About one-half of cervical cancer patients whose tumors recur after radical surgery have negative lymph nodes and clear resection margins. We evaluated 95 patients with squamous cell tumors who underwent radical hysterectomy and pelvic lymphadenectomy between January 1975 and December 1985 and who were thought to be at low risk for recurrence to see whether other clinical or histopathologic factors were predictive of tumor recurrence. Detailed retrospective record review and complete pathology review were accomplished for each case. The 5-year actuarial survival rate was 89%. Nine patients developed recurrent disease (9.5%), of whom eight died. Several clinical features were evaluated as possible prognostic factors: patient age (P = .26), patient race (P = .60), cervical diameter (P = .24), extent of gross cervical involvement (P = .36), and presence of contact bleeding (P = .82). Histopathologic features were examined: depth of invasion (P = .31), number of mitoses (P = .42), character of the tumor-stromal border (P = .15), histologic differentiation (P = .02), lymph-vascular space invasion (P = .56), and width of tumor (P = .23). Depth of invasion did correlate with increasing tumor width (P less than .001). Once node- and margin-positive patients are excluded, differentiation may be the only feature useful in identifying patients at risk for recurrence. Because almost one-half of our patients had poorly differentiated tumors, sole use of this feature as a criterion for adjuvant therapy would have resulted in overtreatment of low-risk patients.