• Vol. 34 No. 11, 671–677
  • 15 December 2005

2005 Galloway Memorial Lecture: Breast Phyllodes Tumours – Morphology and Beyond

142



142 Views
23 Downloads

Download PDF

ABSTRACT

Objectives: The aims of this study were to evaluate the predictive role of morphology in a series of breast phyllodes tumours (PTs) in Asian women, and to determine the utility of immunohistochemical expression of p53 and CD117 in the stromal component. Materials and Methods: Breast PTs, diagnosed between January 1992 and December 2002 at the Department of Pathology, Singapore General Hospital, were classified into benign, borderline and malignant categories. Surgical margins were evaluated as complete or diffusely involved. Patient follow-up was obtained from casenotes and the Singapore Cancer Registry. Tissue microarrays (TMAs) were constructed using the 2-mm punch on the Beecher arrayer. p53 and CD117 immunohistochemistry was applied on 4-µm sections cut from TMA blocks. Immunostaining intensity was graded as 0, 1+, 2+ or 3+, for nil, weak, moderate and strong reactivity. The proportion of stromal cells decorated was assessed. Statistical analysis utilised the software SPSS for windows 11.5. Survival curves were plotted using the Kaplan-Meier method, while multivariate analysis was accomplished using the stepwise Cox proportional hazards model. A P value of <0.05 was considered a significant result. For verification of protein expression results, a pure stromal population derived from laser capture microdissection was subjected to real-time polymerase chain reaction to determine p53 and CD117 mRNA upregulation. Results: Three hundred thirty-five women diagnosed with PT were aged 16 to 69 years (median, 42 years). Tumour size ranged from 0.9 to 25 cm (median, 4 cm). Histologic classification revealed 250 (74.6%) benign, 54 (16.1%) borderline and 31 (9.3%) malignant PTs. Surgical margins were focally involved in 186 (55.5%) cases, diffusely affected in 9 (2.9%) cases and complete in 139 (41.5%) cases. Stromal cells positively stained ranged from 1% to 80% (mean, 15%; median, 5%) for p53, and 1% to 25% for CD117 (mean, 8%; median, 3%). p53 and CD117 staining was associated with PT grade (P = 0.004, P <0.001). Forty-three (12.8%) women suffered 57 recurrences (mean and median follow-up of 30.3 and 20.4 months respectively). Nine (2.7%) died during follow-up, 7 from malignant disease (mean and median survival duration 37.6 and 23.6 months respectively). Complete margins (P = 0.033) resulted in reduced recurrence risk by 51.7%. Though tumour grade was associated with an increased hazard of 1.63, it was not significant (P = 0.28). Immunoexpression of p53 was not associated with recurrence (P = 0.447), while CD117 was (P = 0.001). Upregulation of p53 and CD117 mRNA was found in cases in which there was protein overexpression. Conclusions: Involved surgical margins remain a key prognostic parameter in breast PT. Death from disease occurred in women with malignant PTs at presentation, underscoring the need to manage this group more aggressively. p53 staining can be used to corroborate malignancy in PT. CD117 was predictive of recurrence, and if further validated, its expression can be explored for therapeutic purposes. Preliminary molecular studies verify mRNA upregulation in p53 and CD117 overexpressed cases.


The clinical behaviour of phyllodes tumours (PTs) of the breast is difficult to predict.1-5 While histological features have been traditionally used as predictors of clinical outcome, specific parameters that define recurrent likelihood are still not universally agreed upon.2 Though grade,6,7 stromal overgrowth,8,9 tumour necrosis and heterologous stromal elements,10 or a combination of histologic features,11 have been variously found to be prognostic, a substantial number of reports have concluded that adequacy of surgical margins is of paramount importance,5,12-15 and that histologic factors have an inconsistent influence on biologic behaviour.

This article is available only as a PDF. Please click on “Download PDF” to view the full article.