Purpose: The aim of this study was to evaluate the clinical features and prognostic implications of patients with recurrent cervical cancer.
Methods: By reviewing the medical records we evaluated all patients with cervical cancer at stages IA to IVA who started treatment at a specialized hospital in the Southeast region of Brazil from 2007 to 2009. Recurrence episodes were categorized according to location of disease and information was collected regarding the type of treatment and survival of these patients. The sample was characterized by descriptive statistics and association analyses were performed using Fisher's exact test.
Results: Fifty cases of recurrence were identified among 469 selected records, with 31 patients being symptomatic at diagnosis of recurrence (62%); and 19 being asymptomatic (38%). Among women with symptoms, eight requested anticipation of the previously scheduled appointment because of the presence of clinical complaints. Patients with symptoms at the diagnosis of recurrence had lower rates of overall two-year survival (39.4 versus 67.6%) (p=0.081). None of the patients with recurrence at distance received curative intent treatment, but all received surgical treatment or radiotherapy aiming at full remission of the disease. Women who requested anticipation of the appointment because of the presence of symptoms had a significant reduction of overall two-year survival after recurrence (0 versus 60.4%; p<0.001) compared to those who attended the consultation on the scheduled date, and none of them received curative intent treatment. As expected, the patients who underwent palliative treatment with the main objective of improving quality of life and increasing survival but with no perspective of cure had a significant reduction in overall survival compared to those who were treated with curative intent (76.7 versus 35.4%; p<0.001).
Conclusion: The benefit of detecting asymptomatic recurrence of cervical cancer has the potential to improve the prognosis of patients with local and regional recurrence, but studies on larger series are necessary to confirm this possibility.