Predicting late effects of pelvic radiotherapy: is there a better approach?

ABSTRACT

PURPOSE: Significant chronic symptoms following pelvic radiotherapy occur more frequently than commonly realized. Predictive factors for the development of late symptoms are poorly defined. Moderate sustained acute (cumulative) toxicity might predict severe late effects better than peak reaction. METHODS AND MATERIALS To determine prospectively whether peak or cumulative gastrointestinal (GI) acute symptoms better predict late symptoms in patients receiving pelvic radiotherapy. Symptom scores were measured weekly from the start of radiotherapy, and at 1 year using the Modified Inflammatory Bowel Disease Questionnaire-Bowel subset. The possible prognostic impact of patient-related factors was explored. RESULTS: Three hundred and eight patients were recruited. 100 were excluded due to lack of follow-up data at one year resulting from death, too ill, stoma, relapsed, non-response or withdrawal. A further 15 were excluded for incomplete data, leaving 193 patients with evaluable data. Of these, 28 had GI, 101 urological, and 64 gynecological cancers. Patients' median age was 65 years (range, 23-82), and they were treated with median 60 Gy dose for a median of 6 weeks. Univariate analysis revealed a significant association between cumulative acute symptom scores and scores at 1 year (p < 0.001), which was dose-independent (p < 0.001). Acute peak and 1-year scores were not associated (p = 0.431). The correlation coefficient between cumulative acute symptoms and symptoms at 1 year was 0.367 and for peak acute symptoms was weaker at 0.057. Patients with an abnormal body mass index and current smokers were more likely to experience worse symptoms at 1 year. CONCLUSION: Cumulative acute symptoms are more predictive of late symptoms than peak acute changes in score. This association is independent of the radiotherapy dose delivered and is suggestive of a consequential late effect.