[en] Purpose: 1) To define the incidence of isolated lung events in a cohort of women treated with conservative surgery (CS) and radiation therapy (RT) for early stage breast cancer. 2) Among such patients, to define the relative distribution of primary lung cancer, metastatic breast cancer, and indeterminate lesions; and to identify any predictors for a diagnosis of lung vs metastatic breast cancer. 3) To examine the cohort with respect to whether a higher than expected incidence of lung cancer is seen following breast irradiation. Materials and Methods: Between 1968 and 1986, 1865 patients with clinical stage I-II breast cancer were treated with CS and RT; the median follow-up for surviving patients is 129 months. The study population was limited to patients who developed a subsequent isolated lung event as the first site of distant disease. Isolated lung event was defined as disease limited to the thoracic cavity, without evidence of either uncontrolled local breast disease or metastatic disease elsewhere. Diagnosis of the lung event as a primary lung cancer, a metastatic breast lesion, or an indeterminate lesion was documented from the viewpoint of 1) the pathologic analysis and 2) the clinical impression at the time of the lung event. Results: Sixty six of the 1865 patients (3.5%) developed an isolated lung event. The relative distribution of the pathologic and clinical diagnoses is shown below: The 66 lung events were characterized either as a solitary pulmonary nodule (27), multiple nodules (23), pleural effusion alone (10), unknown (2), or miscellaneous other findings (4). Among the 47 patients for whom pathology was available, the diagnosis remained indeterminate for 24 (51%). For patients with a definitive pathologic diagnosis, 69% ((9(13))) of smokers had a new lung cancer compared to 20% ((2(10))) of non-smokers (p=0.036), and 67% ((10(15))) of patients with a solitary pulmonary nodule had lung cancer compared to 14% ((1(7))) for other lung presentations (p=0.027). Factors related to the initial breast cancer which were not predictive for diagnosis of the lung event included nodal status, prior local or regional recurrence, 2 vs 3 field RT technique and machine energy. Time to lung event appeared fairly constant over two year intervals up to 10 years and beyond for patients with both metastatic breast lesions and primary lung cancers. In particular, there did not appear to be a relative increase in the risk of lung cancer after 10 years. Eleven of the initial cohort of 1865 patients (0.6%) had an ultimate pathologic diagnosis of lung cancer. With respect to the irradiated breast, these lung cancers were located in the ipsilateral lung for 5, the contralateral lung for 5 and the location was unknown for one. Conclusions: 1) Isolated lung events following early stage breast cancer are uncommon. 2) Smoking history and presentation as a solitary pulmonary nodule were more likely to be associated with a diagnosis of lung cancer than a metastatic breast lesion. However, even with a biopsy, it is often impossible to distinguish with certainty between the two diagnoses. In these situations, management must be based on clinical judgment. 3) In this cohort, the development of lung cancer was a rare event. There was no demonstrable evidence of radiation-induced lung cancer through this time of follow-up based on laterality or radiation technique