Physiological changes in pulmonary function (PF) as a result of radiation therapy (RT) or radiation therapy plus chemotherapy (RT + CT) for unresectable lung cancer were evaluated in an ongoing prospective study and an attempt was also made to define a guideline which can be used to minimize adverse effect of RT on pulmonary function before RT is given. The study design consisted of: (a) standard overall pulmonary function test (PFT); (b) regional PFT, i.e. a quantitative analysis of regional distribution of ventilation, perfusion and volume using 13N and a positron camera before RT; and (c) follow-up studies of standard PFT every 6 months for 3 years after RT or RT + CT. Predicted post-RT PF prior to RT was calculated by a formula: predicted FEV1 after RT = FEV1 before RT x (1 - an average of the percent of ventilation and perfusion contributed by lung tissue within the RT treatment volume). A total of 267 patients with unresectable, but still potentially curable lung cancer by RT were entered into this study, and 135 patients who were free of recurrence underwent repeat studies. Loss of PF as a result of RT is closely related to the degree of PF reserve prior to RT. Patients with FEV1 > 50% of the predicted showed a statistically significant decrease in FEV1, FVC, MBC, peak expiratory flow rate and DLCO, i.e. a 22% loss of the initial value. Airway resistance was increased by 31%. Two-thirds of this group of patients showed a decrease in PF as predicted by the above formula. For patients with limited PF reserve defined by FEV1 < 50% of the predicted, the pattern of PF loss after RT was quite different. An improvement in PF although it was < or = 10%, contrary to the prediction, was noted in 50% of patients, and another 37% of patients showed a small decrease in PF (< or = 10% of the initial value). Only 13% of patients showed a loss of pulmonary function as predicted by regional PF data. Patients with a significant shift (> 10%) of ventilation and/or perfusion to the uninvolved side of the lung by centrally located primary tumor or involved lymph nodes showed an increase in PF in 60% of patients after RT, and another 20% of patients showed a minor decrease in PF (< 10% of the initial value). Only 20% of these patients showed a decrease in pulmonary function as predicted by regional PF data. Guidelines for minimizing adverse effect of RT on PF, which are based on the initial PF reserve and regional PF data, are presented.