Angiogenesis in Hyperplasia to In-Situ Breast Cancers

Institution: University of California, Irvine
Investigator(s): Min-Ying (Lydia) Su, Ph.D. -
Award Cycle: 2003 (Cycle IX) Grant #: 9WB-0020 Award: $250,000
Award Type: STEP Award
Research Priorities
Biology of the Breast Cell>Pathogenesis: understanding the disease



Initial Award Abstract (2003)
Widespread implementation of mammographic screening has led to earlier detection of breast carcinomas. Today DCIS is estimated to comprise 20-40% of all mammogram-directed biopsies. The advance in screening technology may further shift the diagnosis to even earlier stages of breast cancer progression or even pre-malignancy. One theory is that invasive breast cancer results from a temporal and genetic sequence of several steps: from normal epithelium, through hyperplasia, progression to atypical hyperplasia (ADH), and then DCIS. In terms of eventual progression, DCIS can remain unchanged, or regress to previous normal state, or it may progress to become invasive cancer. However, little is known about the pathways and the involved stimulating factors leading to the transformation process.

Formation of new blood vessels (angiogenesis) within tumors is essential for the growth and spread of cancer. In this study we will investigate the role of angiogenesis on disease progression from pre-malignant hyperplasia to in situ cancers. Many studies have shown that microvessel density in invasive breast cancer correlates with metastatic potential and patient survival. Since the ability of tumor cells to induce a vascular supply occurs early in the cell transformation, pre-malignant breast lesions may also induce a vascular supply to facilitate malignant transformation. We hypothesize that angiogenesis is associated with disease progression, thus it may have diagnostic and prognostic values. Specifically we will test that MRI measurements of angiogenesis can improve the specificity of mammography in differentiating benign hyperplasia from in situ cancers.

We will use dynamic contrast-enhanced magnetic resonance imaging (MRI) and immunohistochemical molecular biomarkers to characterize the angiogenic status from fibrocystic proliferation, atypical hyperplasia, to in situ breast cancer. One hundred patients with probable benign or probable malignant findings (with BIRADS 3 or 4- by mammography) will be enrolled. They will receive MRI examination prior to core biopsy or surgery. Then the specimens will be analyzed for molecular angiogenic markers, including expression of mutant tumor suppressor gene p53, Vascular Endothelial Growth Factor (VEGF), thrombospondin-1 (TSP1), and microvessel density (using CD31 and CD 105 staining). MRI vascular parameters and IHC biomarkers will be correlated with different disease stages to investigate their association with disease progression.

To date, most of the angiogenesis research in breast cancer has been focused on invasive cancers. Little is known about its role on pre-malignant proliferative diseases and in situ cancers. The use of MRI in conjunction with IHC techniques may provide microscopic (IHC) and macroscopic (MRI) measurements of the vascularity status. We feel these methods complement each other, and may provide a more accurate assessment of angiogenesis. More complete clinical evaluation for breast lesions may lead to a better treatment and surveillance protocols for individual patients.


Final Report (2005)
Introduction: Widespread implementation of self-examination and mammographic screening has led to earlier detection of breast cancer. The diagnosis of in-situ cancers has increased rapidly over the past few years. A commonly proposed theory is that invasive cancer results from a sequence of several steps: from normal epithelium to become hyperplastic and progress to atypical hyperplasia (ADH) then DCIS. However, little is known about the pathways and the involved stimulating factors leading to the transformation process. Formation of new blood vessels (angiogenesis) within tumors is essential for the growth and spread of cancer. In this study we will investigate the role of angiogenesis on disease progression from pre-malignant hyperplasia to in-situ cancers, using dynamic contrast enhanced MRI and angiogenic molecular markers.

Topic addressed: MRI and immunohistochemistry are used to characterize the angiogenic status from fibrocystic proliferation, atypical hyperplasia, to in-situ breast cancer, to investigate the role of angiogenesis in the disease continuum, from benign disease to in-situ cancer then to invasive cancer. A better understanding of its limitation in false positive detection of benign diseases as well as false negative detection of in-situ cancers is very important.

Progress towards specific aims: To date 69 patients who had palpable lesions which were occult on mammogram/ultrasound or who had probable benign or probable malignant findings (with BIRADS 3 or 4) were enrolled prospectively. Due to the high sensitivity, many patients who had BIRADS 3 findings but had a negative finding on MRI chose not to go for biopsy. Therefore, all cases with MRI done at our institution in the last 4 years were reviewed retrospectively to identify more histological proven benign cases. In-situ cancers had a lower vascular content compared to invasive cancers thus might not show strong contrast enhancements. A higher intraductal component in an invasive cancer was associated with a lower contrast enhancement and slower wash-in and wash-out. Some benign conditions such as atypical ductal hyperplasia, fibrocystic changes, fibroadenoma, papilloma, and sclerosing adenosis might exhibit strong contrast enhancements, also with rapid wash-in even wash-out. A better characterization of benign enhanced tissues is a major problem for MRI. In addition analysis of contrast enhancement kinetics, we also developed quantitative method to analyze morphological and texture features. The dynamic and morphological information may provide complementary information to improve diagnosis. For the angiogenic molecular marker study, to increase the specimen number we also identified cases from our tumor bank (including hyperplasia, atypical hyperplasia, and DCIS) for immunohistochemical studies.

Future directions and impact: With the sensitivity of MRI established, more and more preoperative MRI studies are performed for surgical planning, also MRI is considered as the optimal screening modality for young women with dense breast. A better characterization of the enhancing lesions noted on MRI is important to avoid unnecessary biopsy or overtreatment. On the other hand MRI may miss DCIS, especially low grade DCIS. More studies are needed to understand the role of MRI for breast imaging. The current study will be able to address these issues.


Final Report (2006)
Note: this grant was extended 1-yr to complete the funding and aims.

As the awareness of breast cancer and the availability of screening programs increases, diagnosis of breast cancer has shifted to earlier stages, such as ductal in-situ cancer (DCIS). Unfortunately, this has led to many “false positives” that result in unnecessary biopsies of non-cancerous breast tissue. The cancer detection problem is how to maintain a high sensitivity, but not at the expense of lowering specificity. The initial aim of this project was to study the role of angiogenesis in progression from pre-malignant diseases to DCIS, then to invasive cancers. We use dynamic contrast enhanced magnetic resonance imaging (MRI) and tumor angiogenic biomarkers for detection and interpretation of results.

Using CBCRP funding we expanded the scope of the project to compare the MR contrast enhancement kinetics in benign diseases, low-grade DCIS, intermediate to high grade DCIS, and also between different types of invasive cancers including invasive ductal cancer (IDC) and invasive lobular cancer (ILC). We measured the enhancement kinetics (i.e., rate at which MR contract agents enter and exit the tumor tissue). A rapid “wash-in” followed by “wash-out” is considered a marker of malignant tumors, whereas the kinetics showing a slow to medium wash-in, then a continuous enhancement pattern, is considered characteristic of a more benign breast abnormality. We found that DCIS is more likely to show the benign type enhancement kinetics, and compared to IDC, ILC is more likely to also show the benign type kinetics. The microvessel density (CD-31 staining = PECAM-1, platelet endothelial cell adhesion molecule) and mutant p53 staining both showed an increasing pattern with disease progression, from hyperplasia, to atypical hyperplasia, to low-grade, then high-grade DCIS.

Although these results were encouraging, substantial overlaps were found between these different disease types, thus no single MRI detection parameter could not serve as reliable diagnostic criteria. On the other hand, more and more evidence has suggested that morphological presentation may provide more useful information compared to enhancement kinetics in differentiating between benign and malignant lesions. We have also conducted studies to characterize and compare the morphological features of benign, in-situ, and invasive cancers. For example, based on breast MRI lexicon (i.e., a uniform standard language and reporting system for breast MRI) we found a high percentage of DCIS (42%) presented with multiple foci of enhancements scattering in a region. This represents a unique feature for diagnosis of DCIS, also seen in ILC, which is more likely to show non-mass-like enhancements compared to IDC. These findings inspired us to develop computer algorithms for analysis of quantitative morphology and texture features for tumor characterization. Since many parameters are analyzed, an artificial neural network has to be applied to select the best descriptors for differential diagnosis. We anticipate that this effort will form the basis for development of a computer-aided-diagnosis (CAD) system for breast MRI. In addition to differential diagnosis, the analysis techniques developed during this investigation can be applied to other research areas as well. For example, we are developing combined MRI and optical imaging system, aiming to improve the specificity using the additional information which the optical imaging provides. The characterization of tumor using biomarkers as well as morphological and kinetic properties can be applied to investigate the response of tumors to neoadjuvant chemotherapy.


Symposium Abstract (2005)
In this study we investigated the treatment efficacy of AC (doxorubicin and cyclophosphamide) followed by Taxane + Carboplatin ± Trastuzumab Regimen, assessed using longitudinal MRI monitoring. Also we investigated whether the morphological pattern of breast cancer is associated with response. At our institution neoadjuvant chemotherapy was offered mainly due to three reasons: 1) with locally advanced disease which was inoperable, 2) with clinically documented lymph node involvement, 3) down-staging disease for breast conservation surgery or for a better surgical outcome. Forty-one patients (32-75 yo, median 47) were included in this study. The patients received 2-4 cycles AC followed by Taxane regimen (TCa ± H, Taxane and Carboplatin, with Herceptin for Her-2/neu positive patients). Dose-dense AC treatment was given every 2 weeks, with growth factor support, unless intolerable. Patients received a baseline MRI before treatment, then several follow-ups during the course of treatment, then a final MRI before surgery.

Based on the 3D morphological pattern of the lesion, all lesions were classified into four types according to Esserman et al. [Ann Surg Oncol. 2001 8(6):549-559]. Type I- circumscribed mass, which has one primary mass with well-defined border; Type II- nodular pattern, which is consisted of connecting nodules; Type III- diffuse pattern, which demonstrates a regional enhancement without differentiable mass or nodule, and Type IV- septal pattern, which involves almost the entire breast. Depending on the post-treatment size compared to the baseline size determined in MRI, those showing less than 30% 1-dimensional reduction was defined as non-responders (NR), greater than 30% and less than 100% as partial responders (PR), and those who did not have detectable residual disease as complete responders (CR).

All 41 patients had completed the AC treatment. Of these, 23 patients had also completed TCa ± H regimen. Firstly, the response to AC was determined for each patient, then for cancers within each morphological type (I-IV), the percentage of responder was calculated. For I-Circumscribed mass 64% (14/22) were responders; 64% (7/11) for II-Nodular pattern; 67% (2/3) for III-Diffuse pattern, and 50% (2/4) for IV-Septal pattern. The morphological pattern of the breast cancer was not associated with response to AC. Fourteen patients received AC+TCa, and 8 of them reached a complete response (among them 7 were responders to AC). Eight patients with Her-2/neu positive cancer received AC+TCaH, and all of them were complete responders. For those patients who did not show respond to AC, 7/9 still responded to the following TCa ± H regimen.

The findings were: 1) The size determined on MRI was highly correlated with final pathological size. However, MRI may not detect residual DCIS or small invasive foci. 2) Cancer with different MRI morphological patterns did not show different responses to the dose-dense AC regimen, or the AC followed by TCa ± H regimen. 3) The response to AC was not associated with response to TCa with Herceptin, but it might enhance the response to TCa without H. 4) The addition of Herceptin in Her-2/neu positive patients enhanced the overall treatment outcome. As more patients elect to receive neoadjuvant chemotherapy, study of optimal treatment regimen has become increasingly important. The reported regimen protocol has achieved a high complete response rate. MRI monitoring may provide an accurate assessment of response for timely adjustment of treatments (e.g. stopping the AC at 2 cycles for non-responders), to avoid unnecessary toxicity without compromising the final outcome.

Acknowledgement: This work was supported in part by California CBCRP grant 9WB-0020 and NIH/NCI CA90437.


Symposium Abstract (2007)
Purpose: The currently favored working hypothesis of human breast cancer evolution suggests sequential stages from normal, hyperplasia, hyperplasia with atypia, in-situ cancer, and ultimately, invasive carcinoma. Fibrocystic change (FCC) is a commonly encountered benign diagnosis. The hyperplasic lesions, especially those with atypia, are associated with a higher risk of developing cancer, and hence are termed as pre-malignant lesions. Ductal carcinoma in situ (DCIS) is a pre-invasive cancer. This study compared MRI features of benign lesions (FCC), pre-malignant hyperplasic lesions, and pre-invasive (DCIS) lesions.

Material and Methods: A retrospective analysis of all breast MRI performed at 1.5T from 2002 to 2005 identified 44 pathological-proven FCC, 16 pre-malignant cases (Hyperplasia, ADH, ALH, LCIS) and 34 DCIS. Their MR morphological features and enhancement kinetics were compared. Feature descriptor was analyzed based on BI-RADS MRI Lexicon. The morphologic criteria included mass type lesion (focus/foci: < 5mm, mass: > 5 mm), and non-mass type lesion (focal area, linear, ductal, segmental, regional, multiple regions, diffuse enhancement). The internal enhancement patterns included punctuate, clumped, dendritic, heterogeneous, and homogeneous. The evaluation of enhancement kinetic curve was based on the initial phase (within the first 2 minutes or when the curve starts to change) as fast, medium, and slow; and late phase (after 2 minutes or after the change) as persistent, plateau, and washout. When the kinetic curve showed initial fast/medium up-slope followed by washout or plateau, it was considered as suspicious of malignancy.

Results: While the pre-malignant and pre-invasive groups had a similar percentage presenting as mass type lesion (50% vs. 41%), FCC is less likely, only 29%. Non-mass type lesion appeared in similar frequency, 50%, 56% and 57% for pre-malignant, pre-invasive lesions, and FCC, respectively. One DCIS (3%) and 6 FCC (14%) were not enhanced. The subtypes of non-mass lesions (diffuse, linear, segmental) between these 3 groups were not significantly different. The analysis of internal enhancement patterns showed that DCIS was more likely to present clumped pattern (10/19, 53%) and less likely punctate pattern, (2/19, 11%); while pre-malignant lesions showed equal frequency presenting these two patterns (37.5%). FCC was more likely to present the heterogeneous enhancement (11/25, 44%). The malignant type enhancement kinetics was only seen in 46% FCC and 46% pre-malignant lesions, whereas it was seen more frequently in 75% DCIS.

Conclusion: There are no MRI features that can clearly distinguish between the benign FCC, hyperplasic lesions and DCIS. Malignant type enhancement kinetics was a good indicator for DCIS, but nearly half of FCC and hyperplasic lesions also showed this kinetic feature. FCC was more likely to show heterogeneous pattern, DCIS was more likely to present clumped pattern; and the pre-malignant lesions were more likely to show punctuate pattern. Correct diagnosis of FCC may avoid unnecessary biopsy; early diagnosis of DCIS will improve patient’s prognosis; also accurate diagnosis of hyperplasia with or without atypia will help determine whether the patient should receive additional excision biopsy. Understanding the differences in their MRI features will aid in a better diagnosis to distinguish between them, and to impact on their management.


Symposium Abstract (2007)
Breast cancer arising from milk ducts is thought to progress through stages from normal ducts to increased but non-malignant growth (ductal hyperplasia, atypical ductal hyperplasia [ADH]), increasing grades of tumor confined to the ducts (ductal carcinoma in situ [DCIS]), and finally to invasive cancer. The mechanisms responsible for this progression are incompletely understood. Angiogenesis, the proliferation of new blood vessels in a growing tumor, is necessary for the mass to obtain oxygenation and nutrients as it grows. The point at which angiogenesis begins in tumor progression is known as the "angiogenic switch." Angiogenesis in invasive breast cancer is well documented, but little is known of the role of angiogenesis in pre-malignant ductal disease, or when the angiogenic switch occurs during the evolution of breast malignancy. It is important to study angiogenesis because anti-cancer drugs that target angiogenesis are available, and because the presence of angiogenesis allows the visualization of lesions by magnetic resonance imaging (MRI). The goal of this study was to document angiogenesis during the progression of ductal diseases by special staining and microscopic evaluation of the markers of new blood vessel formation, CD31 and CD105.

A gene often mutated in breast cancer, p53, may be involved in the process of angiogenesis and was measured in breast duct cells. Other angiogenesis related proteins, thromospondin 1 (TSP1) and vascular endothelial growth factor (VEGF), were also measured. The number of new blood vessels in breast samples from 81 patients was determined, and the staining intensity of 0 to 3+ was evaluated for the other proteins. The vessel counts for CD31 revealed 0.8 +/- 2.2 vessels around normal ducts, and a significant increase in density to 18.1 +/- 5.6 vessels for the progression to increased but non malignant growth (p < 0.0001). The mean vessel density progressively increased to 22.4 +/- 9.0 for ADH, 25.4 +/- 18.1 for low grade DCIS, 27.0 +/- 11.6 for intermediate grade DCIS and 28.9 +/1 8.9 for high grade DCIS. This progressive increase in mean density was significant by the Jonckheere-Terpstra test for ordered alternatives (p=0.0001). A similar increase in vessel density was noted in tissues stained with CD105. The proteins involved in the regulation of angiogenesis, including p53, TSP1 and VEGF showed greater staining of hyperplasia than normal ducts, but only slight increases of these markers as the lesions progressed. These data provide evidence that new blood vessel growth occurs early in the sequence of breast cancer progression, and may be related to the increase in angiogenic regulatory proteins also observed during this progression. This work is important because it offers insights into the role of angiogenesis in breast cancer progression, and it provides an explanation for why these lesions are sometimes detected by MRI.

Pharmacokinetic Parameters Analyzed from MR Contrast Enhancement Kinetics of Multiple Malignant and Benign Breast Lesions Detected in the Same Patients.
Periodical:Technology in Cancer Research and Treatment
Index Medicus:
Authors: Su MY, Yu HJ, Carpenter PM, McLaren CE, Nalcioglu O.
Yr: 2005 Vol: 3 Nbr: Abs: Pg:255-64

Comparison of choline and pharmacokinetic parameters in breast cancer measured by MR spectroscopic imaging and dynamic contrast enhanced MRI.
Periodical:Technology in Cancer Research and Treatment
Index Medicus:
Authors: Su MY, Baik HM, Yu HJ, Chen JH, Mehta RS, Nalcioglu O.
Yr: 2006 Vol: 5 Nbr: 4 Abs: Pg:401-10

Quantification of choline-containing compounds in malignant breast tumors by 1H MR spectroscopy using water as an internal reference at 1.5 T.
Periodical:Magnetic Resonance Materials in Physics, Biology and Medicine
Index Medicus: MAGMA
Authors: Baik HM, Su MY, Yu H, Mehta R, Nalcioglu O.
Yr: 2006 Vol: 19 Nbr: 2 Abs: Pg:96-104

Coregistration of dynamic contrast enhanced MRI and broadband diffuse optical spectroscopy for characterizing breast cancer.
Periodical:Technology in Cancer Research and Treatment
Index Medicus:
Authors: Hsiang D, Shah N, Yu H, Su MY, et al and Tromberg BJ
Yr: 2005 Vol: 4 Nbr: 5 Abs: Pg:549-58