Validation of the Singapore nomogram for outcome prediction in breast phyllodes tumours: an Australian cohort Tze Wei Chng, 1 Jonathan Y H Lee, 2 C Soon Lee, 3 HuiHua Li, 4 Min-Han Tan, 5 Puay Hoon Tan 6 1 Department of Anatomical Pathology, Singapore General Hospital, Singapore 2 School of Health and Science, Western Sydney University, Western Sydney, Australia 3 Discipline of Pathology, School of Medicine, Western Sydney University; Cancer Pathology, Bosch Institute, University of Sydney; Ingham Institute for Applied Medical Research, and South Western Sydney Clinical School, UNSW, Liverpool, New South Wales, Australia 4 Division of Medicine, Singapore General Hospital, Singapore 5 Division of Biodevices and Diagnostics, Institute of Bioengineering and Nanotechnology, Singapore 6 Division of Pathology, Singapore General Hospital, Singapore Correspondence to Dr Puay Hoon Tan, Division of Pathology, Singapore General Hospital, Academia, Level 7 Diagnostics Tower, 20 College Road, Singapore 169856, Singapore; tan.puay.hoon@sgh.com.sg Received 16 June 2016 Revised 7 July 2016 Accepted 8 July 2016 Published Online First 27 July 2016 To cite: Chng TW, Lee JYH, Lee CS, et al. J Clin Pathol 2016;69:1124–1126. ABSTRACT Aim To validate the utility of the Singapore nomogram for outcome prediction in breast phyllodes tumours. Methods Histological parameters, surgical margin status and clinical follow-up data of 34 women diagnosed with phyllodes tumours were analysed. Biostatistics modelling was performed, and the concordance between predicted and observed survivals was calculated. Results Women with a high nomogram score had an increased risk of developing relapse, which was predicted using the parameters defined by the Singapore nomogram. Conclusions The Singapore nomogram is useful in predicting outcome in breast phyllodes tumours when applied to an Australian cohort of 34 women. INTRODUCTION Phyllodes tumours constitute approximately 0.3%– 1% of primary breast tumours. 1 The grading of this relatively rare lesion, even when based on established histological features, 1 such as tumour border, stromal cellularity, stromal atypia, mitotic activity, stromal overgrowth and malignant heterol- ogous elements, is fraught with uncertainty and issues of interobserver reproducibility. 2 Due to intralesional heterogeneity, histological features within a single tumour may span one or more diag- nostic categories notwithstanding grading based on worst areas, and the relative importance of the graded parameters in predicting outcome is uncer- tain ( figure 1). The Singapore nomogram for predicting clinical behaviour of breast phyllodes tumours (http:// mobile.sgh.com.sg/ptrra/), first introduced in 2011 by Tan et al, 3 attempts to define a predictive model using histological parameters and surgical margin status. Integrating the degree of stromal atypia, stromal mitotic count, presence or absence of stromal overgrowth and surgical margin status (AMOS criteria), the nomogram estimates the recurrence-free survival (RFS) of individuals with phyllodes tumours, while circumventing discrete categorisation of the lesion into benign, borderline or malignant subtypes. The utility of the Singapore nomogram has been validated in a cohort of Japanese women, where it was proven to be effective in predicting patient outcome. 4 We now examine whether the nomo- gram may be equally useful in an Australian population. METHODS Surgical resection specimens of breast phyllodes tumours that were retrieved from the archives of the Department of Anatomical Pathology, Liverpool Hospital, Sydney, NSW, Australia, between 1997 and 2015, were included in the analysis. The study was approved by South Western Sydney Local Health District Ethics and Research Governance Office (HREC reference: HREC/12/LPOOL/102). Of these, cases with incomplete follow-up informa- tion and cases where relevant histological para- meters could not be confirmed were excluded. A centralised pathology review was performed to obtain specific pathological information which was not available from the original histology reports, though the original grading where available was used for analysis. A limited number of slides were retrievable for a number of older cases. Kaplan-Meier survival curves were used to esti- mate RFS, which was defined as the time from the date of surgery to the date of first relapse or death from phyllodes tumour, or to the last follow-up date for censored cases ( figure 2). Univariable Cox regression analyses were performed to evaluate the effects of individual predictors in the Singapore nomogram as well as final nomogram score on RFS. The small number of events (disease recur- rences) precluded multivariable analyses. Harrell’s c-index was calculated to evaluate the probability of concordance between predicted and observed survival by means of the Singapore nomogram, with c=0.5 representing random predictions and c=1 representing a perfectly discriminating model. RESULTS Thirty-four women were included in the final ana- lysis. Their ages ranged from 25 to 96 years, with a median age of 54 years. The median duration of follow-up was 1.7 years (range: 0–9.2 years of follow-up). Local recurrences were present in six women. The median time to recurrence for this set of patients was 7.4 years. Phyllodes tumour was the cause of death in three women who had recurrent lesions, each of whom suffered one episode of recurrent disease after the initial diagnosis. Histologically, 38.2% (13 out of 34) of the phyl- lodes tumours were graded as benign, 41.2% (14 out of 34) were graded as borderline and 5.9% (2 out of 34) were graded as malignant (table 1). Five (14.7%) of thirty-four cases were originally either not graded or given alternative forms of grading such as ‘low- grade phyllodes tumour’ (table 1). Three (8.8%) out of thirty-four phyllodes tumours showed mild 1124 Chng TW, et al. J Clin Pathol 2016;69:1124–1126. doi:10.1136/jclinpath-2016-203951 Short report