- Email: [email protected]
Prognostic Value of Preoperative Positron Emission Tomography in Resected Stage I Non-small Cell Lung Cancer
ORIGINAL ARTICLE
Prognostic Value of Preoperative Positron Emission Tomography in Resected Stage I Non-small Cell Lung Cancer Boone Goodgame, MD,* Giancarlo A. Pillot, MD,* Zhiyun Yang, MD,† Jabi Shriki, MD,† Bryan F. Meyers, MD,‡§ Jennifer Zoole, BSN,§ Feng Gao, PhD,‡储 Farrokh Dehdashti, MD,† Alexander Patterson, MD,‡§ Barry A. Siegel, MD,†§ and Ramaswamy Govindan, MD*§
Purpose: Approximately 20 to 40% of patients with surgically resected stage I non-small cell lung cancer (NSCLC) will develop recurrent disease. Positron emission tomography (PET) with 2-[18F] fluoro-2-deoxy-D-glucose (FDG) is used often in staging NSCLC. We conducted this study to determine whether the preoperative maximum tumor standardized uptake value (SUVmax) was associated with recurrence in patients with resected stage I NSCLC. Patients and Methods: We identified consecutive patients who underwent curative surgical resection for stage I NSCLC between 1999 and 2003 who had preoperative FDG-PET imaging. Patients were divided into two cohorts based on SUVmax above or below the median for the group. Recurrence rates were estimated by the Kaplan-Meier method and overall survival was analyzed as a secondary end point. Results: Of 136 patients who met inclusion criteria, 77 (57%) had T1 and 59 (43%) had T2 tumors. The median follow-up time was 46 months and 32 patients had a disease recurrence. The median SUVmax was 5.5. The 5-year estimates of recurrence rates for patients with low and high SUVmax were 14% and 37%, respectively (p ⫽ 0.002), with 5-year overall survivals of 74% and 53%, respectively (p ⫽ 0.006). In multivariate analyses based on SUVmax, T-classification, age, and histology, high SUVmax was independently associated with recurrence (p ⫽ 0.002) and mortality (p ⫽ 0.041). Conclusion: High SUVmax (ⱖ5.5) on preoperative FDG-PET is an independent predictor of relapse and death in resected stage I *Division of Medical Oncology, †Division of Nuclear Medicine, Mallinckrodt Institute of Radiology, ‡Alvin J. Siteman Cancer Center, §Division of Cardiothoracic Surgery, and 储Division of Biostatistics, Washington University School of Medicine, St. Louis, Missouri. Disclosure: Dr. Siegel receives financial support from Radiology Corporation of America (PET Mobile Service Provider) as a Medical Advisory Board Member, lecturer and stockholder. He also serves as a lecturer for Siemens, GE Healthcare, and PETNET. Preliminary data were previously presented at the ASCO Annual Meeting in Atlanta, Georgia, 2006. Address for correspondence: Ramaswamy Govindan, MD, Division of Medical Oncology, 4960 Children’s Place, Suite 108, St. Louis, MO 63110. E-mail: [email protected] Copyright © 2008 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/08/0302-0130
130
NSCLC. Prospective trials of adjuvant chemotherapy in patients with stage I NSCLC and high SUVmax should be considered. Key Words: Non-small cell lung cancer, Prognosis, PET, Imaging, Stage I, Early stage, Surgery outcomes. (J Thorac Oncol. 2008;3: 130 –134)
L
ung cancer is the leading cause of cancer-related death in both men and women, with an estimated 213,380 new cases and 160,390 patients dying from this disease in 2007 in the United States.1 Nearly 87% of lung cancer patients are diagnosed with non-small cell lung cancer (NSCLC), and of these, approximately 30% of patients are candidates for curative resection.2,3 It has now become an accepted practice to offer adjuvant chemotherapy for patients with resected stage IB–III NSCLC based on the results of several prospective studies that have demonstrated a statistically significant and clinically meaningful survival advantage with this approach.4 – 6 Nevertheless, subset analyses of stage IB patients, and a large trial of only stage IB patients, have failed to show a significant survival advantage in this population.7 Identifying stage I patients who are at highest risk of relapse could help determine which patients are most likely to benefit from adjuvant chemotherapy. Positron emission tomography (PET) with the glucose analogue 2-[18F]fluoro-2-deoxy-D-glucose (FDG) has an established role in the treatment of NSCLC. It enhances the accuracy of clinical staging, improving patient selection for surgical treatment.8 –10 Tumor FDG uptake, as measured by the standardized uptake value (SUV), has been demonstrated to correlate with several predictors of outcome in NSCLC. Accordingly, several studies have addressed the relationship between the tumor SUV and outcome in NSCLC; we have recently reviewed the available literature on this topic.11 These studies have in aggregate suggested that the SUV for FDG may be a powerful surrogate marker for outcome in NSCLC. Nevertheless, the patient populations in these studies have typically been heterogeneous, often including subjects with NSCLCs of several different stages and histologies. We thus undertook a review of the Washington University experience to determine whether we could confirm the pre-
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
dictive value of the SUV for FDG in a patients with stage I NSCLC treated uniformly with surgery and no adjuvant chemotherapy or radiation. We excluded patients with bronchioalveolar carcinoma given the low FDG uptake of many of these tumors and the apparently different biologic behavior.12
PATIENTS AND METHODS We reviewed the medical records of all patients with stage I NSCLC of nonbronchioalveolar carcinoma pathology, as staged according to the 1997 revised International System,13 who underwent curative resection from January 1, 1999, to December 31, 2003, and for whom complete pathologic staging data were available. Patients who underwent FDG-PET within 120 days before definitive resection were included in this analysis. Our institutional review board approved this study. We collected data regarding age at diagnosis, timing of FDG-PET, timing of surgery, tumor, node, metastasis staging, the maximum SUV (SUVmax) of the primary tumor,14 extent of surgery, time to recurrence, time to last follow-up, time to death, and cause of death. All original pathology reports were reviewed and data were collected from the review of all available radiologic and pathologic reports; clinic charts from thoracic surgery, radiation oncology, and medical oncology; and inpatient electronic charts. Date of death was confirmed through the Social Security Death Index. PET imaging over the course of this study was conducted on three different scanners, including a Siemens ECAT EXACT scanner (Siemens, Knoxville, Tennessee), a Phillips/ADAC CPET-Plus scanner (Phillips, Milpitas, California), and a Siemens Biograph LSO Duo hybrid PET/ computed tomography scanner (Siemens, Knoxville, Tennessee). PET imaging was performed in routine clinical fashion in accordance with a standard protocol. All of the PET studies were first interpreted clinically and then were subsequently evaluated a second time for this study by a radiologist, fellowship trained in nuclear medicine. The SUVmax of the primary bronchogenic carcinoma was measured by placing a region of interest (ROI) around the visible tumor, with careful attention to avoid inclusion of any myocardium. Standard software tools supplied by the manufacturer of each scanner were used. When the primary tumor was seen on PET, the ROI was placed around the visible tumor, with careful attention to avoid inclusion of any myocardium. For tumors that were not seen on PET, no measurement was made on conventional PET studies performed with either the ECAT EXACT or the CPET, and the patient was assigned to the low-SUV group (see later). For studies performed with the biograph where the lesion was not visible on PET, the ROI was placed around the lesion seen on computed tomography lung windows.
Statistical Analysis To assess the prognostic value of the SUVmax, we subdivided the patients into two cohorts: those with values for SUVmax below the median versus those with SUVmax greater than or equal to the median for the population. This decision was made a priori to avoid introduction of bias into the study. Overall survival (OS) was defined as the interval from sur-
PET-SUV for Prognosis in Stage I NSCLC
gery to death from any cause and those patients who remained alive were censored at the date of last clinical contact. Time to recurrence was defined as time from the date of surgery to the recurrence of lung cancer. Patients without recurrence were censored from the analysis at the time of last negative follow-up encounter, death without evidence of recurrence, or diagnosis of a second primary lung cancer. OS and recurrence rates were estimated by Kaplan-Meier product limit method. The log-rank test was used to detect the potential association between each covariate and survival or recurrence. A multivariate Cox proportional hazards model was fitted to identify independent risk factors and to adjust for the potential effects of known risk factors (age, T-classification, and histology). Interaction terms with SUVmax were entered into the model one at a time. To avoid the assumption of linear relationships as well as for a better interpretation of the results, the continuous variables, age, and SUVmax were dichotomized. Tumor size was not analyzed as a covariate because of its interdependence with T-classification. A multivariate Cox proportional hazards model was fitted to identify independent risk factors. To better understand the relationship between SUVmax and outcomes, the assumption of linearity was also assessed. In this analysis, a smoothed line is imposed on the scatter plot of SUVmax against the martingale residuals, which were obtained from the multivariate Cox model containing all variables but SUVmax. If the linearity assumption is well satisfied, the plot shows an approximately straight line. This plot also provides useful information for choosing the appropriate cut-off point when the linearity assumption is violated. The statistical analysis was performed with the statistical package SAS version 9. A p value under 0.05 was taken to indicate statistical significance and all tests were two-sided.
RESULTS We identified 136 consecutive patients diagnosed with stage I NSCLC between 1999 and 2003 who met our selection criteria. Patient characteristics are shown in Table 1. None of the patients received adjuvant chemotherapy or radiotherapy. Of 136 patients with stage I NSCLC, 77 (57%)
TABLE 1. Characteristics of Resected Stage I NSCLC Subjects, Grouped by SUVmax ⬍5.5 or ⱖ5.5 Low SUVmax Variables Total subjects Age at diagnosis, median T-classification 1 2 Size (cm), median Histology Adenocarcinoma Squamous Other
High SUVmax
n
%
n
%
65 65
51
71 69
49
46 19 1.9
71 29
31 40 3.4
44 56
46 13 6
68 23 10
25 34 12
36 45 16
p 0.004 ⬍0.001
⬍0.001 ⬍0.001
SUVmax, maximum tumor standardized uptake value.
Copyright © 2008 by the International Association for the Study of Lung Cancer
131
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
Goodgame et al.
had T1 and 59 (43%) had T2 tumors. Ten patients with T2 tumors had a tumor size of 3 cm or less but presence of pleural invasion. The median follow-up time was 46 months,
TABLE 2. Univariate Analysis for Predictors of Outcome in Resected Stage I NSCLC Recurrence Rate Variable SUVmax ⬍5.5 ⱖ5.5 T-classification T1 T2 Histology Adenocarcinoma Squamous Others Age ⬍70 ⱖ70
n
5-yr
65 71
14 37
77 59
17 36
71 47 18
27 12 55
79 57
22 28
p
Overall Survival 5-yr
0.002
p 0.006
74 53 0.035
0.002 73 52
0.016
0.19 69 65 34
0.53
0.006 74 52
SUVmax, maximum tumor standardized uptake value.
FIGURE 1. Recurrence rates based on SUVmax where vertical bars within each curve represent the distributions of censored cases. SUV, standardized uptake value; SUVmax, maximum tumor standardized uptake value.
132
and 32 patients (24%) had a recurrence of their disease. Sixteen patients died within 1 year of surgery, only two of these had a disease recurrence. Most patients had adenocarcinoma (71 patients) or squamous cell carcinoma (47), but other histologies were also present with six patients having adenosquamous carcinoma and 10 patients having large-cell carcinoma, with or without neuroendocrine features. In two patients, no histologic subtype could be determined. PET was performed at a median of 15 days before surgery (range 0 –113 days). The median SUVmax for the population was 5.5. Patients with a higher tumor SUVmax were more likely to be older and have larger or T2 tumors, and were less likely to have adenocarcinoma (Table 1). Only 10 patients with a low SUVmax had a recurrence of their disease (14%), whereas 22 patients (33%) in the high-SUVmax group had a recurrence. The estimated 5-year recurrence rate in the low-SUVmax group was 14%, compared with 37% in the high-SUVmax group (p ⫽ 0.002). The 5-year OS was correspondingly higher in the low-SUVmax group (74% versus 53%, p ⫽ 0.006; Table 2 and Figures 1 and 2). Of note, patients of other histologies (not squamous or adenocarcinoma) had particularly poor outcomes with 5-year recurrence rates and mortality rates of 55% and 66%, respectively. In the univariate analysis for other covariates, histology and T-classification were significantly associated with recurrence (Table 2). T-classification and age (⬍70 versus ⱖ70)
FIGURE 2. Overall survival based on SUVmax where vertical bars within each curve represent the distributions of censored cases. SUV, standardized uptake value; SUVmax, maximum tumor standardized uptake value.
Copyright © 2008 by the International Association for the Study of Lung Cancer
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
TABLE 3. Multivariate Analysis for Predictors of Recurrence and Survival in Resected Stage I NSCLC Recurrence Factor SUV ⱖ5.5 Nonsquamous Age ⱖ70 yr T2 tumor
Overall Survival
HR (95% CI)
p
HR (95% CI)
p
3.46 (1.57–7.65) 2.8 (1.18–6.65) 1.24 (0.61–2.5) 1.48 (0.71–3.09)
0.002 0.020 0.550 0.292
1.91 (1.03–3.57) 1.35 (0.73–2.49) 2.12 (1.19–3.79) 2.06 (1.13–3.73)
0.041 0.344 0.011 0.018
HR, hazard ratio; CI, confidence interval; SUV, standardized uptake value.
were significantly associated with OS. In the multivariate analysis, high SUVmax was independently associated with recurrence, having a hazard ratio (HR) of 3.5, with 95% confidence intervals (CI) of 1.6 to 7.7. Nonsquamous histology was also independently associated with recurrence (HR ⫽ 2.8, 95% CI ⫽ 1.2– 6.7; Table 3). High SUVmax was also independently associated with worse survival in the multivariate analysis (HR ⫽ 1.9, 95% CI ⫽ 1.03–3.57). Age and T2 tumor status were also independently associated with worse OS with HRs of 2.12 and 2.06, respectively. No interaction terms between covariates were statistically significant. An analysis of the residuals for recurrence and OS as described in the methods showed that the relationships between SUVmax and recurrence, and SUVmax and survival, were dependent but nonlinear. This analysis confirmed that median SUVmax was an appropriate cut-off point to distinguish between high- and low-risk patients based on our data (data not shown).
DISCUSSION New risk stratification markers and models are urgently needed to evaluate which patients with resected stage I NSCLC are at highest risk for relapse and stand to benefit most from adjuvant chemotherapy. In this series of 136 stage I NSCLC patients, higher preoperative SUVmax was strongly associated with poor outcomes. Several other investigators have also retrospectively evaluated their experience, attempting to correlate SUV with outcomes in patients with NSCLC, and we have recently reviewed these publications.11 Multiple studies have shown a correlation between SUV and stage, survival or disease-free survival. The majority of these publications have grouped all stages of NSCLC. Cerfolio et al. published a large series of 315 patients with resected NSCLC, which included 141 stage I patients. Higher SUV correlated with stage, as well as with recurrence and survival. Multivariate analysis of the entire cohort found that SUV ⱖ10 was the best independent predictor of disease-free survival and survival. A multivariate analysis of the stage I patients was not presented.15 In a more recent study by Ohtsuka et al., SUVmax was correlated with outcomes for 98 stage I patients, limited to those with adenocarcinoma. A value of 3.3 was found to be the optimum cut-off point. Twelve recurrences were identified and 10 of these were in the high-SUVmax group, which approached statistical significance by multivariate analysis (p ⫽ 0.079).16
PET-SUV for Prognosis in Stage I NSCLC
In the current study, the number of patients (136) and recurrences (32) allowed for a multivariate analysis of factors predictive of recurrence and survival in a pure stage I population. Patients with SUVmax ⱖ5.5 (the median for the group) had a statistically significant increase in recurrence rate and mortality, which was independent of other standard risk factors (age, histology, and T-classification) by multivariate analysis. As shown in Figure 1, the time to recurrence in patients with a high SUVmax was shorter than that in patients with a lower SUVmax (median time of 11 months versus 25 months). The precise definitions and implications of high- and low-risk PET findings have yet to be determined. Ohtsuka et al. found that an SUVmax of greater than 3.3 best defines a high-risk population,16 whereas our data suggest that an SUVmax of approximately 5 is more appropriate. Nevertheless, their study was limited to adenocarcinoma, which has been shown to have a lower SUV in other studies17,18 as well as in our own data (median SUVmax of 4.2 versus 8.4 for squamous). What is the next logical step in refining this prognostic tool? It should be fairly straightforward to develop a prospective registry of patients with resected T1 and T2 NSCLC with preoperative PET imaging performed according to standardized protocols. This would allow for more accurate estimates of relapse risk and identification of the most appropriate SUV groupings. Nevertheless, more important than the prognostic question, is the question of whether these patients would benefit from adjuvant chemotherapy. Current evidence from clinical trials does not support the use of adjuvant chemotherapy in unselected stage I NSCLC patients, but based on our retrospective data, stage I patients with high SUVmax may be more than three times more likely to have recurrences of their disease. Higher recurrence rates suggest that a survival benefit from chemotherapy might be easier to establish in this population than in an unselected group. Therefore, prospective clinical trials of adjuvant chemotherapy in patients with stage I NSCLC and SUVmax greater than 5 should be strongly considered. REFERENCES 1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, 2007. CA Cancer J Clin 2007;57:43– 66. 2. Govindan R, Page N, Morgensztern D, et al. Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: analysis of the surveillance, epidemiologic, and end results database. J Clin Oncol 2006;24:4539 – 4544. 3. Domont J, Soria JC, Le Chevalier T. Adjuvant chemotherapy in earlystage non-small cell lung cancer. Semin Oncol 2005;32:279 –283. 4. Douillard JY, Rosell R, De Lena M, et al. Adjuvant vinorelbine plus cisplatin versus observation in patients with completely resected stage IB-IIIA non-small-cell lung cancer (Adjuvant Navelbine International Trialist Association [ANITA]): a randomised controlled trial. Lancet Oncol 2006;7:719 –727. 5. Winton T, Livingston R, Johnson D, et al. Vinorelbine plus cisplatin vs. observation in resected non-small-cell lung cancer. N Engl J Med 2005;352:2589 –2597. 6. The International Adjuvant Lung Cancer Trial Collaborative Group. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N Engl J Med 2004;350:351– 360.
Copyright © 2008 by the International Association for the Study of Lung Cancer
133
Goodgame et al.
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
7. Strauss GM, Herndon J, Maddaus MA, et al. Adjuvant chemotherapy in stage IB non-small cell lung cancer (NSCLC): update of Cancer and Leukemia Group B (CALGB) protocol 9633 (Meeting Abstract). J Clin Oncol 2006;24(18 Suppl):7007. 8. Pieterman RM, van Putten JWG, Meuzelaar JJ, et al. Preoperative staging of non-small-cell lung cancer with positron-emission tomography. N Engl J Med 2000;343:254 –261. 9. Lardinois D, Weder W, Hany TF, et al. Staging of non-small-cell lung cancer with integrated positron-emission tomography and computed tomography. N Engl J Med 2003;348:2500 –2507. 10. van Tinteren H, Hoekstra OS, Smit EF, et al. Effectiveness of positron emission tomography in the preoperative assessment of patients with suspected non-small-cell lung cancer: the PLUS multicentre randomised trial. Lancet 2002;359:1388 –1392. 11. Pillot G, Siegel BA, Govindan R. Prognostic value of fluorodeoxyglucose positron emission tomography in non-small cell lung cancer: a review. J Thorac Oncol 2006;1:152–159. 12. Raz DJ, Odisho AY, Franc BL, Jablons DM. Tumor fluoro-2-deoxyD-glucose avidity on positron emission tomographic scan predicts mortality in patients with early-stage pure and mixed bronchi-
134
13. 14. 15.
16.
17. 18.
oloalveolar carcinoma. J Thorac Cardiovasc Surg 2006;132:1189 – 1195. Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997;111:1710 –1717. Kumar R, Nadig MR, Chauhan A. Positron emission tomography: clinical applications in oncology. Part 1. Expert Rev Anticancer Ther 2005;5:1079 –1094. 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;130:151–159. Ohtsuka T, Nomori H, Watanabe KI, et al. Prognostic significance of [(18)F]fluorodeoxyglucose uptake on positron emission tomography in patients with pathologic stage I lung adenocarcinoma. Cancer 2006; 107(10):2468 –2473. Higashi K, Ueda Y, Arisaka Y, et al. 18F-FDG uptake as a biologic prognostic factor for recurrence in patients with surgically resected non-small cell lung cancer. J Nucl Med 2002;43:39 – 45. Sugawara Y, Quint LE, Iannettoni MD, et al. Does the FDG uptake of primary non-small cell lung cancer predict prognosis? A work in progress. Clin Positron Imaging 1999;2:111–118.
Copyright © 2008 by the International Association for the Study of Lung Cancer
Prognostic Value of Preoperative Positron Emission Tomography in Resected Stage I Non-small Cell Lung Cancer Boone Goodgame, MD,* Giancarlo A. Pillot, MD,* Zhiyun Yang, MD,† Jabi Shriki, MD,† Bryan F. Meyers, MD,‡§ Jennifer Zoole, BSN,§ Feng Gao, PhD,‡储 Farrokh Dehdashti, MD,† Alexander Patterson, MD,‡§ Barry A. Siegel, MD,†§ and Ramaswamy Govindan, MD*§
Purpose: Approximately 20 to 40% of patients with surgically resected stage I non-small cell lung cancer (NSCLC) will develop recurrent disease. Positron emission tomography (PET) with 2-[18F] fluoro-2-deoxy-D-glucose (FDG) is used often in staging NSCLC. We conducted this study to determine whether the preoperative maximum tumor standardized uptake value (SUVmax) was associated with recurrence in patients with resected stage I NSCLC. Patients and Methods: We identified consecutive patients who underwent curative surgical resection for stage I NSCLC between 1999 and 2003 who had preoperative FDG-PET imaging. Patients were divided into two cohorts based on SUVmax above or below the median for the group. Recurrence rates were estimated by the Kaplan-Meier method and overall survival was analyzed as a secondary end point. Results: Of 136 patients who met inclusion criteria, 77 (57%) had T1 and 59 (43%) had T2 tumors. The median follow-up time was 46 months and 32 patients had a disease recurrence. The median SUVmax was 5.5. The 5-year estimates of recurrence rates for patients with low and high SUVmax were 14% and 37%, respectively (p ⫽ 0.002), with 5-year overall survivals of 74% and 53%, respectively (p ⫽ 0.006). In multivariate analyses based on SUVmax, T-classification, age, and histology, high SUVmax was independently associated with recurrence (p ⫽ 0.002) and mortality (p ⫽ 0.041). Conclusion: High SUVmax (ⱖ5.5) on preoperative FDG-PET is an independent predictor of relapse and death in resected stage I *Division of Medical Oncology, †Division of Nuclear Medicine, Mallinckrodt Institute of Radiology, ‡Alvin J. Siteman Cancer Center, §Division of Cardiothoracic Surgery, and 储Division of Biostatistics, Washington University School of Medicine, St. Louis, Missouri. Disclosure: Dr. Siegel receives financial support from Radiology Corporation of America (PET Mobile Service Provider) as a Medical Advisory Board Member, lecturer and stockholder. He also serves as a lecturer for Siemens, GE Healthcare, and PETNET. Preliminary data were previously presented at the ASCO Annual Meeting in Atlanta, Georgia, 2006. Address for correspondence: Ramaswamy Govindan, MD, Division of Medical Oncology, 4960 Children’s Place, Suite 108, St. Louis, MO 63110. E-mail: [email protected] Copyright © 2008 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/08/0302-0130
130
NSCLC. Prospective trials of adjuvant chemotherapy in patients with stage I NSCLC and high SUVmax should be considered. Key Words: Non-small cell lung cancer, Prognosis, PET, Imaging, Stage I, Early stage, Surgery outcomes. (J Thorac Oncol. 2008;3: 130 –134)
L
ung cancer is the leading cause of cancer-related death in both men and women, with an estimated 213,380 new cases and 160,390 patients dying from this disease in 2007 in the United States.1 Nearly 87% of lung cancer patients are diagnosed with non-small cell lung cancer (NSCLC), and of these, approximately 30% of patients are candidates for curative resection.2,3 It has now become an accepted practice to offer adjuvant chemotherapy for patients with resected stage IB–III NSCLC based on the results of several prospective studies that have demonstrated a statistically significant and clinically meaningful survival advantage with this approach.4 – 6 Nevertheless, subset analyses of stage IB patients, and a large trial of only stage IB patients, have failed to show a significant survival advantage in this population.7 Identifying stage I patients who are at highest risk of relapse could help determine which patients are most likely to benefit from adjuvant chemotherapy. Positron emission tomography (PET) with the glucose analogue 2-[18F]fluoro-2-deoxy-D-glucose (FDG) has an established role in the treatment of NSCLC. It enhances the accuracy of clinical staging, improving patient selection for surgical treatment.8 –10 Tumor FDG uptake, as measured by the standardized uptake value (SUV), has been demonstrated to correlate with several predictors of outcome in NSCLC. Accordingly, several studies have addressed the relationship between the tumor SUV and outcome in NSCLC; we have recently reviewed the available literature on this topic.11 These studies have in aggregate suggested that the SUV for FDG may be a powerful surrogate marker for outcome in NSCLC. Nevertheless, the patient populations in these studies have typically been heterogeneous, often including subjects with NSCLCs of several different stages and histologies. We thus undertook a review of the Washington University experience to determine whether we could confirm the pre-
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
dictive value of the SUV for FDG in a patients with stage I NSCLC treated uniformly with surgery and no adjuvant chemotherapy or radiation. We excluded patients with bronchioalveolar carcinoma given the low FDG uptake of many of these tumors and the apparently different biologic behavior.12
PATIENTS AND METHODS We reviewed the medical records of all patients with stage I NSCLC of nonbronchioalveolar carcinoma pathology, as staged according to the 1997 revised International System,13 who underwent curative resection from January 1, 1999, to December 31, 2003, and for whom complete pathologic staging data were available. Patients who underwent FDG-PET within 120 days before definitive resection were included in this analysis. Our institutional review board approved this study. We collected data regarding age at diagnosis, timing of FDG-PET, timing of surgery, tumor, node, metastasis staging, the maximum SUV (SUVmax) of the primary tumor,14 extent of surgery, time to recurrence, time to last follow-up, time to death, and cause of death. All original pathology reports were reviewed and data were collected from the review of all available radiologic and pathologic reports; clinic charts from thoracic surgery, radiation oncology, and medical oncology; and inpatient electronic charts. Date of death was confirmed through the Social Security Death Index. PET imaging over the course of this study was conducted on three different scanners, including a Siemens ECAT EXACT scanner (Siemens, Knoxville, Tennessee), a Phillips/ADAC CPET-Plus scanner (Phillips, Milpitas, California), and a Siemens Biograph LSO Duo hybrid PET/ computed tomography scanner (Siemens, Knoxville, Tennessee). PET imaging was performed in routine clinical fashion in accordance with a standard protocol. All of the PET studies were first interpreted clinically and then were subsequently evaluated a second time for this study by a radiologist, fellowship trained in nuclear medicine. The SUVmax of the primary bronchogenic carcinoma was measured by placing a region of interest (ROI) around the visible tumor, with careful attention to avoid inclusion of any myocardium. Standard software tools supplied by the manufacturer of each scanner were used. When the primary tumor was seen on PET, the ROI was placed around the visible tumor, with careful attention to avoid inclusion of any myocardium. For tumors that were not seen on PET, no measurement was made on conventional PET studies performed with either the ECAT EXACT or the CPET, and the patient was assigned to the low-SUV group (see later). For studies performed with the biograph where the lesion was not visible on PET, the ROI was placed around the lesion seen on computed tomography lung windows.
Statistical Analysis To assess the prognostic value of the SUVmax, we subdivided the patients into two cohorts: those with values for SUVmax below the median versus those with SUVmax greater than or equal to the median for the population. This decision was made a priori to avoid introduction of bias into the study. Overall survival (OS) was defined as the interval from sur-
PET-SUV for Prognosis in Stage I NSCLC
gery to death from any cause and those patients who remained alive were censored at the date of last clinical contact. Time to recurrence was defined as time from the date of surgery to the recurrence of lung cancer. Patients without recurrence were censored from the analysis at the time of last negative follow-up encounter, death without evidence of recurrence, or diagnosis of a second primary lung cancer. OS and recurrence rates were estimated by Kaplan-Meier product limit method. The log-rank test was used to detect the potential association between each covariate and survival or recurrence. A multivariate Cox proportional hazards model was fitted to identify independent risk factors and to adjust for the potential effects of known risk factors (age, T-classification, and histology). Interaction terms with SUVmax were entered into the model one at a time. To avoid the assumption of linear relationships as well as for a better interpretation of the results, the continuous variables, age, and SUVmax were dichotomized. Tumor size was not analyzed as a covariate because of its interdependence with T-classification. A multivariate Cox proportional hazards model was fitted to identify independent risk factors. To better understand the relationship between SUVmax and outcomes, the assumption of linearity was also assessed. In this analysis, a smoothed line is imposed on the scatter plot of SUVmax against the martingale residuals, which were obtained from the multivariate Cox model containing all variables but SUVmax. If the linearity assumption is well satisfied, the plot shows an approximately straight line. This plot also provides useful information for choosing the appropriate cut-off point when the linearity assumption is violated. The statistical analysis was performed with the statistical package SAS version 9. A p value under 0.05 was taken to indicate statistical significance and all tests were two-sided.
RESULTS We identified 136 consecutive patients diagnosed with stage I NSCLC between 1999 and 2003 who met our selection criteria. Patient characteristics are shown in Table 1. None of the patients received adjuvant chemotherapy or radiotherapy. Of 136 patients with stage I NSCLC, 77 (57%)
TABLE 1. Characteristics of Resected Stage I NSCLC Subjects, Grouped by SUVmax ⬍5.5 or ⱖ5.5 Low SUVmax Variables Total subjects Age at diagnosis, median T-classification 1 2 Size (cm), median Histology Adenocarcinoma Squamous Other
High SUVmax
n
%
n
%
65 65
51
71 69
49
46 19 1.9
71 29
31 40 3.4
44 56
46 13 6
68 23 10
25 34 12
36 45 16
p 0.004 ⬍0.001
⬍0.001 ⬍0.001
SUVmax, maximum tumor standardized uptake value.
Copyright © 2008 by the International Association for the Study of Lung Cancer
131
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
Goodgame et al.
had T1 and 59 (43%) had T2 tumors. Ten patients with T2 tumors had a tumor size of 3 cm or less but presence of pleural invasion. The median follow-up time was 46 months,
TABLE 2. Univariate Analysis for Predictors of Outcome in Resected Stage I NSCLC Recurrence Rate Variable SUVmax ⬍5.5 ⱖ5.5 T-classification T1 T2 Histology Adenocarcinoma Squamous Others Age ⬍70 ⱖ70
n
5-yr
65 71
14 37
77 59
17 36
71 47 18
27 12 55
79 57
22 28
p
Overall Survival 5-yr
0.002
p 0.006
74 53 0.035
0.002 73 52
0.016
0.19 69 65 34
0.53
0.006 74 52
SUVmax, maximum tumor standardized uptake value.
FIGURE 1. Recurrence rates based on SUVmax where vertical bars within each curve represent the distributions of censored cases. SUV, standardized uptake value; SUVmax, maximum tumor standardized uptake value.
132
and 32 patients (24%) had a recurrence of their disease. Sixteen patients died within 1 year of surgery, only two of these had a disease recurrence. Most patients had adenocarcinoma (71 patients) or squamous cell carcinoma (47), but other histologies were also present with six patients having adenosquamous carcinoma and 10 patients having large-cell carcinoma, with or without neuroendocrine features. In two patients, no histologic subtype could be determined. PET was performed at a median of 15 days before surgery (range 0 –113 days). The median SUVmax for the population was 5.5. Patients with a higher tumor SUVmax were more likely to be older and have larger or T2 tumors, and were less likely to have adenocarcinoma (Table 1). Only 10 patients with a low SUVmax had a recurrence of their disease (14%), whereas 22 patients (33%) in the high-SUVmax group had a recurrence. The estimated 5-year recurrence rate in the low-SUVmax group was 14%, compared with 37% in the high-SUVmax group (p ⫽ 0.002). The 5-year OS was correspondingly higher in the low-SUVmax group (74% versus 53%, p ⫽ 0.006; Table 2 and Figures 1 and 2). Of note, patients of other histologies (not squamous or adenocarcinoma) had particularly poor outcomes with 5-year recurrence rates and mortality rates of 55% and 66%, respectively. In the univariate analysis for other covariates, histology and T-classification were significantly associated with recurrence (Table 2). T-classification and age (⬍70 versus ⱖ70)
FIGURE 2. Overall survival based on SUVmax where vertical bars within each curve represent the distributions of censored cases. SUV, standardized uptake value; SUVmax, maximum tumor standardized uptake value.
Copyright © 2008 by the International Association for the Study of Lung Cancer
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
TABLE 3. Multivariate Analysis for Predictors of Recurrence and Survival in Resected Stage I NSCLC Recurrence Factor SUV ⱖ5.5 Nonsquamous Age ⱖ70 yr T2 tumor
Overall Survival
HR (95% CI)
p
HR (95% CI)
p
3.46 (1.57–7.65) 2.8 (1.18–6.65) 1.24 (0.61–2.5) 1.48 (0.71–3.09)
0.002 0.020 0.550 0.292
1.91 (1.03–3.57) 1.35 (0.73–2.49) 2.12 (1.19–3.79) 2.06 (1.13–3.73)
0.041 0.344 0.011 0.018
HR, hazard ratio; CI, confidence interval; SUV, standardized uptake value.
were significantly associated with OS. In the multivariate analysis, high SUVmax was independently associated with recurrence, having a hazard ratio (HR) of 3.5, with 95% confidence intervals (CI) of 1.6 to 7.7. Nonsquamous histology was also independently associated with recurrence (HR ⫽ 2.8, 95% CI ⫽ 1.2– 6.7; Table 3). High SUVmax was also independently associated with worse survival in the multivariate analysis (HR ⫽ 1.9, 95% CI ⫽ 1.03–3.57). Age and T2 tumor status were also independently associated with worse OS with HRs of 2.12 and 2.06, respectively. No interaction terms between covariates were statistically significant. An analysis of the residuals for recurrence and OS as described in the methods showed that the relationships between SUVmax and recurrence, and SUVmax and survival, were dependent but nonlinear. This analysis confirmed that median SUVmax was an appropriate cut-off point to distinguish between high- and low-risk patients based on our data (data not shown).
DISCUSSION New risk stratification markers and models are urgently needed to evaluate which patients with resected stage I NSCLC are at highest risk for relapse and stand to benefit most from adjuvant chemotherapy. In this series of 136 stage I NSCLC patients, higher preoperative SUVmax was strongly associated with poor outcomes. Several other investigators have also retrospectively evaluated their experience, attempting to correlate SUV with outcomes in patients with NSCLC, and we have recently reviewed these publications.11 Multiple studies have shown a correlation between SUV and stage, survival or disease-free survival. The majority of these publications have grouped all stages of NSCLC. Cerfolio et al. published a large series of 315 patients with resected NSCLC, which included 141 stage I patients. Higher SUV correlated with stage, as well as with recurrence and survival. Multivariate analysis of the entire cohort found that SUV ⱖ10 was the best independent predictor of disease-free survival and survival. A multivariate analysis of the stage I patients was not presented.15 In a more recent study by Ohtsuka et al., SUVmax was correlated with outcomes for 98 stage I patients, limited to those with adenocarcinoma. A value of 3.3 was found to be the optimum cut-off point. Twelve recurrences were identified and 10 of these were in the high-SUVmax group, which approached statistical significance by multivariate analysis (p ⫽ 0.079).16
PET-SUV for Prognosis in Stage I NSCLC
In the current study, the number of patients (136) and recurrences (32) allowed for a multivariate analysis of factors predictive of recurrence and survival in a pure stage I population. Patients with SUVmax ⱖ5.5 (the median for the group) had a statistically significant increase in recurrence rate and mortality, which was independent of other standard risk factors (age, histology, and T-classification) by multivariate analysis. As shown in Figure 1, the time to recurrence in patients with a high SUVmax was shorter than that in patients with a lower SUVmax (median time of 11 months versus 25 months). The precise definitions and implications of high- and low-risk PET findings have yet to be determined. Ohtsuka et al. found that an SUVmax of greater than 3.3 best defines a high-risk population,16 whereas our data suggest that an SUVmax of approximately 5 is more appropriate. Nevertheless, their study was limited to adenocarcinoma, which has been shown to have a lower SUV in other studies17,18 as well as in our own data (median SUVmax of 4.2 versus 8.4 for squamous). What is the next logical step in refining this prognostic tool? It should be fairly straightforward to develop a prospective registry of patients with resected T1 and T2 NSCLC with preoperative PET imaging performed according to standardized protocols. This would allow for more accurate estimates of relapse risk and identification of the most appropriate SUV groupings. Nevertheless, more important than the prognostic question, is the question of whether these patients would benefit from adjuvant chemotherapy. Current evidence from clinical trials does not support the use of adjuvant chemotherapy in unselected stage I NSCLC patients, but based on our retrospective data, stage I patients with high SUVmax may be more than three times more likely to have recurrences of their disease. Higher recurrence rates suggest that a survival benefit from chemotherapy might be easier to establish in this population than in an unselected group. Therefore, prospective clinical trials of adjuvant chemotherapy in patients with stage I NSCLC and SUVmax greater than 5 should be strongly considered. REFERENCES 1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, 2007. CA Cancer J Clin 2007;57:43– 66. 2. Govindan R, Page N, Morgensztern D, et al. Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: analysis of the surveillance, epidemiologic, and end results database. J Clin Oncol 2006;24:4539 – 4544. 3. Domont J, Soria JC, Le Chevalier T. Adjuvant chemotherapy in earlystage non-small cell lung cancer. Semin Oncol 2005;32:279 –283. 4. Douillard JY, Rosell R, De Lena M, et al. Adjuvant vinorelbine plus cisplatin versus observation in patients with completely resected stage IB-IIIA non-small-cell lung cancer (Adjuvant Navelbine International Trialist Association [ANITA]): a randomised controlled trial. Lancet Oncol 2006;7:719 –727. 5. Winton T, Livingston R, Johnson D, et al. Vinorelbine plus cisplatin vs. observation in resected non-small-cell lung cancer. N Engl J Med 2005;352:2589 –2597. 6. The International Adjuvant Lung Cancer Trial Collaborative Group. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N Engl J Med 2004;350:351– 360.
Copyright © 2008 by the International Association for the Study of Lung Cancer
133
Goodgame et al.
Journal of Thoracic Oncology • Volume 3, Number 2, February 2008
7. Strauss GM, Herndon J, Maddaus MA, et al. Adjuvant chemotherapy in stage IB non-small cell lung cancer (NSCLC): update of Cancer and Leukemia Group B (CALGB) protocol 9633 (Meeting Abstract). J Clin Oncol 2006;24(18 Suppl):7007. 8. Pieterman RM, van Putten JWG, Meuzelaar JJ, et al. Preoperative staging of non-small-cell lung cancer with positron-emission tomography. N Engl J Med 2000;343:254 –261. 9. Lardinois D, Weder W, Hany TF, et al. Staging of non-small-cell lung cancer with integrated positron-emission tomography and computed tomography. N Engl J Med 2003;348:2500 –2507. 10. van Tinteren H, Hoekstra OS, Smit EF, et al. Effectiveness of positron emission tomography in the preoperative assessment of patients with suspected non-small-cell lung cancer: the PLUS multicentre randomised trial. Lancet 2002;359:1388 –1392. 11. Pillot G, Siegel BA, Govindan R. Prognostic value of fluorodeoxyglucose positron emission tomography in non-small cell lung cancer: a review. J Thorac Oncol 2006;1:152–159. 12. Raz DJ, Odisho AY, Franc BL, Jablons DM. Tumor fluoro-2-deoxyD-glucose avidity on positron emission tomographic scan predicts mortality in patients with early-stage pure and mixed bronchi-
134
13. 14. 15.
16.
17. 18.
oloalveolar carcinoma. J Thorac Cardiovasc Surg 2006;132:1189 – 1195. Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997;111:1710 –1717. Kumar R, Nadig MR, Chauhan A. Positron emission tomography: clinical applications in oncology. Part 1. Expert Rev Anticancer Ther 2005;5:1079 –1094. 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;130:151–159. Ohtsuka T, Nomori H, Watanabe KI, et al. Prognostic significance of [(18)F]fluorodeoxyglucose uptake on positron emission tomography in patients with pathologic stage I lung adenocarcinoma. Cancer 2006; 107(10):2468 –2473. Higashi K, Ueda Y, Arisaka Y, et al. 18F-FDG uptake as a biologic prognostic factor for recurrence in patients with surgically resected non-small cell lung cancer. J Nucl Med 2002;43:39 – 45. Sugawara Y, Quint LE, Iannettoni MD, et al. Does the FDG uptake of primary non-small cell lung cancer predict prognosis? A work in progress. Clin Positron Imaging 1999;2:111–118.
Copyright © 2008 by the International Association for the Study of Lung Cancer