Xanthopoulos, Eric P. MD, JD*; Corradetti, Michael N. MD, PhD*; Mitra, Nandita PhD†; Fernandes, Annemarie T. MD*; Kim, Miranda MD, MBA*; Grover, Surbhi MD*; Christodouleas, John P. MD, MPH*; Evans, Tracey L. MD‡; Stevenson, James P. MD‡; Langer, Corey J. MD‡; Lee, Tony T. MD‖; Pryma, Daniel A. MD§; Lin, Lilie L. MD*; Simone, Charles B. MD*; Apisarnthanarax, Smith MD*; Rengan, Ramesh MD, PhD*
This study suggests that positron emission tomography computed tomography (PET-CT) improves staging accuracy and intrathoracic disease identification in limited-stage small-cell lung cancer (LS-SCLC), which translates into an improvement in clinical outcome in these patients. Although retrospective, these data suggest that PET-CT should be routinely used in the staging of newly diagnosed patients with SCLC. These data need to be further confirmed in a prospective clinical trial.
Each year, 13% of all newly diagnosed lung cancer patients are diagnosed with SCLC.1 Approximately 39% of patients with SCLC are diagnosed with LS disease treated with chemotherapy and definitive radiation therapy.1 Staging information is essential because of the high propensity for metastatic disease in SCLC, and the identification of metastases can spare patients from the toxicity associated with thoracic radiotherapy. Furthermore, in those patients who do receive radiotherapy, knowing the exact extent of intrathoracic disease may permit more accurate treatment volume delineation.
Until 2011, the National Comprehensive Cancer Network (NCCN) recommended a (99m)Tc-MDP bone scan as part of the initial evaluation of all newly diagnosed SCLC patients. However, in 2012, the NCCN began recommending 18-fluorodeoxyglucose (18F) PET-CT in lieu of bone scan in its initial workup algorithm. No prospective trial has compared the sensitivity or specificity of PET-CT with (99m)Tc-MDP scanning in the detection of metastatic bone disease in SCLC. In 2012, Lee et al., in the Annals of Nuclear Medicine, retrospectively reported sensitivity and specificity of 100% for metastatic bone disease on a per-patient basis with PET-CT, which compared with 37% and 92%, respectively, with bone scan. The study also observed 100% positive and negative predictive values for PET as opposed to 69% and 24%, respectively, with bone scan.
PET has emerged in the last decade as an important tool in the staging and delineation of disease for conformal radiotherapy planning of non-SCLC. In 2009, Medicare approved the use of PET for the initial staging of SCLC.3 Medicare allows a significantly higher national average fee for whole-body PET ($1020, CPT 78816) relative to bone scan ($206, CPT 78315). Yet PET’s role in SCLC remains unclear.2
It is believed that PET may more accurately detect patients with extensive-stage disease than CT-staging alone.4 This stage migration allows physicians to withhold potentially toxic radiation therapy from poorer prognosis extensive-stage patients who would not benefit from it (Fig. 1A). With better ability to identify patients who will likely respond to treatment, stage-specific survival will improve.5
At the same time, PET may demonstrate with greater accuracy the extent of intrathoracic tumor and nodal disease (Fig. 1B). Improved treatment targeting may result in improved local disease control, which may translate into a real survival increment.We hypothesize that the addition of PET to CT staging is associated with improved progression-free and overall survival in patients with SCLC. To our knowledge, no one has yet reported an association between the use of PET and survival in SCLC.2
PATIENTS AND METHODS
Patient Inclusion Criteria
With the University of Pennsylvania Institutional Review Board approval, we identified 54 LS-SCLC patients consecutively treated with definitive concurrent chemoradiation from January 2002 to August 2010. Patient staging was based on all available pathologic and radiographic studies, including CT, PET, magnetic resonance imaging (MRI), and bone scintigraphy. None of the patients received surgery, CT of the brain (all patients underwent cranial MRI imaging),or a bone marrow biopsy.
Patient Diagnosis and Follow-up
All patients had a complete history, that is a physical examination, abdominothoracopelvic CT with or without PET scan, and a brain CT or MRI, at diagnosis. All patients had pathologically confirmed small-cell lung carcinoma. Since 2002 it has been an institutional practice to include PET in staging, when possible. However, the recently treated patients in this study were more likely to undergo PET staging because of changes in reimbursement for PET in SCLC.
The standard staging approach for SCLC patients is to classify them as having a limited- or extensive-stage disease. Limited-stage disease is defined as disease that can be encompassed within a tolerable radiation portal. For this study, we further assigned each patient a TNM stage, on the basis of the American Joint Committee on Cancer Staging Manual, 7th edition.
N status was determined primarily by size and morphologic criteria as interpreted by the lead radiologist. None of the patients in the study underwent mediastinoscopy. In general, suspicious nodes were defined as those clearly visible above the mediastinal background on PET. These nodes were then compared with the CT for benign findings, such as nodes that had a very small short axis or fatty hilum. Although no rigorous criteria were used, maximum standard uptake value corrected for body weight greater than 2 or 3 and node axes greater than 1 cm prompted closer scrutiny. Ultimately, clinical judgment was applied for marginal cases. Two radiologists reviewed each case and were in agreement on the staging.
After treatment, standard chest CT and/or PET scans were used to evaluate for disease recurrence. Patients were followed up approximately every 3 months for the first 1 to 3 years after treatment and every 6 to 12 months thereafter. All patients had a minimum of 6 months’ follow-up postradiotherapy.
Treatment Techniques and Parameters
The gross tumor volume (GTV) consisted of all known sites of primary and nodal thoracic disease based on imaging findings and pathologic staging. For patients who underwent four-dimensional (4D)-CT planning, an internal GTV was generated based on the motion envelope of the GTV. A clinical target volume (CTV) expansion of 0.8 cm was added to the internal GTV to treat regions at risk for microscopic spread of tumor. For patients who did not undergo 4D-CT planning, the GTV was expanded by 1.5 cm for lower lobe tumors and 1.0 cm for upper lobe tumors, to generate the CTV. A uniform planning target volume expansion of 0.5 cm was then applied to the CTV to account for setup error.
Most patients (85%) received 0.150 Gy twice-daily fractions to 45 Gy, and patients who could not tolerate twice-daily treatments, received 0.180 Gy once a day to 56-72 Gy. We delivered the maximally safe dosage to the planning target volume while constraining the lung to the following parameters: V20 35% or lesser, V5 60% or lesser, and mean lung dosage less than or equal to 20 Gy. In patients for whom once-daily treatment was delivered, the final prescription dosage was based upon several factors, including dosimetric constraints and performance status.
All patients received at least one cycle of either concurrent or sequential chemotherapy. Patients usually received cisplatin or carboplatin in combination with another chemotherapeutic agent, typically etoposide (Table 1).
Local Control Evaluation
Tumor burden and response were scored using the guidelines established by the revised Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1.
Patient comorbidities were assessed using the Charlson comorbidity index. The index predicts the 10-year mortality for patients based on their comorbidity profile. We assigned one, two, three, or six points to more than 20 conditions, based on the risk of dying from each one, using the scale below. We then summed the total score to assess mortality.
- One point: myocardial infarction, congestive heart failure, peripheral vascular disease, dementia, uncomplicated diabetes, cerebrovascular disease, chronic lung disease, connective tissue disease, ulcer, and chronic liver disease without portal hypertension.
- Two points: hemiplegia, moderate or severe kidney disease, diabetes with end organ disease, local malignancy, and lymphoma.
- Three points: moderate or severe liver disease.
- Six points: metastatic tumor or acquired immunodeficiency syndrome.
Each subject’s overall survival was computed from the date of diagnosis by imaging (or date of first treatment for a parallel analysis) until death or date of last follow-up when they were censored. We used Kaplan–Meier and log-rank testing to compare unadjusted survival profiles. Cox proportional hazards regression controlled for potential confounders. Any variable with p value of 0.10 or lesser in the univariate model was eligible for multivariable Cox modeling. A threshold of p value lesser than or equal to 0.05 established significance on multivariate analysis.
We examined local and distant control rates using two different methodologies. First, we report unadjusted local and distant control rates on any event basis. This approach includes all patients in the control analysis, regardless of whether they previously had a distant failure and vice versa. Second, we report the local and distant control rates on a first-event basis while adjusting for the competing risk of death using nonparametric cumulative incidence functions.6
We used Cox regression to identify individual covariates associated with survival and to adjust for potential confounders. Any covariate with a univariate p value of 0.10 or lesser was included in the model. We used the Schoenfeld residuals test to verify the underlying proportional hazards assumption.
We compared the distribution of patient characteristics between the two treatment groups with Pearson’s χ2 test, Fisher’s exact test, or nonparametric equality of medians test.
Descriptive statistics were performed on STATA/IC (version 11.0 for Mac OS X; StataCorp, College Station, TX). Competing risk analysis was conducted using R (version 2.15.0 Mac OS X; R Foundation for Statistical Computing, Vienna, Austria).
Our study included 54 LS-SCLC patients. Table 1 shows the distribution of patient characteristics among patients staged with and without 18F-fluorodeoxyglucose (FDG) PET. The two cohorts had similar age, race, sex, performance status, T-stage, lung lobe involvement, tumor laterality, comorbidities, proportion staged with bone scan, median time from diagnosis to start of radiation, percentages of patients who started radiation while hospitalized, chemotherapy regimens and cycles, and proportion of patients who received prophylactic cranial and (postmetastatic) whole-brain radiation therapy. Both groups received similar prescription dosages (median 4500 cGy) and number of daily fractions.
PET-staged patients had significant differences in N-staging at diagnosis compared with those staged with CT alone. PET-staged patients were more likely to have any nodal metastasis (74% versus 50%) and advanced N3 nodal metastasis (10% versus 0%) at diagnosis. There was a trend for recently treated patients to undergo PET staging (PET-staged median treatment year = 2008 versus CT-stage median= 2006; p = 0.12).
Survival and Local Control
Median follow-up was 38 months (range, 14–97 months) in the 19 surviving PET-staged patients and 40 months in the two surviving non-PET-staged patients (range, 29–51 months; p = 0.59). LS-SCLC patients who underwent PET as a part of initial staging had significantly improved survival compared with those who did not. PET-staged patients had a higher median overall survival of 32 months (95% confidence intervals [CI]: ≥24) compared with 17 months (95% CI: 11–32) in all patients not staged with PET (p = 0.03) (Fig. 2). Median survival did not change (32 versus 17 months; p = 0.03) even after excluding the five non-PET-staged patients who did not receive a bone scan either. Patients staged with PET also had an improved 1-year (75% versus 64%) and 3-year (47% versus 19%) overall survival rates (p = 0.03) from time of diagnosis.
PET staging was associated with longer time-to-distant failure. On an unadjusted any-event basis, PET-staged patients had median time-to-distant failure of 29 months (95% CI ≥18) compared with 11 months (95% CI: 4–16) in patients not staged with PET (p < 0.01) (Fig. 3A). After adjusting for the competing risk of death on a first event basis, PET-staged patients had median time-to-distant failure of 29 months compared with 12 months in patients not staged with PET (p = 0.04) (Fig. 4A).
PET staging was also associated with time-to-local failure. On an unadjusted any event basis, PET-staged patients did not reach their median time-to-local failure compared with 26 months (95% CI ≥9) in patients not staged with PET (p = 0.05) (Fig. 3B). After adjusting for the competing risk of death on a first event basis, PET-staged patients still did not reach their median time-to-local failure compared with 16 months in patients not staged with PET (p = 0.04) (Fig. 4B).
On multivariate analysis, PET staging (odds ratio [OR]= 0.24; p = 0.04), performance status (OR = 1.89; p = 0.05), and N-stage (OR = 4.94; p <0.01) were significantly associated with overall survival. Table 2 shows all the covariates adjusted for in the Cox model.
Pretreatment PET staging of LS-SCLC was associated with improved survival. PET-staged patients had an improved 3-year overall survival from diagnosis (47% versus 19%; p = 0.03) compared with those with LS-SCLC who were not staged with PET. They also had longer times-to-local and distant failures. PET’s ability to (1) foster stage migration that accurately distinguishes limited- from extensive-stage patients (Fig. 1A), and (2) more
thoroughly identify intrathoracic disease for radiation treatment planning (Fig. 1B) may, in part, explain these observed differences between the two treatment groups.
To our knowledge, this is the first study to report on overall and disease-free survival in SCLC patients staged with and without PET. A number of earlier studies have shown that PET added to conventional staging has better sensitivity detecting sites of local- and distant-disease involvement than conventional staging alone.2,4,7–10 Prior studies have also shown high occurrences of patient management changes attributed to PET.4,7,8,11,12 But these studies have not shown that an association between PET staging and management changes are associated with a survival benefit. Additional, prospective investigation would help confirm the survival benefit of PET staging demonstrated in this study. Although a number of prospective trials—including the Ontarion Clinical Oncology Group (OCOG) Impact of Positron Emission Tomography in Stage III Non-Small Cell Lung Cancer: A Prospective Randomized Trial (PET START) by Ung et al.13—are or will be investigating the role of PET in non-SCLC,13–17 none have been commissioned for SCLC.
In addition to PET staging, several other factors are associated with survival. Echoing earlier series, poorer performance status (OR = 1.89; p = 0.05),18,19 and nodal metastases at diagnosis (OR = 4.54; p < 0.01)19 were associated with worse survival in this study. Other factors—including age, race, sex, and the proportion of patients with bone and head scans—were well balanced across the treatment groups and did not readily explain the observed local control, distant control, and survival differences (Table 1).
We next investigated whether confounders could explain the observed survival difference. Most characteristics were well balanced across the treatment groups and did not readily explain the survival gap. But the two groups did differ in one respect: They had a significantly different distribution of nodal disease at diagnosis (p = 0.05), with PET-staged patients more likely diagnosed with any nodal metastasis (74% versus 50%) and advanced N3 nodal metastasis (10% versus 0%). This, however, harmonized with previous reports, indicating that staging without PET underdiagnosed mediastinal lymph node disease, and most likely did not drive the survival decrement in the two groups.2,7–9
Besides being associated with overall survival, PET staging was associated with increased local and distant progression-free survival. PET-staged patients experienced distant failure less rapidly than those who were not (median 29 versus 12 months on a first-event basis adjusting for the competing risk of death; p = 0.04) (Fig. 4A). Several studies have shown PET’s superior ability to identify distant disease, particularly osseous metastases (Fig. 1).2,20,21 Some patients not staged with PET may have had microscopic or otherwise occult distant disease not detected during their initial staging with CT alone. As a result, these patients seem to develop distant progression faster than their PET-staged counterparts.
Likewise, PET-staged patients progressed to local failure less quickly than those who were not staged with PET (median not reached versus 16 months on first-event basis adjusting for competing risk of death; p = 0.04) (Fig. 4B). PET’s ability to identify more accurately the full extent of disease at diagnosis may explain why. Numerous studies have shown the difficulties of identifying the primary tumor’s full extent without PET staging (Fig. 2).2,4,7–10,22 If hard-to-detect regions like these were not specifically targeted for treatment, local failure rates might increase, and as uncontrolled disease progressed, it would give the illusion of a shorter time to local failure in patients not staged with PET.
As in any retrospective study, selection bias because of unmeasured confounders may explain the potentially worse survival rate in patients staged without PET. For example, although more recently diagnosed patients were more likely to undergo PET staging (PET-staged median treatment year 2008) than patients treated in the more distant past (CT-stage median 2006; p = 0.12), there was some overlap. It is unclear how insurance coverage—a possible proxy for socioeconomic status—compared across the two cohorts. At the same time, although the fundamental therapeutic approach has remained the same, improvements in technology over the 6-year study span may have confounded the observed association between PET staging and survival: Beginning in 2008, we began to use four-dimensional treatment planning in many of our LC patients to account for tumor respiratory motion during radiation delivery. But as noted earlier, patients treated after 2008 also had a significantly greater likelihood of undergoing PET staging. Separately, other possible indicia of disease severity, such as serum alkaline phosphatase and lactate dehydrogenase, were not sufficiently available for comparison across the two groups. Finally, although most patients in both groups received a platin with etoposide, it is impossible in a retrospective study to exactly match chemotherapy across the treatment groups. These potential biases mean that interpretation of our data should be viewed as preliminary, and larger studies are needed to confirm our findings.
The addition of PET in the routine staging of SCLC seems to be associated with improved time-to-local and distant failure and overall survival. These findings are likely attributed to more accurate extra- and intrathoracic staging of SCLC with PET, and underscore its value in staging of patients with SCLC.
1. 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
2. Thomson D, Hulse P, Lorigan P, Faivre-Finn C. The role of positron emission tomography in management of small cell lung cancer. Lung Cancer. 2011; 73:121–126
3. In: Services DoHaH Medicare National Coverage Determinations Manual. 2012; Chapter 1, part 4:Centers for Medicare and Medicaid Services pp 200–300
4. Seidenfeld JS, Samson DJ, Bonnell C, Ziegler KM, Aronson N. In: Services USDoHaH Managementof small cell lung cancer. 2006; Washington, D.C. Agency for Healthcare Research and Quality (AHRQ)
5. Feinstein AR, Sosin DM, Wells CK. The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med. 1985; 312:1604–1608
6. Scrucca LS, Aversa F. Competing risk analysis using R: an easy guide for clinicians. Bone Marr Transpl. 2007; 40:381–87
7. Kamel EM, Zwahlen D, Wyss MT, Stumpe KD, von Schulthess GK, Steinert HC. Whole-body (18)F-FDG PET improves the management of patients with small cell lung cancer. J Nucl Med. 2003; 44:1911–1917
8. Bradley JD, Dehdashti F, Mintun MA, Govindan R, Trinkaus K, Siegel BA. Positron emission tomography in limited-stage small-cell lung cancer: a prospective study. J Clin Oncol. 2004; 22:3248–3254
9. Senan S, De Ruysscher D. Critical review of PET-CT for radiotherapy planning in lung cancer. Crit Rev Oncol Hematol. 2005; 56:345–351
10. Shen YY, Shiau YC, Wang JJ, Ho ST, Kao CH. Whole-body 18F-2-deoxyglucose positron emission tomography in primary staging small cell lung cancer. Anticancer Res. 2002; 22:(2B)1257–1264
11. Blum R, MacManus MP, Rischin D, Michael M, Ball D, Hicks RJ. Impact of positron emission tomography on the management of patients with small-cell lung cancer: preliminary experience. Am J Clin Oncol. 2004; 27:164–171
12. Brink I, Schumacher T, Mix M, et al. Impact of [18F]FDG-PET on the primary staging of small-cell lung cancer. Eur J Nucl Med Mol Imag. 2004; 31:1614–1620
13. Ung Y GK, Sun R, et al. An Ontario Clinical Oncology Group (OCOG) randomized trial (PET START) of FDG PET/CT in patients with stage III non-small cell lung cancer (NSCLC): predictors of overall survival. J Clin Oncol. 2011; 29:2011
supp; abstr 7018)
14. Using FDG-PET during radiation therapy in non-small cell lung cancer, (HUM15709). University of Michigan Cancer Center.
ClinicalTrials.gov identifier NCT01190527. Recruiting as of March 22, 2011
15. Study of positron emission tomography and computed tomography in guiding radiation therapy in patients with stage III non-small cell lung cancer. National Cancer Institute (NCI).
ClinicalTrials.gov identified NCT01507428. Recruiting as of March 26, 2013
16. Study using induction chemotherapy and intensity-modulated radiation therapy guided by combined CT and PET imaging for patients with non-small cell lung cancer. Alberta Health Services.
ClinicalTrials.gov identifier NCT00128999. Completed February 8, 2010
17. Study of 18F-fluorodeoxyglucose (FluGlucoScan) in patients receiving a treatment planning study of 3 dimensional conformal radiation therapy guided by breath held CT and PET imaging for patients with non-small cell lung cancer. Alberta Health Services.
ClinicalTrials.gov identifier NCT00123747 Completed March 15, 2012
18. Kawahara M, Fukuoka M, Saijo N, et al. Prognostic factors and prognostic staging system for small cell lung cancer. Jap J Clin Oncol. 1997; 27:(3)158–65
19. Thatcher N, Anderson H, Burt P, Stout R. The value of anatomic staging and other prognostic factors in small cell lung cancer management: a view of European studies. Semin Radiat Oncol. 1995; 5:19–26
20. Gayed I, Vu T, Johnson M, Macapinlac H, Podoloff D. Comparison of bone and 2-deoxy-2-[18F]fluoro-D-glucose positron emission tomography in the evaluation of bony metastases in lung cancer. Mol Imaging Biol. 2003; 5:26–31
21. Cheran SK, Herndon JE 2nd, Patz EF Jr. Comparison of whole-body FDG-PET to bone scan for detection of bone metastases in patients with a new diagnosis of lung cancer. Lung Cancer. 2004; 44:317–325
22. Vinjamuri M, Craig M, Campbell-Fontaine A, Almubarak M, Gupta N, Rogers JS. Can positron emission tomography be used as a staging tool for small-cell lung cancer? Clin Lung Cancer. 2008; 9:30–34