| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2004;78:1734-1741
© 2004 The Society of Thoracic Surgeons
a Laboratory of Epithelial Cancer Biology, Memorial Sloan-Kettering Cancer Center, New York, New York, USA
b Department of Thoracic Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York, USA
c Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York, USA
d Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, New York, USA
e Department of Thoracic Oncology, Memorial Sloan-Kettering Cancer Center, New York, New York, USA
f Department of Head and Neck Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York, USA
Accepted for publication May 17, 2004.
* Address reprint requests to Dr Singh, Memorial Sloan-Kettering Cancer Center, Head and Neck Surgery, 1275 York Ave, New York, NY 10021, USA
singhb{at}mskcc.org
Presented at the Fortieth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 26–28, 2004.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
METHODS: We classified 150 tumors as pure BAC, BAC with focal invasion, or adenocarcinoma with BAC features. A tissue microarray was constructed from areas of benign lung, BAC, and invasive adenocarcinoma in these tumors. Squamous cell carcinoma–related oncogene expression was graded by immunohistochemistry from 0 to 3 (absent, low, moderate, or high), with positive SCCRO phenotype defined as grade 3. Squamous cell carcinoma–related oncogene specificity was determined by Wilcoxon rank test and area under the receiver-operator curve, survival by the Kaplan-Meier method, and correlation with prognostic factors by log-rank test.
RESULTS: Of the 86.0% (129 of 150) of specimens suitable for analysis, positive SCCRO phenotype was seen in 16.3% (21 of 129) and was 100.0% specific for tumor versus benign tissue (area under receiver-operator curve, 0.92). Positive SCCRO phenotype was greater in tumors with increasing degrees of invasive histologic type (7.0% pure BAC, 13.6% BAC with focal invasion, and 28.6% adenocarcinoma with BAC features; p = 0.02). Low-level SCCRO expression was present in 83.9% (99 of 118) of benign tissues and correlated with tobacco use and poor survival (p = 0.05).
CONCLUSIONS: Squamous cell carcinoma–related oncogene is a marker of invasive tumor progression in BAC. Low-level expression in adjacent benign lung predicts worse survival, and may represent field cancerization or host–tumor effects.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
Although noninvasive lesions as defined by the World Health Organization do occur, the BAC growth pattern is often seen with focal or even predominant invasive adenocarcinoma. This has led some investigators to hypothesize that the continuum of disease from pure BAC to BAC variants with differing degrees of invasion and, ultimately, to pure invasive adenocarcinoma may represent a paradigm of tumor progression, with sequential accumulation of oncogenetic derangements resulting in increasingly invasive tumor phenotypes [2].
Our previous work has shown that overexpression of squamous cell carcinoma–related oncogene (SCCRO) is found in 50% of non–small cell lung cancer (NSCLC), and is associated with tumor progression and an aggressive clinical course in this disease [3]. In this study we constructed a tissue microarray using pure BACs and BAC variants from different stages in the invasive spectrum of this disease, and used it to determine the putative role of SCCRO protein expression as a marker of histopathologic progression and outcomes in these tumors.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
All cases were reviewed again and reclassified on the basis of the degree of the invasive component by one of the reference pathologists (M.F.Z.) according to previously published criteria [4]. Briefly, tumors were classified as (1) pure BAC if there was no evidence of invasion (as per the World Health Organization definition), (2) BAC with focal invasion (BWFI) if the predominant growth pattern was noninvasive with less than 10% comprising invasive adenocarcinoma, or (3) adenocarcinoma with BAC features (AWBF) if the predominant growth pattern was invasive carcinoma with less than 15% of the tumor mass consisting of the BAC growth pattern.
Clinical Data Acquisition
Clinical data were obtained by retrospective review of patient medical records and included demographic data (age and sex), cancer risk data (smoking history and prior history of cancer), recurrence data (relapse or new cancers), clinical disease pattern (unifocal, multifocal, or pneumonic), and surgical-pathologic information (date and pathologic stage at first resection). Outcome data (vital status) was obtained by a search of the Social Security Death Index for all patients to June 1, 2003. Corresponding clinical data could not be obtained for one tumor specimen. This specimen was included in the analysis of SCCRO tumor sensitivity and specificity characteristics, but was excluded from all clinical and survival analyses.
Our criteria for staging and assessment of recurrence (relapse or new primary cancers) have been previously well described [4]. Briefly, all tumors were staged according to the fifth edition of the American Joint Committee on Cancer Staging Handbook [5]. Of note, tumors presenting with a pneumonic pattern of disease and indeterminate tumor size were classified as Tx and overall stage Px. Nodal N status was largely determined by results of routine mediastinal lymph node dissection. In cases of multiple tumors, the Martini-Melamed criteria were used to determine synchronous tumors from intrapulmonary metastases [6]. Relapse was defined as locoregional if recurrence occurred along a staple line or the pleura, or distant if recurrence occurred within bone, brain, or the adrenals. Cases with diffuse and clearly metastatic disease within the lung with lesions too numerous to count were also considered distant relapse. New cancers were defined as one or more new lesions within the lungs at sites distinct from the original tumor. Classification of clinical disease pattern, staging, and second events was made by joint review of patient records by V.W.R. and at least one of two other authors (I.S.S. and M.I.E.).
Tissue Microarray Construction
For all specimens, areas of tumor-adjacent benign lung, BAC, and invasive adenocarcinoma (for BWFI and AWBF variants) were marked. Triplicate 0.6-mm biopsy cores were taken from these areas and arrayed onto master paraffin blocks using a manual tissue arrayer (Beecher Instruments, Inc, Sun Prairie, WI; Fig 1). Presence of tumor in cores was confirmed by hematoxylin and eosin–stained sections of the master blocks.
|
To assess the specificity of the anti-SCCRO polyclonal antibody in NSCLC, we incubated the primary antibody with a 4x excess concentration of immunizing peptide (SCCRO Peptide DTW1 blocking peptide, developed in our laboratory) overnight at 4°C before incubation on tissue microarray and NSCLC sections. Immunohistochemistry performed on tissue microarray and NSCLC sections using antibody preincubated with blocking peptide showed abrogation of staining compared with no preincubation, demonstrating specific binding of the antibody to SCCRO protein.
In specimens with two or more cores suitable for analysis, the level of SCCRO expression was scored with an immunohistochemistry (IHC) grade of 0 (absent), 1 (low), 2 (moderate), or 3 (high) by two investigators blinded to the clinical data, with a greater than 95% concordance rate between the two readers. Only cores with a clear IHC grade of 3 (high-level expression) were considered positive for SCCRO phenotype (Fig 2), with tissues and cell lines known to express the antigen under study used as positive controls. Staining for SCCRO was predominantly cytoplasmic, with rare nuclear expression seen in a scant number of tumors.
|
Statistical Analysis
The Wilcoxon rank test and area under the receiver-operator curve were calculated to determine the sensitivity and specificity of SCCRO tumor expression compared with matched benign lung cores, and also to determine expression differences between BAC and adenocarcinomatous regions from within the same tumor. Survival was determined by the Kaplan-Meier method and assessment of prognostic indicators by log-rank. The 
2 and Fisher's exact tests were used for all other correlations. All analyses were performed by the study biostatisticians (S.C. and E.S.V.) using SAS statistical software (SAS Institute, Inc, Cary, NC).
This study was approved by both the Human Tissue Utilization Committee and Institutional Review Board of Memorial Sloan-Kettering Cancer Center for retrospective investigation of clinical markers in pathologic specimens in December 2002. All data were collected and maintained on a Health Insurance Portability and Accountability Act of 1996–compliant database.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
Eighty-six percent (129 of 150) of tumor specimens were suitable for SCCRO expression and were used in these analyses. Matched tumor-adjacent benign lung was available for 91.5% (118 of 129) of these specimens. The most common reason for loss of specimens from analysis was because of loss of cores during cutting of the master paraffin blocks and was comparable to rates of core loss reported in the literature [7, 8]. A small number of specimens were also lost to analysis as a result of absence of tumor tissue within the arrayed cores.
Table 1 summarizes the demographics for the 129 tumors used in this analysis. All tumor types were represented equally, with one third each of pure BAC, BWFI, and AWBF. Most tumors presented as unifocal lesions (58%), and as pathologic stage I or II (67%). There was a greater predominance of women to men (2:1), and patients with a negative tobacco history (20%) compared with that expected for lung adenocarcinomas as a whole. The overall recurrence rate was approximately 40%. These demographics are representative of our previously reported experience with BAC [4].
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
Molecular and genetic studies have supported this putative BAC progression model, with greater degrees of genomic aberrations in invasive BAC variants compared with noninvasive tumors. A significant rise in allelic loss at 3p, 17p, 18q, and 22q has been demonstrated with malignant progression in mixed BAC tumors [2]. Survivin expression and ras codon 12 mutations are altered in a significantly greater number of BACs compared with atypical adenomatous hyperplasia, but less so when compared with pure adenocarcinomas [12–14]. Similarly, studies of p53 mutation in NSCLC have found a significantly greater prevalence of genomic abnormalities in tumors with increased invasive components, although no definitive correlation with survival or other clinical variables between these markers, including p16, Rb, cyclin D1, and cyclin E [15, 16].
Squamous cell carcinoma–related oncogene is a novel oncogene first identified while investigating 3q amplification in squamous cell carcinomas of the head, neck, and lung. Squamous cell carcinoma–related oncogene is overexpressed in greater than 50% of NSCLC, including 10% of lung adenocarcinomas by mRNA analysis, relative to matched, histologically normal tissue [3, 17, 18]. Squamous cell carcinoma–related oncogene has been shown to correlate with a history of tobacco use and with tumor progression from in situ to invasive cancer in squamous cell carcinoma of the head and neck [3]. In addition, overexpression of SCCRO in human NSCLC, both squamous cell and adenocarcinomas, is associated with worse clinical outcomes.
For this study, we chose to use tissue microarray technology. Tissue microarrays allow for high-throughput screening of multiple tumor tissues. The use of tissue microarray has been validated for these types of studies, showing a strong correlation between the results of immunohistochemical staining of tissue arrays (with at least three cores per specimen) and corresponding whole-tumor section. Using this methodology, we wanted to determine (1) the expression patterns of SCCRO in BAC and its variants within a tumor progression model, and (2) its usefulness as a prognostic marker in this disease. Given the association of SCCRO with invasive tumor phenotypes and poor clinical outcomes, we hypothesized that SCCRO expression would be significantly greater in invasive BAC variants compared with noninvasive BAC tumors, and that SCCRO overexpression would be a strong predictor of outcome.
We first confirmed that SCCRO expression was indeed highly specific for tumors by comparing expression patterns between neoplastic and matched benign cores (specificity 100%, area under the receiver-operating curve, 0.93). We then compared the amount of positive SCCRO phenotype among our different tumor types and found it to be significantly correlated with increasingly invasive BAC variants, doubling in incidence from pure BAC to BWFI (7% to 13.6%), and again to AWBF (28.6%). However, no correlation was seen with overall survival. This finding is congruent with a previous report from our institution, from a subset of this patient cohort, in which survival was shown not to be dependent on the degree of invasive phenotype within these tumors [4]. In contrast, other authors have reported significantly worse outcomes with more-invasive BAC lesions, although the lack of a uniform system for classifying these tumors complicates comparisons between these conflicting findings [10, 11]. Further work needs to be done to elucidate the true role of the invasive BAC phenotype in predicting survival.
Interestingly, low-level expression SCCRO was seen in 84% of tumor-adjacent benign lung tissues, and was significantly greater in patients with a smoking history when compared with never-smokers (p = 0.05), suggesting dysregulation of SCCRO is an early event in the pathogenesis of these tumors and may be augmented by tobacco use. In addition, low-level expression of SCCRO in the adjacent histologically benign lung tissue was associated with a worse overall survival (p = 0.05). We believe this may represent host–tumor interactive effects or possible field cancerization of the surrounding lung induced by oncogenic insults (ie, tobacco). Similar genetic changes in benign lung have been demonstrated with a variety of other markers, including p53, ras, epidermal growth factor receptor, and Ki-67 [19–22]. However, none of these studies have thus far shown a correlation with survival and genetic abnormalities in tumor-adjacent benign tissues in NSCLC. These findings suggest that detection of SCCRO expression in tumor-adjacent benign tissues may be a useful tool in predicting clinical outcomes in these patients. However, studies with larger sample sizes are needed to validate this finding as an independent predictor of survival in multivariate analyses.
In summary, SCCRO is a novel marker in BAC with a high specificity for tumor versus benign tissues. It may be a marker of tumor progression in BAC, with significantly greater degrees of high-level expression in increasingly invasive histologic phenotypes. Low-level expression of SCCRO in tumor-adjacent benign tissue is associated with tobacco use and worse survival in patients undergoing resection of BAC, and may be a marker of molecular margin status in these cases. This finding also raises interesting questions regarding host–tumor effects and field cancerization in lung cancer. Further investigation to clarify the molecular and cellular mechanisms driving the observed associations between SCCRO and BAC may help to elucidate the oncogenetic processes involved in neoplastic progression.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
DR SARKARIA: We believe SCCRO is part of a growing list of oncodevelopmental genes, with regulated expression playing a critical role in development and maintenance, and uncontrolled expression driving malignant transformation.
Most of what we know about SCCRO is through its established function as an oncogene in the dysregulated state. Structural analysis of the SCCRO gene leads us to believe it is a transcription factor of the basic helix-loop-helix leucine zipper family. In concordance with this initial observation, we have shown that SCCRO does in fact function as a transcription factor regulating the expression of a number of genes involved in oncogenesis, including GLI1, VEGF-A, and possibly mmp2.
Less is known about the normal function of SCCRO. However, a number of lines of evidence, including expression patterns in mouse embryos, suggest SCCRO functions in early development. One of the strongest pieces of evidence supporting the role of SCCRO in development comes from our work with transgenic mice, in which the great majority of animals overexpressing the inserted SCCRO transgene die in utero, with animals surviving into the postnatal period showing significant growth retardation.
Other investigations suggest that the activity of SCCRO is dependent, at least in part, on sonic hedgehog (Shh) signaling, a well-described pathway vital to early embryonic development. We have demonstrated that SCCRO translocates to the nucleus from the cytoplasm in the presence of activated Shh signaling. Furthermore, immunohistochemical staining for SCCRO in mouse embryos shows expression in a number of systems, with a good deal of overlap in tissues in which Shh signaling is known to play a significant role.
To summarize, although we are still in the process of characterizing this gene as a whole, our studies to date suggest SCCRO plays a significant role not only in oncogenesis, but also in early development, possibly through Shh-mediated mechanisms.
DR MICHAEL T. JAKLITSCH (Boston, MA): I congratulate you on a very nice paper. One of the problems that we have had trying to characterize bronchioloalveolar cell carcinoma (BAC) is the intratumor variability. In fact, that spectrum of three histologic slides that you showed, I would not be surprised if you told me that they came from the exact same tumor. So how do you account for this when you are coding these into your database? I would suggest to you that what you have found is a novel gene marker for invasive adenocarcinoma.
DR SARKARIA: To address your first question, yes, if you examine discrete points within the same specimen, you are going to see different relative combinations of invasive versus noninvasive components of the tumor. To overcome this, we tried to get that best overall picture of the entire tumor by examining numerous sections from the pathologic specimens. In this manner, we categorized the tumors as pure BAC, BAC with focal invasion, or adenocarcinoma with BAC features based on percentage of the invasive component within the entire tumor, not just within one small area.
Your second question highlights an excellent point, and has generated a good deal of discussion in our internal meetings regarding this work. The hypothetical model of tumor progression we have suggested, from pure BAC to BAC variants, and ultimately to invasive adenocarcinoma, is a construct we are using within the scope of a larger project to investigate the biology of these tumors. Within this framework, we are trying to determine the differential genetic profiles involved in progression not only between tumor types (pure BAC, BAC with focal invasion, adenocarcinoma with BAC features, and pure adenocarcinoma) but also between individual cell populations (invasive and noninvasive) from within the same tumors. Therefore, we interpret our current findings within this model, and suggest that SCCRO is a marker of invasive tumor progression in BAC. However, if the overall hypothetical model does not prove to hold true, then you would likely be correct in saying that SCCRO is simply acting as a marker of invasive adenocarcinomas in general. The definitive answer to this question will rely on the findings from our long-term project.
DR ARA A. VAPORCIYAN (Houston, TX): I want to congratulate you on your work. That is a large amount of work.
I just wanted to clarify one thing and ask you to comment on it. At the beginning when you were describing SCCRO, you mentioned that it is present in 50% of non–small cell lung cancers, yet only 9% of adenocarcinomas, and yet, as your BAC progressed from a pure BAC to an invasive adenocarcinoma, you saw an increase in expression. Could you comment on that?
DR SARKARIA: At a glance, the prevalence of SCCRO mRNA expression in lung adenocarcinomas (9% to 10%) described in our previous studies is actually quite comparable to the overall prevalence of expression (16%) found by immunohistochemistry in the BAC subtype of adenocarcinoma described in the current study. However, as your question points out, how do we explain the high prevalence of expression seen with more-invasive BAC variants (almost 30% in adenocarcinoma with BAC features), which suggests an even higher prevalence in purely invasive adenocarcinomas over the 9% originally described. There are a few ways to address this discrepancy.
First, the initial cohort of adenocarcinomas used to characterize SCCRO expression was not subdivided by type. Therefore, we do not know what proportion of these tumors may or may not have been BAC or BAC variants. The prevalence of expression initially described suggests the original cohort of tumors may have represented a rather diverse cross-section of lung adenocarcinomas consisting of invasive as well as noninvasive variants. In this case, the prevalence of SCCRO expression in purely invasive adenocarcinomas would likely have been underestimated.
Second, the initial experiments looking at expression in lung adenocarcinomas was based mainly on analysis of SCCRO mRNA levels, not protein expression. As there may not be a one-to-one correlation between mRNA and protein levels, the different approaches to the detection of SCCRO expression may yield different results.
Taken all together, these points suggest that the true prevalence of SCCRO expression in lung adenocarcinomas may be higher than what we originally described.
DR CHING TZAO (Taipei, Taiwan): There are an increasing number of centers that perform large-scale screening for lung cancer. Do you have such a program in your institute? If yes, do you see cases of atypical adenomatoid hyperplasia, which is believed to be probably a precancerous lesion for adenocarcinoma? Could you extend your hypothesis to that population or do you have information on that?
DR SARKARIA: Several reports by other investigators have shown that there is increased genetic variability, including dysregulation of p53, survivin, and a variety of other genes between atypical adenomatous hyperplasia and BAC. We have not looked at these lesions in the context of our BAC to invasive adenocarcinoma tumor progression model. It is, however, an important component in the cancer progression model to incorporate into our future investigations.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDISCUSSIONAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  I. Akagi, M. Miyashita, H. Makino, T. Nomura, N. Hagiwara, K. Takahashi, K. Cho, T. Mishima, T. Takizawa, and T. Tajiri SnoN Overexpression is Predictive of Poor Survival in Patients with Esophageal Squamous Cell Carcinoma Ann. Surg. Oncol., October 1, 2008; 15(10): 2965 - 2975. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. S. Sarkaria, M. F. Zakowski, D. Pham, M. Hezel, M. I. Ebright, S. Chuai, E. S. Venkatraman, M. G. Kris, V. W. Rusch, and B. Singh Epidermal Growth Factor Receptor Signaling in Adenocarcinomas With Bronchioloalveolar Components Ann. Thorac. Surg., January 1, 2008; 85(1): 216 - 223. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. B. van den Broek, V. B. Wreesmann, M. W.M. van den Brekel, C. R.N. Rasch, A. J.M. Balm, and P. H. Rao Genetic Abnormalities Associated with Chemoradiation Resistance of Head and Neck Squamous Cell Carcinoma Clin. Cancer Res., August 1, 2007; 13(15): 4386 - 4391. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. Raz, B. He, R. Rosell, and D. M. Jablons Current concepts in bronchioloalveolar carcinoma biology. Clin. Cancer Res., June 15, 2006; 12(12): 3698 - 3704. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
