ATTACHMENT 6 RESEARCH REPORT-MEDICAL DIVISION Oak Ridge Associated Universities FOR THE YEAR ENDING DECEMBER 31, 1966 Operating Under Contract With The United States Atomic Energy Commission NASA Study on Radiation Effects Retrospective Study on Deleterious Radiation Effects in Man The great need for more precise knowledge about the radiosensitivity of man gave rise to a retrospective search for human radiation- exposure data that could be analyzed by rigidly imposed statistical methods. This study, supported jointly by the National Aeronautics and Space Administration and the U. S. Atomic Energy Commission is now in its third year. The commencement of the Apollo program for human exploration of the moon's surface has created deadline pressures for completion of the part of the study focused on the effects of single and repeated exposures within an eight-day period. Previously we reported two such studies in which first we used only data obtained from the hospital records of 100 patients treated at the Medical Division of Oak Ridge Associated Universities, and then combined that study with similar data obtained from the University of Cincinnati and the City of Hope Hospital, Duarte, Calif. We are now in process of expanding the study to include a total of about 1000 such hospital observations. We anticipate that this large amount of data will increase the accuracy of our statistical analysis so that greater reliance can be placed upon the correctness of the estimates now underlying shielding requirements of space vehicles. We also expect that correlation of medical and nursing observations of untoward changes in this large group of patients after the various levels of radiation exposure will enable predictions to be made of the kinds and severity of physiologic responses to be expected in normal man in certain radiation fields. We hope that such knowledge will supply on-ground monitors a basis for deciding whether or not an astronaut showing symptoms of distress in space is suffering from radiation exposure rather than from some other stress. In the interest of giving our results to date the widest and earliest dissemination, the dose-response relations developed for symptoms of the human prodromal response were incorporated in Chapter 5.2.2 "Prodromal Response" and Chapter 6.0 "Lethality" in Space Radiation Hazards, a book to be published soon (March 1967) by the National Research Council. National Academy of Sciences under the editorship of W. H. Langham and Douglas Grahn (C. C. Lushbaugh and G. A. Andrews). 1 The responsibility for data storage, retrieval and programming analysis has been transferred from the Mathematics Division of Oak Ridge National Laboratory to personnel of the Data Processing Center of Oak Ridge Associated Universities. The computer facilities of ORNL are still required to handle the large volume of data and perform the required analyses and statistics evaluations (A. S. Gloster and Donna Bibler). The data under study are from 27 participating institutions that have provided 1226 usable patient charts out of 1274 obtained (Evalyn Reppiinger and Dorothy Voian). At present, approximately one-half million dose-correlated reservations are in the study, although not all information from all charts has been completely encoded. This volume of data required development of an efficient data management system to obtain analytic results in reasonable time periods. This development has paralleled a search for more appropriate statistical methods of data analysis for this kind of study. Probit analysis, the method used so far, was neither designed nor intended to be used in retrospect so that deviations from implied assumptions are not ascertainable with it. Even so, our use of probit analysis appears justified by its widespread use in experimental radiobiology so that relative comparisons can be made between sick and well persons and various species of small and large animals. We are also applying methods used in testing of inanimate engineering components for reliability and durability. The Weibuil density function that describes wear-out failure rather than chance failure, predicting the time that any rate of failure will be evident, seems most promising, since most symptoms and signs of biological distress can be considered as evidence for "wear-out" of a homeostatic physiologic system (E. Frome). The following abstract of one of our studies is included because it illustrates one of the many complex problems that must be solved as a basis for this program. During the year, further refinement has been achieved in the estimates of doses required for production of specific prodromal effects. These new estimates will not be reported in detail here because they will be refined still further in the future as larger numbers of patients are analyzed. Dosimetry Used in Equating Various Radiation Exposures for Determining Dose-Response Relationships W. Beck, R. Cloutier and F. V. Comas The factor that can affect most the validity of our retrospective determination of human dose-response relations is the accuracy of each dose estimate. The statistical analysis commonly used in such studies contains the assumption that the independent variable, dose, is known precisely while the biological response may not be accurately measured. 3 The retrospective study was divided into four phases. Phase I was a pilot study of only 100 single exposures of less than 24 hr in patients at the Medical Division.1 Phase II was an extension of this study to test the feasibility of using data from other hospitals, and also included only single radiation exposures.2 Phase III, now being completed, includes single and fractionated exposures (greater than one day and less than eight days). Phase IV is planned to study protracted exposures where exposure time was greater than eight days. Phase I. Phase I was composed of 93 single-exposure cases of therapeutically administered total-body irradiation (TBI) and seven cases of accidental TBI. Eighty-four of these were treated with the Division's medium-exposure-rate total-body irradiator,3 and the other nine were treated with the Division's cobalt-60 temporary irradiator.4 The dose estimates for the 93 patients were based on a dosimetric study by Hayes, et al.4 The dosimetric system consisted of an aqueous ferrous sulfate solution (Fricke dosimeter) within a plexiglass phantom. The absorbed dose to the solution was taken as the patient's dose when he was exposed to the same amount of radiation. Using this technique we determined the whole-body average dose (WBAD), and the midepigastric dose for all 93 patients. The WBAD is an estimate of the average energy absorbed per gram of tissue, where the average is taken for every gram of tissue in the body. The midepigastric dose is an estimate of the average energy absorbed per gram in the upper abdominal compartment. The dose estimates for the seven accident cases were determined by a mock-up of the accident at its site.5 Phase II. Phase II extended the pilot study by the addition of 11 new ORAU cases, 29 from Cincinnati General Hospital, Cincinnati, Ohio, and 23 from the City of Hope Medical Center, Duarte, Calif., bringing the total number of patients in the study to 163. The dose estimates for the new ORAU cases were made with the same technique described in Phase I. However, all the ORAU dose estimates were modified for patient size. The size-correction factors were determined from the experimental measurements of the absorbed doses in three phantoms representative of an adult, an adolescent and a child. The dose estimates for the Cincinnati and the City of Hope cases were based on a comparison of phantom depth-dose measurements made at each facility. 3