This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schlechter, B. L. Articles by Rosenberg, C. L. Search for Related Content PubMed PubMed Citation Articles by Schlechter, B. L. Articles by Rosenberg, C. L. Social Bookmarking                            What's this?

Quantitative DNA Fingerprinting May Distinguish New Primary Breast Cancer From Disease Recurrence

From the Departments of Medicine, Neurology, and Pathology and Laboratory Medicine, Boston University School of Medicine; Department of Biostatistics, Boston University School of Public Health; and Boston Medical Center, Boston, MA

Address reprint requests to Carol L. Rosenberg, MD, Department of Medicine, Boston University Medical Center, 650 Albany St, EBRC-4, Boston, MA 02118; e-mail: crosenbe{at}bu.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

 
PURPOSE: Approximately 10% of women with breast cancer develop a second breast tumor, either a new primary or a recurrence. Differentiating between these entities using standard clinical and pathologic criteria remains challenging. Ambiguous cases arise, and misclassifications may occur. We investigated whether quantitative DNA fingerprinting, based on allele imbalance (AI) or loss of heterozygosity (LOH), could evaluate clonality and distinguish second primary breast cancer from recurrence.

METHODS: We developed a scoring system based on the AI/LOH fingerprints of 20 independent breast tumors and generated a decision rule to classify any breast tumor pair as related or unrelated. We validated this approach on eight related tumors (cancers and synchronous positive lymph nodes). Finally, we analyzed paired tumors from 13 women (bilateral cancers, primary tumors and contralateral positive axillary lymph nodes, or two ipsilateral tumors). Each pair's genetic classification was compared with their clinical diagnosis and outcome.

RESULTS: Each independent cancer had a unique fingerprint. Every tumor pair's relationship was quantifiable. Six of eight related tumor pairs were genetically classified correctly, two were indeterminate, and none were misclassified. Among the 13 women with two cancers, four of five clinically indeterminate pairs could be classified genetically. In three of 13 women, the pair's classification contradicted the clinical diagnosis. These women had bilateral cancers genetically classified as related and disease progression. This challenges the paradigm that bilateral cancers represent independent tumors. Overall, women with tumors genetically classified as related had poorer outcomes.

CONCLUSION: Quantitative AI/LOH fingerprinting is a potentially valuable tool to improve diagnosis and optimize treatment for the growing number of second breast malignancies.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

 
Advancing age and prior breast cancer diagnosis are two of the most important risk factors for development of a new breast cancer.^1–5 As the population ages and breast cancer treatments improve, a growing number of women will be at risk for the development of second breast cancers. These cancers either represent new primary breast cancers or disease recurrence. Presently, there are no reliable morphologic or clinical criteria to distinguish second primary breast cancer from recurrent disease. Bilateral cancers are generally considered to represent independent primaries, whereas ipsilateral disease is usually regarded as tumor recurrence. The presence of carcinoma-in-situ, different histologic type, or better differentiation of the second tumor may help in identifying independent tumors.⁶ However, cases are often ambiguous, and misclassification may occur.

Because new primaries are associated with improved survival^7,8 and metastatic disease is rarely curable,⁵ differentiating between these entities is imperative for staging, prognosis, and therapeutic decisions and may also contribute to our understanding of breast cancer biology.⁹ The goals of the current study are to develop an easily applicable, quantitative genetic approach to evaluate the clonal relationships between two invasive breast cancers arising in a single individual and to improve differentiation of new primary breast cancers from disease recurrence.

Molecular techniques used previously to differentiate second primary breast cancer from disease recurrence include evaluation of clonality using X-inactivation^10,11 and examination of TP53 mutations^10,12 or protein expression.¹³ However, the clinical utility of these approaches is limited by a relatively low degree of informativity and nonspecific results. Alternatively, use of transcriptional profiling has suggested that multicentric tumors share patterns of gene expression.¹⁴ This approach is promising but requires fresh tissue, which is not available in the vast majority of second breast cancer cases, and assumes that RNA or protein expression profiles remain constant.

An optimal assay for differentiating second primary breast cancer from disease recurrence should use archival tissue (because many second breast tumors are found months or years after initial diagnosis) and should evaluate abnormalities common to all breast cancers, yet somehow distinct in each primary tumor. Evaluation of clonal DNA abnormalities would be advantageous because DNA is generally well preserved in fixed, archival specimens and may be less mutable than RNA or protein profiles. Furthermore, histopathology can be heterogeneous within a case. Numerous DNA abnormalities characterize breast cancers, but no single, or signature, abnormality defines the disease.¹⁵ However, genomic imbalances are found in essentially all breast tumors and analysis of allele imbalance (AI), commonly described as loss of heterozygosity (LOH), is a well-established technique.^16–18 AI/LOH reflects abnormal allele ratios at a specific chromosome locus, resulting either from amplification or loss of one of the two existing heterozygous alleles. The mechanisms leading to AI/LOH are uncertain, but AI/LOH is thought to result in dysregulation of oncogene(s) or tumor suppressor gene(s) near the sites of imbalance.¹⁹ Presumably, this provides a growth advantage to the cell in which it occurs and contributes to tumorigenesis. AI/LOH is generally evaluated using polymerase chain reaction (PCR) of heterozygous microsatellite markers that reveal a substantial change in signal intensity of one of the two heterozygous alleles. It can be evaluated in archival tissue, from small numbers of cells, and although it characterizes nearly all breast cancers,^16,17 the specific alleles involved in individual tumors appear highly distinctive.^16,20,21 When a tumor recurs or metastasizes, it generally retains the pattern of AI/LOH in the parent tumor,^19,22,23 although additional sites of AI/LOH may develop.^24,25 All of these features render a tumor's AI/LOH pattern at a large panel of microsatellite markers, referred to as its AI/LOH fingerprint, a good candidate assay to differentiate second primary breast tumors from recurrent disease.

A limited number of studies have used this technique to investigate clonality in melanoma²⁶ and in ovarian,²⁷ liver,²⁸ and lung²⁹ cancer. Several reports examining invasive bilateral breast cancers find relatively few concordant sites of AI/LOH and suggest that most bilateral cancers represent independent tumors.^21,30–32 However, these studies do not analyze allelic AI/LOH, do not examine AI/LOH in control groups of unrelated and related tumors, do not present clinical information to correlate with the genetic results, and do not provide a statistical analysis of the data.

Taken together, these findings lead us to hypothesize that a comprehensive, quantitative approach to AI/LOH fingerprinting could be used to evaluate the clonal relationship of a patient's breast tumor pair and to distinguish a new primary tumor from disease recurrence. We speculate that a new primary breast tumor should have a distinct AI/LOH fingerprint, whereas recurrent or metastatic disease will maintain most of the original tumor's fingerprint (although additional abnormalities may develop). Furthermore, this relationship should be quantifiable using a statistical approach. To test this hypothesis, we analyzed the AI/LOH fingerprints of 20 unrelated tumors (ie, from separate individuals) and developed a scoring system to quantitate the difference between tumors. The range of scores from these tumor pairs was used to approximate the scores of new primary cancers. We validated this approach by analyzing eight related tumor pairs (ie, a primary cancer and synchronous, positive, cancer-containing lymph nodes [LNs]). Finally, we performed a pilot study analyzing 13 breast tumor pairs from single individuals and compared the tumors' genetic classification to the clinical diagnosis and outcome.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

 
Breast Cancer Cases
After obtaining institutional review board approval, the archived tissue blocks of 20 independent breast cancer cases were evaluated. Each case consisted of a primary breast cancer and either a positive ipsilateral LN, a subsequent local recurrence, a contralateral breast tumor, or a positive contralateral axillary LN without an evident breast tumor on that side. Identifiers were removed, a study identification number was assigned, and the following limited clinical information was collected: dates of diagnoses, pathologic description and immunohistochemical staining of the tumors, treatment and interval follow-up, and clinical diagnosis of the second cancer as either a new primary tumor or recurrent disease.

Clinical Diagnosis
Because no algorithm or definitive criteria exist to distinguish a new primary cancer from metastasis or disease recurrence, we considered strong indicators of a new primary to be negative metastatic work-up, the presence of carcinoma-in-situ in the second tumor (or both tumors, when synchronous), and node negativity in both tumors. Different hormone receptor status, changes in grade or histology, and increased interval between tumors were considered suggestive factors indicating a new primary cancer. Using these criteria, a clinical diagnosis was made by the treating oncologists at the time of the second cancer's detection. An independent breast oncologist also reviewed each case's clinical information. If disagreement or uncertainty existed, the case was labeled indeterminate (see Appendix).

Tissue and DNA Acquisition
A pathologist (A.d.l.M.) identified control tissue (uninvolved LNs, normal breast tissue, and skin) and tumor in each case. Laser microdissection was performed as previously described by our laboratory (Pixcell; Arcturus Engineering, Mountain View, CA; Fig 1). ^20,33 Because of intratumoral heterogeneity, we microdissected three to four samples from each tumor, each containing more than 1,000 cells. DNA was extracted using Instagene Matrix (Bio-Rad, Hercules, CA) as previously described by our laboratory.^20,33

View larger version (60K): [in this window] [in a new window]   Fig 1. Laser capture microdissection. (A) Invasive ductal carcinoma, hematoxylin and eosin–stained section; (B) Invasive ductal carcinoma, high magnification; (C) Unstained section adjacent to carcinoma in A; (D) Same section, after microdissection; (E) Tissue on cap. Magnification for B, x200; magnification for all other sections, x40. Scale bar, 100 µm.

PCR and Electrophoresis
Seven multiplexed PCRs were performed using 5 to 15 µL of the 200-µL DNA solution as a template in a 50-µL reaction, 30 to 35 cycles of amplification, incorporation of ³²P-deoxycytidine triphosphate, and annealing temperatures between 55°C and 60°C. One third of the amplified products was electrophoresed through 7% acrylamide-urea-formamide denaturing gels that were then exposed to autoradiography film.

Determination of AI/LOH
The normal pattern at each microsatellite marker in each individual was defined as that seen in uninvolved LNs or skin. AI/LOH was assessed visually by two observers (B.L.S. and C.L.R.). Most AI/LOH were unambiguous, but when uncertainty existed, the films were scanned by a laser densitometer (Molecular Dynamics, Sunnyvale, CA), and allele ratios were calculated. AI/LOH was defined at heterozygous loci as an imbalance of allele intensities greater than 33% [ie, when (n1)(t2)/(n2)(t1) > 1.50 or < 0.67; where n1 = normal samples' larger allele, n2 = normal samples' smaller allele, t1 = test sample's larger allele, and t2 = test sample's smaller allele]. This degree of reproducible AI indicates that a substantial proportion of the cells within a sample contains the same DNA abnormality and likely represents the presence of a clonal population.^34,35 Abnormal results were demonstrated at least twice with equivalent results.

Scoring and Statistical Analysis
Each sample's AI/LOH fingerprint was defined using the 29-marker panel. To quantify the difference in AI/LOH fingerprints between a pair of tumors, we assigned each informative site a score as follows: –1.0 (ie, both tumors share AI/LOH of the same allele, representing similarity), 0.0 (ie, both tumors lack AI/LOH), +1.0 (ie, one tumor has AI/LOH, but the other does not, representing difference), and +2.0 (ie, both tumors have AI/LOH but of opposite alleles, representing two distinct events). Uninformative sites could not be scored. To minimize bias when comparing a pair of tumors, each tumor sample was treated independently and compared with every other sample in the other tumor. (For example, if tumor A had three samples and tumor B had four samples, then 12 comparisons were made.) To obtain an overall measure of difference between any two samples, we first averaged the scores of informative markers on the same chromosomal arm to obtain a score for each arm. We then averaged the scores from all arms. Because different numbers of markers were typed on different chromosomal arms and markers on the same arm may not have independent AI/LOH, averaging by arms first could avoid bias caused by assigning excessive weights to scores from chromosome arms with more markers than others. To obtain the final score for a tumor pair, we averaged the scores of that pair's samples. A Wilcoxon rank sum test was performed to evaluate the difference in score distribution between related and unrelated tumors pairs.

Tumor Comparisons
Using the scoring system, three different comparisons between tumors were made. First, each of 20 independent primary tumors was compared with the other 19 primaries (if a case had > one tumor, her first tumor or, in synchronous cases, her right-sided tumor was used). Because informative sites vary between individuals, comparisons between tumors with less than 10 informative markers in common were excluded to reduce error. Second, unilateral primary tumors were compared with their synchronous, ipsilateral, positive axillary LNs. The primary tumors were included in the independent tumor group, but the axillary LNs were not. Third, two tumors from single cases were compared with each other. The initial or right-sided tumor was included in the independent tumor group, but the other tumor was not. Thus, each set of comparisons is completely separate from the others, and the analyses are independent.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

 
From 20 cases, 124 tumor and 71 normal samples were microdissected, fingerprinted, and scored. On average, three samples per tumor were analyzed (range, one to six samples), and 16 of 29 markers were informative. Representative examples of AI/LOH and an illustration of scoring are shown in Figure 2. The samples were divided into several categories, and we analyzed the data within each.

View larger version (72K): [in this window] [in a new window]   Fig 2. Comparison between tumor A and tumor B. Representative examples of scoring tumor pairs for quantitative allele imbalance (AI)/loss of heterozygosity (LOH) fingerprinting. (A) No AI/LOH in either tumor; (B) AI/LOH of both tumors' upper alleles; (C) no AI/LOH in tumor A, but AI/LOH of tumor B's upper allele. (D) AI/LOH of tumor A's lower allele and tumor B's upper allele. (E) Both tumors are homozygous and cannot be scored.

View larger version (30K): [in this window] [in a new window]   Fig 3. Distribution of final scores of tumor pairs. (A) Final scores of 112 unrelated tumor pairs (gray bars). Cumulative percent line illustrates the decision rule; 99% of unrelated tumor pairs' final score was > 0.14%, and 90% was > 0.30; (B) final scores of eight related tumor pairs (black bars); (C) final scores of tumor pairs from 13 women with two cancers (black diamonds). Indet, indeterminate.

Definition of Decision Rule
We constructed a decision rule based on the distribution of the final scores of the independent tumors. On the basis of this distribution, we divided the range of scores into three intervals, each associated with a genetic classification. If a tumor pair's final score was more than 0.30, it was classified as unrelated, with a 90% chance of correct classification. If a tumor pair's final score fell between 0.30 and 0.14, it was classified as indeterminate, with a 9% chance of lacking evidence for definite classification. If a tumor pair's final score was less than 0.14, it was classified as related, with a 1% chance of misclassification.

Testing of Decision Rule
We tested how well the decision rule, generated from the 20 independent tumors, applied to eight pairs of related tumors (ie, a unilateral primary tumor and a synchronous, ipsilateral, positive axillary LN). The comparisons in these related tumor pairs are completely separate from the comparisons in independent tumors. The related tumor pairs had a significantly lower median score than the unrelated tumors (P < .00000001, Wilcoxon rank sum test). In six related pairs, the second tumor was classified correctly as recurrent or metastatic disease, and in two pairs, the genetic classification was indeterminate. No related tumor pair was misclassified.

Comparison of Two Breast Tumors Within One Subject
In a pilot study, the tumors of 13 subjects with two breast cancers were analyzed. Seven cases were synchronous (< 1 year between diagnoses), and six were metachronous ( 1 year between diagnoses). Nine of 13 subjects had bilateral breast cancers, three of 13 had a single breast cancer and a contralateral, positive axillary LN, and one of 13 had an ipsilateral tumor after substandard therapy of her first cancer (she refused radiation after lumpectomy). We applied the decision rule to the AI/LOH fingerprints of each subject's tumor pair.

Genetic classification. In six (46%) of 13 cases, the genetic score suggested two primary cancers; in four (31%) of 13 cases, the score suggested recurrent or metastatic disease; and in three (23%) of 13 cases, the scores fell into the indeterminate interval, limiting confidence in the classification of these pairs (Fig 3C and Table 1). Intratumoral heterogeneity in these 13 tumor pairs was similar to what was seen in the 20 independent tumors; 19 (73%) of 26 tumors (from 12 of 13 pairs) contained a mean of two of 17 differences between samples within a tumor. Two tumors seemed more heterogeneous. In subject 7002, one right breast sample contained a core set of four AI/LOH, which was also seen in the other right- and left-sided tumor samples, plus four additional sites of AI/LOH. In subject 7012, one right breast sample had a fingerprint that was similar to the left-sided tumor samples, whereas the other right-sided samples had a different fingerprint. Because the final genetic score averages the scores of all samples, a single sample's impact is diluted.

Genetic classification versus clinical diagnosis. As shown in Table 1, among the eight cases in which the clinical diagnosis of the second tumor seemed clear, three cases' genetic classification agreed with the diagnosis (subjects 7001, 7005, and 7034); two cases' genetic scores were in the indeterminate interval, so classification was uncertain (subjects 7071 and 7015); and three cases' genetic classification contradicted the clinical diagnosis (subjects 7002, 7007, and 7014). Among the five cases in which the clinical diagnosis of the second tumor was indeterminate, the genetic scores of four cases suggested a classification (either of independent tumor, subjects 7011, 7013, and 7065; or recurrent disease, subject 7003), and the score of one case was indeterminate (subject 7012).

Bilateral breast cancers. Bilateral breast cancers are generally considered to represent independent primaries because breast cancer is believed to metastasize uncommonly to the contralateral breast.³⁸ Consistent with this, seven (78%) of nine bilateral cases were diagnosed as independent primary cancers, and two (22%) of nine cases were clinically indeterminate (subjects 7012 and 7065; see Genetic classification versus clinical diagnosis). Surprisingly, the genetic classification of three (33%) of nine bilateral cases suggested recurrent or metastatic disease (subjects 7002, 7007, and 7014). The genetic classification of the remaining six (67%) of nine cases either suggested new primary cancer (subjects 7001, 7005, 7034, and 7065) or were indeterminate (subjects 7012 and 7071).

Contralateral positive axillary LNs. There were three cases consisting of a primary breast cancer with a contralateral positive axillary LN. In two cases, the genetic score indicated two unrelated cancers (subjects 7011 and 7013). Presumably, the positive LN arose from an occult contralateral primary breast cancer. In one case, the score classified the LN as a metastasis, consistent with a history of multiple chest wall recurrences progressively closer to the contralateral axilla (subject 7003).

Ipsilateral disease. The single case of a second tumor developing in the same breast as the original cancer (subject 7015) was clinically diagnosed as a local recurrence. This score fell into the indeterminate interval.

Clinical Outcome
From the 13 test cases analyzed, three had discordance between clinical diagnosis and genetic classification. All cases had clinical diagnoses of new primary but genetic classifications of disease recurrence, and all showed signs of disease progression. Because these three were bilateral cases, we considered the outcome of all nine bilateral cases. According to clinical diagnosis, in all nine cases the second tumor was either a new primary or indeterminate. With an average follow-up of 53 months, five of nine cases had died or had evidence of disease progression, and four of nine had no evident disease. According to genetic classification, three of nine bilateral cases represented recurrent or metastatic disease. With an average follow-up of 50 months, all three cases had died or shown signs of disease progression. The remaining six of nine bilateral cases were genetically classified as new primaries or indeterminate. With an average follow-up of 54 months, these cases had a better outcome; two cases had died, and four had no evident disease. These results suggest that our genetic approach may be better in predicting disease outcome than the clinical diagnosis; however, the sample size was too small to draw a statistically significant conclusion.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

 
We investigated whether AI/LOH fingerprinting could be used to evaluate breast cancer clonality and, therefore, be useful as a quantitative genetic test to differentiate new primary breast tumors from disease recurrence. Differentiating between these two entities remains a clinical challenge with important implications. We compared the AI/LOH fingerprints of 20 independent cancers and generated a decision rule to determine whether any pair of tumors represents two primary cancers or recurrent disease. To validate AI/LOH fingerprinting and the decision rule's accuracy for related tumor pairs, we analyzed eight related pairs (a primary breast cancer and ipsilateral, positive axillary LNs). The difference between the median scores of the unrelated and related populations was highly significant (P < .00000001). Six related tumor pairs were classified correctly, two were indeterminate, and none were misclassified. Finally, in a pilot study analyzing 13 tumor pairs from women with two breast cancers, we contrasted each pair's genetic classification to its clinical diagnosis and outcome. Tumors classified as related were associated with poorer outcome. Quantitative AI/LOH fingerprinting may be a potentially valuable tool to improve diagnosis and optimize treatment of the growing number of women with second breast malignancies.

Important technical advantages of quantitative AI/LOH fingerprinting include reliable and effective use in archival specimens, potential modification of the microsatellite panel as information develops about oncogenes or tumor suppressor genes important to breast tumorigenesis, and feasibility of the approach, once automated, for routine use. This approach seems to be potentially sensitive and specific because we found that all breast tumors were informative, all were characterized by an individual fingerprint (confirming previous reports^16,21), and all relationships between tumor pairs were quantifiable. The final scores incorporated not only AI/LOH at individual alleles, but also average AI/LOH on each chromosome arm, because AI/LOH at multiple markers on a single arm may not reflect independent events. This approach will minimize bias caused by variable numbers of informative markers per arm.

Potential concerns do exist, however, about genetic misclassification of two tumors from a single individual. Independent tumors from a single subject could be more similar than independent tumors from different subjects. This may be a consequence of environmental or host factors, leading to a predisposition to activate particular pathways of tumorigenesis.^39,40 The result could be similar genetic abnormalities (or similar outcomes) in independent tumors. This possibility is exemplified in a recent study showing that the breast cancers of two identical triplets with germline BRCA1 mutations had similar loci undergoing AI/LOH.⁴¹ However, 46 of the markers reported had imbalances at different alleles, whereas only 21 had imbalances at the same allele. This suggests that, although chromosomal sites of AI/LOH may recur, the particular allele lost differs sufficiently often to distinguish between tumors and highlights the specificity provided by considering alleles, as well as loci, lost. Developing a more refined distribution of related tumor pair scores, by testing additional tumor pairs, should clarify this issue.

Another concern about misclassification would be whether, over time, a tumor could undergo clonal evolution and acquire additional AIs. If many additional imbalances occurred, the genetic score could be high, and the two tumors could be misclassified as unrelated. This could explain why the scores of two related tumor pairs, consisting of a primary cancer and its ipsilateral nodal metastasis, were indeterminate (it is notable, however, that none of the related tumor pairs were genetically misclassified). Another explanation could be incomplete sampling of either the tumor or node. Excessive contamination with normal cells, masking AI/LOH, is unlikely because AIs were seen in both tumors and nodes. Intratumoral heterogeneity could potentially confuse the findings; nevertheless, our data suggest that it is unlikely to account for discordance between independent tumors. To address these issues, future studies will consider the specific loci undergoing AI/LOH (in addition to the number of AI/LOH) because a late-occurring recurrence or metastasis should contain a core complement of the original tumor's abnormalities.^22–25 Serial sampling of an individual tumor will determine whether clonal evolution will present a substantial problem.

Despite these concerns, we believe that evaluation of tumor pairs from subjects with two breast cancers suggests that quantitative AI/LOH fingerprinting may provide information unavailable from standard clinical and pathologic assessments. In five tumor pairs, the clinical diagnosis of the second tumor was indeterminate; but in four (80%) of five cases, a genetic classification was possible. Thus, AI/LOH fingerprinting might assist in diagnosing clinically ambiguous cases. In eight tumor pairs, the clinical diagnosis of the second tumor seemed certain; but in three (38%) of eight cases, the genetic classification contradicted the clinical diagnosis. All three cases' genetic classification suggested recurrence or metastasis, and all developed progressive disease. Because these cases were bilateral breast cancers, they challenge the existing paradigm that bilateral cancers are independent primary tumors.

The discrepancies between clinical diagnosis and genetic classification raise the possibility that a proportion of second breast cancers are currently misclassified. Although these findings are surprising, it is possible that a proportion of bilateral cancer cases may actually represent metastatic disease. Consistent with this, several studies find that the genetic changes in a fraction (13% to 18%) of bilateral breast cancers could represent metastases.^21,32,42 Conventionally, this is considered unlikely, but we posit that the genetic data may be detecting an unappreciated relationship and that enough clinical uncertainty and histologic heterogeneity exists that standard criteria may not be sufficiently sensitive to distinguish a second primary breast cancer from disease recurrence. Alternatively, it is possible that AI/LOH fingerprints sufficiently similar to classify a pair of tumors as related actually reflects independent tumors using the same pathway of tumorigenesis. In addition, a proportion of ipsilateral second tumors may also be misclassified. Some,^7,43 although not all,⁴⁴ studies conclude that a proportion of second ipsilateral cancers may represent new primaries and not disease recurrence as is usually assumed. Analysis of additional cases should clarify this issue, which will become increasingly important as more lumpectomies are performed.

In conclusion, our results suggest that quantitative AI/LOH fingerprinting is a promising approach to improve diagnosis and, consequently, to predict prognosis and optimize treatment among the growing number of women developing second breast cancers. Its use of a genetic abnormality found in nearly all breast cancers and assessable in archival specimens makes it particularly attractive. In addition to establishing quantitative AI/LOH fingerprinting as a potentially valuable tool worthy of additional investigation, our findings also raise the possibility that a proportion of second breast cancers may be misclassified using standard clinical and pathologic criteria. After further testing and validation, quantitative AI/LOH fingerprinting could enhance currently available approaches to distinguish a new primary cancer from recurrent disease and may improve treatment of second breast malignancies.

	Appendix

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

 
The appendix is included in the full-text version of this article, available on-line at http://www.jco.org. It is not included in the PDF (via Adobe® Acrobat Reader®) version.

	Authors' Disclosures of Potential Conflicts of Interest

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

	Acknowledgment

 
We thank Drs L. Kachnic and I.N. Rosenberg for their thoughtful comments on the manuscript.

	NOTES

 
Supported by the Department of Defense Breast Cancer Research Program grant No. DAMD17-99-1-9573 and the Susan G. Komen Foundation grant No. BCTR02-1535.

Presented in part at the 38th Annual Meeting of the American Society of Clinical Oncology, Orlando, FL, May 18–21, 2002.

Authors' disclosures of potential conflicts of interest are found at the end of this article.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION Appendix Authors' Disclosures of... REFERENCES

 
1. Rosen PP, Groshen S, Kinne DW, et al: Contralateral breast carcinoma: An assessment of risk and prognosis in stage I (T1N0M0) and stage II (T1N1M0) patients with 20-year follow-up. Surgery 106:904–910, 1989[Medline]

2. Gogas J, Markopoulos C, Skandalakis P, et al: Bilateral breast cancer. Am Surg 59:733–735, 1993[Medline]

3. Anastasiadis PG, Liberis VA, Koutlaki NG, et al: Incidence and detection of contralateral breast cancer. Breast J 6:178–182, 2000[CrossRef][Medline]

4. Heron DE, Komarnicky LT, Hyslop T, et al: Bilateral breast carcinoma: Risk factors and outcomes for patients with synchronous and metachronous disease. Cancer 88:2739–2750, 2000[CrossRef][Medline]

5. Ries LAG, Eisner MP, Kosary CL, et al: SEER Cancer Statistics Review, 1973–1999. Bethesda, MD, National Cancer Institute, 2002

6. Chaudary MA, Millis RR, Hoskins EO, et al: Bilateral primary breast cancer: A prospective study of disease incidence. Br J Surg 71:711–714, 1984[Medline]

7. Smith TE, Lee D, Turner BC, et al: True recurrence vs. new primary ipsilateral breast tumor relapse: An analysis of clinical and pathologic differences and their implications in natural history, prognoses, and therapeutic management. Int J Radiat Oncol Biol Phys 48:1281–1289, 2000[CrossRef][Medline]

8. Huang E, Buchholz TA, Meric F, et al: Classifying local disease recurrences after breast conservation therapy based on location and histology: New primary tumors have more favorable outcomes than true local disease recurrences. Cancer 95:2059–2067, 2002[CrossRef][Medline]

9. van der Sijp JR, van Meerbeeck JP, Maat AP, et al: Determination of the molecular relationship between multiple tumors within one patient is of clinical importance. J Clin Oncol 20:1105–1114, 2002[Abstract/Free Full Text]

10. Noguchi S, Motomura K, Inaji H, et al: Differentiation of primary and secondary breast cancer with clonal analysis. Surgery 115:458–462, 1994[Medline]

11. Shibata A, Tsai YC, Press MF, et al: Clonal analysis of bilateral breast cancer. Clin Cancer Res 2:743–748, 1996[Abstract]

12. Janschek E, Kandioler-Eckersberger D, Ludwig C, et al: Contralateral breast cancer: Molecular differentiation between metastasis and second primary cancer. Breast Cancer Res Treat 67:1–8, 2001[CrossRef][Medline]

13. Ackerman J, Baunoch DA, Gimotty P, et al: The role of p53 mutations in bilateral breast carcinoma. Mod Pathol 8:244–248, 1995[Medline]

14. Unger MA, Rishi M, Clemmer VB, et al: Characterization of adjacent breast tumors using oligonucleotide microarrays. Breast Cancer Res 3:336–341, 2001[CrossRef][Medline]

15. Dickson RB, Lippman ME: Cancer of the breast, in Rosenberg SA (ed): Cancer Principles and Practice of Oncology (ed 6). Philadelphia, PA, Lippincott, Williams & Wilkins, 2001, pp 1633–1647

16. Kerangueven F, Noguchi T, Coulier F, et al: Genome-wide search for loss of heterozygosity shows extensive genetic diversity of human breast carcinomas. Cancer Res 57:5469–5474, 1997[Abstract/Free Full Text]

17. Osborne RJ, Hamshere MG: A genome-wide map showing common regions of loss of heterozygosity/allelic imbalance in breast cancer. Cancer Res 60:3706–3712, 2000[Abstract/Free Full Text]

18. Shen CY, Yu JC, Lo YL, et al: Genome-wide search for loss of heterozygosity using laser capture microdissected tissue of breast carcinoma: An implication for mutator phenotype and breast cancer pathogenesis. Cancer Res 60:3884–3892, 2000[Abstract/Free Full Text]

19. Chen LC, Kurisu W, Ljung BM, et al: Heterogeneity for allelic loss in human breast cancer. J Natl Cancer Inst 84:506–510, 1992[Abstract/Free Full Text]

20. Larson PS, De Las Morenas A, Bennett SR, et al: Loss of heterozygosity or allele imbalance in histologically normal breast epithelium is distinct from loss of heterozygosity or allele imbalance in co-existing carcinomas. Am J Pathol 161:283–290, 2002[Abstract/Free Full Text]

21. Imyanitov EN, Suspitsin EN, Grigoriev MY, et al: Concordance of allelic imbalance profiles in synchronous and metachronous bilateral breast carcinomas. Int J Cancer 100:557–564, 2002[CrossRef][Medline]

22. Lininger RA, Fujii H, Man YG, et al: Comparison of loss heterozygosity in primary and recurrent ductal carcinoma in situ of the breast. Mod Pathol 11:1151–1159, 1998[Medline]

23. Lichy JH, Dalbegue F, Zavar M, et al: Genetic heterogeneity in ductal carcinoma of the breast. Lab Invest 80:291–301, 2000[Medline]

24. Fujii H, Marsh C, Cairns P, et al: Genetic divergence in the clonal evolution of breast cancer. Cancer Res 56:1493–1497, 1996[Abstract/Free Full Text]

25. Kuukasjarvi T, Karhu R, Tanner M, et al: Genetic heterogeneity and clonal evolution underlying development of asynchronous metastasis in human breast cancer. Cancer Res 57:1597–1604, 1997[Abstract/Free Full Text]

26. Nakayama T, Taback B, Turner R, et al: Molecular clonality of in-transit melanoma metastasis. Am J Pathol 158:1371–1378, 2001[Abstract/Free Full Text]

27. Abeln EC, Kuipers-Dijkshoorn NJ, Berns EM, et al: Molecular genetic evidence for unifocal origin of advanced epithelial ovarian cancer and for minor clonal divergence. Br J Cancer 72:1330–1336, 1995[Medline]

28. Chen YJ, Yeh SH, Chen JT, et al: Chromosomal changes and clonality relationship between primary and recurrent hepatocellular carcinoma. Gastroenterology 119:431–440, 2000[CrossRef][Medline]

29. Shimizu S, Yatabe Y, Koshikawa T, et al: High frequency of clonally related tumors in cases of multiple synchronous lung cancers as revealed by molecular diagnosis. Clin Cancer Res 6:3994–3999, 2000[Abstract/Free Full Text]

30. Kollias J, Man S, Marafie M, et al: Loss of heterozygosity in bilateral breast cancer. Breast Cancer Res Treat 64:241–251, 2000[CrossRef][Medline]

31. Kung FY, Tse GM, Lo KW, et al: Metachronous bilateral mammary metaplastic and infiltrating duct carcinomas: A molecular study for clonality. Hum Pathol 33:677–679, 2002[CrossRef][Medline]

32. Agelopoulos K, Tidow N, Korsching E, et al: Molecular cytogenetic investigations of synchronous bilateral breast cancer. J Clin Pathol 56:660–665, 2003[Abstract/Free Full Text]

33. Larson PS, de las Morenas A, Cupples LA, et al: Genetically abnormal clones in histologically normal breast tissue. Am J Pathol 152:1591–1598, 1998[Abstract]

34. Diaz-Cano SJ, Blanes A, Wolfe HJ: PCR techniques for clonality assays. Diagn Mol Pathol 10:24–33, 2001[CrossRef][Medline]

35. Tomlinson IP, Lambros MB, Roylance RR: Loss of heterozygosity analysis: Practically and conceptually flawed? Genes Chromosomes Cancer 34:349–353, 2002[CrossRef][Medline]

36. Aubele M, Mattis A, Zitzelsberger H, et al: Intratumoral heterogeneity in breast carcinoma revealed by laser-microdissection and comparative genomic hybridization. Cancer Genet Cytogenet 110:94–102, 1999[CrossRef][Medline]

37. Glockner S, Buurman H, Kleeberger W, et al: Marked intratumoral heterogeneity of c-myc and cyclinD1 but not of c-erbB2 amplification in breast cancer. Lab Invest 82:1419–1426, 2002[Medline]

38. Sterns EE, Fletcher WA: Bilateral cancer of the breast: A review of clinical, histologic, and immunohistologic characteristics. Surgery 110:617–622, 1991[Medline]

39. Tirkkonen M, Johannsson O, Agnarsson BA, et al: Distinct somatic genetic changes associated with tumor progression in carriers of BRCA1 and BRCA2 germ-line mutations. Cancer Res 57:1222–1227, 1997[Abstract/Free Full Text]

40. Hedenfalk I, Duggan D, Chen Y, et al: Gene-expression profiles in hereditary breast cancer. N Engl J Med 344:539–548, 2001[Abstract/Free Full Text]

41. Wistuba II, Tomlinson GE, Behrens C, et al: Two identical triplet sisters carrying a germline BRCA1 gene mutation acquire very similar breast cancer somatic mutations at multiple other sites throughout the genome. Genes Chromosomes Cancer 28:359–369, 2000[CrossRef][Medline]

42. Pandis N, Teixeira MR, Gerdes AM, et al: Chromosome abnormalities in bilateral breast carcinomas: Cytogenetic evaluation of the clonal origin of multiple primary tumors. Cancer 76:250–258, 1995[CrossRef][Medline]

43. Haffty BG, Carter D, Flynn SD, et al: Local recurrence versus new primary: Clinical analysis of 82 breast relapses and potential applications for genetic fingerprinting. Int J Radiat Oncol Biol Phys 27:575–583, 1993[Medline]

44. Regitnig P, Moser R, Thalhammer M, et al: Microsatellite analysis of breast carcinoma and corresponding local recurrences. J Pathol 198:190–197, 2002[CrossRef][Medline]

Submitted May 19, 2003; accepted February 23, 2004.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				B. Banelli, I. Casciano, A. Di Vinci, B. Gatteschi, A. Levaggi, F. Carli, C. Bighin, S. Salvi, G. Allemanni, P. Ghiorzo, et al. Pathological and molecular characteristics distinguishing contralateral metastatic from new primary breast cancer Ann. Onc., October 29, 2009; (2009) mdp470v1. [Abstract] [Full Text] [PDF]

				M. A. Bollet, N. Servant, P. Neuvial, C. Decraene, I. Lebigot, J.-P. Meyniel, Y. De Rycke, A. Savignoni, G. Rigaill, P. Hupe, et al. High-Resolution Mapping of DNA Breakpoints to Define True Recurrences Among Ipsilateral Breast Cancers J Natl Cancer Inst, January 2, 2008; 100(1): 48 - 58. [Abstract] [Full Text] [PDF]

				D. J. Brenner, I. Shuryak, S. Russo, and R. K. Sachs Reducing Second Breast Cancers: A Potential Role for Prophylactic Mammary Irradiation J. Clin. Oncol., November 1, 2007; 25(31): 4868 - 4872. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schlechter, B. L. Articles by Rosenberg, C. L. Search for Related Content PubMed PubMed Citation Articles by Schlechter, B. L. Articles by Rosenberg, C. L. Social Bookmarking                            What's this?