Dose-response relationship for radiation therapy of subclinical disease.

PURPOSE: To determine the dose-response relationship for elective treatment of subclinical metastatic deposits and validate a model for metastatic tumor cell burden. METHODS AND MATERIALS: The incidence of overt metastases in electively irradiated potential sites of spread from carcinomas of the head and neck, breast, cervix, ovary, lung, and testis, and from melanomas and osteosarcomas, was compared with the incidence in patients not receiving elective irradiation. The reduction in incidence of metastases was analyzed as a function of radiation dose.
RESULTS: The dose-response curve for control of subclinical metastases is linear and shallow, and extrapolates to a dose intercept not demonstrably different from zero. A small threshold may reflect growth of residual micrometastases between treatment for the primary and elective irradiation. The shallow linear dose response reflects interpatient heterogeneity in metastatic tumor cell burden, ranging from 1 to M cells, where M is the upper limit of clinical undetectability. While a dose of 50 Gy in 2 Gy fractions is necessary to achieve an overall 90% reduction in the incidence of metastases, the metastatic cell burden in a proportion of patients can be eliminated by low doses. Thus, worthwhile rates of control can still be achieved when "tolerance" dictates lower than optimal doses, evidenced by the linearity and lack of significant threshold in the dose-response curve. This is an important difference from treatment of gross disease. The biological effectiveness of elective treatment is measured directly by the percent reduction in failure rate. Although it depends upon the log cell kill, it relates only to that proportion of patients harboring subclinical disease, and, therefore, is not well described by the increase in the cure rate for the total patient population. The linear dose-response relationship for reduction in failure rate is independent of the "natural" (untreated) incidence of subclinical metastasis, and, therefore, of site, histology, growth rate, stage, or other characteristics of the tumor. Conversely, the clinical effectiveness of elective treatment is measured by increase in tumor control rate and depends upon the "natural" incidence of metastasis: the higher it is, the greater the absolute increase in cure rate from a constant biological effect (log cell kill).
CONCLUSIONS: (a) High control rates for subclinical metastases require doses of about 50 Gy in 2 Gy fractions, but worthwhile benefits can be achieved by lower doses if necessitated by reduced tolerance; (b) elective treatment of subclinical metastases should be instituted close to the time of treatment of the primary; (c) the biological effectiveness of elective radiation (or chemotherapy) should be measured by the percentage decrease in metastasis, not by improvements in the rate of control; and (d) demonstration of success in clinical trials of adjuvant therapy is more likely the higher the incidence of metastases in untreated controls.