Intracystic Papillary Carcinoma of Breast: Interrelationship With in Situ and Invasive Carcinoma and a Proposal of Pathogenesis

Array Comparative Genomic Hybridization Study of 14 Cases

Thaer Khoury; Qiang Hu; Song Liu; Jianmin Wang

Disclosures

Mod Pathol. 2014;27(2):194-203. 

In This Article

Results

Cases

A total of 22 intracystic papillary carcinoma cases were identified in the pathology database at Roswell Park Cancer Institute between 1992 and 2010. Twelve cases had concurrent ductal carcinoma in situ, whereas nine cases had concurrent invasive ductal carcinoma. Concurrent intracystic papillary carcinoma, ductal carcinoma in situ and invasive ductal carcinoma were identified in four cases (Figure 1). Array comparative genomic hybridization test was successful in 14 intracystic papillary carcinoma (5 grade I and 9 grade II), 6 ductal carcinoma in situ (4 grade I and 2 grade II with comedo-type necrosis) and 4 invasive ductal carcinoma (all modified Scarff Bloom Richardson grade I). The median size and range of intracystic papillary carcinoma was 17 mm (6–47 mm). All invasive tumors were of invasive ductal carcinoma type with the median and range of tumor size 5 mm (1–7 mm). Although five cases had array comparative genomic hybridization results on intracystic papillary carcinoma with concurrent ductal carcinoma in situ, these results were available for intracystic papillary carcinoma with concurrent invasive ductal carcinoma in two cases. Metastatic carcinoma to sentinel lymph node was seen in 1 of 13 cases. All cases had negative surgical resection margins, with one case having close margin (1 mm). All cases were estrogen receptor positive and progesterone receptor positive. Four (18%) cases had nuclear grade I and 18 (82%) had nuclear grade II.

Figure 1.

Hematoxylin and eosin (a) and smooth muscle myosin (SMM) (b) immunostain figures of intracystic papillary carcinoma, invasive ductal carcinoma and ductal carcinoma in situ (in circles) ( × 4).

All patients were women except for one (patient no. 1). All women were postmenopausal except for two who were perimenopausal. The patients' age ranged from 51 to 94 years with a median of 73 years. All tumors were surgically removed (excision or mastectomy). Therapy modality was available for all patients except for four. One patient had chemotherapy. However, this case was excluded from the analysis because of the unsuccessful array comparative genomic hybridization. The time of follow-up range was 1–117 months with a median of 36.5 months. Two patients were lost to follow-up. Three patients developed local recurrence, two contralateral in the form of invasive ductal carcinoma and one ipsilateral in the form of papillary carcinoma. All patients who developed local recurrence had concurrent ductal carcinoma in situ (Table 1).

Array Comparative Genomic Hybridization Findings

Genome Profile and Variation of Intracystic Papillary Carcinoma. Intracystic papillary carcinoma showed 16p gain, 16q loss and 1q gain. In addition, it revealed some novel large-scale copy number variations, 7q loss (4 of 14 cases), 8p23.3–21.1 loss (5 of 14 cases), 11q14.1–22.1 loss (4 of 14 cases), 11q22.1–25 loss (5 of 14 cases) and 9p13.1–12 loss (5 of 14 cases) (Figure 2).

Figure 2.

Frequency plot of copy number gains and losses in intracystic papillary carcinoma cases. The proportion of tumors in which each bacterial artificial chromosome (BAC) clone is gained (green bars) or lost (red bars) is plotted (y axis) for each BAC clone according to its genomic position (x axis).

The genomic variation between intracystic papillary carcinoma with ductal carcinoma in situ (n=6) vs without ductal carcinoma in situ (n=8) showed that intracystic papillary carcinoma without ductal carcinoma in situ had 5q35.3 gain (from 176 474 585 to 180 175 485, in 3 of 4 cases) vs 1 of 10 intracystic papillary carcinoma with ductal carcinoma in situ (P=0.041). Intracystic papillary carcinoma without ductal carcinoma in situ had 8q24.3 gain (from 142 015 488 to 145 957 473, in 3 of 4 cases) vs 1 of 10 intracystic papillary carcinoma with ductal carcinoma in situ (P=0.041). Finally, intracystic papillary carcinoma without ductal carcinoma in situ had 21q13.2–13.31 gain (from 42 127 232 to 44 695 209 in 3 of 4 cases) vs 0 of 10 intracystic papillary carcinoma with ductal carcinoma in situ (P=0.011) (Figure 3). The genes involved are listed in Supplementary Table 1.

Figure 3.

Frequency plot of copy number gains and losses in intracystic papillary carcinoma (ICPC) with ductal carcinoma in situ (DCIS) comparing with ICPC without DCIS. Significant changes included 5q35.3 gain, 8q24.3 gain and 21q13.2–13.31 gain in ICPC without DCIS.

The genomic variation between intracystic papillary carcinoma with invasive ductal carcinoma (n=6) vs without invasive ductal carcinoma (n=8) showed that the latter had 11q22.1–23.3 loss in 6 of 8 cases vs 0 of 6 in the former (P=0.031). Similarly, chr5 gain was seen in four of eight cases (an additional case also had gain of part of chr5) in intracystic papillary carcinoma without invasive ductal carcinoma vs 0 of 6 in intracystic papillary carcinoma with invasive ductal carcinoma with borderline significance (P=0.085) (Figure 4). The involved genes are listed in Supplementary Table 2. The following genes are of interest, matrix metalloproteinases (MMP) 1, 3, 7, 8, 10, 12, 13 and 27.

Figure 4.

Frequency plot of copy number gains and losses in intracystic papillary carcinoma (ICPC) with invasive ductal carcinoma (IDC) comparing with ICPC without IDC. Significant changes included chr5 gain and 11q22.1–23.3 loss in ICPC without IDC.

There was no genomic variation with relation to the nuclear grade, patient age or the risk for local recurrence (Figure 5).

Figure 5.

Unsupervised hierarchical clustering analysis of intracystic papillary carcinoma (ICPC) with concurrent invasive ductal carcinoma (IDC) and ductal carcinoma in situ (DCIS). Hierarchical cluster analysis performed with array comparative genomic hybridization categorical states (ie gains, losses) using the Euclidean distance metric and the Wards algorithm. Note the clustering of ICPC with the concurrent DCIS or IDC. No clustering was noted in relation to tumor recurrence, grade or age.

Overall Genomic Variation Between Intracystic Papillary Carcinoma, Ductal Carcinoma in Situ and Invasive Ductal Carcinoma. Hierarchical agglomerative clustering of samples using probe log 2 ratio shows that the three types could not be separated into different clusters. Four of five concurrent intracystic papillary carcinomas with ductal carcinoma in situ clustered together. Both invasive ductal carcinomas clustered with the concurrent intracystic papillary carcinoma (Figures 5 and 6). Pearson's correlation between foci of the sample patient using log 2 ratio ranges from 0.49 to 0.84 with a median value of 0.73.

The total genome copy number change (gain or loss) for intracystic papillary carcinoma (n=14), ductal carcinoma in situ (n=6) and invasive ductal carcinoma (n=4) was 315 segments, 100 segments and 60 segments, respectively, whereas the median and range was 14.5 (4–68), 13.5 (11–29) and 5 (3–42), respectively, with no statistically significant difference. The median and range for the proportions of changed genome was 8% (3.7–19.1%), 6.2% (5.9–11.9%) and 2.2% (0.09–10.9%), respectively (P=0.06 Wilcoxon's rank-sum test). The calculations excluded chromosomes X and Y. All three entities shared many common copy number variations, particularly those with large-scale variation. However, distinct copy number variations were present in each entity (Figure 6). The detailed copy number variation is listed in Supplementary Table 3.

Figure 6.

Clustering intracystic papillary carcinoma (ICPC) with concurrent ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC) using (a) cluster dendogram and (b) principal component analysis.

The major discordant copy number variations for intracystic papillary carcinoma were on chr7. More specifically, one sample (no. 13) had focal loss at 7p15.1–14.1 and p12.1–q11.21, whereas two samples had p arm gain. The same sample (no. 13) had q21.11 gain, whereas three samples q arm loss.

Specific Genome Variation Between Intracystic Papillary Carcinoma and Ductal Carcinoma in Situ. The mean probability of copy number changes for ductal carcinoma in situ was similar to intracystic papillary carcinoma with the following additional focal changes that was seen in ductal carcinoma in situ compared with intracystic papillary carcinoma: 8q22.2–22.3 gain (from 100 114 836 to 102 268 083, in 3 of 6 cases), 8q24.3 gain (from 142 202 743 to 145 957 473, in 3 of 6 cases), 11q25 loss (from 131 565 893 to 133 557 293, in 3 of 6 cases) and 12q24.31–24.33 loss (from 124 302 840 to 129 818 768, in 2 of 6 cases). Supplementary Table 4 lists all the changes with the involved genes. Interestingly, many of these genes are involved in the cell adhesion and cell motility (Table 2).

There was one strong focal change within 1q gain: 1q21.3–23.1 (8 of 14 cases) (from 153 183 579 to 155 186 038) in intracystic papillary carcinoma. This focus was not seen in concurrent ductal carcinoma in situ. The genes located in this region are presented in Supplementary Table 5. Two genes are of interest, papillary renal cell carcinoma (PRCC) and neutrotophic tyrosine kinase receptor type 1 (NTRK1).

processing....