Received 6 March 2008; received in revised form 7 July 2008; accepted 14 July 2008; Available online 21 August 2008

Transcription

1 European Journal of Cardio-thoracic Surgery 34 (2008) Clinical implication and prognostic significance of standardised uptake value of primary non-small cell lung cancer on positron emission tomography: analysis of 176 cases Nael Al-Sarraf a,b,, Kathy Gately a, Julie Lucey c, Rashid Aziz a, Kishore Doddakula a, Lorraine Wilson c, Eillish McGovern a, Vincent Young a a Department of Cardiothoracic Surgery, St. James s Hospital, Dublin 8, Ireland b Department of Thoracic Surgery, Chest Disease Hospital, Kuwait c Department of Nuclear Medicine, Blackrock Clinic, Dublin, Ireland Received 6 March 2008; received in revised form 7 July 2008; accepted 14 July 2008; Available online 21 August 2008 Abstract Objective: We sought to assess the clinical implication and prognostic significance of maximum standardised uptake value (SUV max ) of primary non-small cell lung cancer (NSCLC) staged by integrated PET-CT. Methods: A retrospective review was carried out on 176 consecutive patients with histologically proven NSCLC who underwent staging with integrated PET-CT prior to curative intent surgical resection. SUV max of primary NSCLC were measured and correlated with tumour characteristics, lymph node involvement, surgical stage, type of surgical resection and survival following resection. Results: SUV max was significantly higher in centrally located tumours, tumours 4.0 cm, squamous cell subtype, poorly differentiated tumours, advanced T stage, advanced nodal stage, pleural invasion, and patients requiring complex surgical resection. SUV max value of 15 was the best discriminative cut-off value for survival generated by log-rank test. When patients were stratified based on this value, those with SUV max >15 were more likely to have centrally located tumours, squamous cell subtype, advanced T stage, advanced nodal stage, advanced American Joint Committee on Cancer (AJCC) stage, larger tumour size and required more advanced surgical resections than a simple lobectomy. Overall survival was significantly longer for patients with SUV max 15 than those with SUV max >15. Furthermore, nodal stage specific survival following resection (i.e. non-n2 and N2) were significantly better in patients with SUV max 15 than SUV max >15. Conclusion: SUV max correlates with tumour characteristics, surgical stage and prognosis following resection. SUV max may be a useful preoperative tool, in addition to other known prognostic markers, in allocating patients with potentially poor prognosis preoperatively to neoadjuvant chemotherapy prior to resection in order to improve their overall survival. Prospective and randomised trials are warranted. # 2008 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved. Keywords: Positron emission tomography; Non-small cell lung cancer; Standardized uptake value; Survival 1. Introduction Lung cancer remains the leading cause of death from all cancer-related mortalities and carries a dismal prognosis. Despite great advances in chemotherapy, surgery remains the only curative method. Therefore, adequate preoperative staging should be emphasised to determine the best treatment approach and to avoid inappropriate thoracotomies. Positron emission tomography (PET) is superior to conventional radiological staging such as CT in the preoperative staging of non-small cell lung cancer (NSCLC) [1]. In PETstudies, tumour metabolism is reflected by a semi quantitative measurement Presented in a poster form at the18th WSCTS World Congress, Kos Island, Greece, April 30 May 3, Corresponding author. Address: Department of Thoracic Surgery, Chest Disease Hospital, Al-Deya, P.O. Box 15179, Kuwait, Kuwait. Tel.: ; fax: address: trinityq8@hotmail.com (N. Al-Sarraf). referred to as standardised uptake value (SUV) which is calculated by a standard formula [2]. In this paper, we sought to assess the clinical implication and prognostic significance of maximum standardised uptake value (SUV max ) of primary NSCLC patients undergoing curative intent surgical resection. The advantage of preoperatively identifying patients with poor prognostic indicators is that these cases could be treated more aggressively in the first instance by neo-adjuvant chemotherapy prior to aggressive type surgical resection in the hope of improving their poor long-term outcome. 2. Patients and methods 2.1. Patients Between September 2004 and January 2007, 176 consecutive patients with histologically proven NSCLC who underwent /$ see front matter # 2008 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved. doi: /j.ejcts

2 N. Al-Sarraf et al. / European Journal of Cardio-thoracic Surgery 34 (2008) lung resection and systematic lymph node dissection (SLND) in the department of cardiothoracic surgery at St. James s hospital were retrospectively reviewed. All these patients underwent staging with integrated PET-CT(as an adjunct to CT) 4 weeks prior to lung resection. All clinical, operative, radiological and pathological findings were analysed retrospectively through the review of patients charts, pathological reports and radiological imaging. SUVs were measured prospectively in all patients. Histological classification of NSCLC was based on that of WHO classification. Pre- and postoperative staging was based on the tumour-node metastasis (TNM) staging system [3]. Diabetic patients with BM 8mmoll 1 (n = 16) and those who received neoadjuvant/induction chemotherapy (n = 18) over the study period were excluded. Only patients who underwent surgical resection were enrolled. This study was approved by the institutional review board PET-CT We have previously reported our technique in PET-CT scanning [4,5]. 18 FDG-PET-CTwas performed using integrated PET-CT scanner (Discovery ST, GE Medical systems). Patients were fasting for 6 h prior to the procedure and the images were obtained 1 h after intravenous administration of 18 FDG. Patients remained resting for an uptake period of 1 h. Images were obtained from base of the skull to mid-thigh level. Attenuation corrected images were obtained and interpreted by an experienced nuclear medicine physician. Primary lesions were considered pathological (i.e. positive) if a definite localised area of higher 18 FDG uptake than the surrounding normal tissue exists (excluding physiological uptake). Nodal uptake with a standardised uptake value (SUV max ) >2.5 were considered positive. For determination of SUV, a cylindrical region of interest (ROI) was manually placed over the tumour site on the hottest trans-axial slice. The activity concentration within the ROI was determined and expressed as the SUV, where SUV is the ratio of the activity in the tissue to the decay corrected activity injected into the patient. All SUV measurements were normalised for patient body weight. The maximum SUV within a region of interest (SUV max ) was used as our reference measurement Surgical resection and pathological examination Lung resection with SLND was carried out through posterolateral thoracotomy. The extent of SLND was previously reported [5,6]. Cervical mediastinoscopy was employed in cases where positive uptake is detected in mediastinal nodes (N2 or N3) with nodal SUV max >2.5 as previously reported [5,6]. Resected lung tissue and lymph nodes were examined by pulmonary histopathologist for the presence or absence of malignancy using standard techniques with immunohistochemistry performed as per pathologist discretion. Vascular invasion refers to the presence of malignant cells in the blood vessel lumen detected histologically following resection Data analysis The relationship between SUV max of primary NSCLC and the different variables were analysed using appropriate nonparametric statistics namely, the Mann Whitney, Kruskal Wallis, Fisher s exact test, or the chi-square test. ANOVA was used where appropriate. The log-rank test was used to assess the best discriminative cut-off value of SUV max for survival. Survival curves were generated using Kaplan Meier method and were measured from the date of resection. The software package SPSS version 11.0 (SPSS Inc., Chicago, IL, USA) was used to perform the statistical analysis. Results were considered statistically significant when a p value was < Results Our cohort consisted of 105 men and 71 women. Age ranged from years with a mean (1SD) of 65.1 years (8.8). The mean SUV max of primary NSCLC (1SD) in our series was 11.8 (5.9). Tumour size ranged from 1 cm to 12 cm with a mean (1SD) of 3.9 cm (1.9). A summary of patient characteristics, pathological characteristics and type of surgical resection with relation to SUV max is outlined in Tables 1 3. Two types of analysis were performed. The first examined the relation of different variables to SUV max (Tables 1 3). The second analysis stratified patients into two groups based on the result of the log-rank test (Fig. 1) which identified SUV max of 15 as a cut-off value for survival and is summarised in Table 4. Survival curves using Kaplan Meier methods (Fig. 2) were generated and stratified based on this cut-off value (i.e. SUV max of 15). Survival curves for nodal specific stage (i.e. non-n2 and N2) are shown in Fig. 3 (parts A and B, respectively) SUV max and tumour characteristics Centrally located tumours had a higher mean SUV max than peripheral ones (16.4 vs 10.5, p < ). There was no Table 1 Patient population and tumour characteristics with relation to SUV max Number of patients SUV max (mean SD) p value Overall N/A Range Gender Male Female Age (years) < Site Right Left Size (cm) < < Tumour location Central Peripheral < Lobar distribution Lower lobes Upper lobes Middle lobe (R) N/A: not applicable. SD: standard deviation.

3 894 N. Al-Sarraf et al. / European Journal of Cardio-thoracic Surgery 34 (2008) Table 2 Pathological characteristics of primary NSCLC with relation to SUV max Number of patients SUV max (mean SD) p value NSCLC subtype Adenocarcinoma Squamous Others < Differentiation Well Moderate Poor T stage T T T3/T < Node status Node negative (N0) Node positive (N1/N2) Nodal stage N N N Overall stage (AJCC) I II III + IV SD: standard deviation. significant difference in SUV max and the lobar distribution of tumours or their site (i.e. right vs left) (Table 1). Tumour size showed a correlation with SUV max. Larger tumours (4 cm) had a higher mean SUV max than smaller (<4 cm) tumours (14.4 vs 9.7, p < ). Adenocarcinoma subtype had a lower mean SUV max than squamous cell subtype (8.6 vs 14.8) and other subtypes (8.6 vs 11.8, p < ). Poorly differentiated tumours had higher mean SUV max than Table 3 Operative procedures and vascular invasion with relation to SUV max Number of patients SUV max (mean SD) p value Intratumoral vascular invasion Positive Negative Pleural invasion Uninvolved Parietal pleura Visceral pleura Surgical procedure Lobectomy/wedge Other resections a No resection b < Chest wall resection Yes No SD: Standard deviation. a Refers to four cases of open/close thoracotomy due to invasion into unresectable structures. b Refers to pneumonectomy/sleeve resection/bi-lobectomy. Fig. 1. Log-rank analysis for the best cut-off value of SUV max for survival. moderately and well-differentiated ones (12.8 vs 11.5 vs 8.8, p = 0.045) SUV max and tumour stage T1 lesions had lower mean SUV max than T2 (8.2 vs 12.9) and a lower mean SUV max than T3/T4 lesions (8.2 vs 14.2, p < ). Primary NSCLC with negative nodal disease had a lower mean SUV max than tumours with positive lymph node involvement (10.9 vs 13.1, p = 0.006). In addition, there was a trend of a steady increase in mean SUV max of primary NSCLC in N0 versus N1 (10.9 vs 14.2) and in N2 versus N0 (12.09 vs 10.9, p = 0.004). When applying the AJCC classification, stage I NSCLC had lower mean SUV max than stage II (10.7 vs 13.2) and III/IV (10.7 vs 12.9, p = 0.014) SUV max and operative procedures Tumours requiring complex surgical resections (including sleeve resection and pneumonectomy) had higher mean SUV max than those requiring simple lobectomy (15.2 vs 10.3, p < ). Open/close thoracotomy (by virtue of invasion into unresectable structures) occurred in four patients overall in our series, and their mean SUV max was significantly higher than patients requiring simple lobectomy (15.0 vs10.3, p < ). Chest wall resection was employed in seven cases overall and the distribution of SUV max in relation to chest wall invasion showed no significant difference SUV max and prognostic indicators We found no significant difference in mean SUV max in tumours with vascular invasion than those without it (Table 3). Parietal pleural invasion had higher mean SUV max than tumours with uninvolved pleura (14.2 vs 11.6, p = 0.018). Log-rank test was used to assess the best discriminative SUV max value in relation to survival (Fig. 1). Based on this analysis, a cut-off value of 15 was used. Survival curves of overall survival following resection showed that patients with SUV max 15 had significantly longer survival than those with SUV max >15 (Fig. 2). Furthermore, survival curves based on nodal stages (non N2 and N2) also showed a similar pattern when the cut-off value is used (Fig. 3 A and B, respectively).

4 N. Al-Sarraf et al. / European Journal of Cardio-thoracic Surgery 34 (2008) Table 4 Variables stratification based on cut-off SUV max of 15 Variable SUV max 15 (n = 128) SUV max >15 (n = 48) p value Status Alive 100 (78%) 28 (58%) Dead 28 (22%) 20 (42%) Tumour location Central 14 (11%) 25 (52%) Peripheral 114 (89%) 23 (48%) < Lobar distribution Lower 41 (32%) 9 (19%) Upper 78 (61%) 36 (75%) Middle (right) 9 (7%) 3 (6%) Subtype Adenocarcinoma 72 (56%) 6 (13%) Squamous 46 (36%) 39 (81%) < Others 10 (8%) 3 (6%) Fig. 2. Kaplan Meier survival curve for all patients under the study (n = 176). Location Right 73 (57%) 24 (50%) Left 55 (43%) 24 (50%) Differentiation grade Well/moderate 88 (69%) 29 (60%) Poor 40 (31%) 19 (40%) 0.37 Tumour (T) stage T1 47 (37%) 2 (4%) T2 67 (52%) 29 (60%) < T3/T4 14 (11%) 17 (36%) Nodal stage (N) N0 83 (65%) 20 (42%) N1 19 (15%) 17 (35%) N2 26 (20%) 11 (23%) AJCC stage I/II 97 (76%) 28 (58%) III/IV 31 (24%) 20 (42%) Tumour size (cm) Mean SD Chest wall resection Yes 5 (4%) 2 (4%) No 123 (96%) 46 (96%) Surgical resection Lobectomy 102 (80%) 19 (40%) Complex resection a 23 (18%) 28 (58%) No resection 3 (2%) 1 (2%) < Pleural invasion Involved b 47 (37%) 18 (38%) Uninvolved 81 (63%) 30 (62%) 1.00 Vascular invasion Absent 72 (56%) 26 (54%) Present 56 (44%) 22 (46%) a b Denotes bi-lobectomy, Pneumonectomy and sleeve resections. Denotes either parietal or visceral pleural invasion. 4. Discussion Recent evidence showing improved survival in NSCLC patients treated with neoadjuvant chemotherapy has emerged [7]. This might be the way of replacing the conventional practice of adjuvant chemotherapy following resection. However, no data are currently available on preoperative prognostic indicators that might dictate when patients should receive adjuvant or neoadjuvant chemother- Fig. 3. (A) Kaplan Meier survival curve for non-n2 patients (n = 139). (B) Kaplan Meier survival curve for N2 patients (n = 37).

5 896 N. Al-Sarraf et al. / European Journal of Cardio-thoracic Surgery 34 (2008) Table 5 Previous studies correlating SUV to outcome First author (year) Number Main findings Van Baardwijk [8] (2007) 102 Two cut-off values used 1. SUV 8 worse survival than SUV <8 2. SUV 11 worse survival than SUV <11 3. Only stage I and II studied 4 Higher SUV associated with higher HIF-1a a and GLUT-1 expression Vesselle [9] (2007) No significant association between SUV and survival Cerfolio [10] (2005) High SUV predicts survival for stages IB-IIIA 2. High SUV predict disease-free survival for IB and II 3. SUV correlates with lymphovascular invasion and stage Downey [11] (2004) SUV correlates with histology (Sqcc > AC) b 2. SUV >9 worse 2-year survival post resection 3. SUV >9 and tumours >3 cm had worst survival Sasaki [12] (2003) SUV >5 worse 1-year disease free survival Jeong [13] (2002) SUV 7 worse survival 2. SUV 7 associated with 6.3-fold increased mortality 3. SUV correlates with histology (Sqcc > AC) Higashi [14] (2002) SUV > 5 worse disease-free survival 2. Patients stage I disease and SUV >5 had worse survival than those with SUV 5 Dhital [15] (2000) SUV correlates with 1-year survival 2. SUV >20 carried 4.7-fold increased hazards of death 3. No correlation between histology and SUV Vansteenkiste [16] (1999) SUV >7 worse survival Sugawara [17] (1999) No prognostic association of SUV Ahuja [18] (1998) SUR c >10 poor survival 2. SUR >10 and tumour size >3 cm had worst survival a b c Hypoxia inducible factor. Squamous cell carcinoma and adenocarcinoma. Standardised uptake ratio. apy. The rationale for identifying NSCLC patients with poor prognostic indicators is that these patients could be treated more aggressively in the first instance with neo-adjuvant chemotherapy while those with favourable prognostic indicators could be treated postoperatively with adjuvant type chemotherapy based on their surgical pathological stage. PET-CT has proved useful in the preoperative staging of NSCLC. The prognostic significance of SUV has been widely studied with most data supporting its role (Table 5). The cutoff values in SUV measurements however vary widely reflecting a variety of tumour types, surgical resection, follow-up times and the influence of neo-adjuvant chemotherapy [8 18]. SUV max is a semi-quantitative method of measuring 18 FDG uptake by tumour cells. Tumour subtype reflects the differential avidity of 18 FDG uptake by the cancer cells. Adenocarcinomas have lower expression of Glut-1 transporters which in turn have lower 18 FDG uptake and lower SUV max measurements than squamous cell type [19]. The advancing T stages and the recorded increase in SUV max can be explained in part by tumour size and volume. AJCC classification also correlated with SUV max whereby the most curable stage (stage I) had lowest mean SUV max than the other stages. This is a combination of factors such as size, local invasion and nodal involvement. NSCLC with pleural invasion showed a higher mean SUV max than those without it. Pattern of surgical resection correlated with mean SUV max as patients requiring more aggressive and technically demanding resection had lower mean SUV max than patients requiring a simple lobectomy. This in part is dependant on tumour size and central location as opposed to peripherally located tumours. Survival following resection is dependant on inherent biological aggressiveness of the tumours and the stage which in turn depends on lymph node involvement and local invasion. Our overall survival was better for patients with SUV max 15 than those with SUV max >15 (Fig. 2). The same finding was also observed when patients were stratified into various lymph nodal stages (non-n2 and N2, Figs. 3A and 3B). All these prognostic factors do correlate with the degree of 18 FDG uptake by cancer cells. Our data remains consistent with previously published work, a summary of which is depicted in Table 5. One interesting point shown in this table is the difference in the cut-off values used for survival. This has varied across all the studies and various reasons can potentially explain it. First, the different prevalence of NSCLC subtype and whether or not carcinoids have been studied as well. Second, the pattern of surgical resection might have been the same, but the means whereby lymph nodes were surgically staged differed (i.e. sampling vs SLND, preoperative mediastinoscopy selectively vs routinely). Third, different studies used different follow-up time frames and some have included neo-adjuvant chemotherapy which in itself could have altered both the staging and the prognosis. In our work we relied only on patients with documented NSCLC diagnosis

6 N. Al-Sarraf et al. / European Journal of Cardio-thoracic Surgery 34 (2008) who underwent a curative intent surgical resection and we intentionally excluded all neoadjuvant chemotherapy patients, patients with diabetes and those with carcinoids (which can interfere with SUV max measurements). In addition, we relied on SLND as a gold standard technique for pathological assessment of lymph nodes following resection thus giving us a higher standard of pathological staging than sampling would normally give. Various conventional prognostic indicators have been previously described [20,21]. A major limitation of these was their postoperative nature. The difficulty arises when attempting to use them preoperatively and since SUV max of primary NSCLC does correlate with survival and is of prognostic significance, one might argue the case of using SUV max preoperatively in allocating patients to different treatment strategies such as neoadjuvant versus adjuvant versus no chemotherapy at all. However, the best cut-off value that could be used universally remains unknown. Such a strategy certainly warrants a multicentre collaboration in a prospective randomised controlled trial. 5. Limitations of the study Our study remains a retrospective work and as such we can only report an association. We also elected to exclude diabetics and patients who received neoadjuvant chemotherapy which in itself might have caused a selection bias. However, due to the relatively large number of patients studied, such bias may not become significant. 6. Conclusion Integrated PET-CT is a valuable investigation tool in preoperative staging of NSCLC. In addition to its superiority to CT in the preoperative staging, it provides additional information regarding tumour characteristics, aggressive tumour behaviour and serves as a prognostic indicator. Such information may help in selecting patients preoperatively into receiving neoadjuvant chemotherapy or not. Prospective and randomised trials are warranted. Acknowledgements We are grateful for the assistance of Ms Rita Luddy and Ms Mary Devlin in the data acquisition. References [1] Reed CE, Harpole DH, Posther KE, Woolson SL, Downey RJ, Meyers BF, Heelan RT, MacApinlac HA, Jung SH, Silvestri GA, Siegel BA, Rusch VW. Results of the American college of surgeons oncology group Z0050 trial: the utility of positron emission tomography in staging potentially operable nonsmall cell lung cancer. J Thorac Cardiovasc Surg 2003;126(6): [2] Woodard HQ, Bigler RE, Freed B. Expression of tissue isotope distribution. J Nucl Med 1975;16(10): [3] Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997;111: [4] Al-Sarraf N, Aziz R, Doddakula K, Gately K, Wilson L, McGovern E, Young V. Factors causing inaccurate staging of mediastinal nodal involvement in non-small cell lung cancer patients staged by positron emission tomography. Interact Cardiovasc Thorac Surg 2007;6(3): [5] Al-Sarraf N, Gately K, Lucey J, Wilson L, McGovern E, Young V. Lymph node staging by means of positron emission tomography is less accurate in nonsmall cell lung cancer patients with enlarged lymph nodes: analysis of 1145 lymph nodes. Lung Cancer 2008;60(1):62 8. [6] Al-Sarraf N, Aziz R, Gately K, Lucey J, Wilson L, McGovern E, Young V. Pattern and predictors of occult mediastinal lymph node involvement in non-small cell lung cancer patients with negative mediastinal uptake on positron emission tomography. Eur J Cardiothorac Surg 2008;33(1): [7] Depierre A, Milleron B, Moro-Sibilot D, Chevret S, Quoix E, Lebeau B, Braun D, Breton JL, Lemarie E, Gouva S, Paillot N, Brechot JM, Janicot H, Lebas FX, Terrioux P, Clavier J, Foucher P, Monchatre M, Coetmeur D, Level MC, Leclerc P, Blanchon F, Rodier JM, Thiberville L, Villeneuve A, Westeel V, Chastang C, French thoracic cooperative group. Preoperative chemotherapy followed by surgery compared with primary surgery in resectable stage I (except T1N0), II, and IIIa non-small-cell lung cancer. J Clin Oncol 2002;20: [8] Van Baardwijk A, Dooms C, Van Suylen RJ, Verbeken E, Hochstenbag M, Dehing-Oberije C, Rupa D, Pastorekova S, Stroobants S, Buell U, Lambin P, Vansteenkiste J, De Ruysscher D. The maximum uptake of 18F-deoxyglucose on positron emission tomography scan correlates with survival, hypoxia inducible factor-1a and GLUT-1 in non-small cell lung cancer. Eur J Cancer 2007;43: [9] Vesselle H, Freeman JD, Wiens L, Stern J, Nguyen HQ, Hawes SE, Bastian P, Salskov A, Vallières E, Wood DE. Fluorodeoxyglucose uptake of primary non-small cell lung cancer at positron emission tomography: new contrary data on prognostic role. Clin Cancer Res 2007;13(11): [10] Cerfolio RJ, Bryant AS, Ohja B, Bartolucci AA. The maximum standardized uptake values on positron emission tomography of a non-small cell lung cancer predict stage, recurrence, and survival. J Thorac Cardiovasc Surg 2005;130: [11] Downey RJ, Akhurst T, Gonen M, Vincent A, Bains MS, Larson S, Rusch V. Preoperative F-18 fluorodeoxyglucose-positron emission tomography maximal standardized uptake value predicts survival after lung cancer resection. J Clin Oncol 2004;22: [12] Sasaki R, Komaki R, Macapinlac H. SUV by FDG-PET predicts outcomes of NSCLC. Int J Radiat Oncol Biol Phys 2003;57(Suppl. 2):S166. [13] Jeong HJ, Min JJ, Park JM, Chung JK, Kim BT, Jeong JM, Lee DS, Lee MC, Han SK, Shim YS. Determination of the prognostic value of fluorodeoxyglucose uptake by using positron emission tomography in patients with non-small cell lung cancer. Nucl Med Commun 2002;23: [14] Higashi K, Ueda Y, Arisaka Y. 18F-FDG uptake as a biologic prognostic factor for recurrence in patients with surgically resected non-small cell lung cancer. J Nucl Med 2002;43: [15] Dhital K, Saunders CAB, Seed PT, O Doherty MJ, Dussek J. 18F-Fluorodeoxyglucose positron emission tomography and its prognostic value in lung cancer. Eur J Cardiothorac Surg 2000;18: [16] Vansteenkiste JF, Stroobants SG, Dupont PJ, De Leyn PR, Verbeken EK, Deneffe GJ, Mortelmans LA, Demedts MG, Leuven Lung Cancer Group. Prognostic importance of the standardized uptake value on (18) F-fluoro- 2-deoxy-glucose-positron emission tomography scan in non-small-cell lung cancer: an analysis of 125 cases. J Clin Oncol 1999;10: [17] Sugawara Y, Quint LE, Iannettoni MD, Orringer MB, Russo JE, Recker BE, Saran PA, Wahl RL. Does the FDG uptake of primary non-small cell lung cancer predict prognosis? A work in progress. Clin Positron Imag 1999;2(2): [18] Ahuja V, Coleman RE, Herndon J, Patz Jr EF. The prognostic significance of fluorodeoxyglucose positron emission tomography imaging for patients with non small cell lung carcinoma. Cancer 1998;83: [19] Ito T, Noguchi Y, Saton S, Hayashi H, Inayama Y, Kitamura H. Expression of facilitative glucose transporter isoforms in lung carcinomas: its relation to histological type, differentiation grade, and tumour stage. Mod Pathol 1998;11: [20] Suzuki K, Nagai K, Yoshida J, Nishimura M, Takahashi K, Yokose T, Nishiwaki Y. Conventional clinicopathologic prognostic factors in surgically non-small cell lung carcinoma: a comparison of prognostic factors for each pathologic TNM stage based on multivariate analyses. Cancer 1999;86: [21] Kessler R, Gasser B, Massard G, Roeslin N, Meyer P, Wihlm JM, Morand G. Blood vessel invasion is a major prognostic factor in resected non-small cell lung cancer. Ann Thorac Surg 1996;62(5):