Micrometastatic Tumor Cells in the Bone Marrow of Patients With Non-Small Cell Lung Cancer

Transcription

1 Micrometastatic Tumor Cells in the Bone Marrow of Patients With Non-Small Cell Lung Cancer Akira Ohgarni, MD, Tetsuya Mitsudomi, MD, Kenji Sugio, MD, Tohru Tsuda, MD, Tsunehiro Oyama, MD, Kinue Nishida, Toshihiro Osaki, MD, and Kosei Yasumoto, MD Department of Surgery II and Institute of Ecological Sciences, University of Occupational and Environmental Health, School of Medicine, Kitakyushu, Japan Background. This study was designed to evaluate the incidence and clinical implications of detection of micrometastatic cancer cells in bone marrow aspirates of patients with operable non-small cell lung cancer. The relationship between micrometastatic cells and p53 overexpression in the primary tumor was also assessed. Methods. Thirty-nine patients with stages I through III non-small cell lung cancer who underwent curative resection were entered into this study. Immunohistochemistry with monoclonal antibody to cytokeratin 18 was used to detect tumor cells in bone marrow. Immunostaining of p53 protein in the corresponding primary tumors was also done. Results. Cytokeratin 18-positive cells were detected in 15 (39%) of the 39 patients. Overexpression of p53 was associated with positivity of the tumor cells in the bone marrow at borderline significance (14/29 versus 1/10; p = ). The patients with cytokeratin 18-positive cells in bone marrow demonstrated a significantly earlier recurrence than those without such cells (p = , log-rank test). Conclusions. Micrometastatic cancer cells were frequently present in bone marrow of patients with operable non-small cell lung cancer and may be a significant predictor of early recurrence. Further evaluation of this method may be useful in identifying patients with nonsmall cell lung cancer who are most likely to benefit from adjuvant chemotherapy. (Ann Thorac Surg 1997;64:363-7) 1997 by The Society of Thoracic Surgeons ung cancer is the leading cause of cancer-related L deaths in North America, and it is expected to soon become the leading cause also in Japan [1, 2]. Lung cancer is divided into two morphologic types: small cell lung cancer and non-small cell lung cancer (NSCLC) [2]. About 30% of patients with NSCLC have localized disease, and successful surgical management with longterm disease control is generally restricted to this group of early-stage patients [2]. However, even after a "curative" resection for stage I disease, ie, tumors with no invasion of the adjacent structures and no metastases to the regional lymph nodes or distant organs, about 30% of patients have recurrence and eventually die of the disease [2, 3]. About two thirds of such recurrences are in distant organs such as the brain, contralateral lung, and bone as a result of the hematogenous spread of cancer cells [3, 4]. In other words, recurrence of the cancer is simply a clinical manifestation of metastatic disease that is present but occult and undetected at the time of surgical intervention. Therefore, it is of great importance to develop a technique for detecting these micrometastatic cancer cells that are missed by conventional diagnostic methods to be able to provide a more precise prognosis. Over the past 10 years, several attempts have been made to detect micrometastatic cells in lymph nodes [5], Accepted for publication March 10, Address reprint requests to Dr Ohgami, Department of Surgery II, University of Occupational and Environmental Health, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 807, Japan (e-maih gamisan@med.uoehu.ac.jp). bone marrow (BM) [6-9], or peripheral blood [10, 11] by immunohistochemistry or a polymerase chain reactionbased assay for various types of human cancers. In about 30% to 50% of patients with operable cancer of the breast [8], colon [12], or lung [13-15], occult metastatic cells are detected in BM. Most investigators [8, 12-15] report that the detection of these micrometastatic cells is a poor prognostic indicator. In this study, we examined the incidence and prognostic implications of micrometastatic cancer cells in the iliac BM of 39 patients with stages I through III NSCLC. We took advantage of immunohistochemical staining of cytokeratin 18 (CK18) as a panepithelial marker [14-16]. We also examined primary tumors for any abnormal accumulation of p53 protein to see if a mutation of this protein is related to the detection of micrometastatic cancer cells in the BM, as we [17, 18] have found that a mutation of the p53 tumor suppressor gene is associated with a poor prognosis in these patients. Material and Methods Thirty-nine patients with histologically proven NSCLC who underwent a potentially curative pulmonary resection at the Department of Surgery II, University of Occupational and Environmental Health University Hospital, form the study group. Twenty-two patients had adenocarcinoma, 16 had squamous cell carcinoma, and 1 patient had large cell carcinoma. After extensive diagnostic procedures, including chest roentgenography, computed tomography of the chest and brain, ultrasonogra by The Society of Thoracic Surgeons /97/$17.00 Published by Elsevier Science Inc PII S (97)

2 364 OHGAMI ET AL Ann Thorac Surg BONE MARROW MICROMETASTASES IN NSCLC 1997;64:363-7 phy of the abdomen, and a bone scan, clinical stages were determined according to the TNM classification of the International Union Against Cancer (modified in 1986) [19]. Twenty-two patients had stage I disease, 4 had stage II, 9 had stage IIIA, and 4 had stage IIIB. As negative controls, we studied 5 patients who were suspected of having lung cancer preoperatively but later benign lung disease (tuberculosis in 4 and pulmonary abscess in 1) was diagnosed pathologically. For the postoperative follow-up, the patients were asked to visit our clinic to undergo chest roentgenography and tumor marker examinations every month for the first year, every other month for the second year, and every 3 months thereafter. Computed tomographic scanning and a bone scintiscan were performed every 6 months postoperatively. The median follow-up was 140 days (range, 30 to 910 days). Specimens Informed consent was obtained from all patients. Approximately 5 ml of BM aspirate was obtained from one site in the upper iliac crest within a week before operation and yielded an average of 1.0 x 107 nucleated cells. Mononuclear cells were isolated by density centrifugation with FICOLL-HYPAQUE (Mono-Poly Resolving Medium; Dainippon Pharmaceutical Co, Ltd, Osaka, Japan). The interface cells were washed by centrifugation and then resuspended in phosphate-buffered saline solution. Finally, the cells were cytocentrifuged at 1,300 rpm on glass slides in a cytocentrifuge (Auto Smear; Sakura Seiki Co, Tokyo, Japan) after the number of ceils was counted (average number of cells per slide, 3.6 x 105). The slides were either stained immediately or stored at -80 C after overnight air-drying and 4% paraformaldehyde-acetone fixation for 90 seconds. Five slides were routinely examined for each patient. Resected lung cancer specimens were fixed in 20% formalin for 3 days and embedded in paraffin. For the histologic study, sections were stained with hematoxylin and eosin. As a positive control in 17 patients, touch preparations of malignant cells for staining of CK18 were made by lightly pressing the cut surface of a block of fresh tissue to the slide. Immunohistochemistry of BM Aspirates or Touch Preparations for CK18 The slides of the BM aspirates or touch preparations were stained for CK18 with the alkaline phosphatase-antialkaline phosphatase method using a commercially available kit (DAKO, Tokyo, Japan). The primary antibody was CK2 (Boehringer Mannheim Biochemica, Mannheim, Germany), 2.0/~g/mL, raised against CK18. Staining was done according to the procedures recommended by the manufacturer. Counterstaining was briefly performed with Mayer's hematoxylin. Immunohistochemistry of Primary Tumors for p53 Protein Immunohistochemistry for p53 protein was performed as described previously [18]. Briefly, the sections were im- mersed in a citrate buffer and irradiated five times for 5 minutes each in a domestic microwave oven for antigen retrieval. They were then stained with the mouse monoclonal antibody DO-1 (Oncogene Science Inc, Cambridge, MA) against the p53 protein using a Labeled Streptavidin Biotin kit (DAKO). The sections were subsequently examined for nuclear staining under a light microscope. Five hundred tumor cells in the primary section were examined to estimate the percentage of p53-positive cells. Statistical Analysis A comparison of the proportion was done using the two-tailed Fisher's exact test. The Kaplan-Meier method was used to estimate the probability of survival as a function of time, and survival differences were analyzed by the log-rank test. For patients with lung cancer recurrence, the time to the first recurrence was calculated as the number of days from pulmonary resection to first documentation of recurrence. Results Detection of CK18-Positive Cells in BM Cytokeratin 18-positive cells were detected in densities ranging from 1/ nucleated cells to 1/ nucleated cells in BM aspirates from 15 (39%) of the 39 patients with NSCLC. None of the 5 patients with benign lung disease had any CK18-positive cells in their BM. All 17 touch preparations of the primary lung cancer specimens were positive for CK18. Table 1 summarizes the relationships between the presence of CK18-positive cells in BM and various clinical features. No association was found between the detection of CK18-positive cells and age, sex, histology or stage of disease. However, it was noteworthy that 6 (27%) of 22 patients with stage I disease had CK18- positive cells in iliac BM. Cytokeratin 18-positive cells were also detected in BM of 2 of 4 patients in stage II and 7 (54%) of 13 in stage III. Moreover, the incidence of CK18-positive cells in BM aspirates tended to be associated with the N factor. In patients with NO disease, the incidence was 6 (26%) of 23 versus 9 (56%) of 16 patients with N1 through N3 disease (p = ). Association Between Micrometastatic Cells and p53 Staining We examined resected lung tumors for p53 protein expression with immunohistochemistry. Of the 39 tumor specimens, 29 (74%) were positive for p53. The patients with a p53-positive tumor had a higher incidence of CK18-positive ceils in BM (14/29, 48%) than did those without p53 overexpression (1/10, 10%) at borderline significance (p = ). Prognostic Significance of Detection of CK18-Positive Cells We then analyzed the prognostic significance of the detection of CK18-positive ceils. The patients who had CK18 cells in BM demonstrated recurrent disease signif-

3 Ann Thorac Surg OHGAMI ET AL ;64:363-7 BONE MARROW MICROMETASTASES IN NSCLC icantly earlier than those who were free from these cells (p = ) (Fig 1A). In the subset of 26 patients with apparently earlier disease (stage I and II), the presence of occult CK18-positive cells in BM was also associated with earlier recurrence (p = ) (Fig 1B). However, in this cohort, p53 overexpression was not predictive of earlier recurrence (p = ) (Fig 1C). We also noted that the presence of CK18 cells in BM was predictive of the pattern of recurrence. Four patients with recurrence through hematogenous spread (contralateral lung in 2, brain in 1, and bone in 1) had CK18-positive cells in BM, whereas 2 patients who had regional lymph node recurrence were CK18 negative. The remaining patient (CK18-positive cells in BM) showed an abnormal elevation of the tumor marker (carcinoembryonic antigen) after operation, and we thus regarded that as recurrence, although the evidence had not yet been revealed by other methods. Comment An important question is how specific is CK18 staining for micrometastatic cancer cells in BM. Pantel and coauthors [20] reported that of 215 patients without clinical evidence of malignancy, only 6 were positive for CK18 cells. Our results showed that none of the 5 patients with benign lung disease had any CK18-positive ceils in BM. On the basis of these results, we think it reasonable to assume that most CK18-positive cells in BM, if not all, are micrometastatic cancer cells. We detected micrometastatic cancer cells in patients with apparently localized lung cancer who were thus lo0- Estimated 80- disease free ' rate (%) 60- B 20- Oq 100" 80- Estimated disease free rate (%) " Estimated disease free 80" rate (%) ] CK (-)in BM (n=24) 1 CK (+) in BM (n=15) *p= Cox-Mantel Test CK (-) in BM (n=18) 3 [3 o o L CK (+) in BM (n=8) i *,,..,,,,.,,., *p= Cox-Mantel Test ~4~_ L = a p53(-) in the primary tumor (n=lo) [] [] N.S. ~3 ooo o o o p53 (+) in the primary tumor (n=29) 20- Table 1. Relationship Between Detection of Cytokeratin 18- Positive Cells in Bone Marrow and Various Clinical Features a No. of No. With CK Variable Patients Positive-Cells p Value b Age (y) < (50) > (29) Sex Male (45) Female 10 2 (20) Histology Squamous 16 7 (44) Nonsquamous 23 8 (35) Stage I and II 26 8 (31) III 13 7 (54) T status > T1 and T (38) T3 and T4 7 3 (43) N status NO 23 6 (26) N1 through N (56) Numbers in parentheses are percentages. tained with the two-tailed Fisher's exact test. CK = cytokeratin. b The p value was ob-,...,..,..,,.., Fig 1. Disease-free interval in patients with non-small cell lung cancer who underwent pulmonary resection. (A) Detection of cytokeratin 18 (CK)-positive cells in bone marrow (BM) (stage I through stage III). (B) Detection of CK18-positive cells in BM (stages I and II). (C) Overexpression of p53 protein in primary tumor. (N.S. = not significant.) candidates for a potentially curative resection. In this study, 8 (31%) of the 26 patients in stages I and II had micrometastatic cancer cells in BM compared with 7 (54%) of 13 patients in stage III. Cote and associates [13] reported similar results in that micrometastatic cells in BM were detected in 5 (29%) of 17 patients with stage I or II NSCLC and in 12 (46%) of 26 with stage III disease. The incidence of micrometastatic cell detection in BM appears to be dependent on extension of the tumor. Concerning lymph node involvement, the incidence of CK18-positive cells was lower in patients with NO disease (6/23, 26%) than in those with N1 through N3 disease (9/16, 56%) in our study. However, Pantel and colleagues [20] found no correlation between the detection of CK18 cells in BM and lymph node involvement (54% in NO versus 65% in N1 or N2). They speculated that

4 366 OHGAMI ET AL Ann Thorac Surg BONE MARROW MICROMETASTASES IN NSCLC 1997;64:363-7 different mechanisms exist for the homing of tumor cells to BM tissue compared with lymphoid tissue. Another important question is whether or not the detection of these micrometastatic cancer cells in the BM is predictive of early disease recurrence or poor survival of the patient. We found that patients who had CK18- positive cells in iliac BM had a significantly shorter disease-free interval and that this was also the case in the subset of patients in stages I and II; this finding is in accordance with previous reports showing that patients with NSCLC who display cytokeratin-positive cells in BM demonstrate a significantly shorter disease-free interval. However, Heiss and associates [21l reported that the detection of micrometastatic cells alone was not a poor prognostic sign but that an increasing number of such cells by serial BM aspirates, accompanied with an expression of urokinase-type plasminogen activator receptor, was significantly correlated with a poor survival. In CK18-positive patients in our study, the pattern of recurrence seemed to be hematogenous, and only 1 patient of 4 patients with recurrence through hematogenous spread had bone metastasis as an initial site of failure. The existence of micrometastatic cells in BM may not directly represent a very early phase of bone metastasis, but CK18 cells in BM may be a predictor of distant metastases. It is known that the metastatic process consists of several stages (ie, invasion into the vessels, migration, adhesion to the endothelium, extravasation, and colonization), all of which are indispensable for the development of clinically overt metastases. The detection of CK18-positive cells in BM may indicate that some micrometastatic cancer cells have escaped from the primary tumor and are circulating systemically. Further study is needed concerning the qualitative evaluation of micrometastatic cancer cells with respect to metastatic potential, eg, the expression of oncogene product. It is now clear that human cancer develops and progresses through an accumulation of multiple genetic changes. In NSCLC, alterations in various oncogenes or tumor suppressor genes such as ras, my, p53, rib, p16, and p15 have been reported [22]. We [17, 18] have shown that the mutation of the p53 gene or overexpression of the p53 protein is associated with a poor prognosis in a subset of patients with NSCLC, although other investigators [23, 24] report opposite results. Overexpression of the p53 protein is not considered to be associated with stage of the disease determined by conventional methods [17, 18, 25], and therefore the mechanism or mechanisms through which the overexpression of p53 contributes to the poor outcome remains to be clarified. In this study, patients with tumors showing p53 overexpression had a tendency toward a higher incidence of micrometastatic cancer cells in BM (48% versus 10%; p = ). This may at least partially account for the effect of p53 mutation on patient survival. In conclusion, micrometastatic cancer cells in iliac BM were found in a significant proportion of patients with operable NSCLC. The detection of these cells in the BM appeared to be predictive of early recurrence, particularly bloodborne metastases, and thus of a poor prognos- tic outcome as well. Overexpression of the p53 protein in primary tumors appeared to be correlated with an increased incidence of these micrometastatic cells in the BM. The use of systemic chemotherapy after operation has been shown to be of limited usefulness in the treatment of NSCLC [26]. However, by monitoring the micrometastatic cancer cells in BM, it would be possible to select patients who might benefit from adjuvant chemotherapy. To determine the treatment of the patient with lung cancer, the micrometastatic status as well as the conventional TNM system would be considered. For example, adjuvant therapy for stage I CK18-positive patients or aggressive surgical resection in stages IIIA and IIIB CK18-negative patients should be advocated. This work was supported in part by grants-in-aid for scientific research (no and no ) from the Ministry of Education, Science and Culture, Japan. We thank Dr Tetsuo Hamada and the staff of the Department of Surgical Pathology, University of Occupational and Environmental Health, University Hospital, for their technical advice and assistance. References 1. Statistics and Information Department. Vital statistics 1990 Japan. Tokyo, Japan: Ministry of Health and Welfare, 1991: Minna JD, Pass H, Glatstein E, Ihde DC. Cancer of the lung. In: DeVita VT, Hellman S, Rosenberg SA, eds. Cancer: principles and practice of oncology. Philadelphia: JB Lippincott, 1989: Little AG, DeMeester TR, Ferguson MK, et al. Modified stage I (TINOM0, T2NOM0), nonsmall cell lung cancer: treatment results, recurrence patterns, and adjuvant immunotherapy. Surgery 1986;100: Martini N, Bains MS, Burt ME, et al. Incidence of local recurrence and second primary tumors in resected stage I lung cancer. J Thorac Cardiovasc Surg 1995;109: Chen ZL, Perez S, Holmes EC, et al. Frequency and distribution of occult micrometastasis in lymph nodes of patients with non small cell lung carcinoma. J Natl Cancer Inst 1993; 85: Schlimok G, Funke I, Holzmann B, et al. Micrometastatic cancer cells in bone marrow: in vitro detection with anticytokeratin and in vivo labeling with anti-17-1a. Proc Natl Acad Sci USA 1987;84: Schlimok G, Funke I, Pantel K, et al. Micrometastatic tumor cells in bone marrow of patients with gastric cancer: methodological aspects of detection and prognostic significance. Eur J Cancer 1991;27: Redding HW, Coombes RC, Monaghan P. Detection of micrometastases in patients with primary breast cancer. Lancet 1983;2: Pantel K, Schlimok G, Braun S, et al. Differential expression of proliferation-associated molecules in individual micrometastatic carcinoma cells. J Natl Cancer Inst 1993;85: Burchill SA, Bradbury MF, Pittman K, et al. Detection of epithelial cancer cells in peripheral blood by reverse transcriptase-polymerase chain reaction. Br J Cancer 1995;71: Moreno JG, Croce CM, Fischer R, et al. Detection of hematogenous micrometastasis in patients with prostate cancer. Cancer Res 1992;52: Lindemann F, Schlimok G, Dirschedl P, Witte J, Riethmuller G. Prognostic significance of micrometastatic tumour cells in

5 Ann Thorac Surg OHGAMI ET AL ;64: BONE MARROW MICROMETASTASES IN NSCLC bone marrow of colorectal cancer patients. Lancet 1992;340: Cote RJ, Beattie EJ, Chaiwun B, et al. Detection of occult bone marrow micrometastases in patients with operable lung carcinoma. Ann Surg 1995;222: Pantel K, Izbicki JR, Angstwurm M, et al. Immunocytological detection of bone marrow micrometastasis in operable nonsmall cell lung cancer. Cancer Res 1993;53: Pantel K, Dickmanns A, Zippelius A, et al. Establishment of micrometastatic carcinoma cell lines: a novel source of tumor cell vaccines. J Natl Cancer Inst 1995;87: Pantel K, Schlimok G, Angstwurm M, et al. Methodological analysis of immunocytochemical screening for disseminated epithelial tumor cells in bone marrow. J Hematother 1994;3: Mitsudomi T, Oyama T, Kusano T, et al. p53 Gene mutations as a predictor of poor prognosis in patients with non-small cell lung cancer. J Natl Cancer Inst 1993;85: Mitsudomi T, Oyama T, Nishida K, et al. p53 Nuclear immunostaining and gene mutations in non-small cell lung cancer and their effect on patient survival. Ann Oncol 1995;6(Suppl 3):$ Mountain CF. A new international staging system for lung cancer. Chest 1986;89(Suppl):225S-33S. 20. Pantel K, Izbicki J, Passlick B, et al. Frequency and prognostic significance of isolated tumour cells in bone marrow of patients with non-small-cell lung cancer without overt metastases. Lancet 1996;347: Heiss MM, Alligayer H, Gruetzner KU, et al. Individual development and upa-receptor expression of disseminated tumour cells in bone marrow. Nature Med 1995;1: Minna JD. The molecular biology of lung cancer pathogenesis. Chest 1993;103:445S-56S. 23. Lee JS, Yoon A, Kalapurakal SK, et al. Expression of p53 oncoprotein in non-small-cell lung cancer: a favorable prognostic factor. J Clin Oncol 1995;13: Passlick B, Izbicki JR, Haussinger IG Thetter O, Pantel K. Immunohistochemical detection of P53 protein is not associated with a poor prognosis in non-small-cell lung cancer. J Thorac Cardiovasc Surg 1995;109: Nishio M, Koshikawa T, Kuroishi T, et al. Prognostic significance of abnormal p53 accumulation in primary, resected non-small-cell lung cancers. J Clin Oncol 1996;14: Non-Small Cell Lung Cancer Collaborative Group. Chemotherapy in non-small cell lung cancer: a meta-analysis using updated data on individual patients from 52 randomised clinical trials. BMJ 1995;311: INVITED COMMENTARY This study by Ohgami and associates can be added to a growing literature in non-small cell lung cancer and a large literature in other cancers regarding the clinical usefulness of the detection of occult bone marrow metastasis in patients with early-stage disease. The current study has shown once again that patients with operable non-small cell lung carcinoma who demonstrate the presence of occult metastasis in bone marrow have significantly earlier times to recurrence. Importantly, this association was also true in patients with stages I and II non-small cell lung cancer, the group for which nonsurgical therapeutic decisions is most difficult. In addition, this study shows that the presence of occult bone marrow metastasis is associated with p53 tumor suppressor status of the primary tumor. Thus, occult bone marrow metastasis appears to be associated with molecular alterations known to be important in tumor progression. Finally, the study by Ohgami and co-workers suggests that patients with occult bone marrow metastasis show a pattern of overt metastasis consistent with hematogenous or systemic (as opposed to regional) tumor spread. Adjuvant chemotherapy is presumably useful because it can control occult systemic disease; thus, the association of occult bone marrow metastasis with overt systemic metastasis may serve to further indicate that patients with occult bone marrow metastasis are those who might most benefit from adjuvant chemotherapy. This study and all other studies regarding the clinical significance of occult bone marrow metastasis in patients with non-small cell lung carcinoma have used immunohistochemistry to identify the tumor cells. This method takes advantage of the differential expression of antigens by epithelial (tumor) cells versus normal bone-marrow elements and also uses morphologic differences between tumor and normal cells. Molecular methods, in particular amplification of epithelial-specific RNA message using the polymerase chain reaction (reverse transcriptasepolymerase chain reaction), have been touted as superior alternatives to immunohistochemical methods for detecting occult tumor dissemination. However, with the possible exception of prostate carcinoma, reverse transcriptase-polymerase chain reaction methods for the detection of occult bone marrow metastasis have not been shown to provide clinically relevant data in patients with epithelial tumors. Studies from our group and others indicate that this is due to the lack of specificity of the reverse transcriptase-polymerase chain reaction assay for those epithelial markers tested to date. The next steps are clear. The detection of occult bone marrow metastasis in patients with non-small cell lung cancer (and other cancers) has been demonstrated to provide clinically relevant data and should be considered when it might have an appropriate impact on treatment decisions. Thus, the detection of occult bone marrow metastasis may be particularly relevant in the assessment of patients with stage I or II non-small cell lung carcinoma. Clinical trials randomizing patients with earlystage lung carcinoma on the basis of bone marrow status will provide definitive assessment of the general applicability of this potentially important technique. Richard J. Cote, MD Department of Pathology and Laboratory Medicine University of Southern California School of Medicine USC/Norris Comprehensive Cancer Center 1441 Eastlake Ave Los Angeles, CA by The Society of Thoracic Surgeons /$17.00 Published by Elsevier Science Inc PII S (97)