at NCI and later here—sort of in a broad sense. Both Tom and J were research-oriented people that realized that someone like myself could help them. Tape 1, Side B

Edmund A. Gehan, PhD:

In a way, they should have invited to all to come to that [7th Foundations of Clinical Cancer Research Symposium].

Lesley W. Brunet:

We don’t get invited to those things.

Edmund A. Gehan, PhD:

I talked some with Thomas (?) and he said to me that he is going to send you up to Michigan to see Dr. Hickey. Oh, you didn’t know that?

Lesley W. Brunet:

I am glad it is getting to be Spring.

Edmund A. Gehan, PhD:

Time to go.

Lesley W. Brunet:

That’s good.

Edmund A. Gehan, PhD:

Freireich is great. How many people did they have at that dinner? Fifty? Something like that. I don’t know, but different tables. Somewhere along the line, Freireich gets up and says, “The cost of your coming to this dinner is to get to give a speech.” He sort of presses people to give some kind of a talk, so people do. Of the people last night, Tom Frei gave a nice talk and told a very good joke. I gave a talk. Michael Keating and Freireich were kind of co-MC’s [Master of Ceremonies]. They got Dr. Levy [Robert], a very fine man, and of course a lot of the talks were in reference to Gottlieb [Jeffery A.], but this is the 28th [Annual Award]. My talk was more about Frei and Freireich last night. Rowinsky [Eric], a younger guy, getting the Freireich Award, he talked later. Was it about they are all honored to be with Frei and Freireich? What is it about them? And he mentioned the passion—the passion. And he was right. There was a certain amount of camaraderie, but also passion for doing better. This is part of the theme. You know, mathematics is really kind of dry, but statistics is… I have always turned on by sort of being a member of the team where the statistical part is an important part of the progress. You are a member of the team and this is the part that you are contributing to. I have just given you a statistical example. You could say, “I don’t really think that Gehan has showed that… and therefore we can give…” It was broader than that. I certainly wouldn’t claim that certain mathematical arguments would suggest that combination chemotherapy was better. But, it can be seen as part of the justification anyway.

Lesley W. Brunet:

When did you first join that Frei/Freireich team at NCI?

Edmund A. Gehan, PhD:

Well, I started in January 1958. I worked some with Freireich. I would have to review the larger curriculum vitae. I remember working with Freireich, in particular, on several projects. It was in 1958. But, you know Freireich was kind of a bear to people. He would tend to put people down. I guess I can be thankful that I was in a different field. He wasn’t going to put me down in my field, and I think he could see that I could of help him. I basically started working with both of them then, where it says [curriculum vitae] “Acting Head” and then “Head of the Biometrics Section, Cancer Chemotherapy National Service Center. Another good man at NCI was a guy named Marvin Schneiderman. He left to get his Ph.D. in England, so they needed somebody to replace him. In 1959, I am 30 years old then, a youngster, they appoint me Acting Head and then later Head of this Biometrics Section. Well I am sure you have heard of these cooperative groups, including the Southwest [Oncology] Group and the Eastern [Oncology] Group. This section kind of oversaw all of those statistical groups. Some of the groups, in particular the Acute Leukemia Group B, I was the statistician for that group, and we had a statistical clerk that worked on that. Tom Frei was the chairman of the group and J Freireich was one of the members, so beginning then I worked very closely with them as part of the Acute Leukemia Group B, which is now the Cancer and Leukemia Group B, which is still in existence. I told a couple of stories even last night. In some ways it was more fun. I remember a particular meeting at Mt. Sinai in New York. Maybe the conference table was bigger than this, maybe one and a half times [bigger]. But the whole group is around the table. One of the stories I told… (Laughter) Tom is just an amazing guy. He is the chairman, so he was making certain announcements at the beginning. He has an index card and he starts, “This is what we are going to do in the meeting,” and this and that. As he talks he keeps referring to this index card and saying something about the meeting and what’s going to happen. How in the hell did he have all that stuff on this index card? So I looked at it later. The index card was blank. (Laughter) He didn’t have anything on the index card. The other thing about it was that how were ideas for clinical studies developed? They were developed at the blackboard. Dr. X (?) was offering Dr. Freireich, “I think the treatment should be this.” Then they would write it out. Then the other group should be this. At that time, Frei was the moderator. He is a fantastic moderator—great, I think. And J was very outspoken. On one side, Jim [James] Holland, I am sure you have heard that name before, was very outspoken. By the time all of the arguments [were presented] for this or that approach, everybody had a pretty good understanding of what it is we are finally agreeing to. What would happen in the course of the discussion, somebody would go to the blackboard and erase, “No, it shouldn’t be that—it should be this.” Then various ideas would be put forward. Eventually it would be agreed to, and Tom was the one for bringing people to agreement. Then they would say, “OK, well let’s do this group for the next three months and see how things go.” [There was] no written protocol, no IRB, no approval mechanisms; it was just done. There was a certain level of excitement in what was being done. One of my analogies is it was more fun back then. I haven’t been to a Southwest Oncology Group meeting for a long time, but one of my statistical theories… Everyone is trying to find the cure for cancer. What is the chance that you will find the cure for cancer if you are a member of a team and you don’t even know all of the people on this team. (Laughter) It’s just gotten so big. It was more fun back then. Is it better now? In many ways it is better, but I think it was more fun [back then]. I became good friends of them there. For myself, from 1962 to 1964, I was a special fellow at Birkbeck College of London [England]. Perhaps one of the most famous people in statistics is Sir David R. Cox. In 1956, he was a visiting professor in Chapel Hill, so I first met him in 1956. He is now Sir David Cox, but I could see that he was a brilliant man. I was immersed a lot in data, but an important decision for me—I was kind of interested in doing some research work on things that came up at NIH, so I applied for this fellowship to go to London and I was fortunate enough to get this fellowship. So I was working in Professor Cox’s department, which at that time was at Birkbeck College of London. Birkbeck College is a night school of the University of London. It is not one of the better colleges. He is at Oxford University now though. He eventually became the Warden at Meffield (?) College, which is a very prestigious position, and he is still at Oxford. I’ll think of the other school in London. I worked with Professor Cox for two years. I married my wife Brenda in part because we were going to have to go to London, and you have to either fish or cut bait at that point, you know, in terms of personal relationships. I came back to NIH, but I wasn’t too happy for reasons that I don’t want to [talk about]. It just would take too long to talk about.

Conditions Governing Access


Chapter 03: Work at NCI/NIH, the Role of Statistics in Medical Research, and its Application in Sequential and Combination Cancer Treatment Evaluation