Clinicopathologic characteristics and molecular subtypes of invasive papillary carcinoma of the breast: a large case study

Abstract

Purpose

Although patients with invasive papillary carcinoma (IPC) often have favorable prognoses, it remains unclear whether this special type of breast cancer represents a distinct morphological entity with its own biological features and clinical behavior distinct from those of invasive ductal carcinoma (IDC) and whether its four molecular subtypes are associated with different prognoses.

Methods

The study is a retrospective analysis of a large patient cohort from a single institution. 284 IPC samples were collected from January 2000 to May 2011. 300 IDC cases were selected randomly from 13,428 cases of IDC during the same periods. We assessed the clinicopathologic characteristics, molecular features, and prognostic value of IPC (n = 284) and compared them to those of IDC (n = 300). Clinicopathologic features and survival status of the four subtypes of IPC were also evaluated.

Results

IPC differed from IDC with respect to age upon diagnosis, tumor grade, lymph node status, and menopausal status (P < 0.05). IPC was associated with a better 5-year overall survival rate (OS) (92.77 vs. 87.95 %) and disease-free survival rate (DFS) (87.95 vs. 80.72 %) than IDC. Tumors of the luminal A subtype had a better 5-year OS (97.78 %) and DFS (95.56 %) than other subtypes.

Conclusions

The biologic behavior of IPC is more favorable to patient outcome than that of IDC. The chance of pure IPC causing death without an intervening event of a different histologic type is exceptionally low. Luminal A subtypes have better outcomes when compared to the other subtypes.

Introduction

Breast cancer is a heterogeneous disease with varying morphology, behavior, and response to therapy. In 2003, the World Health Organization (WHO) classification of tumors of the breast and female genital organs defined invasive papillary carcinoma (IPC) of the breast as a type as invasive mammary carcinoma. Papillary carcinomas of breast are either noninvasive (intraductal) or invasive. When papillary carcinomas become invasive, they generally assume the pattern of an infiltrating duct carcinoma. The overall incidence of IPC is rare, accounting for less than 1.5 % of all newly diagnosed cases of breast cancer (Fisher et al. 1980; Gentile and Becette 1996). It is most commonly seen in females and is exceedingly rare but not unknown in males. IPC tends to affect older women than more common breast cancers and papillomas. Papillary tumors in women aged 70 years or more have been found to be the most likely to represent IPC (Koerner 2010). Patients with IPC may present with a palpable mass, bloody nipple discharge, or a radiographic abnormality (Louwman et al. 2007). Histologic features of this tumor include proliferations of cells arranged around fibrovascular cores. This conclusion is based on the fact that they lack a myoepithelial cell layer within their papillary structures. The differences between IPC and noninvasive cancers are very important because each entity is associated with a different prognosis.

Because of its rarity, there is a paucity of literature regarding this type of tumor. Most previous studies are from small case series. In this report, we evaluated 284 cases of IPC of the breast and evaluated the clinicopathologic findings, molecular immunophenotyping, and overall disease-related outcomes of patients treated at a single institution.

Materials and methods

Study population

The cohort included patients diagnosed with breast IPC. This cohort was derived from the archival paraffin-embedded breast cancer samples collected from the Tianjin Medical University Cancer Institute and Hospital between January 2000 and May 2011. Our exclusion criteria were as follows: postoperative deaths, non-tumor-mediated mortality, and women over the age of 90. Invasive ductal carcinoma (IDC) of the breast cases during the corresponding time period was selected randomly for use as matched controls. This was a well-characterized series of patients who had undergone long-term follow-up organized by a single institution. In IPC cases, all the patients underwent preoperative mammography and ultrasound of the breast. All patients involved in this study had their diagnoses of IPC confirmed histologic re-examination of the paraffin-embedded slides by two senior pathologists. All patients received the standard surgical treatment of either mastectomy or wide local excision. Adjuvant systemic treatment was performed on the basis of prognostic and predictive factors, such as patient age, tumor grade, tumor size, lymph node (LN) stage, and estrogen receptor (ER) status.

Clinicopathological evaluation

We retrospectively evaluated conventional clinicopathological factors between IPC and IDC. The variables recorded included patients’ demographics (age, menopausal status, and family history), tumor characteristics (grade, size, LN status, and subtype), type of adjuvant systemic treatment (chemotherapy, radiotherapy, and endocrine therapy), recurrence, and survival information. The tumors were classed by histological grade as low, intermediate, and high. The nuclear grade (low to high) of the monotonous population of cells was based on nuclear hyperchromasia, pleomorphism, and nuclear/cytoplasmic ratio. Patients that had been treated before May 2006 were subjected to detailed review to assess survival data, including overall survival (OS) time, disease-free survival (DFS) time, local and regional contralateral recurrence, and development of distant metastases. The study protocol was approved by the Human Ethical Committee of Tianjin Medical University Cancer Institute and Hospital. Informed consent was obtained from all patients before their surgery and before examination of the specimens.

Molecular subtype criteria

Immunohistochemical staining was performed for the biomarkers of ER, progesterone receptor (PR), and human epidermal growth factor receptor 2 (Her2) in each case. ER and PR were considered positive if nuclear staining was present in more than 1 % of the tumor cells. Scoring of Her2 immunoreactivity followed the American Society of Clinical Oncology and College of American Pathologists guidelines (Wolff et al. 2007). IPC can be classified into four clinically relevant subtypes on the basis of gene expression patterns: luminal A (ER+ and/or PR+ and Her2−), luminal B (ER+ and/or PR+ and Her2+), Her2-overexpressing (ER−, PR−, and Her2+), and triple-negative (ER−/PR−, Her2−).

Statistical analysis

All statistical analyses were carried out using the SPSS 15.0 software program (SPSS Inc., Chicago, IL, USA). The χ ² test was used to assess relationships between variables. The Kaplan–Meier product limit method was used to estimate OS and DFS. OS is here defined as each breast cancer patient’s chance of remaining alive. DFS is used to denote each patient’s chance of staying free of disease after a particular treatment. The log-rank test was used to compare differences in survival among histologic subtypes. Cox proportional hazard models were used to calculate relative risk accounting for covariates. P values are here derived from two-tailed tests, and P < 0.05 was considered statistically significant.

Results

Patient cohort

This study included 284 patients, accounting for about 0.68 % of all cases of breast cancer diagnosed in this hospital during this time (41,765). In addition, 300 cases (2.23 %) were selected randomly from 13,428 cases of IDC. The median age of each IPC patient was 65 years (range 25–90). The clinicopathologic features are presented in Table 1. Significant differences between IPC and IDC were obvious upon diagnosis with respect to age, tumor grade, LN status, and menopausal status (P < 0.05). As shown in Table 1, low- and intermediate-grade tumors were dominant in IPC, accounting for about 91.19 % of the total. The rate of axillary node metastases was 17.25 %, which was also lower than IDC. There were no clear differences in tumor size, family history, chemotherapy, radiotherapy, endocrine therapy, or subtype (P > 0.05).

Table 1.

Patients’ demographics and clinicopathologic characteristics of IPC compared with IDC

Characteristic	IPC (n = 284)		IDC (n = 300)		χ2	P
	No.	%	No.	%
Age at diagnosis (years)
<50	59	20.77	183	61.00	97.28	<0.05
≥50	225	79.23	117	39.00
Tumor size (cm)
<2	81	28.52	72	24.00	1.87	>0.05
2–5	179	63.03	205	68.33
>5	24	8.45	23	7.67
Tumor grade
Low	101	35.56	61	20.33	18.05	<0.05
Intermediate	158	55.63	197	65.67
High	25	8.81	42	14.00
Lymph node status
Negative	235	82.75	153	51.00	65.94	<0.05
Positive	49	17.25	147	49.00
Menopausal status
Postmenopausal	211	74.30	105	35.00	90.72	<0.05
Premenopausal	73	25.70	195	65.00
Family history
Yes	41	14.44	52	17.33	0.91	>0.05
No	243	85.56	248	82.67
Chemotherapy
Yes	252	88.73	273	91.00	0.83	>0.05
No	32	11.27	27	9.00
Radiotherapy
Yes	86	30.28	109	36.33	2.40	>0.05
No	198	69.72	191	63.67
Endocrine therapy
Yes	179	63.03	187	62.33	0.03	>0.05
No	105	36.97	113	37.67
Subtype
Luminal A	145	51.06	162	54.00	4.37	>0.05
Luminal B	49	17.25	34	11.34
Her-2	37	13.03	40	13.33
Triple-negative	53	18.66	64	21.33

IPC invasive papillary carcinoma, IDC invasive ductal carcinoma

Diagnosis

Most cases of IPC demonstrated cystic changes with a mural nodule or intracystic papillary lesion ranging in size from 0.8 to 6.3 cm. The invasive component was also characterized by an infiltrative appearance with extension beyond the fibrous capsule of the lesion (Fig. 1a). The tumor appeared pink-tan and exceptionally fragile.

Fig. 1.

a Specimen grossing showing the cystic tumor mass with solid area of invasion. b Invasive papillary carcinoma of the breast. HE staining; magnification ×40. c Malignant papillary lesion. HE staining; magnification ×100. d Immunohistochemical staining of SMA revealed 90 % unstained cytoplasm in IPC, original magnification ×40. e Immunohistochemical staining of p63 revealed 90 % unstained nucleus in IPC, original magnification ×40. f Immunohistochemical staining of calponin revealed 90 % unstained cytoplasm in IPC, original magnification ×40

Invasive papillary carcinoma is a morphologically heterogeneous group of lesions, all of which share a growth pattern characterized by the presence of arborescent fibrovascular stalks lined by epithelial cells and with papillary architecture (Fig. 1b, c). The carcinoma cells displayed cytologic atypia and showed stratification. We found low and intermediate nuclear grades typical of these lesions (91.19 %), but high-grade nuclear atypia was rare (8.81 %). All our IPC cases lacked an intact myoepithelial cell layer within the fibrovascular papillae. This helped distinguish malignant from benign intraductal papillomas. There were 27 cases that were not distinguishable from benign papillary lesions, so we used special immunohistochemical staining markers smooth muscle antibody (SMA), p63, and calponin to identify myoepithelial cells (Fig. 1d–f). The invasive foci showed a papillary pattern or a papillary pattern with an infiltrating ductal carcinoma component, but this component showed a predominately papillary pattern.

Molecular immunophenotype

We determined the intrinsic molecular subtype in 284 IPC tumors. Of these, 145 (51.06 %) patients were classified as luminal A, 49 (17.25 %) as luminal B, 37 (13.03 %) as Her2-overexpressing, and 53 (18.66 %) as triple-negative. The distributions of clinicopathological characteristics among these immunophenotypic subtypes were compared (Table 2). We found significant differences in age, tumor grade, and menopausal status among the subtype cohorts (P < 0.05). There were no differences in tumor size, lymph node status, or family history (P > 0.05).

Table 2.

Distribution of clinicopathological characteristics among the various breast cancer intrinsic subtypes

Characteristic	Luminal A (n = 145)		Luminal B (n = 49)		Her-2 (n = 37)		Triple-negative (n = 53)		χ2	P
	No.	%	No.	%	No.	%	No.	%
Age at diagnosis (years)
<50	22	15.17	9	18.37	12	32.43	16	30.19	8.85	<0.05
≥50	123	84.83	40	81.63	25	67.57	37	69.81
Tumor size (cm)
<2	39	26.90	12	24.49	11	29.73	19	35.85	4.32	>0.05
2–5	95	65.52	34	69.39	21	56.76	29	54.72
>5	11	7.58	3	6.12	5	13.51	5	9.43
Tumor grade
Low	67	46.21	23	46.94	5	13.51	6	11.32	38.44	<0.05
Intermediate	73	50.34	22	44.90	25	67.57	38	71.70
High	5	3.45	4	8.16	7	18.92	9	16.98
Lymph node status
Negative	123	84.83	40	81.63	29	78.38	43	81.13	1.07	>0.05
Positive	22	15.17	9	18.37	8	21.62	10	18.87
Menopausal status
Postmenopausal	116	80.00	40	81.63	22	59.46	33	62.26	12.13	<0.05
Premenopausal	29	20.00	9	18.37	15	40.54	20	37.74
Family history
Yes	21	14.48	5	10.20	6	16.22	9	16.98	1.08	>0.05
No	124	85.52	44	89.80	31	83.78	44	83.02

Her-2 human epidermal growth factor receptor 2

Disease-specific outcomes

During the period lasting from January 2000 to May 2006, 83 patients diagnosed with IPC underwent long-term follow-up. We chose these 83 IDC cases out of 13,428 patients who all had similar follow-up time and were treated during the same period as the selected 83 IPC cases. They were matched according to age, menopausal status, lymph node status, tumor size, and tumor grade. The median follow-up time was 89 months for patients with IPC, and 86 months for patients with IDC. IPC was associated with a higher 5-year OS than IDC (92.77 vs. 87.95 %) and with better DFS than IDC (87.95 vs. 80.72 %) (Fig. 2a, b). For the IPC cases, the actuarial 5-year OS rates of luminal A, luminal B, Her2-overexpressing, and triple-negative subtypes were 97.78, 92.31, 81.82, and 85.71 %, respectively (Fig. 2c), and the actuarial 5-year DFS rates were 95.56, 84.62, 72.73, and 78.57 %, respectively (Fig. 2d). Multivariate Cox regression analysis including age, tumor size, tumor grade, LN status, Her2 status, and subtypes showed that only LN status (P = 0.036; HR = 4.512; 95 % CI, 1.104–18.440 and P = 0.009; HR = 4.485; 95 % CI, 1.448–13.895) and molecular subtypes (P = 0.033; HR = 1.878; 95 % CI, 1.051–3.354 and P = 0.042; HR = 1.588; 95 % CI, 1.016–2.482) were independent predictors of OS and DFS (Table 3). However, age, tumor size, tumor grade, and Her2 status were not significantly associated with OS or DFS in these cases (results not shown). This may be because our study population was not still large enough.

Fig. 2.

a Kaplan–Meier survival curve showing the comparison of overall survival between invasive papillary carcinomas (IPC; upper curve) and invasive ductal carcinomas (lower curve χ² = 4.148, P = 0.042). b Kaplan–Meier survival curve showing the comparison of disease-free survival between invasive papillary carcinomas (IPC; upper curve) and invasive ductal carcinomas (lower curve χ² = 4.367, P = 0.037). c Kaplan–Meier survival curve showing the comparison of overall survival among molecular subtype of IPC; when luminal A subtypes compared with Her2-overexpressing and triple-negative subtype, there were significant differences (χ² = 4.679, P = 0.031 and χ² = 5.135, P = 0.023). d Kaplan–Meier survival curve showing the comparison of disease-free survival among molecular subtype of IPC; when luminal A subtypes were compared to Her2-overexpressing and triple-negative subtypes, there were significant differences (χ² = 6.134, P = 0.013 and χ² = 4.681, P = 0.030)

Table 3.

Multivariate Cox regression analysis of factors associated with overall survival and disease-free survival

Variable	Overall survival (n = 72)			Disease-free survival (n = 67)
	HR	95 % CI	P	HR	95 % CI	P
Age (years) (<50 vs.  ≥50)	0.947	0.249–3.601	0.937	1.285	0.399–4.138	0.674
Tumor size (cm) (<2 vs. ≥2)	0.266	0.052–1.363	0.112	0.288	0.076–1.100	0.069
Tumor grade (low vs. middle vs. high)	1.238	0.480–3.195	0.659	1.225	0.551–2.727	0.619
Lymph node status (negative vs. positive)	4.002	1.043–15.363	0.043	4.174	1.382–12.610	0.011
Her2 status (positive vs. negative)	0.468	0.126–1.733	0.256	0.504	0.179–1.422	0.196
Subtypes (luminal A vs. luminal B vs. Her2 overexpressing vs. triple-negative)	1.848	1.022–3.340	0.042	1.570	1.003–2.458	0.048

CI confidence interval, HR hazard ratio, Her2 human epidermal growth factor receptor 2

Among these cases of IPC, two recurrences were observed in the contralateral breast, five in the ipsilateral breast, and four in the regional LN. Among patients who died, seven were shown by invasive foci to have IDC, invasive lobular carcinoma (ILC) or invasive micropapillary carcinoma (IMPC). Only four cases displayed pure IPC (Table 4).

Table 4.

Outcomes of patients who developed DFBC

Patient no.	Type of mixture component	Recurrence before DFBC	Grade of recurrence	Time to recurrence (months)	DM before DFBC	Time to DM (months)	Time to DFBC (months)
1	No	LRa	Intermediate	34	Yes	39	58
2	IDC	LR + LNR	Intermediate	59	Yes	67	85
3	IDC	No	No	0	Yes	31	54
4	IDC	CLBC	Intermediate	50	Yes	66	78
5	No	No	No	0	Unknow	Unknow	90
6	No	LR + LNR	High	36	Yes	39	43
7	IMPC	LR	High	45	Yes	56	76
8	No	No	No	0	Yes	25	56
9	ILC	CLBC + LNR	Intermediate	74	Yes	77	100
10	IDC	LR + LNR	Intermediate	40	Yes	40	50
11	IDC	No	No	0	Yes	15	25

DFBC death from breast cancer, IDC invasive ductal carcinoma, DCIS ductal carcinoma in situ, ILC invasive lobular carcinoma, IMPC invasive micropapillary carcinoma, LR local recurrence, CLBC contralateral breast cancer, LNR lymph nodes recurrence, DM distant metastases

^a13 months after LR, patient developed CLBC

Discussion

Previous studies have produced limited data regarding IPC epidemiology. Only a small number of studies have focused on outcome. In this study, we used a large cohort of patients from a single institution and made a detailed comparison between patients with IPC and those with IDC to determine the relative prognosis of IPC through a rigorous histologic review. We discovered that patients with pure IPC had significantly more favorable prognoses than those with IDC. However, IPC incorporating components of invasive ductal or lobular carcinoma was typically associated with a poorer outcome. In this study, the large number of patients and the single institution nature of the study population enhance the reliability of the results. The main weakness of this study is its retrospective nature, with its inherent bias.

Malignant papillary neoplasms of the breast comprised a number of microscopically distinct lesions, including ductal carcinoma in situ (DCIS) arising in an intraductal papilloma, papillary DCIS, encapsulated papillary carcinoma, solid papillary carcinoma, and IPC (Mulligan and O’Malley 2007; Collins and Schnitt 2008; Ueng et al. 2009). Elverici reported that more than 75 % of the cases included a ductal DCIS component (Elverici et al. 2007). Schneider found that IPC may arise as a single nodule in the central portion of the breast or as multiple nodules that extend out from the subareolar area to the periphery of the breast (Schneider 1989). Tumor has a tendency to present as a large mass due to the common presence of bulky cystic components. For this reason, the size of the invasive component is small (Elverici et al. 2007). In this series, IPC was defined according to internationally accepted criteria using a well-defined constellation of histologic features, including more than 90 % fibrovascular papillae lack an intact myoepithelial cell layer. SMA and p63 are the most commonly used markers of myoepithelial cells which have different sensitivities and specificities (Mulligan and O’Malley 2007; Nassar et al. 2006; Shah et al. 2006; Hill and Yeh 2005). In some of our cases, the presence of myoepithelial cells was investigated using SMA, p63, and calponin. Pathologists must remain vigilant because IPC often takes the form of well-defined, smoothly contoured nests that can simulate the appearance of ducts overrun by noninvasive papillary carcinoma (Koerner 2010). The distinction between invasive papillaries and micropapillary carcinoma has significant clinical implications because the latter is considered an aggressive form of mammary carcinoma frequently associated with lymph-vascular invasion and axillary lymph node metastases (Walsh and Bleiweiss 2001; Nassar et al. 2001; Pettinato et al. 2004). The results in this study emphasize the importance of high-quality pathology assessment of tumor characteristics in distinguishing IPC from IDC.

In our study, IPC was associated with absence of axillary node metastases. Compared with IDC, IPC was found usually to affect patients of postmenopausal females, which is similar to that reported previously. In addition, the tumor grade of IPC mostly appears with low or intermediate while IDC shows intermediate and high. No significant difference between both tumor types was found with regard to association with family history or molecular subtype. These studies are inconsistent with that of Kader et al., who have reported all breast papillary carcinoma including encapsulated, solid, and IPC showed ER-positive and HER2-negative (luminal A type). However, there is only 12 (19.0 %) IPC cases in that paper, but in our article, there are a total of 284 IPC cases, which can better reflect the overall situation.

Invasive papillary carcinoma is low-grade, slow-growing cancers with few common axillary node metastases, and the prognosis remains favorable even for women with axillary metastasis. In our case cohort, low and intermediate nuclear grades were common, and high-grade nuclear atypia was rare. The rate of axillary node metastasis was lower than that of IDC. Our study shows that the OS and DFS rates of IPC were higher than those of IDC, which is similar to that reported previously (Soo et al. 1995). Mitnick et al. (1990) reported the 5-year DFS rate of IPC to be about 90 %. When addressing the issue of prognosis in IPC, we found it difficult to distinguish recurrences of the primary tumor from development of new primary tumors. Moreover, we found that in a different series of 1,000 metastatic breast carcinoma patients, only 5 patients had an initial diagnosis of IPC. In all 5 patients, there were intervening recurrences as IDC. This indicates that the chance of primary IPC developing distant metastasis or causing death without an intervening event of different histologic type is exceptionally low. Many IPCs do not spread very far beyond their original site (Pervez and Khan 2007). However, some cases showed metastatic disease or pursued an aggressive clinical course (Coyne 2007). The incidences of local recurrence, distant metastasis, and cancer-related death were relatively low. In the 11 patients who died from breast cancer, 7 showed mixtures of other invasive components, and only 4 presented with pure IPC. This indicated that the chance of pure IPC causing death was exceptionally low. One explanation for this may be that IPC is less biologically aggressive than other forms of cancer. Another may be that the majority of patients with IPC were older and may have succumbed to competing cause of deaths due to comorbid illnesses.

In this study, the majority of IPC patients were positive for ER and PR and negative for Her2. In the four molecular subtypes, luminal A was very common, and the OS and DFS were clearly better than other subtypes. Of particular concern was the low rate of OS and DFS observed in the Her2 and triple-negative subgroups. Our study revealed that triple-negative and Her2 tumors were closely associated with high risk of recurrence and death from breast cancer. In the four molecular subtypes, there were significant differences in age, tumor grade, and menopausal status among the subtype cohorts. Our data show there is no statistically significant difference among the various IPC subtypes among tumor size, lymph node status, and family history. Maybe the data were still limited in this study. Our study showed high-risk factors such as more lymph nodes involved and subtypes were statistically significant independent predictors of OS and DFS. These results are concordant with repeated observations. However, relatively few studies have performed to find an association between IPC molecular subtypes and relapse or death. We have demonstrated the prevalence of distinct breast cancer subtypes in China through immunohistochemical markers. The classification was found to be prognostic on multivariable analysis. However, the biology underlying these observations remains poorly understood. We propose that the survival patterns reflect the underlying molecular heterogeneity of breast cancer, and this heterogeneous biology reflects the fact that breast cancers can initiate in different cell types, either breast epithelial stem cells or their progeny.

There have been dramatic advances in mammographic screening, which has seen increasingly widespread use. This had led to the increased number of diagnoses of IPC seen in recent years. IPC can be managed through local treatments, such as surgery and adjuvant radiation therapy. Adjuvant systemic chemotherapy can cause significant patient morbidity and can be detrimental to low-risk patients. Although hormone therapy and targeted therapy are applied widely now, the long-term effects and the relative benefits of these therapies merit further investigation. We here have a hypothesis that IPC seems to have a favorable prognosis with adequate local therapy alone, although most patients in our cohort were given chemotherapy, hormonal therapy, or targeted therapy. It was only a conjecture, because we have no data to support it. To provide the greatest benefit to patients, it is important that oncologists determine the appropriate therapeutic regimen for each distinct tumor subtype based on the likelihood of recurrence. Further research is needed to determine the role for systemic therapies in the treatment of IPC.