Predictors for pneumonitis during locoregional radiotherapy in high‐risk patients with breast carcinoma treated with high‐dose chemotherapy and stem‐cell rescue

BACKGROUND To study the predictive value of serial pulmonary function testing (PFT) for toxicity in patients who have received high‐dose chemotherapy (HDCT) and stem‐cell rescue for breast carcinoma. These patients are at risk of developing therapy‐related pneumonitis (TRP) during or after radiotherapy (RT). METHODS Sixty‐eight patients who received induction chemotherapy (CT) and consolidation HDCT (cyclophosphamide, cisplatin, carmustine) underwent serial PFTs before induction CT, after HDCT, and before locoregional RT. The rate of TRP, i.e., pulmonary complications of Grade 2 or higher (World Health Organization classification), was studied during and 2 months after RT. We analyzed the time‐course of changes in the diffusing capacity of carbon monoxide (DLCO) and forced expiratory volume at one second (FEV1) and studied the differences between patients who developed TRP and those who did not. RESULTS The incidence of TRP was 46%. There were marked reductions in DLCO and FEV1 at the time of RT compared with baseline (Wilcoxon signed rank test: P < 0.001). However, pre‐RT PFT values did not predict subsequent development of TRP. Instead, the ratio of pre‐RT DLCO to the minimum post‐ HDCT DLCO, i.e., trend of improvement, predicted the development of TRP in patients (logistic regression analysis: P = 0.048). At a cutoff level of 1, the positive and negative predictive values for this ratio were 61% and 87%, respectively. There was an association between this ratio and a longer interval between HDCT and RT (Spearman rank correlation: P = 0.002). CONCLUSIONS The results suggest that the directional trend of DLCO after HDCT, i.e., no recovery from nadir values, is a predictor for TRP. TRP patients have a shorter median interval between HDCT and RT than asymptomatic patients. To minimize the occurrence of TRP, one should consider either delaying RT beyond 2 months following carmustine‐based HDCT to allow the PFTs to partly recover, or confirm apositive directional trend for improvement of DLCO at the start of RT compared to the post‐HDCT nadir value. Cancer 2002;94:2821–9. © 2002 American Cancer Society. DOI 10.1002/cncr.10573 Pulmonary toxicity during radiotherapy is seen in 50% of high‐risk patients with breast carcinoma who were treated with high‐dose carmustine‐ based chemotherapy (HDCT) and stem‐cell rescue. Serial pulmonary function testing (PFT) is used to monitor therapy‐induced pulmonary toxicity. A patient's nadir and subsequent PFT values are record