Why Measure Lung Structure?

EWALD R. WEIBEL

Department of Anatomy, University of Berne, Berne, Switzerland

	ARTICLE

TOP ARTICLE

I have chosen to discuss here the background of the first publication of the morphometry of the human lung that appeared as an article in Science in 1962 under the title "Architecture of the Human Lung" and was authored by me and Domingo M. Gomez (Science 1962;137;577-585). The work that led to this paper was the crucial turning point that made me change my career plans completely, and thus determined my future path.

A preface on my background is needed. After completing Medical School in Zürich, my original ambitions were to pursue a career in academic medicine, combining clinical and research work. However, at that time the prospects for such a career were quite poor. It was customary and expected that any young ambitious doctor who wished to pursue a career in a University Hospital Department should work as a volunteer at no pay for two to three years. This I could not afford, as I was not from an affluent background. My father was a typewriter mechanica highly skilled profession as we were still in the time when clever lever systems (rather than electric power) were used to magnify finger strokes. He could help me through medical school, but after that I had to be on my own. So I tried an indirect track by obtaining a qualification in research and basic sciences in the Department of Anatomy, where I had done a thesis on the development of the inner ear. After that I was to go to pathology, and from there on to surgery. In planning this track the Chairman of Pathology offered me a paid job if and only if I had done some preliminary work on the anatomy of lung vessels, specifically on the question of whether there were anastomoses between the bronchial and the pulmonary arteries. So with the consent of my Professor of Anatomy I studied this topic, obtaining some interesting and partly controversial results.

These studies made me eligible for a postdoctoral fellowship awarded by the Swiss Academy of Medical Sciences that would allow me to go abroad in 1958. In view of this I wrote to Professor André Cournand at Columbia University in New York. I wanted to test my ideas on the anatomy of these anastomoses by a physiological approach, and his group had just published interesting studies on collateral circulation to the lung. Furthermore, the year before André Cournand had been awarded the Nobel Prize for his work on the pulmonary circulation by cardiac catheterization. So he was a very famous man. But my airmail letter to him returned several months later with the postal remark "addressee unknown." Not having heard from him for weeks I had concluded he was not interested in receiving me and contacted Professor Averill Liebow at Yale University who immediately accepted me for a two-year fellowship to study the development of bronchopulmonary arterial anastomoses.

When my letter to Cournand was returned I wrote him again explaining what had happened. I received an immediate angry reply that he could not understand why I was so impatient (the obvious reason was a deadline) and that I had accepted a post inferior to what he could have offered. But he invited me anyway to present a seminar at Bellevue Hospital in early 1959, half a year after my arrival at Yaleit was to be my first seminar in English. After that lecture Cournand asked me to come with him to the office of Dickinson W. Richards (there is an interesting parallel to this in the article in this series by Alfred P. Fishman [Am J Respir Crit Care Med 2000; 161:692-693]). There I was offered a position by these two Nobel laureates as Research Associate in the cardiopulmonary laboratory, and they said they would pay an additional $3,000 per year as a supplement to whatever fellowship I had. I might mention that my fellowship from Switzerland amounted to $3,250 per year. So that was a significant offer. When I asked them what they expected me to do, Cournand answered bluntly: "Do anything on the structure of the lung that is of interest to physiology." That was of course too good an offer to refuse. I immediately accepted, went back to Yale to work frantically so as to finish my two-year experimental project in one year and write a paper that was published in Circulation Research. This pacified Liebow who had, of course, been very upset about my change of plans.

When I arrived at Bellevue Hospital in the summer of 1959 I was given funds to set up a small histology lab, and then I listened around to find out what could be "of interest for physiology." I received all sorts of suggestions and questions, but there was one sort of question that clicked. It was asked by the Cuban refugee Domingo Gomez, who had fled Fidel Castro's reign of terror and had been given refuge in Cournand's laboratory just a few months before I arrived. Gomez asked me questions such as "How many alveoli are there in the human lung?" When I asked him why he wanted to know this he said he wanted to calculate the alveolar surface area, which he needed for model calculations.

The background was clear. With the development of new techniques, including cardiac catheterization, enormous advances had been made in the measurement of human lung function. In order to interpret the mechanisms behind gas exchange there was an increasing need for comparable measurements of lung structure. However, a search of the literature showed that there were no accurate quantitative data on lung anatomy, and no comprehensive review of the design of the lung as an organized integrated system. To obtain this type of information would clearly be of interest for physiology, so Gomez and I started to undertake it. However, the methods to do this did not exist. Anatomists traditionally rejoiced in producing beautiful pictures and in describing structures in all their fascinating details. But they saw no purpose in obtaining reliable measurements on their observations. This type of information was only of interest for those studying the lung from a physiological perspective, exactly what I had been asked to do.

Domingo Gomez was a mathematical physiologist and a hot-tempered genius. His particular fascination was to describe functional relations in more or less complicated systems of equations. These he would develop on piles of yellow pad sheets, most of which he would dump in his waste paper basket, shoving those of importance under his blotting pad. In these equations he would also need quantitative information on structural design, such as the alveolar surface or capillary volume, but such information would always have to relate to the entire lung. That was a problem because the structures of interest could only be seen and measured in the microscope on samples much smaller than the size of the lung.

Under the guidance of Domingo Gomez I started to take a completely new approach to the study of lung structure. Together we developed an integrated concept of the organization of the lung, in terms of both the gas exchange region and the airways and blood vessels leading the air and the blood to where oxygen could be exchanged. We also had to develop a new methodology including sampling schemes to account for the structure of the entire lung studied on microscopic sections, and had to find methods to obtain accurate measurements in the microscope. I collected a number of normal human lungs in the Pathology Department of Bellevue Hospital, mostly from victims of criminal assaults, not so rare in those years. I finally had five satisfactorily prepared total lung specimens on which I could start with these studies.

The next problem was how we could obtain the necessary measurements on three-dimensional structures when in the microscope we could study only two-dimensional thin sections where the alveolar surface area occurred as a linear trace. Such methods were not readily available. The first approach was to develop a method for counting alveoli on sections and then calculate the surface on the basis of a model of the alveolus. It may be significant that the method Domingo Gomez and I developed for counting structures such as alveoli was published in the Journal of Applied Physiology (1962;17:343- 348), and not in an anatomical journal. Applying this method to my lung specimens we found that the human lung contains 300 million alveoli. From that we calculated a surface area of about 60 m². We then found out that geologists had developed a method of measuring the surface area directly on sections by using linear probes of known length and counting the number of intersections, the method that is often called the mean linear intercept method. We found that this method gave very similar results to the geometrical approach.

The next problem was to estimate the volume and surface of the alveolar capillaries. For this we developed a very complicated geometrical method which required hours of computation on the giant computer at Columbia's Watson Laboratory. This was necessary because the light microscopic preparations we had available did not give us adequate resolution for direct measurements of structures as small as capillaries. Clearly, we had reached a limit of what could be done with the methods at hand in 1959. It is only later that these investigations could be significantly refined when I could use the electron microscope to measure the structural parameters of the pulmonary gas exchanger.

Since my studies had to be "useful for physiology," the logical next step would have been to collect lungs of patients who had died in respiratory failure such as in emphysema. Together with good physiological data this would have allowed us to make correlations between lung function and quantitative lung design. However, I experienced serious "logistic" problems in procuring such lungs because of a lack of cooperation on the part of some colleaguesa not infrequent experience, I suppose. This is why my studies on the morphometry of the human lung were based on no more than five lungs. I don't know whether one could get by with such small samples today!

But maybe this was a lucky situation because, meanwhile, the appetite of the physiologists at Cournand's laboratory for quantitative morphological information had grown very significantly. At the time, the interest of the group focused on the distribution of ventilation and perfusion in the lung. So, this raised questions about the design of the airway and blood vessel trees as they lead into the gas exchange region. My former mentor at Yale, Averill A. Liebow, had produced excellent plastic casts of the bronchial tree. He lent me a particularly good specimen on which I could do, with the help of a dedicated young technician, some detailed measurements on how the bronchial path changed its dimensions from the trachea out to the peripheral airways. In this analysis and in the formulation of the bronchial tree model, the collaboration with Domingo Gomez was again very crucial. The general result we obtained was very interesting: we found that the bronchial diameter decreases with each generation according to a systematic rule. This rule provides for optimal design in the conducting airways, where mass flow of air is the dominant functional factor. We then found that the dimensions change according to a different rule in the acinar airways, where diffusion of oxygen in the air phase is more important than mass air flow.

The article in Science put all these different measurements into a coherent perspective without giving much detail on the evidence or on the tortuous ways in which it had been obtained. All this work had been done in the two years that I worked in this fascinating environment in the Cardiopulmonary Laboratory of Bellevue Hospital. I believe it was most important that I interacted during that period mostly with physiologists, with pulmonary physicians, and with the superb analytical mind of Domingo Gomez. It may well have been important that I was "shielded" from the critique by hard-core anatomists who proved to have very little understanding of this type of approach, which seemed to neglect all the beautiful fine details in order to draw up a coarse picture of the lung as an integrated system. Whatever recognition this work has received came from physiology and not from anatomy, although this is where I earned my bread for the rest of my career in Switzerland.

	Footnotes

Correspondence and requests for reprints should be addressed to Ewald R. Weibel, M.D., Murtenstrasse 41, POB 620, CH-3000 Bern 9, Switzerland. E-mail: weibel{at}mem.unibe.ch