ABSTRACT: Inflammatory breast cancer (IBC) is a rare and aggressive subtype of locally advanced breast cancer (LABC). Its diagnosis is primarily clinical; however, a pathological confirmation of invasive cancer is required. Historically, IBC was a uniformly fatal disease. A major advance in the last three decades has been the introduction of a multidisciplinary approach to the management of this aggressive disease, incorporating pre-operative chemotherapy, surgery, and radiation therapy; this approach has significantly improved survival. Our review focuses on the progress made in the field of IBC research over the last decade, with particular attention to advances in the areas of epidemiology, molecular biology, and clinical management.
Over the past decade, inflammatory breast cancer (IBC), a rare and aggressive subtype of locally advanced breast cancer (LABC), has received much attention at the level of public awareness as well as at the level of research. The diagnosis of IBC is primarily clinical, although a pathological confirmation of invasive cancer is required. The most widely used definition of IBC is that which has been put forward by the American Joint Committee on Cancer; this definition states in part that IBC is “a clinicopathologic entity characterized by diffuse erythema and edema of the breast, often without an underlying palpable mass.”[1] The incidence of IBC varies across different geographic locations but typically accounts for approximately 1% to 5% of all newly diagnosed breast cancers in the United States.[2] The rarity of the disease makes large prospective clinical trials difficult, with the result that much of the information we have on IBC is based on retrospective studies. Furthermore, the subjective nature of the clinical diagnostic criteria has led to wide variability in clinical reports at both the epidemiological and molecular levels.
Historically, IBC was a uniformly fatal disease, with a 5-year actuarial overall survival of less than 5%, a median survival of only 15 months,[3] and local recurrence rates as high as 50% when treated with surgery alone or in combination with radiation therapy.[4,5] A major advance of the last three decades has been the introduction of a multidisciplinary approach to the management of this aggressive disease that incorporates pre-operative chemotherapy, surgery, and radiation therapy. As a result, survival outcomes have improved significantly, with 15-year survival rates of 20% to 30% reported.[6,7] However, despite this progress, most women in whom IBC is diagnosed eventually experience a recurrence and die from the disease—highlighting the need for further research.
Specifically, current research is focused on the following objectives:
• Refining the clinical diagnostic parameters to decrease subjectivity.
• Delineating diagnostic markers to enhance diagnostic accuracy.
• Developing predictive and prognostic markers to aid in treatment planning.
• Characterizing IBC at the molecular level.
• Developing targeted therapies based on an enhanced understanding of the biology of IBC, with the hope that such an approach will positively impact prognostic outcomes.
High-throughput molecular analysis has provided great insight into the understanding and characterization of non-IBC breast tumors.[8] This has perhaps been one of the biggest advances in the field of breast cancer. Such technology has rarely been used in the realm of IBC, in part because of IBC's comparative infrequency and thus the small number of diagnostic samples. However, the last decade has seen an upsurge in the understanding of IBC at both a clinical and a molecular level.
Epidemiology and Risk Factors
Epidemiological research on IBC has not enjoyed the kind of progress seen with other subtypes of breast cancer. This lag in progress is primarily due to the rarity of the disease, which has resulted in small numbers of patients with IBC being available to any single institution; it is also due to the lack of an agreed upon case definition for IBC, which has made it difficult for institutions to collaborate. Nonetheless, data have emerged from large population-based studies. One important observation that has been made is that, unlike with non-IBC breast tumors and despite the overall low incidence of IBC (it accounts for only 1% to 5% of all newly diagnosed breast cancer cases), the incidence of IBC in the United States appears to be increasing.[1] This may in part be the result of increased awareness among both patients and oncologists.
It is interesting to note that there is striking geographic variation in the incidence of IBC, with the lowest incidence reported in North America[1] and a higher incidence reported in North African countries—especially Morocco, Algeria, Tunisia, and Egypt, where the incidence of IBC has been reported to be between 10% and 15%.[9] However, it is uncertain in these countries whether case registration is complete and whether strict definitions are used for IBC detection.
Variations in incidence have also been reported among different ethnic groups in the United States. In a large population-based study of 3,626 women with IBC diagnosed between 1994 and 1998, Wingo et al[10] reported an IBC rate of 1.3 per 100,000 for all races combined, with African Amer-ican women having the highest risk (1.6 per 100,000) and Asian and Pacific Islander women having the lowest risk (0.7 per 100,000). Other key observations concerning women in whom IBC has been diagnosed include their younger age at presentation compared with women with non-IBC, their lower median survival compared with patients with LABC (2.9 years vs 6.4 years),[1] and the significantly in-creased odds of IBC developing in women with a high body mass index.[11] Currently, the mammary tumor virus is being investigated as a potential risk factor for IBC based on the observation that in North America a higher incidence of the viral sequence (71%) has been observed in women with IBC than in women with sporadic breast cancers; a similar incidence has also been observed in IBC tumor samples from Tunisia.[12]
Advances in the Molecular Characterization of IBC
Molecular characteristics
As stated earlier, IBC is diagnosed primarily on the basis of a set of clinical characteristics, with pathological confirmation of invasive carcinoma also required. Characteristic pathological findings that can aid in the diagnosis of IBC include the presence of dermal lymphatic invasion by invasive tumor emboli, leading to obstruction of the lymphatic drainage and resulting in a clinical appearance of erythema and edema mimicking an inflammatory process.[13,14] However, the absence of a dermal lymphatic invasion does not preclude a diagnosis of IBC. Molecular alterations that have been reported with high incidence in IBC include negative hormone receptor status,[15] overexpression and/or gene amplification of human epidermal growth factcor receptor 2 (HER2),[16] overexpression of epidermal growth factor receptor (EGFR), high S-phase fraction, high-grade p53 mutation,[16] overexpression of E-cadherin,[17] dysfunction of mucin 1, overexpression of RhoC GTPase (a gene involved in cytoskeletal reorganization),[18] loss of the LIBC (lost in inflammatory breast cancer) gene,[18] and a high level of angiogenesis.[19] Studies have also demonstrated high levels of expression of multiple chemokine receptors in IBC. In a recent study, Cabliogu et al looked at the prognostic implications of the chemokine receptors CXCR4 and CCR7, as well as HER2 and EGFR, in IBC.[20] Working with a sample of 44 cases of IBC, the authors reported high levels of expression of CXCR4, EFGR, and HER2 amplification that were associated with a higher risk of recurrence and death. However, the molecular alterations described here are not specific to IBC. Several studies looking at these molecular markers have shown that in multivariate analysis, IBC status itself still remains an independent adverse prognostic feature; this indicates that other molecular differences are probably present that distinguish IBC from non-IBC.
Gene expression profiling
To better define the differences between IBC and non-IBC that are present at the molecular level, high-throughput molecular technologies have been utilized that allow for the analysis of several thousand genes in a tumor sample simultaneously. Over the past decade, several studies have specifically looked at expression profiling. To the best of our knowledge, six groups have profiled clinical samples of IBC and reported their results.[21-30] One of the earliest publications, from Bertucci et al in 2004,[21] used cDNA microarrays containing approximately 8000 genes to profile 37 IBC and 44 non-IBC prechemotherapy tumor samples. The authors made a number of important observations, including their finding of the presence of extensive transcriptional heterogeneity in IBC, and their demonstration that IBC exhibited overexpression of basal, immune, and vascular gene clusters and exhibited underexpression of luminal gene clusters compared with non-IBC tumors. In addition, they identified a 109-gene signature that discriminated between the two sample types. Using the same data set, Bertucci et al further demonstrated[22] that similar to non-IBC, five molecular subtypes (defined according to the Stanford intrinsic gene set)—including luminal A, luminal B, basal, HER2-positive, and normal-like—could be demonstrated in IBC. Higher frequencies of basal and HER2-positive IBC samples were observed compared with the frequencies of these two subtypes in non-IBC samples.
Van Laere et al in 2005[23] used cDNA microarrays containing approximately 6000 genes to profile 16 IBC and 18 non-IBC prechemotherapy samples. Using supervised analysis, the authors identified 953 genes that exhibited a 1.5-fold difference in expression between the two sample types. Of these genes, 756 (retained after the exclusion of genes related to the expression of estrogen receptor 1 and HER2) resulted in a perfect segregation of IBC and non-IBC samples. Using the same data set, the investigators further identified five different cell-of-origin subtypes in the two sample sets, with the basal (37%) and HER2-positive (13%) subtypes representing 50% of the IBC samples.[24] Using Affymetrix technology on a set of 19 IBC and 40 non-IBC samples, the investigators confirmed their initial findings, identifying 26% of the IBC samples as basal and 42% as HER2-positive.[25]
Bièche et al[26] used real-time polymerase chain reaction (RT-PCR) to quantify the mRNA expression level of 538 candidate genes in 8 IBC and 8 non-IBC samples; they demonstrated a more than two-fold change in a set of 48 genes between the two sample types. The investigators further identified a three-gene signature composed of epiregulin (EREG), v-myc myelocytomatosis viral-related oncogene (MYCN), and sonic hedgehog (SHH). This gene signature was shown to distinguish between three subgroups of women with IBC with different recurrence rates.
Nguyen et al,[28] using Affymetrix U133A microarrays, defined the expression profiles of prechemotherapy samples from 13 patients with IBC and 12 patients with stage III non-IBC. Because of considerable heterogeneity across both sample types, a robust signature could not be identified; however, the investigators were able to demonstrate a higher expression of genes associated with increases in metabolic rate, lipid signaling, and cell turnover in the IBC samples than in the non-IBC samples. They also confirmed the presence of three molecular subtypes of IBC, including luminal, HER2-positive, and basal subtypes.
The most recent study comes from the group at the National Cancer Institute in Bethesda;[29] these investigators used Affymetrix U133A microarrays to profile 15 IBC samples and 35 non-IBC samples. Eighty seven percent of the IBC samples and 23% of the non-IBC samples were obtained postchemotherapy, and each sample was obtained via laser capture microdissection. The investigators were unable to identify a significant signature in the epithelia that could distinguish between the two sample types, but they were able to identify a stromal signature that could distinguish IBC from non-IBC. They further demonstrated that the 109-gene signature originally identified by Bertucci et al[21] performed better in classifying IBC when the tumor stroma was used instead of the tumor epithelial; this finding lends evidence to the hypothesis that gene expression in the stroma contributes to the IBC phenotype.
There are several drawbacks to these studies, including the facts that they involved very small numbers of IBC and non-IBC samples, that the definition of IBC varied across studies, that some studies used prechemotherapy samples while others used postchemotherapy samples, and that the methods used to collect samples also varied. Despite these drawbacks, we can derive valuable information from the data described. First, the transcriptional heterogeneity of IBC appears to be as extensive as that for non-IBC. Second, at the clinical level, the existence of molecular subtypes within IBC may in the future prove to have a prognostic significance similar to that seen in non-IBC subtypes. Third, the genetic signatures identified may also be used to divide patients with IBC into several prognostic groups. Bertucci et al[21] identified an 85-gene set that divided patients with IBC into two groups with different pathological complete response rates (pCR), while the three-gene signature identified by Bièche et al[26] categorized patients with IBC into good, intermediate, and poor outcome groups. Fourth, a signature identified for IBC may reflect the aggressiveness of breast cancer generally and thus may play a prognostic role if also identified within non-IBC samples. Van Laere et al[30] applied a gene signature predictive of IBC to a set of 1157 non-IBC samples and showed that non-IBC samples with an IBC-like phenotype had a significantly shorter relapse-free survival than samples that did not have an IBC-like phenotype.
