O-229 Prognostic value of cytokeratin-positive cells in the bone marrow and lymph nodes of patients with resected non-small cell lung cancer: a multicenter prospective study

Oral Sessions/Limits of the TNM Staging System purpose of this study was to determine the relationship between tumor size, and survival in patients with pathologic stage IA NSCLC. Methods: We conducted a retrospective review of 246 consecutive, surgically treated patients with pathologic stage IA NSCLC. Eligible patients were identified from the tumor registries and pathology records. Follow-up was obtained from the surgical database and primary physicians’ records, Results: Eighty six patients had tumors 5 1.5 cm while 160 patients had tumors 1.6 cm-30 cm. The median follow-up time is 60 months, The patients withpawith the tUmOrS i i .5 Cm had an improved outcome when compared tients with tumors 1.6-3.0 cm. The S-year recurrence free survival was 81.5% and 70.9% respectively (p=.O3) and the 5-year overall survival was 85.5% and 78.6% respectively (p=.O5). In the multivariate analysis, tumor size was an independent prognostic factor for survival, Parameter Tumor Size Differentiation Histology Gt?lldW Mediastinoscopy

Variables 4= 1scm I-6-3.0 cm Well Moderate P00r Adenocarcinoma Squamous cell Large Cell Male Female Performed Not Performed

%5-yr RFS (n) 81.5 (80) 70.9 (149) 84.9 (33) 77.5 (105) 67.6 (91) 75.2 (162) 73.1 (37) 72.2 (22) 74.4 (135) 74.8 (94) 77.3 (95) 73.8 (117)

,, .03 .08 .96 .59 .76

%5-yr OS (n) 85.5 (85) 78.6 (157) 88.1 (34) 86.3 (112) 75.5 (96) 82.6 (172) 74.9 (38) 82.5 (24) 80.6 (139) 81.7 (103) 84.4 (99) 79.1 (126)

p .05 .08 .85 230 .25

I0 228

S67

Analysis of telomerase expression in circulating non-small cell lung cancer cells in relation to patient survival

Jennifer L. Hess’, Laurence Austin Dovle’, David M. Lu’ , Chaunfo Guo’ , Jaime C. Hey2, Petr F. Hausner’, Martin J. Edelman’ , Mark J. Krasna’ , William E. Highsmith’. ’ Greenebaum Cancer Center, University of Maryland Medical Center, Baltimore, USA; 2 Greenebaum Cancer Center, University of Mary/and Media/ Cenfer, Baltimore, USA Analysis of telomerase expression in circulating non-small cell lung cancer cells in relation to patient survival. J.L. Hess, L.A. Doyle, D.M. Lu, C. Guo, J.C. Hey, P. Hausner, M.J. Edelman, M.J. Krasna, W.E. Highsmith. Greenebaum Cancer Center, Baltimore, MD We have analyzed telomerase activity in epithelial cell samples isolated from peripheral blood from 85 patients with lung cancer and 50 controls. Epithelial cells were isolated from peripheral blood mononuclear cells with anti-EpCAM coated magnetic beads. Telomerase activity from cell lysates was determined using the telomerase repeat amplification protocol (TRAP) assay. Classification as positive or negative was by visual inspection of polyacrylamide gels for the presence of the characteristic 6 base pair TRAP ladder. To avoid PCR contamination, each sample was analyzed in triplicate, with one of the replicates heated to 85°C for 15 minutes to destroy telomerase activity as a sample blank. Using five or more TRAP bands as a positive result, none of the 50 samples from non-cancer disease controls expressed telomerase. Telomerase expression was seen in 37/85 lung cancer patients at the time of diagnosis (44%), but in up to 79% of patients throughout the course of their disease. The rates of positive telomerase expressed were similar between lung cancer samples from patients with adenocarcinoma versus squamous histology. There was a trend by chi-square test between positive telomerase activity and advanced stages of non-small cell lung cancer (P=O.O7). The median time to relapse after initial therapy was 6.1 months in telomerase-positive patients versus 11.4 months in telomerase-negative patients (P=O.O5). The median survival time of 37 lung cancer patients with a positive initial test result was 10.5 months, while the median survival time of 48 patients with a negative test result was 20.5 months (P=O.O5). The one-year survival was 48% for telomerase-positive patients and 67% for telomerase-negative patients. The three-year survival was 20% for telomerase-positive patients and 42% for telomerase-negative patients. Multivariate analysis of overall survival using Cox’s proportional hazard model showed a non-significant hazard ratio of 1 .I6 for telomerase expression. Supported by NCI EDRN grant UOI CA84988

I0 229

Prognostic value of cytokeratin-positive cells in the bone marrow and lymph nodes of patients with resected non-small cell lung cancer: a multicenter prospective study

Kenii Suaio’, Toshihiro Osaki’, Yoh Watanabe3, Harufumi Kato4, Tetsuya Mitsudomi5, Akira Nagashima s, Kosei Yasumoto’. ‘2nd Dept. of Surgery, University of Occuptional and Environmental Health, Kifakyushu, Japan; ‘2nd Dept. of Surgery, University of Occupational and Environmental Health, Kitakyushu, Japan; 3 Debt. of Surgery I, Kanazawa University, Kanazawa, Japan; 4 Dept of Surgery Tokyo Medical University: Tokyo, Japan; 5 Dept. of Thoracic Surgery, Aichi Cancer Center, Nagoya, Japan; 6 Department of Chest Surgery, Kitakyushu Municipal Medical Center, Kitakyushu, Japan: 72nd Dept. of Surgery University of Occupational and Environmental Health, and Chairman of the Cooperative Project NO.24 Group of the Japanese Foundation for Multidisciplinary Treatment of Cancer(JFMC), Kitakyushu, Japan

Conclusions: Tumor size is a prognostic factor in pathologic stage IA NSCLC treated with adequate surgical resection. Such improvement, not observed in some earlier reports, may represent the effect of improved imaging and preoperative surgical techniques. This finding requires confirmation in prospective series, but is important since it makes a case for further subdividing Tl lesions in resected NSCLC.

Purpose: Occult micrometastatic tumor cells, which are not detected by current clinical staging examinations and conventional histopathologic methods such as HE staining, have already spread to the distant mesenchymal organs or the regional lymph nodes, at the time of surgery. This study was designed to prospectively substantiate the prognostic value of cytokeratin (CK)-positive cells in the bone marrow (EM) and regional lymph nodes (LNs) in resected non-small cell lung cancer (NSCLC) patients from a large population within a multicenter study. Patients and Methods: This prospective study was performed by the Cooperative Project NO.24 Group of the Japanese Foundation for Multidisciplinary Treatment of Cancer. The study population consisted of 351 patients with stage I-IIIA NSCLC, from 15 Japanese institutes, between February 1997 and September 1998. BM aspirates were stained immunocytochemically with the anti-CK Antibody, CK2. The hilar and mediastinal LNs of 216 patients with stage I NSCLC were stained immunohistochemically with the anti-CK antibody, AEI/AE3. Results: CK-positive cells were detected in 112 (31.9%) of the 351 BM aspirates. The frequEhcy of CK-positive cells showed no differences among pathological stages. The patients with CK-positive cells in the BM had a tendency to have shorter survival periods than those without CK-positive cells (p=O.O76). Although the presence of CK-positive cells in the BM of patients with stage I

S68

Oral Sessions/Systematic

Reviews of Evidence

did not allow the prediction of overall survival, it reduced the overall survival significantly in patients with stage II-IIIA (p=O.O47). CK-positive cells in the LNs were detected in 34 (15.7%) of 216 patients with stage I. The patients with CKpositive cells in the LNs had a poor prognosis by both univariate (p=O.O04) and multivariate analyses (p=O.O18). Conclusion: The presence of CK-positive cells in the BM was related to a poor prognosis for patients with stages II-IIIA NSCLC; however, it did not predict the prognosis of patients with stage I. For stage I NSCLC, the detection of CKpositive cells in the LNs implied a poor prognosis for the patients.

in both arms, 5 which tested dose intensity in general, and 17 that were confounded by later cross-over of treatments, or different second-line treatments. Conclusion: Clinical research in SCLC has been addressed in a largely ad hoc manner by individual research teams, with little co-ordination between groups globally. The large proportion of confounded trials has led to a confused picture, hampering a coherent synthesis of current knowledge. A global, coordinated research strategy is needed to optimise resource use and ensure that research builds on what is known by addressing the most relevant questions in soundly designed trials.

0 230 I..:.

I 0 232

Lymphatic metastatic patterns and its prognostic lmpllcatlons in NSCLC: A retrospective analyses in n=2205 curatively resected patients

H. Dienemann,

Heidelberg

Germany,

H. Biilzebru, Th. Muley, H. Hoffmann.

Background: The proper extent of lymph node dissection, the anatomical definition of lymph node stations, and the prognostic significance of specific lymph node involvement continues to be a controversial topic in surgery for lung cancer. Methods: The lymphatic metastatic patterns according to site and extent were retrospectively analyzed in n=2205 curatively resected (RO) patients with NSCLC who underwent lung resection with systematic hilar and mediastinal lymph node dissection. Results: One thousand two hundred twenty fife patients had lymph node involvement: n=679 at Ni nodes, n=406 at N2 nodes, and n=140 at N3 nodes. Nine hundred eighty patients were node negative. The prevalence of lymph node metastases did not differ significantly with tumor histology (Squamous: n=1035, adeno: n=926, large cell: n=131, others: n=113), but was higher in 1 T2 tumors (59.5%) as compared to Tl tumors (42.7%). Tumor localization showed a predominance of upper lobe tumors (n=1265,57%), as compared to lower lobe tumors (n=626, 28.4%), middle lobe tumors (n=76, 3.5%) and centrally located tumors (n=238, 10.8%). Lobe specific analyses demonstrated Nl disease in 27.8% of upper lobe tumors, 28.9% of middle lobe tumors, in 31.3% of lower lobe tumors, and in 45.8% of centrally located tumors. N2 disease was found in 18.0% of upper lobe tumors, 17.1% of middle lobe tumors, in 16.9% of lower lobe tumors, and in 24.8% of centrally located tumors. The 5-year survival rates correlated with the extent of lymph node involvement. Comment: Systematic nodal dissection appears to be the most important component of staging in NSCLC. The results of our retrospective analyses support the validity of the internationally approved lymph node map.

Can survival differences be found among cisplatin, carboplatin, and non-platinum regimens? Results of a comprehensive review of chemotherapy in advanced non-small-cell lung cancer (NSCLC) randomized trials including 12181 patients from 1882 -2002

Harrv Raftopoulos’ , Richard J. GraIla’, Fabrizzio Nelli’, Giulliana D’Auria’, Emilio Bria’ ’ Columbia University, New York, USA; 2 New York Lung Cancer Alliance, New York, USA in treatOur previous analysis using this methodology assessed the progress ing NSCLC since the Collaborative Group Meta-analysis (Proc Am Sot C/in Oncol 2002, Abs# 1284). The prior analysis demonstrated that cisplatin (DDP) based regimens yielded a 220% improvement in median survival over DDP as a single agent (with the exception of etoposide regimens). However, that initial analysis did not assess the impact of DDP dose intensity (DI). In a more recent analysis, both too low and too high DDP DI has been found. In nearly 9000 patients, the longest median survivals are achieved in studies with a DDP DI of 21-30mg/m*/week (most commonly delivered as 75-80mg/m’ q 3 weeks or lOOmg/m’ q 4weeks). Lower and higher DI ranges were associated with lesser survivals (9 vs 7 month median survivals). Current inclusion criteria for analy sis are: 1) randomized phase Ill trials with >I00 patients per arm, published from 1992 to 2002, 2) DDP single-agent or platinum-based regimens regarded as standard with old (etoposide, teniposide, vindesine, vinblastine, MIC, MVP) and newer (paclitaxel, gemcitabine, vinorelbine, docetaxel) drugs, 3) only trials with DDP DI regarded as optimal (21-30 mg/m2/week), 4) carboplatin and non-platinum combinations were included. 33 trials met inclusion criteria (12181 patients). Negative prognostic factors (stage IV, males, poor PS) were similar among the groups. Results are presented in the table below for older agents, or newer agents added to DDP or carboplatin, or for newer combination regimens without a platinum drug.

Regimen

THURSDAY, 14 AUGUST 2003

Systematic

I0 231

Reviews

of Evidence

The state of current randomised trials of chemotherapy for extensive small cell lung cancer (SCLC)

Elinor Thompson’, Yolanda Agra*, Marta Pelayo3, Montse Sacristan’, Fergus Macbeth4, Xavier Bonfill’. ’ lberoamerican Cochrane Centre, Barcelona, Spain; ‘Area 1, IMSALUD, Madrid, Madrid, Spain; 3Area IO, SERVASA, Valencia, Valencia, Spain; 4 Velindre Hospital, Cardie Wales, UK

Background: Over 120 randomised controlled trials (RCTs) have investigated chemotherapy for extensive SCLC. Most trials have found only modest differences between regimens so pooling of data from different trials is necessary to determine effects more conclusively. Grouping of similar trials is complex since trial objectives vary greatly and the use of terms such as dose intensity, dose and cumulative dose has been inconsistent. Objective: To document the range of comparisons tested in RCTs of chemotherapy for extensive SCLC and to develop a system by which similar trials may be grouped for systematic review and meta-analysis. Methods: An exhaustive search was undertaken for all RCTs of chemotherapy (without chest radiotherapy) in extensive SCLC. Data on agents, doses and schedules used in each trial arm were extracted and the comparisons tested between arms grouped. Trials comparing more than one variable between arms were considered confounded. Results: Many combinations of agents, doses, regimens and schedules of chemotherapy have been compared. Preliminary analysis of 82 RCTs showed that 45 (55%) were confounded. Homogeneous groups of unconfounded trials compared: i) different aspects of dose intensity (17), ii) an alternating regimen against the same regimens given separately (8), iii) substitution of one agent or regimen for another (9), iv) number of agents or addition of agents (3). Confounded trials included 23 that differed in both dose and types of agents used

No. of patients

Median Survival

DDP single-agent DDP + etoposide DDP + old agents

794 379 769

6.6 8.3 8.8

New agents + DDP New agents + Carboplatin New agents without platins

6140 3080 1019

9.4 9.0 8.0

(months)

Concerning newer agents, with at least 1000 patients per group, median survivals are in the 8 to 9+ month range, with carboplatin results intermediate between DDP and non-platinum regimens. All combinations provide a >15% benefit over single agent DDP; when DDP is used in the optimal DI range, the survival benefit is at least 20%. By using a comprehensive review method, a timely analysis of results without positive publication bias issues (Simes et a/ Stat Med, 1987) is provided that illustrates the relative survival impact of commonly used regimens.

El0 233

Platinum-based versus non-platinum-based chemotherapy in advanced non-small cell lung cancer (NSCLC): A meta-analysis of the published literature

Giannicola D’Addario’, Melania Pintilie’, Thomas Cerny’ , Ronald Feld’, Natasha B. Leigh12, Frances A. Shepherdz. ’ Onkologie Kantonsspital, St, Gal/en, Switzerland; p Princess Margaret Hospital Toronto, Canada

Background: Platinum-based combination chemotherapy is the standard of care in the palliative treatment of good performance status patients with stage IIIB/IV NSCLC. The superiority of platinum-based regimens was shown in randomized trials and meta-analyses that mainly included the second-generation chemotherapy regimens. Third-generation agents such as the taxanes, gemcitabine and vinorelbine, may offer single-agent activity comparable to that of platinum compounds. This meta-analysis was undertaken to determine whether platinum compounds are still a mandatory part of chemotherapy in the treatment of advanced NSCLC. Methods: During October 2002, a search of Medline, Cancerlit and Embase for randomized trials comparing platinum-based with non-platinum-based