Discrimination of multiple primary lung cancers from intrapulmonary metastasis based on the expression of four cancer-related proteins

Authors

  • Kenji Ono MD,

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author
  • Kenji Sugio MD,

    Corresponding author
    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    • Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 807-8555, Japan
    Search for more papers by this author
    • Fax: (011) 81-93-692-4004

  • Hidetaka Uramoto MD,

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author
  • Tetsuro Baba MD,

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author
  • Yoshinobu Ichiki MD,

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author
  • Mitsuhiro Takenoyama MD,

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author
  • Takeshi Hanagiri MD,

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author
  • Tsunehiro Oyama MD,

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author
  • Kosei Yasumoto MD

    1. Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
    Search for more papers by this author

Abstract

BACKGROUND:

Because distinguishing between multiple primary lung cancers and intrapulmonary metastasis is often difficult when the tumor histology is same, the feasibility of analyzing differential protein expression profiles to distinguish multiple primary lung cancers from intrapulmonary metastasis was evaluated.

METHODS:

This study enrolled 50 patients, with multiple primary lung cancers demonstrating the same histology, who underwent surgery between April 1994 and March 2006 and 20 patients who were diagnosed to have intrapulmonary metastasis during the same period. Thirty patients with lymph-node metastasis were selected for comparison purposes. The sum value of the differences in the expression ratio of 4 proteins (p53, p16, p27, and c-erbB2) was evaluated in immunohistochemically stained specimens among multiple primary lung cancers and intrapulmonary metastasis.

RESULTS:

None of the 30 patients with lymph-node metastasis showed a sum value of the differences between primary tumor and lymph-node metastasis in the 4 protein expression ratios >90. Therefore, when the difference between 2 tumors exceeded 90, the 2 tumors were considered to be different from each other, ie, multiple primary lung cancers. Forty-one of 50 (82%) patients who were clinically diagnosed to have multiple primary lung cancers showed a sum value of the differences in their protein expression ratios >90, 4 of 20 (20%) patients who were clinically diagnosed to have intrapulmonary metastasis showed a sum value >90. Among the patients who were clinically diagnosed to have multiple primary lung cancers and intrapulmonary metastasis, the 5-year survival of 70 patients who had a sum value of the differences in their 4 protein expression ratios, either >90 (newly classified multiple primary lung cancers) and ≤90 (newly classified intrapulmonary metastasis), were 81.1% and 40.2%, respectively (P = .002).

CONCLUSIONS:

The profile of protein expression in cancer-related genes is, thus, considered to be a useful tool for distinguishing multiple primary lung cancers from intrapulmonary metastasis and for determining the appropriate biological staging of lung cancer. Cancer 2009. © 2009 American Cancer Society.

Lung cancer is the leading cause of cancer-related deaths in Japan, and it kills more than 60,000 people per year. It is a malignancy with poor prognosis, and only 50% of patients achieve a 5-year survival, even with a complete surgical resection.1 Recent developments in various imaging and therapeutic modalities have increased the ability to identify either synchronous or metachronous multiple primary lung cancers. The incidence of multiple primary lung cancers has been reported to range from 0.7% to 15% of patients with lung cancer.2-6 Distinguishing multiple primary lung cancers from intrapulmonary metastasis is difficult in a substantial number of cases, thus hindering the administration of appropriate therapeutic strategies. Although surgical treatment can also be a curative option for multiple primary lung cancers as well as single lung cancer, the 5-year survival of patients with synchronous multiple primary lung cancers has been reported to range from 0% to 44%7-10 despite an early diagnosis. One of the reasons for such a wide range in this survival data is that patients who were clinically diagnosed with multiple primary lung cancers may include some patients with intrapulmonary metastasis. Criteria proposed by Martini and Melamed7 has been widely used in the diagnosis of multiple primary lung cancers in the clinical setting; however, clearly distinguishing multiple primary lung cancers from intrapulmonary metastasis is often difficult when the diagnosis is based solely on the clinical findings. However, several authors have reported that they were able to distinguish multiple primary lung cancers by using a gene-mutational analysis of tumors.11-13

Several cancers have been known to develop through the serial, multistage accumulation of genetic and epigenetic alterations.14, 15 The expression of p53, ras, and c-erbB2 and negative or reduced expression of p16 and p27 have been reported to be independently unfavorable prognostic factors for nonsmall cell lung cancer (NSCLC) patients,16-20 thus suggesting that changes in the expression of these genes may depend on either the stage or the malignant potential of the tumor. Therefore, the variability in the expression of these genes between cancers could be used for the fingerprinting of the clonal origin of cancer cells if multiple oncogenes are combined. Therefore, the differential expression of 4 proteins between either 2 tumors or primary tumor and lymph-node metastasis from the same patient was evaluated by immunohistochemical (IHC) staining to evaluate whether these proteins can be used for discriminating between multiple primary lung cancers and intrapulmonary metastasis.

MATERIALS AND METHODS

Patients and Clinical Features

Among 851 consecutive patients with primary lung cancer who had undergone a surgical resection between April 1994 and March 2006 at the Second Department of Surgery at University of Occupational and Environmental Health, Kitakyushu, Japan, 84 (9.9%) patients were diagnosed with multiple primary lung cancers according to criteria proposed by Martini and Melamed.7 Of these, 65 patients were diagnosed as having multiple primary lung cancers with the same histological type, and 19 patients were diagnosed with different histological types. This study was aimed at discovering whether an additional nodule of the same histological type as the primary tumor represents either a second primary tumor or metastasis. Among 65 multiple primary lung cancers patents with the same histological type, 39 patients were diagnosed with synchronous multiple primary lung cancers, and 26 patients were diagnosed with metachronous multiple primary lung cancers. Fifteen patients were excluded from this study because of a lack of surgical specimens (11 patients, nonsurgical treatment for the second tumor; 4 patients, the first lesion had been resected at another hospital). Finally, 50 patients (29 patients, synchronous multiple primary lung cancers; 21 patients, metachronous multiple primary lung cancers) with multiple primary lung cancers in whom paraffin sections of all tumors were available were enrolled in this study. Twenty patients diagnosed with intrapulmonary metastasis during the same period were included in this study. Thirty patients with lymph-node metastasis were selected for comparison of the expression of the 4 proteins between the primary and metastatic lymph-node tumors. Furthermore, 11 patients diagnosed to have multiple primary lung cancers with different histological types in whom paraffin sections of all tumors were available were included as a negative control. The clinicopathological data were obtained from a retrospective chart review. The tumor stage was classified according to the 1997 revisions in the international system for staging lung cancer.21

The characteristics of the patients with multiple primary lung cancers are shown in Table 1. The 20 patients with intrapulmonary metastasis included 14 men and 6 women from 54 to 79 years of age. Sixteen patients were diagnosed with adenocarcinoma, and 4 were diagnosed with squamous cell carcinoma. The pathological stages of the cancer were stage IIIB in 9 patients and stage IV in 11. Initially, 30 patients with lymph-node metastasis were analyzed for the difference in the expression in the 4 proteins between the primary tumor and the metastatic lymph-node lesion as a control. They included 23 men and 7 women from 48 to 78 years of age. Eighteen patients were diagnosed with adenocarcinoma, 9 with squamous cell carcinoma, 2 with large cell carcinoma, and 1 with adenosquamous carcinoma. The pathological stages of the cancers were stage IIA in 3 patients, stage IIB in 4, stage IIIA in 21, stage IIIB in 1, and stage IV in 1.

Table 1. Characteristics of Multiple Primary Lung Cancers of the Same Histological Type
CharacteristicsSynchronous Multiple Lung CancersMetachronous Multiple Lung Cancers Intrapulmonary MetastasisLymph Node Metastasis
No. (%) of PatientsNo. (%) of Patients No. (%) of PatientsNo. (%) of Patients
Patients on study2921 2030
Sex     
 Men19 (65.5)14 (66.7) 14 (70.0)23 (76.7)
 Women10 (34.5)7 (33.3) 6 (30.0)7 (23.3)
Age, y     
 Mean [range]70.5 [43-84]68.1 [46-82] 66.4 [54-79]67.3 [48-78]
No. of tumors     
 223 (79.3) 
 35 (17.2) 
 61 (3.4) 
Interval after the first tumor, mo     
 Median [range]51.0 [12.5-147] 
Histological type     
 Adenocarcinoma24 (82.8)17 (81.0) 16 (80.0)18 (60.0)
 Squamous cell carcinoma5 (17.2)4 (19.0) 4 (20.0)9 (30.0)
 Others0 (0)0 (0) 0 (0)3 (10.0)
p-stageMaximum StageFirst TumorSecond Tumor  
 IA13 (44.8)12 (57.1)16 (76.2)
 IB8 (27.6)5 (23.8)3 (14.3)
 IIA1 (3.4)0 (0)0 (0)3 (10.0)
 IIB2 (6.9)1 (4.8)1 (4.8)4 (13.3)
 IIIA4 (13.8)2 (9.5)1 (4.8)21 (70.0)
 IIIB1 (3.4)1 (4.8)0 (0)9 (45.5)1 (3.3)
 IV0 (0)0 (0)0 (0)11 (55.5)1 (3.3)

Immunohistochemical (IHC) Staining

Four proteins (p53, p16, p27, and c-erbB2) were selected for evaluation of the differential diagnosis of multiple primary lung cancers and intrapulmonary metastasis. IHC staining was conducted by using serial sections from the same paraffin-embedded blocks. All tissue were formalin-fixed and processed similarly, according to standard histology practices. A 3 μm thick formalin-fixed paraffin-embedded tissue section was prepared from each specimen. All specimens were stained with hematoxylin and eosin (H & E) for the histological diagnosis. IHC staining was performed by the streptavidin-biotin-peroxidase complex method (Histofine SAB kit; Nichirei, Tokyo, Japan). The sections were briefly immersed in citrate buffer (0.01 mol/L citric acid [ph 6.0]) and then were incubated for 2 5-minute intervals at 100°C in a microwave oven for antigen retrieval. They were then incubated with the anti-p53 mouse monoclonal antibody (DO-1; Siemens Healthcare Diagnostics: Oncogene Science, Cambridge, Mass) diluted at 1:100, the anti-p16 rabbit polyclonal antibody (C-20; Santa Cruz Biotechnology, Santa Cruz, Calif) diluted at 1:200, anti-p27 mouse monoclonal antibody (1B4; Novocastra Laboratories, Newcastle upon Tyne, UK) diluted at 1:250, anti-c-erbB2 mouse monoclonal antibody (5A2; Novocastra Laboratories) diluted at 1:100 overnight in a cold room using a Labeled Streptavidin Biotin kit (CA930 13, DAKO LSAB kit; Dako, Carpinteria, Calif). These antibodies were diluted in phosphate-buffered saline (PBS) containing 2% bovine serum albumin (BSA).

Evaluation of the Stained Specimens

Appropriate positive and negative controls were used throughout the experiment, and the slides were independently reviewed by 2 of the authors (K. O. and K. S.) who were blinded to the clinicopathological data. After the IHC detection of p53, p16, p27, and c-erbB2 in each specimen, the percentage of immunoreactive tumor cells in 5 fields (400× original magnification) selected randomly from 1 slide was recorded, and the final value of positive tumor cells was determined as the average of the positive immunostained cells. The extent of immunostaining was scored on the basis of the percentage of positive cells for each tumor specimen as follows: 0, no staining; 10, 1% to 10% staining; 20, 11% to 20%; 30, 21% to 30%; 40, 31% to 40%; 50, 41% to 50%; 60, 51% to 60%; 70, 61% to 70%; 80, 71% to 80%; 90, 81% to 90%; 100, 91% to 100%. Positive cells for p53, p16, and p27 were defined to be cells with nuclear staining, whereas positive cells for c-erbB2 were defined to be cells with cell-membrane staining.

Statistical Analysis

In the prognostic analysis, the cumulative survival rate was estimated by using Kaplan-Meier survival curves with the date of surgery (the date of second operation in metachronous multiple primary lung cancers) and the date of last follow-up or death of the patient defined as the beginning points and endpoints, respectively. The significance of differences was assessed by using the log-rank test, where P-values <.05 were considered statistically significant. All data were analyzed with the use of Survival Tools for the StatView software package (Abacus Concepts, Berkeley, Calif).

RESULTS

First of all, the difference in the expression of p53, p16, p27, and c-erbB2 proteins were compared in primary lung tumor and the metastatic tumor in the lymph nodes in 30 patients as shown in Table 2 to discriminate multiple primary lung cancers from intrapulmonary metastasis. With p53, the maximal value of the difference in the protein expression ratio was 40% in 1 patient, and the number of patients in whom the value of the difference in the protein expression ratio was ≤20% was 24 of 30. With p16, the maximal value of difference in the protein expression ratio was 30% in 8 patients, and the number of patients in whom the value of the difference in the protein expression ratio was ≤20% was 22 of 30. With p27, the maximal value of the difference in the protein expression ratio was 40% in 2 patients, and the number of patients in whom the value of the difference in the protein expression ratio was ≤20% was 23 of 30. For c-erbB2, the maximal value of difference in the protein expression ratio was 30% in 6 patients, and the number of patients in whom the value of the difference in the protein expression ratio was ≤20% was 24 of 30. Finally, the maximal sum value of the differences in the 4 protein expression ratios was 90 in 3 patients as shown in Figure 1, and the average value was 62 with a range of 30-90. The limit of the sum value of differences in the 4 protein expression ratios diagnosed as identical was ≤90. In contrast, 11 patients diagnosed to have multiple primary lung cancers with different histological types showed sum values ranging from 110 to 220 for the difference in the 4 protein expression ratios. Therefore, when the difference between 2 tumors exceeded 90, it indicated that the 2 tumors were different, ie, multiple primary lung cancers. On the basis of these data, when the difference between 2 tumors exceeded 90, the tumor was considered to be a newly classified multiple primary lung cancer. When the difference was ≤90, the tumor was considered to be identical, ie, a newly classified intrapulmonary metastasis.

Figure 1.

The sum value of the difference in the protein expression ratio among 4 cancer-related genes in patients who were pathologically diagnosed to have lymph-node metastasis.

Table 2. Protein Expressions of Four Genes by IHC in Primary Tumors and Metastatic Lymph Nodes
Case No.Age, ySexHistologyp-Stagep53p16p27c-erbB2Sum of the Differences*
% (Primary Tumor/Metastatic Lymph Node)
  • Ad indicates adenocarcinoma; sq, squamous cell carcinoma; la, large cell carcinoma; ad-sq, adenosquamous carcinoma.

  • *

    Sum of the differences is the value used to add up a difference of protein expression ratio among 4 cancer-related genes.

  • Difference in percentage of positive-staining cells between primary tumor and metastatic tumor in the lymph node.

164MadIIA0 (90/90)20 (90/70)10 (70/80)10 (10/20)40
266MadIIA20 (40/60)30 (40/70)20 (80/60)10 (20/10)80
372MadIIB20 (10/30)10 (60/70)10 (80/90)20 (30/10)60
476WadIIB10 (20/10)20 (90/70)10 (60/70)20 (70/50)60
565MadIIB30 (40/70)30 (30/60)10 (80/90)0 (10/10)70
664MadIIB40 (80/40)10 (70/60)10 (80/90)20 (50/30)80
760WadIIIA10 (60/70)30 (30/60)0 (70/70)0 (10/10)40
866MadIIIA0 (10/10)0 (90/90)20 (90/70)20 (30/10)40
973MadIIIA10 (0/10)10 (40/30)10 (70/80)10 (10/0)40
1068MadIIIA10 (70/80)0 (80/80)30 (40/70)10 (40/30)50
1157MadIIIA20 (30/10)20 (80/60)10 (10/20)30 (60/30)80
1263MadIIIA10 (0/10)30 (10/40)30 (20/50)10 (0/10)80
1367WadIIIA30 (40/70)10 (90/80)30 (40/70)10 (10/20)80
1469MadIIIA30 (30/60)30 (30/60)20 (70/90)10 (10/0)90
1574MadIIIA30 (50/80)30 (20/50)20 (40/60)10 (30/20)90
1672MsqIIIA0 (0/0)10 (10/20)10 (80/70)10 (50/40)30
1771MsqIIIA20 (20/0)10 (10/0)10 (60/70)10 (30/40)50
1864MsqIIIA20 (90/70)20 (40/60)10 (70/80)10 (10/20)60
1966WsqIIIA10 (60/70)20 (60/80)10 (50/60)20 (20/40)60
2066WsqIIIA30 (50/80)10 (80/70)10 (60/70)20 (60/80)70
2172MsqIIIA10 (80/70)20 (70/50)10 (70/80)30 (50/20)70
2277MsqIIIA0 (0/0)20 (20/40)20 (70/90)30 (50/20)70
2366WsqIIIA10 (60/70)10 (70/60)30 (50/80)30 (60/30)80
2467MsqIV20 (90/70)30 (80/50)10 (70/80)10 (20/10)70
2568MlaIIIA10 (80/70)10 (80/70)10 (80/70)30 (40/10)60
2678MlaIIIA0 (40/40)10 (90/80)40 (40/80)10 (20/10)60
2774MlaIIIA10 (60/70)10 (70/80)40 (40/80)10 (20/30)70
2854MlaIIIB20 (30/10)20 (10/30)20 (60/80)10 (0/10)70
2972Mad-sqIIA20 (20/40)20 (70/50)30 (20/50)20 (30/50)90
3048Wad-sqIIIA10 (0/10)30 (90/60)10 (80/90)30 (50/20)80

Among the 29 patients with synchronous multiple primary lung cancers who were clinically diagnosed on the basis of criteria proposed by Martini and Melamed, 24 (82.8%) patients had a sum value of differences in protein expression ratio of >90 (Fig. 2). Among the 21 patients with metachronous multiple primary lung cancers who were clinically diagnosed on the basis of criteria proposed by Martini and Melamed, 17 (81.0%) patients had a sum value of differences in the protein expression ratio >90 (Fig. 3). According to this criterion, 41 of 50 (82.0%) patients were rediagnosed as having newly classified multiple primary lung cancers, and 9 patients were rediagnosed as having newly classified intrapulmonary metastasis. The number of patients with 50 multiple primary lung cancers in whom differences in the protein expression ratio was >40% (the maximal value of difference in the p53 and p27 protein expression ratios between primary and metastatic tumors in the lymph node) was 21 and 10 in p53 and p27, respectively. The number of patients with multiple primary lung cancers in whom differences in the protein expression ratio was >30% (the maximal value of difference in the p16 and c-erbB2 protein expression ratio between primary and metastatic tumor in the lymph node) was 19 and 19 in p16 and c-erbB2, respectively.

Figure 2.

The sum value of the difference in the protein expression ratio among 4 tumor-related genes in patients who were clinically diagnosed to have synchronous multiple lung cancers of the same histological type. A: newly classified intrapulmonary metastasis, B: newly classified multiple primary lung cancers.

Figure 3.

The sum value of the difference in the protein expression ratio among 4 tumor-related genes in patients who were clinically diagnosed to have metachronous multiple lung cancers of the same histological type. A: newly classified intrapulmonary metastasis, B: newly classified multiple primary lung cancers.

In a similar way, among the 20 patients with intrapulmonary metastasis who were clinically diagnosed according to criteria proposed by Martini and Melamed, 16 (80.0%) patients had a sum value of the differences of ≤90 in the expression ratios of the 4 proteins. According to the same criterion, 16 of the 20 patients were rediagnosed with newly classified intrapulmonary metastasis, and 4 patients were rediagnosed as having newly classified multiple primary lung cancers (Fig. 4). In total, 45 patients were, therefore, rediagnosed to have newly classified multiple primary lung cancers, and 25 patients were rediagnosed to have newly classified intrapulmonary metastasis.

Figure 4.

The sum value of the difference in the protein expression ratio among 4 cancer-related genes in patients who were clinically diagnosed to have intrapulmonary metastasis. A: newly classified intrapulmonary metastasis, B: newly classified multiple primary lung cancers.

Survival Analysis

Two survival analyses were compared among the 70 patients who were clinically diagnosed with either multiple primary lung cancers or intrapulmonary metastasis by Martini and Melamed's criteria and by the classification as newly classified multiple primary lung cancers and newly classified intrapulmonary metastasis. The 5-year survival rates of the 50 multiple primary lung cancer patients and the 20 intrapulmonary metastasis patients based on Martini and Melamed's criteria were 61.7% and 46.2%, respectively (P = .069). The 5-year survival rate in the 45 patients with newly classified multiple primary lung cancers was 81.8%, whereas the survival rate in 25 patients with newly classified intrapulmonary metastasis was 40.2%, which showed a statistically significant difference (P = .002; Fig. 5A). Among the 29 synchronous multiple primary lung cancer patients (based on Martini and Melamed's criteria), a survival analysis was performed in 21 patients with p-stage I. The patients with newly classified multiple primary lung cancers (n = 16) showed more favorable prognosis than the patients with newly classified intrapulmonary metastasis (n = 5; P = .015; Fig. 5B).

Figure 5.

The Kaplan-Meier survival curve of patients divided into 2 groups by the sum value of the difference in protein expression among 4 cancer-related genes in patients. (A) 45 patients were rediagnosed to have newly classified multiple primary lung cancers, and 25 patients were rediagnosed to have newly classified intrapulmonary metastasis. A significant difference was shown between these 2 groups (P = .002). (B) In the patients with p-stage I of synchronous multiple primary lung cancers, 16 patients were in newly classified multiple primary lung cancers and 5 were in newly classified intrapulmonary metastasis (P = .015).

DISCUSSION

It may be easy to diagnose multicentric primary lung cancers in multiple lung lesions when their histological types are different. However, if they show the same histological type, it is often difficult to discriminate multiple primary lung cancers from intrapulmonary metastasis. In addition, patients who have been clinically diagnosed with multiple primary lung cancers sometimes show extremely poor 5-year survival (0%-44%) even at stage I disease.7-10 These results suggest that some of the patients who have been clinically diagnosed to have multiple primary lung cancers may, therefore, have metastatic lesions. This indicates a possible limitation in the criteria proposed by Martini and Melamed, which have widely been used for the diagnosis of multiple primary lung cancers in clinical settings. On the basis of analyses of the overall survival among pathologically staged cases, nodules within the same lobe as the primary lesion are categorized by the International Union Against Cancer (UICC, 6th edition) as T4 and those located in the different lobe (whether it is ipsilateral or contralateral side) as M1. The International Association for the Study of Lung Cancer (IASLC) lung cancer-staging project recommended changes in the T classification to subclassify additional nodules in the same lobe as T3 and nodules in the ipsilateral different lobe as T4 and those in a contralateral lobe as M1.22 Their different biologic behaviors may, thus, be responsible for prognostic differences. Because some of the patients with intrapulmonary metastasis show poor prognosis, it is necessary to discriminate intrapulmonary metastasis from multiple primary lung cancers. Therefore, biological analyses are considered to be one of the useful tools for distinguishing multiple primary lung cancers from intrapulmonary metastasis and for determining the appropriate biological staging of lung cancer. To overcome this problem, several authors have reported the feasibility of clonal analyses between tumors to discriminate multiple primary lung cancers from intrapulmonary metastasis. A multiple-gene analysis to identify the clonality in a combination of multiple-gene mutations, such as a p53 gene mutation, K-ras mutation, and/or loss of heterozygosity has been reported.12, 23-31 Mitsudomi13 analyzed the p53 gene mutation in 16 patients with multiple primary lung cancers. Among those patients, 7 were not informative because of the absence of the p53 mutation in both tumors. Among 9 patients who had at least 1 p53 mutation in their pair of tumors, 6 were suggested to be of different clonal origin and were diagnosed as multiple primary lung cancers. Matsuzoe11 reported that 7 of 20 patients who were clinically diagnosed to have intrapulmonary metastasis showed different p53 gene mutations between the 2 lesions, thus indicating these lesions to be multiple primary lung cancers. These reports indicate that the somatic mutations of the p53 gene may be a suitable biological factor to identify multiple primary lung cancers. Although there is a positive expression of the p53 due to a prolonged half-life time induced by a somatic mutation of the p53 gene, the concordance rate between p53 gene mutations and immunopositivity in NSCLC is reported to be 60% to 70%.32, 33 In the present study, we have used IHC staining to distinguish between multiple primary lung cancers and intrapulmonary metastasis.

The IHC expressions of p53,16 p16,17 p2718, and c-erbB220 have been reported by us as significant prognostic factors for lung cancer. p53 is a transcription factor activated in response to DNA damage, associated with regulation of the cell cycle, induction of apoptosis, and DNA repair, but mutated p53 does not have these functions, causing inactivation of certain genes due to either a point mutation or a deletion. Either gene mutation or overexpression of p53 is observed in approximately 80% of small cell lung cancer patients and in approximately 50% of NSCLC patients, and studies report mutation or overexpression of p53 to be an unfavorable prognostic factor,34 although this is still controversial. p16, a CDK inhibitor belonging to the INK4 family, specifically combines with CDK4 to prevent Rb phosphorylation by CDK; thus, p16 regulates the cell cycle. A negative proportion of p16 expression is observed in 30% to 70% of NSCLCs,35 and patients with negative p16 staining have a less favorable prognosis than those patients with positive p16 staining.17 p27 is a CDK inhibitor belonging to the CIP-KIP family, associated with the regulation of the cell cycle, and suppresses the G1 to S cell-cycle progression by inhibiting Rb phosphorylation. A reduced proportion of p27 expression is observed in 70% to 80% of NSCLCs, and this reduced-expression group show poorer prognosis in comparison to the positive-expression group.18, 36, 37 Either gene amplification or the overexpression of c-erbB2 has been reported to be a poor prognostic factor for breast cancer patients. The positive expression of c-erbB2 has been reported to occur in 28% to 38% of adenocarcinomas of the lung, and it also correlates with an unfavorable prognosis.19, 38-40

There is a problem related to tumor heterogeneity in IHC staining. To minimize the heterogeneity problem, the percentage of positive immunostaining cells was evaluated by averaging 5 randomly selected microscopic fields (×400) of the tumor specimen. Moreover, we chose 4 cancer-related gene products (p53, p16, p27, and c-erbB2) to make up for a weakness in IHC staining, which is its poor ability to distinguish multiple primary lung cancers from intrapulmonary metastasis. The difference in protein expression was hypothesized to be larger in multiple primary lung cancers because they have different clonal origins, whereas protein expression would be smaller in intrapulmonary metastases because they have the same clonal origins. In practice, the sum of the differences of the expression ratios of the 4 proteins did not exceed 90 between the primary tumor and metastatic tumors in the lymph node. This result indicated that when the difference between the 2 tumors exceeds 90, they should be newly classified as multiple primary lung cancers. Eleven patients diagnosed to have multiple primary lung cancers with different histological types also showed a sum value of the difference in the 4 protein expression ratios of ≤90. On the contrary, when the difference remains within 90, they should then be newly classified as intrapulmonary metastasis. Therefore, 50 of the multiple primary lung cancers and 20 of the intrapulmonary metastasis (based on Martini and Melamed's criteria) were reclassified by these new criteria. Forty-one of the 50 multiple primary lung cancers were classified as newly classified multiple primary lung cancers and 9 of them as newly classified intrapulmonary metastasis. Furthermore, 16 of 20 intrapulmonary metastasis were classified as newly classified intrapulmonary metastasis and 4 of them as newly classified multiple primary lung cancers. In our preliminary study, we examined the genomic mutations of the epidermal growth factor receptor (EGFR) and K-ras in 29 of the 50 patients with multiple primary lung cancers enrolled in this study. In 9 patients, we identified the different genotypes of the 2 genes in their tumors, which indicated these tumors to be multiple cancers with a different clonal origin, and they had a sum value of the difference in the protein expression ratio of between 130 and 220. The analyses of genomic mutations are, therefore, considered to be more reliable for distinguishing multiple primary lung cancers from intrapulmonary metastasis; however, there is a limitation, as the number of informative cases was small.

The overall survival also showed a significant difference between the patients with newly classified multiple primary lung cancers and those with newly classified intrapulmonary metastasis as shown in Figure 5. Conversely, in patients with their clinical diagnosis based on Martini and Melamed's criteria, the 5-year survival rate in 50 patients with multiple primary lung cancers was 61.7%, whereas it was 46.2% in 20 with intrapulmonary metastasis, and showed no statistical significance (P = .069). These results suggest that the present proposal for the discrimination of multiple primary lung cancers from intrapulmonary metastasis based on the difference in the expression of the 4 proteins is applicable in clinical setting. However, because the number of cases evaluated in this study was relatively small, further study in larger cohorts is recommended to validate these results.

Conclusion

The results of this study suggest that not only mutational analyses of genes but also immunohistochemical analyses of differential protein expression profiles of multiple genes can be used to distinguish multiple primary lung cancers from intrapulmonary metastasis, which, in turn, can be used for the selection of patients with a high risk of a poor prognosis.

Acknowledgements

We thank Dr. Brian Quinn for critical comments in preparing this manuscript, and we also thank Dr. Masaru Morita, Ms. Yukiko Goto, and Misako Fukumoto for critical comments and technical assistance.

Conflict of Interest Disclosures

This work was supported in part by Grants-in-Aid for Scientific Research (Nos. 17390388, 17591451, and 18659410) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

Ancillary