Between May 1991 and December 1995, follow-up after primary therapy for 472 consecutive patients with localized breast cancer was randomly assigned to assess the efficacy of routine chest X-ray in detecting intrathoracic relapse as the first metastatic event. One group had regular chest X-rays while the other group had chest radiographs only when clinically needed (spontaneous). Patients were followed up until December 1999 or death. In the routine group, 243 patients had 1429 chest X-rays (mean 5.9 chest X-rays per patient). In the spontaneous group, 229 patients had 411 chest X-rays (mean 1.8 chest X-rays per patient). Both sensitivity and specificity were somewhat higher in the spontaneous arm compared to the routine arm. Patient sensitivity was 30% in the regular arm and 36% in the spontaneous arm. Film sensitivity was 11% in the regular arm and 20% in the spontaneous arm. Patient specificity was 85% in the regular arm and 86% in the spontaneous arm. The predictive values were practically the same in both arms. Patient positive predictive values were 22% in both arms and film positive predictive values were 21% in the regular and 22% in the spontaneous arm. Patient negative predictive values were 90% in the regular and 92% in the spontaneous arm. Film negative predictive values were 93% in the regular and 89% in the spontaneous arm. The differences in accuracy were not statistically significant between the arms. There were no significant differences in disease free survival or overall survival. The 5-year disease free survival was 86% in the regular and 89% in the spontaneous arm and the overall survival was 88 and 85% in the regular versus spontaneous arm, respectively. Routine chest X-ray in the follow-up of breast cancer patient is evidently of little use and is not likely to be of benefit to the patient in terms of disease free survival or overall survival as compared to the spontaneous schedule.