Pathologic findings from the national surgical adjuvant breast project (protocol 6). Relation of local breast recurrence to multicentricity

Abstract
One hundred ten local breast recurrences were observed in 1108 pathologically evaluable patients enrolled in NSABP protocol 6 who were treated by lumpectomy and followed for 5 to 95 months (average, 39 months). Eighty‐six percent and 95% of all local breast recurrences were noted within 4 and 5 years, respectively, following lumpectomy. Life table analysis revealed their incidence to be 24% for those not and 6% for those receiving lumpectomy and breast irradiation. One hundred four (95%) of the breast recurrences involved the mammary parenchyma and the remaining 6 (5%) involved the skin and/or nipple only. Eleven (10%) of the former were noninvasive. The most common (86%) presentation of breast recurrence appeared to be a localized mass within or close to the quadrant of the index cancer. In 14% the recurrence not only involved the same quadrant, but was more diffuse within the breast extending to remote areas as well. This type was characterized pathologically by marked intralymphatic extension as well as involvement of the overlying skin and/or nipple after the fashion of so‐called inflammatory or occult inflammatory breast cancer. The recurrences noted in the skin and/or nipple only were also pathologically characterized by intralymphatic involvement at these sites in the majority of instances. These two forms of breast recurrences appear to reflect the localized growth of highly aggressive invasive breast cancers. The concordance of histologic types and grades of the index and recurrent cancers implies that such events represent growth of overlooked tumor, a deficiency attendant with lumpectomy due to the extreme multifocal nature (not multicentricity) of some breast cancers and/or inadequacies in evaluating the lines of resection of lumpectomy specimens. Sources of error in regard to this latter are identified and guidelines for the examination of such specimens, as well as the assessment of margins, are presented. The observation that local breast recurrences noted following lumpectomy occurred within or close to the same quadrant as the index cancer, despite the presence of multicentric noninvasive cancers in 10% of the patients treated by total mastectomy, minimizes the biological and clinical significance of multicentric foci of cancer present in some breast cancers. Cancer measuring ≥2.0 cm, having high histologic and nuclear grades, or intralymphatic extension, were found to have a statistically significant association with local breast recurrence in all patients following lumpectomy. A converse relationship was noted with tubular and scar cancers of types 1 and 4. Only intralymphatic extension was found to exhibit a strong trend as a discriminant for local breast recurrences in the lumpectomy and irradiation group. Considerations indicate, however, that this finding does not represent a contraindication to lumpectomy and irradiation.