- Email: [email protected]
Tumor Volume Is a Prognostic Factor in Non–Small-Cell Lung Cancer Treated With Chemoradiotherapy
Int. J. Radiation Oncology Biol. Phys., Vol. 79, No. 5, pp. 1381–1387, 2011 Copyright Ó 2011 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/$–see front matter
doi:10.1016/j.ijrobp.2009.12.060
CLINICAL INVESTIGATION
Lung
TUMOR VOLUME IS A PROGNOSTIC FACTOR IN NON–SMALL-CELL LUNG CANCER TREATED WITH CHEMORADIOTHERAPY BRIAN M. ALEXANDER, M.D., M.P.H.,* MEGAN OTHUS, M.S.,y HALE B. CAGLAR, M.D.,z z AND AARON M. ALLEN, M.D. *Harvard Radiation Oncology Program, Harvard Medical School, Boston, MA; yBiostatistics Core, Dana-Farber Cancer Institute, Boston, MA; and zDepartment of Radiation Oncology, Dana-Farber Cancer Institute/Brigham and Women’s Hospital, Boston, MA Purpose: To investigate whether primary tumor and nodal volumes defined on radiotherapy planning scans are correlated with outcome (survival and recurrence) after combined-modality treatment. Methods and Materials: A retrospective review of patients with Stage III non–small-cell lung cancer treated with chemoradiation at Brigham and Women’s Hospital/Dana-Farber Cancer Institute from 2000 to 2006 was performed. Tumor and nodal volume measurements, as computed by Eclipse (Varian, Palo Alto, CA), were used as independent variables, along with existing clinical factors, in univariate and multivariate analyses for association with outcomes. Results: For patients treated with definitive chemoradiotherapy, both nodal volume (hazard ratio [HR], 1.09; p < 0.01) and tumor volume (HR, 1.03; p < 0.01) were associated with overall survival on multivariate analysis. Both nodal volume (HR, 1.10; p < 0.01) and tumor volume (HR, 1.04; p < 0.01) were also associated with local control but not distant metastases. Conclusions: In addition to traditional surgical staging variables, disease burden, measured by primary tumor and nodal metastases volume, provides information that may be helpful in determining prognosis and identifying groups of patients for which more aggressive local therapy is warranted. Ó 2011 Elsevier Inc. Chemoradiotherapy, Non–small-cell lung cancer, Prognosis.
INTRODUCTION
distant failure by addressing micrometastatic disease (8). Radiotherapeutic control of lung cancer may depend on the volume of disease treated; cell-kill models predict probability of tumor control based on an estimation of the cells that are present (9). Furthermore, a larger volume of disease, especially nodal disease, may be a surrogate for metastatic disease burden, thereby shifting the balance of the relative importance of local control (radiation) and distant control (chemotherapy). The current staging system has very little information that directly addresses these issues. Planning for the delivery of chemoradiotherapy increasingly takes advantage of modern imaging modalities such as computed tomography (CT) and positron emission tomography (PET) (10, 11). Three-dimensional conformal planning in this manner provides the opportunity to answer questions related to the interaction of disease burden and outcome in the setting of nonsurgical disease. We used our experience in the treatment of locally advanced lung cancer to investigate whether measuring the disease volume in both the nodal stations and the primary tumor could possibly aid in designing future therapeutic paradigms by providing information with
Combination chemoradiation is used in the treatment of locally advanced non–small-cell lung cancer (NSCLC) either as definitive therapy or as adjuvant therapy when combined with surgery (1, 2). The current TNM staging for lung cancer was developed as a surgical-pathologic system and not specifically with respect to combined-modality treatment (3). As such, the surgery-based TNM system may not provide sufficient prognostic information when surgery is not the primary mode of treatment. Indeed, studies that examine the TNM staging system in the setting of chemoradiation have found additional prognostic factors (tumor volume, number of positive nodal stations, performance status) that may be useful in making treatment decisions (4–7). Chemoradiotherapy involves two components of treatment. Radiotherapy is used as a local, definitive treatment that is expected to decrease local recurrence rates and possibly limit metastatic progression in a disease model that assumes an orderly stage progression (8). Chemotherapy may be used as a radiosensitizer or to decrease the rates of Reprint requests to: Aaron M. Allen, M.D., Department of Radiation Oncology, Davidoff Center, Rabin Medical Center, Petach Tikvah, Israel. Tel: 972-39377960; Fax: 972-39377962; E-mail: [email protected]
Conflict of interest: none. Received June 23, 2009, and in revised form Oct 26, 2009. Accepted for publication Dec 22, 2009. 1381
1382
I. J. Radiation Oncology d Biology d Physics
respect to the risks of distant and local recurrence for a given clinical scenario. METHODS AND MATERIALS A retrospective review of patients with Stage III NSCLC treated with concurrent chemoradiation at Brigham and Women’s Hospital/Dana-Farber Cancer Institute from 2000 to 2006 was performed with institutional review board approval. Staging was based on the sixth edition of the American Joint Committee on Cancer’s atlas (12). The clinical results have been previously published (13). In brief, all patients were treated with three-dimensional conformal radiation therapy (median dose, 60 Gy). Patients were evaluated for surgical resection before the start of treatment as well as after chemoradiation therapy (CRT) and were considered for resection when they
Volume 79, Number 5, 2011
had disease that was considered to be technically resectable and were medically able to undergo the procedure. Patients were followed up with CT and PET/CT scans starting at 6 weeks after the completion of therapy and then every 8 to 12 weeks for the first year of follow-up and every 12 to 16 weeks in Years 2 and 3. Thereafter patients continued to be followed up every 4 to 6 months until death or until lost to follow-up. Images from the original planning CT scans were retrieved from patients in the database, and tumor and nodal volumes were recontoured for the purpose of this analysis on the Eclipse Treatment Planning System (Varian, Palo Alto, CA). Computed tomography planning was performed on a GE LightSpeed helical scanner (GE Medical Systems, Milwaukee, WI). Tumor and nodal volume measurements, as computed by Eclipse, were entered into the database for analysis along with existing clinical factors (age, race, gender,
Table 1. Patient characteristics (staging based on AJCC sixth edition) Patient characteristics
Entire group (n = 107)
Surgery (n = 31)
No surgery (n = 76)
Age [median (range)] (y) Gender Female Male Race White African American Asian Other/unknown Performance status (WHO) 0 1 2 Histology NSCLC (NOS) Squamous Adenocarcinoma Large cell Stage IIIA IIIB T stage 1 2 3 4 N stage 0 1 2 3 FDG–PET staging Yes No Mediastinoscopy Yes No Nodal volume [median (range)] Tumor volume [median (range)] Stage IIIA Nodal volume [median (range)] Tumor volume [median (range)] Stage IIIB Nodal volume [median (range)] Tumor volume [median (range)]
60 (33–81)
59 (33–74)
61 (46–81)
50 (47%) 57 (53%)
16 (52%) 15 (48%)
34 (45%) 42 (55%)
92 (86%) 10 (9%) 1 (1%) 3 (3%)
27 (87%) 3 (10%) 0 1 (3%)
65 (86%) 7 (9%) 1 (1%) 2 (3%)
13 (12%) 89 (83%) 5 (5%)
6 (19%) 25 (81%) 0
7 (9%) 64 (84%) 5 (7%)
38 (36%) 27 (25%) 38 (36%) 4 (4%)
7 (23%) 6 (19%) 18 (58%) 0
31 (41%) 21 (28%) 20 (26%) 4 (5%)
46 (43%) 61 (57%)
24 (77%) 7 (23%)
22 (29%) 54 (71%)
11 55 11 30
7 17 2 5
4 38 9 25
8 3 57 39
5 0 24 2
3 3 33 37
88 (82%) 18 (17%)
26 (84%) 5 (16%)
62 (82%) 13 (17%)
71 (66%) 45 (33%) 17.25 (0–400.60) 24.60 (0–596.60)
28 (90%) 3 (10%) 7.20 (0–89.65) 38.24 (0–306.90)
43 (57%) 32 (42%) 25.67 (0–400.60) 15.20 (0–596.60)
11.96 (0–140.50) 19.55 (0–227.00)
7.39 (0–85.49) 35.70 (0.72–227.00)
16.88 (0–140.50) 8.72 (0–164.50)
25.19 (0–400.60) 31.35 (0–596.60)
3.53 (0–89.65) 48.63 (0–306.90)
30.49 (0–400.60) 22.88 (0–595.60)
Abbreviations: AJCC = American Joint Committee on Cancer; WHO = World Health Organization; NSCLC = non–small-cell lung cancer; NOS = not otherwise specified; FDG = fluorodeoxyglucose; PET = positron emission tomography.
Prognostic value of tumor volume in NSCLC d B. M. ALEXANDER et al.
1383
Table 2. Treatment characteristics Treatment characteristics
Entire group (n = 107) Surgery (n = 31) No surgery (n = 76)
Chemotherapy Induction + concurrent Concurrent Concurrent + consolidation Concurrent chemotherapy Weekly carboplatin + paclitaxel Cisplatin + etoposide every 4 wk Surgery Lobectomy and bilobectomy Pneumonectomy Wedge resection No surgery RT dose Median (Gy) (range) <54 Gy 54 Gy 55–59 Gy 60 Gy 61–65 Gy 66 Gy 68 Gy 70 Gy
37 (35%) 39 (36%) 31 (29%)
9 (29%) 13 (42%) 9 (29%)
28 (37%) 26 (34%) 22 (29%)
70 (65%) 37 (35%)
15 (48%) 16 (52%)
55 (72%) 21 (28%)
54 (54–70) 0 16 (52%) 0 3 (10%) 0 3 (10%) 8 (26%) 1 (3%)
60 (46–70) 1 (1%) 5 (7%) 4 (5%) 25 (33%) 4 (5%) 17 (22%) 14 (18%) 6 (8%)
76 (71%) 60 (46–70) 1 (1%) 21 (20%) 4 (4%) 28 (26%) 4 (4%) 20 (19%) 22 (21%) 7 (7%)
Abbreviation: RT = radiotherapy. and stage). Tumor and nodal volumes were identified by enlargement on CT scan and visual correlation with PET scans. Two radiation oncologists reviewed these new volumes for accuracy and consistency. Nodal volumes were not segregated by nodal station; the volume of nodal disease is reported as a composite. This was done to provide a more distinct measurement to N stage and to limit bias in designating nodal station borders (14). Tumor volumes were contoured primarily by use of lung windows and nodal volumes with mediastinal windows. All contours were reviewed for consistency by the study team. The Kaplan-Meier method was used to determine median overall survival and associated confidence intervals. For both the entire cohort and the subset of patients who were treated with chemoradiation alone (no surgical component), univariate Cox proportional hazards models were used to examine associations between presenting clinical factors and treatment details with the outcomes of local failure,
distant failure, and overall survival. Multivariate Cox proportional hazards were determined by use of a stepwise method. Tumor and nodal volumes were treated as continuous variables with hazard ratios (HRs) assigning risk associated with each 10-cm3 increase in volume. Volumes were also analyzed categorized into quartiles to determine whether any threshold effects could be seen. The Fisher exact test was used to compare the highest quartile volume (tumor and nodal) with the lowest quartile with whether a patient received surgery or not.
RESULTS Clinical characteristics A total of 107 patients received CRT for Stage III NSCLC between January 2000 and October 2006 and had
Table 3. Univariate and multivariate analysis for factors associated with overall survival for overall cohort Univariate analysis
Age (reference <70 y) Stage (reference = IIIA) Gender (reference = female) Race (reference = not white) CT (reference = carboplatin + paclitaxel) Consolidation CT (reference = no) Induction (reference = no) Surgery (reference = no) PET (reference = no) PET follow-up (reference = no) RT dose (reference <60 Gy) Nodal volume (10 cm3) Tumor volume (10 cm3)
Hazard ratio
p Value
1.51 3.09 0.90 0.75 1.32 1.28 0.89 0.13 0.82 0.56 2.19 1.09 1.01
0.82 < 0.01 0.71 0.47 0.35 0.41 0.69 < 0.01 0.56 0.04 0.03 < 0.01 0.47
Multivariate analysis Hazard ratio
p Value
1.09
< 0.01
Abbreviations: CT = computed tomography; PET = positron emission tomography; RT = radiotherapy.
1384
I. J. Radiation Oncology d Biology d Physics
Volume 79, Number 5, 2011
Fig. 1. Kaplan-Meier curve describing (a) overall survival, (b) local control, and (c) distant control of all patients in study (n = 107) separated into quartiles based on nodal volume.
measurements of nodal and tumor volume. Of the patients, 46 (43%) had Stage IIIA disease and 61 (57%) had Stage IIIB disease. Nearly all patients (95%) had an excellent performance status (0 or 1). A majority of patients were men (53%). Of the 107 patients, 76 had chemoradiation alone (without surgery); 22 of these (29%) had Stage IIIA disease, whereas 54 (71%) had Stage IIIB. Clinical and treatment characteristics are shown in Tables 1 and 2, respectively. The median follow-up time of the entire group was 15 months (range, 2–61 months), and the median potential follow-up time was 32 months (range, 12–84 months). The median overall survival time for the entire group was 23 months (95% confidence interval [CI], 16 months to no upper limit). Patients treated with CRT alone had a median overall survival time of 15 months (95% CI, 11–23 months).
only variable associated with overall survival. Both nodal volume (HR, 1.10; 95% CI, 1.05–1.16; p < 0.01) and tumor volume (HR, 1.03; 95% CI, 1.00–1.06; p = 0.04), as well as stage (HR, 3.08; 95% CI, 1.55–6.11; p < 0.01), were associated with local control. Stage (HR, 2.04; 95% CI, 1.13–3.70; p = 0.02) and addition of surgery (HR, 0.50; 95% CI, 0.26– 0.99; p = 0.05) were the only variables associated with distant failure on multivariate analysis. As expected, patients with higher tumor volume (p = 0.02) and nodal volume (p = 0.01) were more likely to undergo chemoradiation alone rather than trimodality therapy. The impact of increased nodal volume on overall survival is represented in Fig. 1 after categorization by quartile. The HR with respect to overall survival of the highest quartile of nodal volume compared with the lowest was 5.18.
Entire cohort On univariate analysis, stage, addition of surgery, radiation dose, and nodal volume were all correlated with overall survival (Table 3). Multivariate Cox regression analysis showed nodal volume (HR, 1.09; 95% CI, 1.05–1.13; p < 0.01) as the
Definitive chemoradiotherapy cohort Because surgical resection was a strong predictor of overall survival in the whole cohort, we redid the analysis removing these patients. Univariate predictors of overall survival are shown in Table 4. On multivariate analysis, both nodal
Prognostic value of tumor volume in NSCLC d B. M. ALEXANDER et al.
1385
Fig. 2. Kaplan-Meier curve describing (a) overall survival, (b) local control, and (c) distant control of patients in study receiving chemoradiation alone (n = 76) separated into quartiles based on nodal volume.
volume (HR, 1.09; 95% CI, 1.05–1.14; p < 0.01) and tumor volume (HR, 1.03; 95% CI, 1.01–1.06; p < 0.01) were associated with overall survival in addition to gender (HR for men, 0.49; 95% CI, 0.258–0.883; p = 0.02). Nodal volume (HR, 1.10; 95% CI, 1.04–1.15; p < 0.01) and tumor volume (HR, 1.04; 95% CI, 1.02–1.08; p < 0.01) were associated with local control on multivariate analysis. Stage (HR, 2.46; 95% CI, 1.18–5.12; p = 0.02) was the only predictor for distant failure on multivariate analysis. No significant relationships could be found when volumes were broken down into quartiles (Fig. 2). DISCUSSION Stage III NSCLC is a heterogeneous disease. Patterns-offailure studies have shown that the most common site of recurrence is distant (13, 15, 16), with the central nervous system being the most common site of metastasis (17). In fact, isolated local recurrences are rare. In a previous article reporting on patterns of failure of the current cohort, only
11% had an isolated local recurrence (13). If most NSCLCs will recur distantly in response to inadequate systemic therapy, then strategies for increasing local control (i.e., radiation dose escalation) may be adding toxicity without commensurate benefit. In our study stage outweighed nodal volume in multivariate analysis using distant recurrence as an endpoint. Nodal volume, however, did remain a predictor of local control and overall survival in multivariate analysis, perhaps reflecting a radiobiological truth not reflected in the current surgical staging system. The TNM staging system currently in place for NSCLC is based on the Mountain surgical data from the MD Anderson Cancer Center (3). Staging segregates patients mostly by the site of the primary tumor and the location of involved lymph nodes. The primary tumor location is extremely important in determining whether the mass is resectable, whereas the node location suggests resectability and provides a surrogate for the likelihood of micrometastatic disease. What is missing from this staging system is a direct measure of disease burden. We know from other surgical series that disease burden
1386
I. J. Radiation Oncology d Biology d Physics
Volume 79, Number 5, 2011
Table 4. Univariate and multivariate analysis for factors associated with overall survival for chemoradiation-alone cohort Univariate analysis Hazard ratio
p Value
1.22 2.20 0.71 0.90 2.78 1.47 0.65 0.75 0.64 1.31 1.07 1.02
0.54 0.03 0.24 0.80 < 0.01 0.21 0.16 0.43 0.12 0.54 < 0.01 0.16
Age (ref <70 y) Stage (reference = IIIA) Gender (reference = female) Race (reference = not white) CT (reference = carboplatin + paclitaxel) Consolidation CT (reference = no) Induction (reference = no) PET (reference = no) PET follow-up (reference = no) RT dose (reference <60 Gy) Nodal volume (10 cm3) Tumor volume (10 cm3)
Multivariate analysis Hazard ratio
p Value
0.49
0.02
1.09 1.03
< 0.01 < 0.01
Abbreviations: CT = computed tomography; PET = positron emission tomography; RT = radiotherapy.
is an important addition to the current staging system—Andre et al. (18) showed that patients with N2 disease evident on preoperative imaging and multiple nodal stations had a much lower 5-year survival rate (3%) than those with N2 disease in only one nodal station that was only found at the time of surgery (34%). This effect of disease burden, not reflected in the TNM staging, may be even more important for patients undergoing chemoradiation alone based on radiobiological principles (9). Disease burden is extremely important with respect to tumor control models of radiation therapy—a given dose induces a log cell kill, meaning that the larger the tumor, the more cells, and thus the more radiation needed for local control (9). In medically inoperable NSCLC for example, gross tumor volume is associated with overall survival (1, 19– 21), and this relationship may be mitigated by increased dose in early-stage patients (20). The treatment of NSCLC involves decisions regarding the utilization and timing of multiple therapeutic modalities. Information relating to the risk of recurrence and the site of likely recurrence would be helpful in making such decisions. In our study both greater nodal volume and tumor volume were associated with decreased local control and overall survival, but not distant metastases. This result may reflect the superiority of the current staging system in identifying those patients in whom distant metastases are more likely to develop (stage was a predictor of distant recurrence in our study) but inferiority in defining parameters that affect local control in patients treated with chemoradiotherapy. When groups were segregated into quartiles by nodal disease burden, there was clearly a group that did poorly with regard to overall survival. This may have been a result of increased local failure as well as functioning as a surrogate for micrometastatic disease. That nodal volume did not correlate with metastatic disease on multivariate analysis in our results may be related to the power of the study and/or colinearity with stage. Patients with high nodal volume and those with
N3 disease likely have a large degree of overlap, but the volume of nodal disease is not captured in the N stage and might provide additional information. Patients in the highest quartile of nodal disease burden did extremely poorly, suggesting that this disease may be metastatic at presentation and not likely to benefit as much from local control in the absence of effective systemic therapy. The prognostic effect of nodal volume was not found in other studies. Dehing-Oberije et al. (4) looked at the effect of gross tumor volume (primary tumor plus all nodal volume) and number of involved nodal stations, as measured by PET– CT, on survival in patients with Stage I–III NSCLC treated with chemoradiotherapy. They did not find a correlation between nodal volume and survival on univariate analysis, even though there were correlations between two other measures of nodes and survival: number of positive lymph node stations (0, 1–2, or $3) and N stage. Gross total tumor volume and number of nodal stations were also prognostic of survival in the final model and obviated the usefulness of N stage (4). The population in this study included Stages I through III in contrast to our study, which looked solely at Stage III patients. Perhaps the more advanced disease in our population provided the necessary background upon which nodal volume functioned in a predictive capacity. Another group had similar findings, showing that primary tumor volume and total tumor volume (primary tumor volume plus nodal volume) were associated with survival, but nodal volume alone was not (1). Again, the study population had closer parity of Stage IIIA and IIIB patients than in our study and reflected less advanced disease overall. Our study does have some limitations, mainly given the sample size and its retrospective nature. We would have liked to have determine whether nodal station level was a surrogate for nodal volume or provided additional biological information. Given the relatively small size of the cohort and the colinearity expected with these variables, we did not attempt to include both variables in a model. By definition in this
Prognostic value of tumor volume in NSCLC d B. M. ALEXANDER et al.
retrospective study, there was some heterogeneity of our cohort, which could have introduced other variables not selected for that may have impacted on the final outcome of our results. CONCLUSION Surgical staging systems offer incomplete information with regard to prognosis for patients treated with chemoradiotherapy. Disease burden, in terms of both primary tumor
1387
and nodal metastases, provides additional information that may be helpful in determining whether chemoradiotherapy will be useful for local control and as a surrogate for the likely presence of micrometastatic disease. Future studies should include a measure of actual disease burden in addition to stage to identify those patients for whom aggressive local therapy and the associated morbidity would not provide additional benefit so that we might better tailor therapy to meet the needs of each patient.
REFERENCES 1. Basaki K, Abe Y, Aoki M, et al. Prognostic factors for survival in Stage III non–small-cell lung cancer treated with definitive radiation therapy: Impact of tumor volume. Int J Radiat Oncol Biol Phys 2006;64:449–454. 2. Kunitoh H, Suzuki K. How to evaluate the risk/benefit of trimodality therapy in locally advanced non-small-cell lung cancer. Br J Cancer 2007;96:1498–1503. 3. Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997;111:1710–1717. 4. Dehing-Oberije C, De Ruysscher D, van der Weide H, et al. Tumor volume combined with number of positive lymph node stations is a more important prognostic factor than TNM stage for survival of non–small-cell lung cancer patients treated with (chemo)radiotherapy. Int J Radiat Oncol Biol Phys 2008;70: 1039–1044. 5. Brundage MD, Davies D, Mackillop WJ. Prognostic factors in non-small cell lung cancer: A decade of progress. Chest 2002; 122:1037–1057. 6. Firat S, Byhardt RW, Gore E. Comorbidity and Karnofsky performance score are independent prognostic factors in Stage III non–small-cell lung cancer: An institutional analysis of patients treated on four RTOG studies. Radiation Therapy Oncology Group. Int J Radiat Oncol Biol Phys 2002;54:357–364. 7. Pfister DG, Johnson DH, Azzoli CG, et al. American Society of Clinical Oncology treatment of unresectable non-small-cell lung cancer guideline: Update 2003. J Clin Oncol 2004;22: 330–353. 8. Eberhardt W, Pottgen C, Stuschke M. Chemoradiation paradigm for the treatment of lung cancer. Nat Clin Pract Oncol 2006;3:188–199. 9. Hall EJ, Giaccia AJ. Radiobiology for the radiologist. 6th ed. Philadelphia: Lippincott Williams & Wilkins; 2006. 10. Marks LB, Sibley G. The rationale and use of three-dimensional radiation treatment planning for lung cancer. Chest 1999;116: 539S–545S. 11. Vansteenkiste JF, Stroobants SG. Positron emission tomography in the management of non-small cell lung cancer. Hematol Oncol Clin North Am 2004;18:269–288. 12. Greene FL, American Joint Committee on Cancer. AJCC cancer staging atlas. New York: Springer; 2006.
13. Caglar H, Baldini EH, Othus M, et al. Outcomes of patients with stage III NSCLC treated with chemotherapy and radiation with and without surgery. Cancer 2009;115:4156–4166. 14. Vallieres E, Shepherd FA, Crowley J, et al. The IASLC Lung Cancer Staging Project: Proposals regarding the relevance of TNM in the pathologic staging of small cell lung cancer in the forthcoming (seventh) edition of the TNM classification for lung cancer. J Thorac Oncol 2009;4:1049–1059. 15. Gandara DR, Chansky K, Albain KS, et al. Consolidation docetaxel after concurrent chemoradiotherapy in stage IIIB nonsmall-cell lung cancer: Phase II Southwest Oncology Group Study S9504. J Clin Oncol 2003;21:2004–2010. 16. Vokes EE, Herndon JE II, Kelley MJ, et al. Induction chemotherapy followed by chemoradiotherapy compared with chemoradiotherapy alone for regionally advanced unresectable stage III non-small-cell lung cancer: Cancer and Leukemia Group B. J Clin Oncol 2007;25:1698–1704. 17. Mamon HJ, Yeap BY, Janne PA, et al. High risk of brain metastases in surgically staged IIIA non-small-cell lung cancer patients treated with surgery, chemotherapy, and radiation. J Clin Oncol 2005;23:1530–1537. 18. Andre F, Grunenwald D, Pignon JP, et al. Survival of patients with resected N2 non-small-cell lung cancer: Evidence for a subclassification and implications. J Clin Oncol 2000;18:2981– 2989. 19. Werner-Wasik M, Swann RS, Bradley J, et al. Increasing tumor volume is predictive of poor overall and progression-free survival: Secondary analysis of the Radiation Therapy Oncology Group 93-11 phase I-II radiation dose-escalation study in patients with inoperable non–small-cell lung cancer. Int J Radiat Oncol Biol Phys 2008;70:385–390. 20. Zhao L, West BT, Hayman JA, et al. High radiation dose may reduce the negative effect of large gross tumor volume in patients with medically inoperable early-stage non–small cell lung cancer. Int J Radiat Oncol Biol Phys 2007;68:103–110. 21. Bradley JD, Ieumwananonthachai N, Purdy JA, et al. Gross tumor volume, critical prognostic factor in patients treated with three-dimensional conformal radiation therapy for non-smallcell lung carcinoma. Int J Radiat Oncol Biol Phys 2002;52: 49–57.
doi:10.1016/j.ijrobp.2009.12.060
CLINICAL INVESTIGATION
Lung
TUMOR VOLUME IS A PROGNOSTIC FACTOR IN NON–SMALL-CELL LUNG CANCER TREATED WITH CHEMORADIOTHERAPY BRIAN M. ALEXANDER, M.D., M.P.H.,* MEGAN OTHUS, M.S.,y HALE B. CAGLAR, M.D.,z z AND AARON M. ALLEN, M.D. *Harvard Radiation Oncology Program, Harvard Medical School, Boston, MA; yBiostatistics Core, Dana-Farber Cancer Institute, Boston, MA; and zDepartment of Radiation Oncology, Dana-Farber Cancer Institute/Brigham and Women’s Hospital, Boston, MA Purpose: To investigate whether primary tumor and nodal volumes defined on radiotherapy planning scans are correlated with outcome (survival and recurrence) after combined-modality treatment. Methods and Materials: A retrospective review of patients with Stage III non–small-cell lung cancer treated with chemoradiation at Brigham and Women’s Hospital/Dana-Farber Cancer Institute from 2000 to 2006 was performed. Tumor and nodal volume measurements, as computed by Eclipse (Varian, Palo Alto, CA), were used as independent variables, along with existing clinical factors, in univariate and multivariate analyses for association with outcomes. Results: For patients treated with definitive chemoradiotherapy, both nodal volume (hazard ratio [HR], 1.09; p < 0.01) and tumor volume (HR, 1.03; p < 0.01) were associated with overall survival on multivariate analysis. Both nodal volume (HR, 1.10; p < 0.01) and tumor volume (HR, 1.04; p < 0.01) were also associated with local control but not distant metastases. Conclusions: In addition to traditional surgical staging variables, disease burden, measured by primary tumor and nodal metastases volume, provides information that may be helpful in determining prognosis and identifying groups of patients for which more aggressive local therapy is warranted. Ó 2011 Elsevier Inc. Chemoradiotherapy, Non–small-cell lung cancer, Prognosis.
INTRODUCTION
distant failure by addressing micrometastatic disease (8). Radiotherapeutic control of lung cancer may depend on the volume of disease treated; cell-kill models predict probability of tumor control based on an estimation of the cells that are present (9). Furthermore, a larger volume of disease, especially nodal disease, may be a surrogate for metastatic disease burden, thereby shifting the balance of the relative importance of local control (radiation) and distant control (chemotherapy). The current staging system has very little information that directly addresses these issues. Planning for the delivery of chemoradiotherapy increasingly takes advantage of modern imaging modalities such as computed tomography (CT) and positron emission tomography (PET) (10, 11). Three-dimensional conformal planning in this manner provides the opportunity to answer questions related to the interaction of disease burden and outcome in the setting of nonsurgical disease. We used our experience in the treatment of locally advanced lung cancer to investigate whether measuring the disease volume in both the nodal stations and the primary tumor could possibly aid in designing future therapeutic paradigms by providing information with
Combination chemoradiation is used in the treatment of locally advanced non–small-cell lung cancer (NSCLC) either as definitive therapy or as adjuvant therapy when combined with surgery (1, 2). The current TNM staging for lung cancer was developed as a surgical-pathologic system and not specifically with respect to combined-modality treatment (3). As such, the surgery-based TNM system may not provide sufficient prognostic information when surgery is not the primary mode of treatment. Indeed, studies that examine the TNM staging system in the setting of chemoradiation have found additional prognostic factors (tumor volume, number of positive nodal stations, performance status) that may be useful in making treatment decisions (4–7). Chemoradiotherapy involves two components of treatment. Radiotherapy is used as a local, definitive treatment that is expected to decrease local recurrence rates and possibly limit metastatic progression in a disease model that assumes an orderly stage progression (8). Chemotherapy may be used as a radiosensitizer or to decrease the rates of Reprint requests to: Aaron M. Allen, M.D., Department of Radiation Oncology, Davidoff Center, Rabin Medical Center, Petach Tikvah, Israel. Tel: 972-39377960; Fax: 972-39377962; E-mail: [email protected]
Conflict of interest: none. Received June 23, 2009, and in revised form Oct 26, 2009. Accepted for publication Dec 22, 2009. 1381
1382
I. J. Radiation Oncology d Biology d Physics
respect to the risks of distant and local recurrence for a given clinical scenario. METHODS AND MATERIALS A retrospective review of patients with Stage III NSCLC treated with concurrent chemoradiation at Brigham and Women’s Hospital/Dana-Farber Cancer Institute from 2000 to 2006 was performed with institutional review board approval. Staging was based on the sixth edition of the American Joint Committee on Cancer’s atlas (12). The clinical results have been previously published (13). In brief, all patients were treated with three-dimensional conformal radiation therapy (median dose, 60 Gy). Patients were evaluated for surgical resection before the start of treatment as well as after chemoradiation therapy (CRT) and were considered for resection when they
Volume 79, Number 5, 2011
had disease that was considered to be technically resectable and were medically able to undergo the procedure. Patients were followed up with CT and PET/CT scans starting at 6 weeks after the completion of therapy and then every 8 to 12 weeks for the first year of follow-up and every 12 to 16 weeks in Years 2 and 3. Thereafter patients continued to be followed up every 4 to 6 months until death or until lost to follow-up. Images from the original planning CT scans were retrieved from patients in the database, and tumor and nodal volumes were recontoured for the purpose of this analysis on the Eclipse Treatment Planning System (Varian, Palo Alto, CA). Computed tomography planning was performed on a GE LightSpeed helical scanner (GE Medical Systems, Milwaukee, WI). Tumor and nodal volume measurements, as computed by Eclipse, were entered into the database for analysis along with existing clinical factors (age, race, gender,
Table 1. Patient characteristics (staging based on AJCC sixth edition) Patient characteristics
Entire group (n = 107)
Surgery (n = 31)
No surgery (n = 76)
Age [median (range)] (y) Gender Female Male Race White African American Asian Other/unknown Performance status (WHO) 0 1 2 Histology NSCLC (NOS) Squamous Adenocarcinoma Large cell Stage IIIA IIIB T stage 1 2 3 4 N stage 0 1 2 3 FDG–PET staging Yes No Mediastinoscopy Yes No Nodal volume [median (range)] Tumor volume [median (range)] Stage IIIA Nodal volume [median (range)] Tumor volume [median (range)] Stage IIIB Nodal volume [median (range)] Tumor volume [median (range)]
60 (33–81)
59 (33–74)
61 (46–81)
50 (47%) 57 (53%)
16 (52%) 15 (48%)
34 (45%) 42 (55%)
92 (86%) 10 (9%) 1 (1%) 3 (3%)
27 (87%) 3 (10%) 0 1 (3%)
65 (86%) 7 (9%) 1 (1%) 2 (3%)
13 (12%) 89 (83%) 5 (5%)
6 (19%) 25 (81%) 0
7 (9%) 64 (84%) 5 (7%)
38 (36%) 27 (25%) 38 (36%) 4 (4%)
7 (23%) 6 (19%) 18 (58%) 0
31 (41%) 21 (28%) 20 (26%) 4 (5%)
46 (43%) 61 (57%)
24 (77%) 7 (23%)
22 (29%) 54 (71%)
11 55 11 30
7 17 2 5
4 38 9 25
8 3 57 39
5 0 24 2
3 3 33 37
88 (82%) 18 (17%)
26 (84%) 5 (16%)
62 (82%) 13 (17%)
71 (66%) 45 (33%) 17.25 (0–400.60) 24.60 (0–596.60)
28 (90%) 3 (10%) 7.20 (0–89.65) 38.24 (0–306.90)
43 (57%) 32 (42%) 25.67 (0–400.60) 15.20 (0–596.60)
11.96 (0–140.50) 19.55 (0–227.00)
7.39 (0–85.49) 35.70 (0.72–227.00)
16.88 (0–140.50) 8.72 (0–164.50)
25.19 (0–400.60) 31.35 (0–596.60)
3.53 (0–89.65) 48.63 (0–306.90)
30.49 (0–400.60) 22.88 (0–595.60)
Abbreviations: AJCC = American Joint Committee on Cancer; WHO = World Health Organization; NSCLC = non–small-cell lung cancer; NOS = not otherwise specified; FDG = fluorodeoxyglucose; PET = positron emission tomography.
Prognostic value of tumor volume in NSCLC d B. M. ALEXANDER et al.
1383
Table 2. Treatment characteristics Treatment characteristics
Entire group (n = 107) Surgery (n = 31) No surgery (n = 76)
Chemotherapy Induction + concurrent Concurrent Concurrent + consolidation Concurrent chemotherapy Weekly carboplatin + paclitaxel Cisplatin + etoposide every 4 wk Surgery Lobectomy and bilobectomy Pneumonectomy Wedge resection No surgery RT dose Median (Gy) (range) <54 Gy 54 Gy 55–59 Gy 60 Gy 61–65 Gy 66 Gy 68 Gy 70 Gy
37 (35%) 39 (36%) 31 (29%)
9 (29%) 13 (42%) 9 (29%)
28 (37%) 26 (34%) 22 (29%)
70 (65%) 37 (35%)
15 (48%) 16 (52%)
55 (72%) 21 (28%)
54 (54–70) 0 16 (52%) 0 3 (10%) 0 3 (10%) 8 (26%) 1 (3%)
60 (46–70) 1 (1%) 5 (7%) 4 (5%) 25 (33%) 4 (5%) 17 (22%) 14 (18%) 6 (8%)
76 (71%) 60 (46–70) 1 (1%) 21 (20%) 4 (4%) 28 (26%) 4 (4%) 20 (19%) 22 (21%) 7 (7%)
Abbreviation: RT = radiotherapy. and stage). Tumor and nodal volumes were identified by enlargement on CT scan and visual correlation with PET scans. Two radiation oncologists reviewed these new volumes for accuracy and consistency. Nodal volumes were not segregated by nodal station; the volume of nodal disease is reported as a composite. This was done to provide a more distinct measurement to N stage and to limit bias in designating nodal station borders (14). Tumor volumes were contoured primarily by use of lung windows and nodal volumes with mediastinal windows. All contours were reviewed for consistency by the study team. The Kaplan-Meier method was used to determine median overall survival and associated confidence intervals. For both the entire cohort and the subset of patients who were treated with chemoradiation alone (no surgical component), univariate Cox proportional hazards models were used to examine associations between presenting clinical factors and treatment details with the outcomes of local failure,
distant failure, and overall survival. Multivariate Cox proportional hazards were determined by use of a stepwise method. Tumor and nodal volumes were treated as continuous variables with hazard ratios (HRs) assigning risk associated with each 10-cm3 increase in volume. Volumes were also analyzed categorized into quartiles to determine whether any threshold effects could be seen. The Fisher exact test was used to compare the highest quartile volume (tumor and nodal) with the lowest quartile with whether a patient received surgery or not.
RESULTS Clinical characteristics A total of 107 patients received CRT for Stage III NSCLC between January 2000 and October 2006 and had
Table 3. Univariate and multivariate analysis for factors associated with overall survival for overall cohort Univariate analysis
Age (reference <70 y) Stage (reference = IIIA) Gender (reference = female) Race (reference = not white) CT (reference = carboplatin + paclitaxel) Consolidation CT (reference = no) Induction (reference = no) Surgery (reference = no) PET (reference = no) PET follow-up (reference = no) RT dose (reference <60 Gy) Nodal volume (10 cm3) Tumor volume (10 cm3)
Hazard ratio
p Value
1.51 3.09 0.90 0.75 1.32 1.28 0.89 0.13 0.82 0.56 2.19 1.09 1.01
0.82 < 0.01 0.71 0.47 0.35 0.41 0.69 < 0.01 0.56 0.04 0.03 < 0.01 0.47
Multivariate analysis Hazard ratio
p Value
1.09
< 0.01
Abbreviations: CT = computed tomography; PET = positron emission tomography; RT = radiotherapy.
1384
I. J. Radiation Oncology d Biology d Physics
Volume 79, Number 5, 2011
Fig. 1. Kaplan-Meier curve describing (a) overall survival, (b) local control, and (c) distant control of all patients in study (n = 107) separated into quartiles based on nodal volume.
measurements of nodal and tumor volume. Of the patients, 46 (43%) had Stage IIIA disease and 61 (57%) had Stage IIIB disease. Nearly all patients (95%) had an excellent performance status (0 or 1). A majority of patients were men (53%). Of the 107 patients, 76 had chemoradiation alone (without surgery); 22 of these (29%) had Stage IIIA disease, whereas 54 (71%) had Stage IIIB. Clinical and treatment characteristics are shown in Tables 1 and 2, respectively. The median follow-up time of the entire group was 15 months (range, 2–61 months), and the median potential follow-up time was 32 months (range, 12–84 months). The median overall survival time for the entire group was 23 months (95% confidence interval [CI], 16 months to no upper limit). Patients treated with CRT alone had a median overall survival time of 15 months (95% CI, 11–23 months).
only variable associated with overall survival. Both nodal volume (HR, 1.10; 95% CI, 1.05–1.16; p < 0.01) and tumor volume (HR, 1.03; 95% CI, 1.00–1.06; p = 0.04), as well as stage (HR, 3.08; 95% CI, 1.55–6.11; p < 0.01), were associated with local control. Stage (HR, 2.04; 95% CI, 1.13–3.70; p = 0.02) and addition of surgery (HR, 0.50; 95% CI, 0.26– 0.99; p = 0.05) were the only variables associated with distant failure on multivariate analysis. As expected, patients with higher tumor volume (p = 0.02) and nodal volume (p = 0.01) were more likely to undergo chemoradiation alone rather than trimodality therapy. The impact of increased nodal volume on overall survival is represented in Fig. 1 after categorization by quartile. The HR with respect to overall survival of the highest quartile of nodal volume compared with the lowest was 5.18.
Entire cohort On univariate analysis, stage, addition of surgery, radiation dose, and nodal volume were all correlated with overall survival (Table 3). Multivariate Cox regression analysis showed nodal volume (HR, 1.09; 95% CI, 1.05–1.13; p < 0.01) as the
Definitive chemoradiotherapy cohort Because surgical resection was a strong predictor of overall survival in the whole cohort, we redid the analysis removing these patients. Univariate predictors of overall survival are shown in Table 4. On multivariate analysis, both nodal
Prognostic value of tumor volume in NSCLC d B. M. ALEXANDER et al.
1385
Fig. 2. Kaplan-Meier curve describing (a) overall survival, (b) local control, and (c) distant control of patients in study receiving chemoradiation alone (n = 76) separated into quartiles based on nodal volume.
volume (HR, 1.09; 95% CI, 1.05–1.14; p < 0.01) and tumor volume (HR, 1.03; 95% CI, 1.01–1.06; p < 0.01) were associated with overall survival in addition to gender (HR for men, 0.49; 95% CI, 0.258–0.883; p = 0.02). Nodal volume (HR, 1.10; 95% CI, 1.04–1.15; p < 0.01) and tumor volume (HR, 1.04; 95% CI, 1.02–1.08; p < 0.01) were associated with local control on multivariate analysis. Stage (HR, 2.46; 95% CI, 1.18–5.12; p = 0.02) was the only predictor for distant failure on multivariate analysis. No significant relationships could be found when volumes were broken down into quartiles (Fig. 2). DISCUSSION Stage III NSCLC is a heterogeneous disease. Patterns-offailure studies have shown that the most common site of recurrence is distant (13, 15, 16), with the central nervous system being the most common site of metastasis (17). In fact, isolated local recurrences are rare. In a previous article reporting on patterns of failure of the current cohort, only
11% had an isolated local recurrence (13). If most NSCLCs will recur distantly in response to inadequate systemic therapy, then strategies for increasing local control (i.e., radiation dose escalation) may be adding toxicity without commensurate benefit. In our study stage outweighed nodal volume in multivariate analysis using distant recurrence as an endpoint. Nodal volume, however, did remain a predictor of local control and overall survival in multivariate analysis, perhaps reflecting a radiobiological truth not reflected in the current surgical staging system. The TNM staging system currently in place for NSCLC is based on the Mountain surgical data from the MD Anderson Cancer Center (3). Staging segregates patients mostly by the site of the primary tumor and the location of involved lymph nodes. The primary tumor location is extremely important in determining whether the mass is resectable, whereas the node location suggests resectability and provides a surrogate for the likelihood of micrometastatic disease. What is missing from this staging system is a direct measure of disease burden. We know from other surgical series that disease burden
1386
I. J. Radiation Oncology d Biology d Physics
Volume 79, Number 5, 2011
Table 4. Univariate and multivariate analysis for factors associated with overall survival for chemoradiation-alone cohort Univariate analysis Hazard ratio
p Value
1.22 2.20 0.71 0.90 2.78 1.47 0.65 0.75 0.64 1.31 1.07 1.02
0.54 0.03 0.24 0.80 < 0.01 0.21 0.16 0.43 0.12 0.54 < 0.01 0.16
Age (ref <70 y) Stage (reference = IIIA) Gender (reference = female) Race (reference = not white) CT (reference = carboplatin + paclitaxel) Consolidation CT (reference = no) Induction (reference = no) PET (reference = no) PET follow-up (reference = no) RT dose (reference <60 Gy) Nodal volume (10 cm3) Tumor volume (10 cm3)
Multivariate analysis Hazard ratio
p Value
0.49
0.02
1.09 1.03
< 0.01 < 0.01
Abbreviations: CT = computed tomography; PET = positron emission tomography; RT = radiotherapy.
is an important addition to the current staging system—Andre et al. (18) showed that patients with N2 disease evident on preoperative imaging and multiple nodal stations had a much lower 5-year survival rate (3%) than those with N2 disease in only one nodal station that was only found at the time of surgery (34%). This effect of disease burden, not reflected in the TNM staging, may be even more important for patients undergoing chemoradiation alone based on radiobiological principles (9). Disease burden is extremely important with respect to tumor control models of radiation therapy—a given dose induces a log cell kill, meaning that the larger the tumor, the more cells, and thus the more radiation needed for local control (9). In medically inoperable NSCLC for example, gross tumor volume is associated with overall survival (1, 19– 21), and this relationship may be mitigated by increased dose in early-stage patients (20). The treatment of NSCLC involves decisions regarding the utilization and timing of multiple therapeutic modalities. Information relating to the risk of recurrence and the site of likely recurrence would be helpful in making such decisions. In our study both greater nodal volume and tumor volume were associated with decreased local control and overall survival, but not distant metastases. This result may reflect the superiority of the current staging system in identifying those patients in whom distant metastases are more likely to develop (stage was a predictor of distant recurrence in our study) but inferiority in defining parameters that affect local control in patients treated with chemoradiotherapy. When groups were segregated into quartiles by nodal disease burden, there was clearly a group that did poorly with regard to overall survival. This may have been a result of increased local failure as well as functioning as a surrogate for micrometastatic disease. That nodal volume did not correlate with metastatic disease on multivariate analysis in our results may be related to the power of the study and/or colinearity with stage. Patients with high nodal volume and those with
N3 disease likely have a large degree of overlap, but the volume of nodal disease is not captured in the N stage and might provide additional information. Patients in the highest quartile of nodal disease burden did extremely poorly, suggesting that this disease may be metastatic at presentation and not likely to benefit as much from local control in the absence of effective systemic therapy. The prognostic effect of nodal volume was not found in other studies. Dehing-Oberije et al. (4) looked at the effect of gross tumor volume (primary tumor plus all nodal volume) and number of involved nodal stations, as measured by PET– CT, on survival in patients with Stage I–III NSCLC treated with chemoradiotherapy. They did not find a correlation between nodal volume and survival on univariate analysis, even though there were correlations between two other measures of nodes and survival: number of positive lymph node stations (0, 1–2, or $3) and N stage. Gross total tumor volume and number of nodal stations were also prognostic of survival in the final model and obviated the usefulness of N stage (4). The population in this study included Stages I through III in contrast to our study, which looked solely at Stage III patients. Perhaps the more advanced disease in our population provided the necessary background upon which nodal volume functioned in a predictive capacity. Another group had similar findings, showing that primary tumor volume and total tumor volume (primary tumor volume plus nodal volume) were associated with survival, but nodal volume alone was not (1). Again, the study population had closer parity of Stage IIIA and IIIB patients than in our study and reflected less advanced disease overall. Our study does have some limitations, mainly given the sample size and its retrospective nature. We would have liked to have determine whether nodal station level was a surrogate for nodal volume or provided additional biological information. Given the relatively small size of the cohort and the colinearity expected with these variables, we did not attempt to include both variables in a model. By definition in this
Prognostic value of tumor volume in NSCLC d B. M. ALEXANDER et al.
retrospective study, there was some heterogeneity of our cohort, which could have introduced other variables not selected for that may have impacted on the final outcome of our results. CONCLUSION Surgical staging systems offer incomplete information with regard to prognosis for patients treated with chemoradiotherapy. Disease burden, in terms of both primary tumor
1387
and nodal metastases, provides additional information that may be helpful in determining whether chemoradiotherapy will be useful for local control and as a surrogate for the likely presence of micrometastatic disease. Future studies should include a measure of actual disease burden in addition to stage to identify those patients for whom aggressive local therapy and the associated morbidity would not provide additional benefit so that we might better tailor therapy to meet the needs of each patient.
REFERENCES 1. Basaki K, Abe Y, Aoki M, et al. Prognostic factors for survival in Stage III non–small-cell lung cancer treated with definitive radiation therapy: Impact of tumor volume. Int J Radiat Oncol Biol Phys 2006;64:449–454. 2. Kunitoh H, Suzuki K. How to evaluate the risk/benefit of trimodality therapy in locally advanced non-small-cell lung cancer. Br J Cancer 2007;96:1498–1503. 3. Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997;111:1710–1717. 4. Dehing-Oberije C, De Ruysscher D, van der Weide H, et al. Tumor volume combined with number of positive lymph node stations is a more important prognostic factor than TNM stage for survival of non–small-cell lung cancer patients treated with (chemo)radiotherapy. Int J Radiat Oncol Biol Phys 2008;70: 1039–1044. 5. Brundage MD, Davies D, Mackillop WJ. Prognostic factors in non-small cell lung cancer: A decade of progress. Chest 2002; 122:1037–1057. 6. Firat S, Byhardt RW, Gore E. Comorbidity and Karnofsky performance score are independent prognostic factors in Stage III non–small-cell lung cancer: An institutional analysis of patients treated on four RTOG studies. Radiation Therapy Oncology Group. Int J Radiat Oncol Biol Phys 2002;54:357–364. 7. Pfister DG, Johnson DH, Azzoli CG, et al. American Society of Clinical Oncology treatment of unresectable non-small-cell lung cancer guideline: Update 2003. J Clin Oncol 2004;22: 330–353. 8. Eberhardt W, Pottgen C, Stuschke M. Chemoradiation paradigm for the treatment of lung cancer. Nat Clin Pract Oncol 2006;3:188–199. 9. Hall EJ, Giaccia AJ. Radiobiology for the radiologist. 6th ed. Philadelphia: Lippincott Williams & Wilkins; 2006. 10. Marks LB, Sibley G. The rationale and use of three-dimensional radiation treatment planning for lung cancer. Chest 1999;116: 539S–545S. 11. Vansteenkiste JF, Stroobants SG. Positron emission tomography in the management of non-small cell lung cancer. Hematol Oncol Clin North Am 2004;18:269–288. 12. Greene FL, American Joint Committee on Cancer. AJCC cancer staging atlas. New York: Springer; 2006.
13. Caglar H, Baldini EH, Othus M, et al. Outcomes of patients with stage III NSCLC treated with chemotherapy and radiation with and without surgery. Cancer 2009;115:4156–4166. 14. Vallieres E, Shepherd FA, Crowley J, et al. The IASLC Lung Cancer Staging Project: Proposals regarding the relevance of TNM in the pathologic staging of small cell lung cancer in the forthcoming (seventh) edition of the TNM classification for lung cancer. J Thorac Oncol 2009;4:1049–1059. 15. Gandara DR, Chansky K, Albain KS, et al. Consolidation docetaxel after concurrent chemoradiotherapy in stage IIIB nonsmall-cell lung cancer: Phase II Southwest Oncology Group Study S9504. J Clin Oncol 2003;21:2004–2010. 16. Vokes EE, Herndon JE II, Kelley MJ, et al. Induction chemotherapy followed by chemoradiotherapy compared with chemoradiotherapy alone for regionally advanced unresectable stage III non-small-cell lung cancer: Cancer and Leukemia Group B. J Clin Oncol 2007;25:1698–1704. 17. Mamon HJ, Yeap BY, Janne PA, et al. High risk of brain metastases in surgically staged IIIA non-small-cell lung cancer patients treated with surgery, chemotherapy, and radiation. J Clin Oncol 2005;23:1530–1537. 18. Andre F, Grunenwald D, Pignon JP, et al. Survival of patients with resected N2 non-small-cell lung cancer: Evidence for a subclassification and implications. J Clin Oncol 2000;18:2981– 2989. 19. Werner-Wasik M, Swann RS, Bradley J, et al. Increasing tumor volume is predictive of poor overall and progression-free survival: Secondary analysis of the Radiation Therapy Oncology Group 93-11 phase I-II radiation dose-escalation study in patients with inoperable non–small-cell lung cancer. Int J Radiat Oncol Biol Phys 2008;70:385–390. 20. Zhao L, West BT, Hayman JA, et al. High radiation dose may reduce the negative effect of large gross tumor volume in patients with medically inoperable early-stage non–small cell lung cancer. Int J Radiat Oncol Biol Phys 2007;68:103–110. 21. Bradley JD, Ieumwananonthachai N, Purdy JA, et al. Gross tumor volume, critical prognostic factor in patients treated with three-dimensional conformal radiation therapy for non-smallcell lung carcinoma. Int J Radiat Oncol Biol Phys 2002;52: 49–57.