Quantification of late complications after radiation therapy.

BACKGROUND: An increasing number of patients survive cancer after having received radiation therapy. Therefore, the occurrence of late normal tissue complications among long-term survivors is of particular concern.
METHODS: Sixty-three patients treated by radical surgery and irradiation for rectal carcinoma were subjected to an unconventional sandwich therapy. Preoperative irradiation was given in four fractions of 5 Gy each applied within 2 or 3 days; postoperative irradiation consisted mostly of 15 x 2 Gy (range, 20-40 Gy). A considerable proportion of these patients developed severe late complications (Radiother Oncol 53 (1999) 177). The data allowed a detailed analysis of complication kinetics, leading to a new model which was tested using data from the literature.
RESULTS: Data on late complications were obtained for eight different organs with a follow-up of up to 10 years. For the various organs, the percentage of patients being free from late complications, plotted as a function of time after start of radiation therapy, was adequately described by exponential regression. From the fit, the parameter p(a) was obtained, which is the percentage of patients at risk in a given year of developing a complication in a given organ during that year. The rate p(a) remained about constant with time. Following sandwich therapy, the annual incidence of complications in the bladder, ileum, lymphatic and soft tissue, and ureters was about the same (p(a)=10-14%/year), whereas complications in bone or dermis occurred at lower rates (4.7 or 7.5%/year, respectively). DISCUSSION: Numerous data sets collected from published reports were analyzed in the same way. Many of the data sets studied were from patients in a series where there was a high incidence of late effects. Three types of kinetics for the occurrence of late effects after radiotherapy were identified: Type 1, purely exponential kinetics; Type 2, exponential kinetics, the slope of which decreased exponentially with time; Type 3, curves composed of two components, a fast initial decline followed by an exponential decrease. For each kind of kinetics, provided that the dose distribution is not too heterogeneous, the incidence of late effects appears to occur at exponential or approximately exponential kinetics, even many years after treatment. This implies that a random process might be involved in the occurrence of late radiation sequelae.
CONCLUSIONS: There might be a lifelong risk of developing late complications, of which patients and clinicians should be aware. It appears worthwhile to try to identify, in follow-up examinations of patients after radiation therapy, what kind of processes might be involved in triggering subclinical residual injury to develop into a clinically manifest late effect.