Tumour size in cancer of the cervix

In a randomized trial comprising 204 patients with operable cervical carcinomas stages I and II, two low-dose rates in gynaecological brachytherapy were compared. Treatment consisted of Cs-137 uterovaginal application followed by surgery (either immediate or delayed). The results for the two dose rates have been published previously. The present paper concerns the correlation between outcome and tumour size. Tumour size was carefully estimated in two ways: by clinical examination under general anaesthesia and by measurements on the customized vaginal mould used for the brachytherapy. Ninety-one patients (45%) were classified as stage I, and 113 were classified as stage II proximal. The mean tumour size was 39 mm (range 15-64 mm). Cox's multivariate analysis indicated that the factors with a poor prognostic value were for survival: node involvement (N +) (p < 0.001), large tumour size (T +) (p < 0.001) and involvement of the endocervix (E +) (p <0.01); for event-free survival: N + (p <0.001), T + (p < 0.001); for local control; N + (p = 0.0001); for metastasis and regional relapse: N + (p < 0.001) and T + (p < 0.001). Stage was not a prognostic factor over the present range in either univariate or multivariate analysis. In this series tumour size is a powerful independent prognostic factor. It is therefore suggested that for the classification of cervical cancer and the indications for surgical staging and adjuvant treatment, tumour size should be taken into account.