Quantification of intracavitary brachytherapy parameters and correlation with outcome in patients with carcinoma of...

To quantify the M. D. Anderson criteria for acceptable implant geometry; to relate our system of intracavitary radiotherapy (ICRT) prescription to Manchester and ICRU reference doses; and to correlate these parameters with outcome measures.

The relationships between intracavitary applicators and normal structures were measured directly from localization films of 808 applications performed in 396 patients who completed definitive treatment for cervical cancer between 1990 and 1994. The distances between applicators and tissue landmarks and the doses to Manchester and normal tissue reference points were correlated with outcome.

The median distance from the tandem to the sacrum was 4.0 cm, or one-third the distance from the pubis to the sacrum. The mean distance between the vaginal ovoids and cervical marker seeds was 7 mm, and the median distance between the tandem and the posterior edge of the ovoids was 50% of the ovoid length. In 92% of insertions, vaginal packing was posterior to or within 5 mm of a line that passed through the posterior edge of the ovoids, parallel to the tandem. The median doses to Point A and rectal, bladder, and vaginal surface reference points were 87 Gy, 68 Gy, 70 Gy, and 125 Gy, respectively. Although these reference doses were not routinely used to prescribe treatment, consistent applicator geometry and source selection resulted in a relatively narrow range of delivered doses. The average ratios between the doses at bladder or rectal reference points and Point A were somewhat greater when smaller vaginal applicators were used. Patients received a median of 5600 mgRaEq-h from ICRT. The total mgRaEq-h were correlated with but were not proportional to the dose at Point A. There were no significant correlations between the doses to standard reference points and the rates of central recurrence or major complications.

When ICRT implants are carefully placed, relatively high paracentral doses can be delivered that yield a high rate of central disease control with an acceptable rate of complications. The narrow range of doses delivered to standard reference points and their inconsistent correlation with the maximum doses delivered to normal tissues probably contributed to a lack of correlation between reference doses and outcome.