RISK-FACTORS FOR LOCAL RECURRENCE AFTER CONSERVATIVE TREATMENT IN STAGE-I BREAST-CANCER - DEFINITION OF A SUBGROUP NOT REQUIRING RADIOTHERAPY

Background: Risk factors for local recurrence after breast-conserving treatment of early breast cancer have not previ ously been evaluated i n settings where mammography has been a major pathway to diagnosis of both primary tumour and recurrences, or in patients treated surgically by a formal sector resection. Patients and methods: Three hundred eig hty-one women with stage I primary breast cancer were randomised after a standardised sector resection to either a course of postoperative r adiotherapy to 54 Gy to the breast (XRT group) or to surgery alone (no n XRT group). At five years, 43 local recurrences, six of them in the XRT group, appeared. Patient characteristics collected from the medica l records, histopathological characteristics determined by re-examinat ion of slides, and mammographic characteristics from the pre-operative mammograms were evaluated as risk factors for recurrence by univariat e and multivariate Cox proportional hazards models. Results are report ed as relative hazards (RH) with 95% confidence intervals (95% CI). Re sults: In the univariate analysis comedo cancer, RH 3.5 (95% CI 1.8-6. 7), lobular cancers RH 2.8 (95% CI 1.1-7.1), mammographic appearance a s circular/oval shaped density RH 2.3 (95% CI 1.1-4.5), and mammograph ic appearance as a stellate lesion with microcalcifications inside the lesion, RH 3.8 (95% CI 1.1-13.0) were identified as risk factors for local recurrence. Age, with a RH of 0.97 (95% CI 0.94-0.99) for each i ncreasing year was inversely associated with risk. A multivariate anal ysis, which also took postoperative radiotherapy into account, only sh owed comedo cancers with a RH 2.6 (95% CI 1.3-5.0) and mammographic ap pearance of a stellate lesion with microcalcification inside the lesio n RH 4.5 (95% CI 1.1-17.6) to be statistically significant. The estima tes for age RH 0.98 (95% CI 0.95-1.0) and lobular cancers RH 2.5 (95% CI 0.98-6.6) were marginally changed, with widened CIs. Patients > 60 years of age, without comedo or lobular carcinomas were found to be at low risk (5.9% at five years in Kaplan-Meyer estimate) of local recur rence, even without postoperative radiotherapy. Conclusion. Low age, c omedo and lobular cancers and mammographic appearance of the tumour as a stellate lesion with microcalcifications inside the lesion indicate an increased risk for local recurrence after sector resection in stag e I tumours at five years. Patients >60 years of age without comedo or lobular cancers are at low risk for local recurrence al five years ev en without postoperative radiotherapy.