- Email: [email protected]
Metabolic tumor burden as marker of outcome in advanced EGFR wild-type NSCLC patients treated with erlotinib
Lung Cancer 94 (2016) 81–87
Contents lists available at ScienceDirect
Lung Cancer journal homepage: www.elsevier.com/locate/lungcan
Metabolic tumor burden as marker of outcome in advanced EGFR wild-type NSCLC patients treated with erlotinib Anne Winther-Larsen a,∗ , Joan Fledelius b , Boe Sandahl Sorensen a , Peter Meldgaard c a b c
Department of Clinical Biochemistry, Aarhus University Hospital, Aarhus, Denmark Department of Nuclear Medicine, Herning Regional Hospital, Herning, Denmark Department of Oncology, Aarhus University Hospital, Aarhus, Denmark
a r t i c l e
i n f o
Article history: Received 28 December 2015 Received in revised form 25 January 2016 Accepted 30 January 2016 Keywords: Positron emission tomography Lung cancer Tyrosine kinase inhibitor EGFR wild-type Metabolic tumor volume
a b s t r a c t Objectives: Accurate estimation of the prognosis of advanced non-small cell lung cancer (NSCLC) patients is essential before initiation of palliative treatment; especially in the second and third-line setting. This study was conducted in order to evaluate tumor burden measured on an 2 -deoxy-2 -[18F] fluoro-Dglucose (F-18-FDG) positron emission tomography/computed tomography (PET/CT) scan as a marker of outcome in advanced epidermal growth factor receptor (EGFR) wild-type patients treated with second or third-line erlotinib. Material and methods: Fifty-one patients were included from a prospectively collected cohort. An F-18FDG-PET/CT scan was conducted prior to erlotinib treatment and tumor burden was measured in terms of metabolic tumor volume (MTV) and total lesion glycolysis (TLG). Median values of MTV and TLG were used for dichotomization of patients. Survival outcome was compared between groups. Results: MTV and TLG could be measured in 49 patients. High values of MTV and TLG were significantly correlated with shorter PFS (p < 0.001 and p = 0.027, respectively) and OS (p < 0.001 and p = 0.002, respectively). In multivariate analyses, including both clinical and imaging data, high MTV and TLG remained strong independent markers of both shorter PFS (MTV, hazard ratio (HR) = 5.44 (95% confidence interval (CI) 2.46–12.02); TLG, HR = 2.17 (95% CI 1.11–4.26)) and OS (MTV, HR = 4.80 (95% CI 2.08–11.06); TLG, HR = 2.76 (95% CI 1.33–5.71)). Conclusion: High MTV and TLG are independently correlated with shorter PFS and OS in advanced EGFR wild-type NSCLC patients treated with second or third-line erlotinib. Metabolic tumor burden is a highly promising clinical tool that may allow better patient selection for palliative treatment in the future. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Background The prognosis of non-small cell lung cancer (NSCLC) patients is heterogeneous. Extent of disease at diagnosis is one of the most useful prognostic markers [1]. All patients are staged by the tumornode-metastases (TNM) classification [2] and treatment strategy is based on the stage of disease. Patients with metastatic, stage IV disease are incurable with short life expectancy and choice of treatment in these patients must be carefully balanced between the possible benefit of treatment and side effects. A precise evaluation of the individual patient’s prognosis is therefore of the greatest importance. However, evaluation of survival time is challenging for clinicians and they have a tendency to overestimate the prognosis
∗ Corresponding author at: Department of Clinical Biochemistry, Aarhus University Hospital, Noerrebrogade 44, 8000 Aarhus C, Denmark. E-mail address: [email protected] (A. Winther-Larsen). http://dx.doi.org/10.1016/j.lungcan.2016.01.024 0169-5002/© 2016 Elsevier Ireland Ltd. All rights reserved.
[3]. This leads to a growing tendency to overtreatment with patients receiving therapy near end of life [4]. New tools to guide prognosis estimation and choice of treatment are therefore highly warranted. In the recent years, growing interest has emerged for the use of 2 -deoxy-2 -[18F] fluoro-d-glucose (F-18-FDG) positron emission tomography/computed tomography (PET/CT) scans as an imaging biomarker of prognosis and treatment effect. The most widely used PET indicator of metabolic activity is the maximal standardized uptake value (SUVmax ). However, the usefulness of SUVmax may be limited since it only represents a single voxel value and not sufficiently reflects the entire tumor burden. Therefore, interest has therefore emerged for the use of volume-based PET measurements since they are capable of measuring the metabolic tumor burden by incorporating both metabolic activity and volumetric data [5]. Metabolic tumor volume (MTV) is solely a volumetric parameter delineated by a specific metabolic threshold while tumor lesion glycolysis (TLG) in addition incorporates the metabolic activity of the
82
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
tumor. In previous studies, MTV and TLG evaluated before first-line of treatment have shown promising results as markers of outcome in stage IV NSCLC patients [6–9]. However, the value of metabolic tumor burden assessment prior to following lines of treatment has not been evaluated. Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR TKIs) are a treatment option for advanced EGFR wild-type (wt) patients in the second or third-line palliative setting [10,11]. Though, response rates are only approximately 8% [12] and the only recommended factor to guide patient selection is performance status (PS) [10–12]. Moreover, these patients have a short life expectancy and better markers are highly needed to prevent use of unnecessary and potentially harmful treatment at the end of life. Thus, the purpose of this study was to evaluate if MTV and TLG could be a markers of outcome in advanced EGFR wt NSCLC patients deemed to start second or third-line treatment with erlotinib.
2. Materials and methods 2.1. Patients and treatment In this prospective, single-centre study, we enrolled advanced NSCLC patients starting treatment with erlotinib in a palliative setting from April 2013 until August 2015 at the Department of Oncology, Aarhus University Hospital, Denmark. Patients were eligible for enrolment if they fulfilled the following criteria for erlotinib treatment: histologically or cytologically proven NSCLC, age ≥18 years, performance status ≤2 and no prior TKI treatment. If patients had uncontrolled diabetes mellitus or severe respiratory problems, they were excluded from enrolment. All patients gave informed written consent before inclusion in the study and the study was approved by the Central Denmark Region Committees on Biomedical Research Ethics (no. 1-10-72-19-12). For the purpose of this study, we included only patients from the total cohort who were EGFR wt and received erlotinib as second or third-line of treatment. All enrolled patients received an initial dose of 150 mg erlotinib daily. If side effects appeared, erlotinib was continued together with symptomatic management of toxicity before any dose reduction. However, if the patient had continuous side effects of grade 2 or more according to the CTC criteria version 4.0, dose-reduction was performed. The patients were treated until radiologic or clinical progression, unacceptable toxicity, or death. All patients had a pre-treatment CT scan of the chest and abdomen performed prior to start of erlotinib. Follow-up was done by CT scans every 12 weeks during the treatment period combined with clinical and biochemical evaluation every fourth week in the first 12 weeks and subsequently every sixth week as standard of care in our institution. At the end of treatment, the patients were evaluated regarding either subsequent treatment with pemetrexed or docetaxel or best supportive care according to standard of care at our institution. Treatment response evaluated on CT scans was defined according to Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1) criteria [13]. Neuroimaging was performed on clinical indication. In patients with adenocarcinoma, EGFR mutation testing was performed as part of the routine diagnostic work-up by use of the Therascreen EGFR RGQ PCR kit (QIAGEN, Manchester, UK) according to the manufacturer’s protocol. Testing for EML4-ALK fusions were incorporated as part of the routine diagnostic work-up in adenocarcinoma patients during the inclusion period and performed by fluorescence in situ hybridization. TNM stage was evaluated at time of inclusion according to the 7th edition of the American Joint Committee on Cancer [2]. Data on baseline characteristics and response was collected from medical files.
Table 1 Basic characteristics for all patients (N = 51). Characteristics
All patients N (%)
Age, years Median (range)
67 (49–83)
Gender Female Male
23 (45) 28 (55)
PS, ECOG 0 1 2
4 (8) 37 (72) 10 (20)
Smoking status Never Formera Current
1 (2) 37 (73) 13 (25)
Stage IIIa IIIb IV
2 (4) 2 (4) 47(92)
Brain metastases Yes No
8 (16) 43 (84)
Histology Adenocarcinoma Squamous cell
43 (84) 8 (16)
EML4-ALK gene fusionb Positive Negative Unknown
0 28 (65) 15 (35)
Erlotinib treatment 2nd line 3rd line
42 (82) 9 (18)
Prior treatment 1st line Carboplatin/vinorelbinec Carboplatin/vinorelbine/bevacizumabd 2nd linee Pemetrexed Docetaxel Palliative radiotherapyf
27 (53) 24 (47) 5 (56) 4 (44) 3 (6)
PS, performance status; ECOG, Eastern Cooperative Oncology Group; EML4-ALK, echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase. a Former smoker was defined as having stopped smoking at time of diagnosis. b Only patients with adenocarcinoma were tested. c Carboplatin day 1 (AUC 5) and vinorelbine day 1 and day 8 (60–80 mg/m2 (p.o.)) every 3 weeks for a maximum of four cycles. d Bevacizumab (7.5 mg/m2 i.v. day 1) was given in combination with chemotherapy. Patients with disease control received subsequent maintenance therapy every 3 weeks until progression or toxicity. e Only including patients treated with erlotinib in 3rd line. f Only including radiotherapy delivered less than four weeks prior to the positron emission tomography/computed tomography scan.
2.2. FDG-PET/CT imaging F-18-FDG-PET/CT scans were performed prior to start of second or third-line erlotinib on a combined PET/CT scanner (Siemens Biograph TruePoint 40) at the Department of Nuclear Medicine and PET-Centre Aarhus University Hospital, Denmark. After a fasting period of at least 6 h, the patients were injected intravenously with mean 350 ± 92 MBq of F-18-FDG (5 MBq per kg +/−10%, with a minimum of 200 MBq and a maximum of 600 MBq). Bedside plasma glucose concentrations were measured prior to injection of F-18FDG and a mean of 6.3 ± 0.8 mmol/l (range; 4.6–8.8 mmol/l) was measured. Scans were obtained at a mean of 59 ± 5 min (range; 51–74 min) after 18 F FDG-injection. The acquisition time was 3 min per bed position. A whole-body low-dose CT scan (50 mAS, 120kVp) was performed for attenuation correction purposes and to deter-
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
83
106 paents screened for enrollment 1 paent failed inclusion criteria 38 paents declined to parcipate 67 paents enrolled 4 paents withdrew content prior to PET scan 10 paents were EGFR mutaon-posive 2 EGFR wild type paents received erlonib as 1st line of treatment 51 paents analyzed
1.00 0.00
0 1
20
0
5
10 Time (months)
15
20
0 0
Number at risk High TLG 25 Low TLG 24
2 4
1 1
1 0
0 0
D
High MTV Low MTV
0.75
1.00
0 2
1.00
0 6
15
P=0.002
0.00
0.25
0.50
Overall survival
P<0.001
High TLG Low TLG
0.75
10 Time (months)
0.50
5
0.00
Overall survival
P=0.027
0.25
0 Number at risk High MTV 24 Low MTV 25
C
0.75
Progression−free survival
0.75 0.25
0.50
P<0.001
High TLG Low TLG
0.50
B High MTV Low MTV
0.00
Progression−free survival
A
0.25
1.00
Fig. 1. Flow diagram of patient selection.
0 Number at risk High MTV 24 Low MTV 25
5
6 18
10 15 Time (months) 2 8
0 4
20
0 2
25
0
5
10 15 Time (months)
20
25
0 0
Number at risk High TLG 25 Low TLG 24
8 16
4 6
0 2
0 0
1 3
Fig. 2. Kaplan–Meier survival curves for progression free survival (PFS) and overall survival (OS) according to MTV and TLG above (high) and below (low) median values (A) PFS according to high and low MTV (B) PFS according to high and low TLG (C) OS according to high and low MTV (D) OS according to high and low TLG. Differences between groups were calculated using the log-rank test.
mine the anatomical location. Following the scan, the images were reconstructed using TrueX algorithm from Siemens (21 subsets and 3 iterations) in a matrix of 168 × 168 (4.0 mm/pixel) and postfiltered with a 3.0 mm FWHM Gaussian filter. All SUV values were normalized to lean body mass (SUL). Same scanner model, protocol for acquisition and reconstruction software was used for all patients.
2.3. Evaluation of FDG-PET/CT imaging All PET scans were analysed using Siemens Syngo.via software by an experienced nuclear medicine physician blinded to the patient outcome. All extra cerebral tumor lesions were evaluated according to the recommendations in the Positron Emission tomography Response Criteria in Solid Tumors (PERCIST 1.0) guide-
84
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
Table 2 Distribution of patient characteristics between groups with low and high MTV and TLG (N = 49). Characteristics
Low MTV (N = 25)
High MTV (N = 24)
P valuea
Low TLG (N = 24)
High TLG (N = 25)
P valuea
Age <67 years >67 years
12 (52) 13 (50)
11 (48) 13 (50)
0.879
10 (43) 14 (54)
13 (57) 12 (46)
0.469
Gender Female Male
14 (64) 11 (41)
8 (36) 16 (59)
0.111
12 (55) 12 (44)
10 (45) 15 (56)
0.482
PS, ECOG 0−1 2
19 (49) 6 (60)
20 (51) 4 (40)
0.524
18 (46) 6 (60)
21 (54) 4 (40)
0.435
Smoking statusb Former or never Current
18 (50) 7 (54)
18 (50) 6 (46)
0.812
17 (47) 7 (54)
19 (53) 6 (46)
0.682
Stage III IV
3 (75) 22 (49)
1 (25) 23 (51)
0.317
3 (75) 21 (47)
1 (25) 24 (53)
0.289
Brain metastases Yes No
2 (25) 23 (56)
6 (75) 18 (44)
0.110
2 (25) 22 (54)
6 (75) 19 (46)
0.136
Histology Adenocarcinoma Squamous cell
20 (49) 5 (62)
21 (51) 3 (38)
0.478
19 (46) 5 (62)
22 (54) 3 (38)
0.403
EML4-ALK gene fusionc Negative Unknown
13 (48) 12 (54)
14 (52) 10 (46)
0.656
11 (41) 13 (59)
16 (59) 9 (41)
0.201
Erlotinib treatment 2nd line 3rd line
21 (53) 4 (44)
19 (47) 5 (56)
0.473
20 (50) 4 (44)
20 (50) 5 (56)
0.527
12 (46) 13 (56)
14 (54) 10 (44)
0.469
11 (42) 13 (56)
15 (58) 10 (44)
0.321
1 (25) 3 (75)
3 (75) 1 (25)
0.486
1 (25) 3 (75)
3 (75) 1 (25)
0.486
Prior treatment 1st line Carboplatin/vinorelbined Carboplatin/vinorelbine/bevacizumabe 2nd linef Pemetrexed Docetaxel
MTV, metabolic tumor volume; TLG, tumor lesion glycolysis; PS, performance status; ECOG, Eastern Cooperative Oncology Group; EML4-ALK, echinoderm microtubuleassociated protein-like 4-anaplastic lymphoma kinase. Low MTV ≤ 100.42 cm3; High MTV > 100.42 cm3; Low TLG ≤ 395.27; High TLG > 395.27. a P value was calculated using 2 Test or Fishers exact test where appropriate. b Former smoker was defined as having stopped smoking at time of diagnosis. c Only patients with adenocarcinoma were tested. d Carboplatin day 1 (AUC 5) and vinorelbine day 1 and day 8 (60–80 mg/m2 (p.o.)) every 3 weeks for a maximum of four cycles. e Bevacizumab (7.5 mg/m2 i.v. day 1) was given in combination with chemotherapy. Patients with disease control received subsequent maintenance therapy every 3 weeks until progression or toxicity. f Only including patients treated with erlotinib in 3rd line.
line [14]; a lesion was considered measurable if SULpeak was at least 1.5 times the mean liver SUL (in a standard 3 cm spherical ROI in normal right lobe of the liver). All measurable lesions were three dimensionally delineated using a threshold approach at liver mean + 2SD using a semi-automated contouring tool. Each tumor’s MTV in cubic centimeters (cm3 ) was automatically calculated by the software. Whole-body MTV was calculated by adding up all the volumes of the malignant lesions in each patient. TLG was calculated as the MTV multiplied with the mean SUV of each delineated lesion. Finally, whole-body TLG was calculated by adding up all the TLGs of all measurable lesions in each patient. 2.4. Statistical analysis For the statistical analyses, clinical variables were grouped into two categories except for age (continuous variable). Median values were used as cut-offs for dichotomisation of MTV and TLG. The association between PET parameters and clinical characteristics were calculated using the 2 test or the Fishers exact test where appropriate. OS was determined as the time from start of erlotinib treatment until death of any course or last follow-up date
(October 2nd, 2015). Progression-free survival (PFS) was defined as the time from first administration of erlotinib to first documentation of either radiological or clinical progression or death. Patients ending erlotinib treatment without progression or death were censored at the time of discontinuation. Patients still being treated with erlotinib on the date of the last follow-up were censored on that day. Estimates of median PFS and OS were calculated using the Kaplan–Meier method and compared by the log-rank test. Univariate and multivariate hazard ratios (HR) were determined using the Cox proportional hazards model. MTV and TLG were analysed in two separate multivariate models (model 1 and model 2) since they were highly correlated (spearman correlation = 0.948, p < 0.001) and multicollinearity may exist. Due to the low number of events, only four variables were included in the multivariate analysis. All tests were two-sided, and p-values less than 0.05 were considered to be statistically significant. Statistical analyses were performed using SPSS statistics version 20.0 for windows (IBM SPSS Statistics, Chicago, IL, USA). STATA version 13 (Stata Corporation, Texas, USA) was used for Kaplan-Meier survival analysis.
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
85
Table 3 Univariate and multivariate Cox regression analyses of progression-free survival (N = 49). Adjusted HR (95% CI) Model 1a
P value
Adjusted HR (95% CI) Model 2b
P value
0.562
0.67 (0.32–1.40) 1.00
0.287
0.68 (0.33–1.40) 1.00
0.300
1.05 (0.44–2.50) 1.00
0.918
1.04 (0.43–2.53) 1.00
0.928
0.88 (0.36–2.13) 1.00
0.772
Stage IV III
0.43 (0.12–1.46) 1.00
0.174
Brain metastases Yes No
3.74 (1.59–8.82) 1.00
0.003
3.14 (1.18–8.37) 1.00
0.022
3.62 (1.44–9.11) 1.00
0.006
Erlotinib treatment 2nd line 3rd line
0.87 (0.41–1.85) 1.00
0.723
MTV, cm3 >100.42 ≤100.42
5.00 (2.36–10.57) 1.00
<0.001
5.44 (2.46–12.02) 1.00
<0.001
TLG >395.27 ≤395.27
2.00 (1.06–3.76) 1.00
0.032
2.17 (1.11–4.26) 1.00
0.024
HR (95% CI)
P value
Age
1.02 (0.98–1.06)
0.433
Gender Female Male
0.64 (0.34–1.19) 1.00
0.158
Histology Adenocarcinoma Squamous cell
0.97 (0.45–2.12) 1.00
0.943
Smoking Never or former Current
0.82 (0.42–1.60) 1.00
PS, ECOG 0−1 2
HR, hazard ratio; CI; confidence interval; MTV, metabolic tumor volume; TLG, tumor lesion glycolysis; PS, performance status; ECOG, Eastern Cooperative Oncology Group. a Model 1 containing the four variables: MTV, Smoking status, PS and brain metastases. b Model 2 containing the four variables: TLG, Smoking status, PS and brain metastases.
3. Results 3.1. Patients and FDG-PET scans Fifty-one patients were included in this study. A flow diagram of inclusion is shown in Fig. 1. Patient characteristics are summarized in Table 1. No patients were lost to follow-up. An F-18-FDG-PET/CT scan was performed in all patients and conducted median 1 day (range; 0–21) before start of erlotinib. Tumor volume assessment was not possible in two patients due to carcinomatosis of the lung. Hence, MTV and TLG values were assessable in 49 patients. Median values of MTV and TLG, used for dichotomisation, were 100.42 cm3 (range; 16–1826) and 395.27 (range; 66–7034), respectively. Clinicopathological factors were well balanced between the two MTV groups and TLG groups (Table 2). 3.2. FDG-PET/CT parameters and survival The median PFS for all patients was 2.7 months (95% CI; 2.5–2.9) and median OS was 5.9 months (95% CI; 4.4–7.4). At last follow-up date, one patient was still undergoing treatment and 10 patients were still alive. Reasons for ending erlotinib treatment were radiological or clinical progression of disease (N = 42), toxicity (N = 7) or death (N = 1). Both MTV and TLG values were significantly correlated to PFS. As shown in Fig. 2A, patients with low MTV had a median PFS of 2.8 months (95% CI; 2.6–3.0) compared with 2.2 months (95% CI; 1.7–2.7) in patients with high MTV (p < 0.001). The same was seen for TLG as shown in Fig. 2B (2.8 months (95% CI; 2.7–2.9) versus 2.3 months (95% CI; 1.8–2.7); p = 0.027). Furthermore, both variables were significantly associated with OS (Fig. 2C and 2D). The
median OS for patients with low MTV was 12.6 months (95% CI; 7.6–17.7) compared with 2.6 months (95% CI; 1.6–3.6) in patients with high MTV (p < 0.001). Patients with low TLG had a median OS of 8.6 months (95% CI; 8.1–9.1) while patients with high TLG had a median OS of 3.1 months (95% CI; 1.6–4.6) (p = 0.002). The association between MTV, TLG and survival was consistently found when tertiles were used as cut-offs (Supplementary Table 1). In univariate analysis, MTV, TLG, and cerebral metastases were found to significantly correlate with both PFS and OS (Tables 3 and 4). To evaluate the independent value of the two PET parameters, a multivariate analysis was conducted. Both high MTV (HR = 5.44 (95% CI; 2.46–12.02)) and high TLG (HR = 2.17 (95% CI 1.11–4.26)) remained independent predictors of shorter PFS (Table 3) and shorter OS (MTV: HR = 4.80 (95% CI 2.08–11.06); TLG: HR = 2.76 (95% CI 1.33–5.71)) (Table 4). Presence of cerebral metastases was also found to be an independent marker for both shorter PFS and shorter OS, although the latter was only observed for model 2 containing TLG (Table 4). 4. Discussion The purpose of this study was to evaluate the association between metabolic tumor burden, assessed as MTV or TLG, and outcome in 51 EGFR wt NSCLC patients treated with erlotinib in second or third-line. We found both PET parameters to be highly significant markers of both PFS and OS independent of other known prognostic factors. Of special interest, we observe a negative impact on outcome of high MTV; all patients with high MTV had non-responding disease with a PFS below 2.9 months, while several patients in the low MTV group had PFS longer than 5 months. Furthermore, we found a remarkable difference in OS between the groups. Patients
86
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
Table 4 Univariate and multivariate Cox regression analyses of overall survival (N = 49). Adjusted HR (95% CI) Model 1a
P value
Adjusted HR (95% CI) Model 2b
P value
0.44 (0.19–1.03) 1.00
0.058
0.45 (0.20–1.06) 1.00
0.068
0.001
2.26 (0.93–5.48) 1.00
0.072
3.10 (1.32–7.31) 1.00
0.010
1.21 (0.53–2.76) 1.00
0.652
1.24 (0.53–2.92) 1.00
0.627
1.00 (0.43–2.34) 1.00
0.999
MTV, cm3 >100.42 ≤100.42
4.94 (2.30–10.64) 1.00
<0.001
4.80 (2.08–11.06) 1.00
<0.001
TLG >395.27 ≤395.27
2.80 (1.43–5.48) 1.00
0.003
2.76 (1.33–5.71) 1.00
0.006
HR (95% CI)
P value
Age
1.01 (0.97–1.05)
0.798
Gender Female Male
0.66 (0.34–1.28) 1.00
0.217
Histology Adenocarcinoma Squamous cell
1.25 (0.52–3.01) 1.00
0.622
Smoking Never or former Current
1.43 (0.68–3.01) 1.00
0.348
PS, ECOG 0−1 2
0.72 (0.33–1.58) 1.00
0.412
Stage IV III
0.58 (0.20–1.66) 1.00
0.306
Brain metastases Yes No
3.75 (1.67–8.43) 1.00
Erlotinib treatment 2nd line 3rd line
HR, hazard ratio; CI; confidence interval; PS, performance status; ECOG, Eastern Cooperative Oncology Group; MTV, metabolic tumor volume; TLG, tumor lesion glycolysis. a Model 1 containing the four variables: MTV, PS, brain metastases and line of erlotinib treatment. b Model 2 containing the four variables: TLG, PS, brain metastases and line of erlotinib treatment.
with low values of MTV had a median OS of more than 12 months which is remarkably long for EGFR wt patients treated in second or third-line. Opposite, patients with high MTV and/or TLG had a very short median OS of approximately the same lengths as the PFS found in these patients (high MTV: 2.2 and 2.6 months; high TLG: 2.3 and 3.1 months). Response rates to second and third-line treatments are poor in EGFR wt patients and the only marker stated in the ESMO [10] and ASCO recommendations [11] to guide treatment decision is PS. Yet, even though our patients were selected for treatment based on their PS, the overall PFS and OS remained short; although comparable with previous studies in this patient group [15,16]. This highlights the need for new potential markers to identify patients who will benefit from treatment. Our data indicates that metabolic tumor burden correlates to tumor aggressiveness and non-responsive disease and is a highly promising clinical tool that may allow better patient selection for palliative treatment in the future. To our knowledge, this is the first study to evaluate MTV and TLG as markers of outcome in NSCLC patients treated in second or thirdline. Two previous studies have evaluated metabolic tumor burden as a marker of outcome after first-line TKI treatment; however, with conflicting results. Keam et al. [17] evaluated a retrospective cohort of 75 EGFR mutation-positive NSCLC patients with stage IIIb or IV and found, in line with us, TLG to be independently predictive of longer PFS. Opposite, Kahraman et al. [18] could not show any significant correlation between either MTV nor TLG and PFS in a prospective study of 30 advanced NSCLC patients treated with erlotinib (EGFR mutation-positive (N = 5), EGFR wt (N = 18), and untested patients (N = 7)). The relative low number of patients and, thereby, lack of power in their study, could explain the contradicting results. Furthermore, metabolic tumor burden has been
evaluated as a prognostic factor. Several studies have correlated metabolic tumor burden measured on the diagnostic PET scan in untreated advanced NSCLC patients with OS [6–9,19,20], and all studies except for one have found a positive correlation between high metabolic tumor burden and short OS. Therefore, it is likely that volumetric PET parameters hold a prognostic value. It is well known that tumor burden when classified according to the TNM classification is a highly useful prognostic marker, and it is likely that metabolic tumor burden can refine the classification in the future. Though, we are not able to determine from our study if metabolic tumor burden is a true predictive marker of effect from second or third-line TKI treatment or a prognostic marker. A study including a control group of non-TKI treated patients would be required to clarify this. Moreover, the usefulness of pre-treatment MTV or TLG before other second and third-line treatment options as docetaxel, pemetrexed and immunotherapy would be important to address in future studies. The majority of the previously published studies have been of retrospective nature with the limitations associated with this. Our study was performed prospectively, which allowed us to perform the F-18-FDG-PET scan immediately at the beginning of the treatment and not several days to weeks before as in the case in the majority of the retrospective studies. Furthermore, we used the same scanner model, protocol for acquisition and reconstruction software in all patients. Thereby, we could reduce the possible inter-individual variability of the scans. Additionally, we had complete clinical data on all our patients, including data on brain metastasis, and could adjust for these variables. On the other hand, our study had some limitations to consider. Firstly, the number of patients was limited; yet, we found highly significant differences in PFS and OS. Secondly, our study was a single-centre study and
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
the results will have to be validated in a multicenter study to show the reproducibility of our findings. 5. Conclusion We demonstrated high pre-treatment MTV and TLG to be independent factors for shorter PFS and OS in advanced EGFR wt NSCLC patients treated with erlotinib in second or third-line. If our result can be validated, this study identifies a promising new tool for improved stratification of this patient cohort, which hopefully can refine the process of optimal patient selection in the future. Conflict of interest The authors have none to declare. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.lungcan.2016.01. 024. References [1] P. Goldstraw, D. Ball, J.R. Jett, T. Le Chevalier, E. Lim, A.G. Nicholson, et al., Non-small-cell lung cance, Lancet (London, England) 378 (2011) 1727–1740. [2] R. Rami-Porta, J.J. Crowley, P. Goldstraw, The revised TNM staging system for lung cancer, Ann. Thorac. Cardiovasc. Surg. 15 (2009) 4–9. [3] P. Glare, K. Virik, M. Jones, M. Hudson, S. Eychmuller, J. Simes, et al., A systematic review of physicians’ survival predictions in terminally ill cancer patients, BMJ 327 (2003) 195–198. [4] P. Pacetti, G. Paganini, M. Orlandi, A. Mambrini, M.C. Pennucci, A. Del Freo, et al., Chemotherapy in the last 30 days of life of advanced cancer patients, Support. Care Cancer 23 (2015) 3277–3280. [5] S.M. Larson, Y. Erdi, T. Akhurst, M. Mazumdar, H.A. Macapinlac, R.D. Finn, et al., Tumor treatment response based on visual and quantitative changes in global tumor glycolysis using PET-FDG imaging. The visual response score and the change in total lesion glycolysis, Clin Positron Imaging 2 (1999) 159–171. [6] S.W. Yoo, J. Kim, A. Chong, S.-Y. Kwon, J.-J. Min, H.-C. Song, et al., Metabolic tumor volume measured by F-18 FDG PET/CT can further stratify the prognosis of patients with stage IV non-small cell lung cancer, Nucl. Med. Mol. Imaging 46 (2012) 286–293 (2010).
87
[7] T.-Y. Ho, P.-C. Chou, C.-T. Yang, N.-M. Tsang, T.-C. Yen, Total lesion glycolysis determined per RECIST 1.1 criteria predicts survival in EGFR mutation-negative patients with advanced lung adenocarcinoma, Clin. Nucl. Med. 40 (2015) e295–9. [8] H. Zhang, K. Wroblewski, D. Appelbaum, Y. Pu, Independent prognostic value of whole-body metabolic tumor burden from FDG-PET in non-small cell lung cancer, Int. J. Comput. Assist. Radiol. Surg. 8 (2013) 181–191. [9] H.W. Chung, K.Y. Lee, H.J. Kim, W.S. Kim, Y. So, FDG PET/CT metabolic tumor volume and total lesion glycolysis predict prognosis in patients with advanced lung adenocarcinoma, J. Cancer Res. Clin. Oncol. 140 (2014) 89–98. [10] M. Reck, S. Popat, N. Reinmuth, D. De Ruysscher, K.M. Kerr, S. Peters, Metastatic non-small-cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up, Ann. Oncol. 25 (Suppl 3) (2014) iii27–39. [11] G.A. Masters, S. Temin, C.G. Azzoli, G. Giaccone, S. Baker, J.R. Brahmer, et al., Systemic therapy for stage IV non-small-cell lung cancer: American Society of Clinical Oncology Clinical Practice Guideline Update, J. Clin. Oncol. 33 (2015) 3488–3515. [12] E. Arriola, Á. Taus, D. Casadevall, Is there a role for epidermal growth factor receptor tyrosine kinase inhibitors in epidermal growth factor receptor wild-type non-small cell lung cancer? World J. Clin. Oncol. 6 (2015) 45–56. [13] E.A. Eisenhauer, P. Therasse, J. Bogaerts, L.H. Schwartz, D. Sargent, R. Ford, et al., New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1), Eur. J. Cancer 45 (2009) 228–247. [14] R.L. Wahl, H. Jacene, Y. Kasamon, M.A. Lodge, From RECIST to PERCIST: evolving considerations for PET response criteria in solid tumors, J. Nucl. Med. 50 (Suppl. 1) (2009) 122S–150S. [15] M.C. Garassino, O. Martelli, M. Broggini, G. Farina, S. Veronese, E. Rulli, et al., Erlotinib versus docetaxel as second-line treatment of patients with advanced non-small-cell lung cancer and wild-type EGFR tumours (TAILOR): a randomised controlled trial, Lancet Oncol. 14 (2013) 981–988. [16] T. Ciuleanu, L. Stelmakh, S. Cicenas, S. Miliauskas, A.C. Grigorescu, C. Hillenbach, et al., Efficacy and safety of erlotinib versus chemotherapy in second-line treatment of patients with advanced, non-small-cell lung cancer with poor prognosis (TITAN): a randomised multicentre, open-label, phase 3 study, Lancet Oncol. 13 (2012) 300–308. [17] B. Keam, S.J. Lee, T.M. Kim, J.C. Paeng, S.-H. Lee, D.-W. Kim, et al., Total lesion glycolysis in positron emission tomography can predict gefitinib outcomes in non-small-cell lung cancer with activating EGFR mutation, J. Thorac. Oncol. 10 (2015) 1189–1194. [18] D. Kahraman, A. Holstein, M. Scheffler, T. Zander, L. Nogova, A.A. Lammertsma, et al., Tumor lesion glycolysis and tumor lesion proliferation for response prediction and prognostic differentiation in patients with advanced non-small cell lung cancer treated with erlotinib, Clin. Nucl. Med. 37 (2012) 1058–1064. [19] S. Liao, B.C. Penney, H. Zhang, K. Suzuki, Y. Pu, Prognostic value of the quantitative metabolic volumetric measurement on 18F-FDG PET/CT in stage IV nonsurgical small-cell lung cancer, Acad. Radiol. 19 (2012) 69–77. [20] A.D. Nygaard, P.C. Holdgaard, K.-L.G. Spindler, N. Pallisgaard, A. Jakobsen, The correlation between cell-free DNA and tumour burden was estimated by PET/CT in patients with advanced NSCLC, Br. J. Cancer 110 (2014) 363–368.
Contents lists available at ScienceDirect
Lung Cancer journal homepage: www.elsevier.com/locate/lungcan
Metabolic tumor burden as marker of outcome in advanced EGFR wild-type NSCLC patients treated with erlotinib Anne Winther-Larsen a,∗ , Joan Fledelius b , Boe Sandahl Sorensen a , Peter Meldgaard c a b c
Department of Clinical Biochemistry, Aarhus University Hospital, Aarhus, Denmark Department of Nuclear Medicine, Herning Regional Hospital, Herning, Denmark Department of Oncology, Aarhus University Hospital, Aarhus, Denmark
a r t i c l e
i n f o
Article history: Received 28 December 2015 Received in revised form 25 January 2016 Accepted 30 January 2016 Keywords: Positron emission tomography Lung cancer Tyrosine kinase inhibitor EGFR wild-type Metabolic tumor volume
a b s t r a c t Objectives: Accurate estimation of the prognosis of advanced non-small cell lung cancer (NSCLC) patients is essential before initiation of palliative treatment; especially in the second and third-line setting. This study was conducted in order to evaluate tumor burden measured on an 2 -deoxy-2 -[18F] fluoro-Dglucose (F-18-FDG) positron emission tomography/computed tomography (PET/CT) scan as a marker of outcome in advanced epidermal growth factor receptor (EGFR) wild-type patients treated with second or third-line erlotinib. Material and methods: Fifty-one patients were included from a prospectively collected cohort. An F-18FDG-PET/CT scan was conducted prior to erlotinib treatment and tumor burden was measured in terms of metabolic tumor volume (MTV) and total lesion glycolysis (TLG). Median values of MTV and TLG were used for dichotomization of patients. Survival outcome was compared between groups. Results: MTV and TLG could be measured in 49 patients. High values of MTV and TLG were significantly correlated with shorter PFS (p < 0.001 and p = 0.027, respectively) and OS (p < 0.001 and p = 0.002, respectively). In multivariate analyses, including both clinical and imaging data, high MTV and TLG remained strong independent markers of both shorter PFS (MTV, hazard ratio (HR) = 5.44 (95% confidence interval (CI) 2.46–12.02); TLG, HR = 2.17 (95% CI 1.11–4.26)) and OS (MTV, HR = 4.80 (95% CI 2.08–11.06); TLG, HR = 2.76 (95% CI 1.33–5.71)). Conclusion: High MTV and TLG are independently correlated with shorter PFS and OS in advanced EGFR wild-type NSCLC patients treated with second or third-line erlotinib. Metabolic tumor burden is a highly promising clinical tool that may allow better patient selection for palliative treatment in the future. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Background The prognosis of non-small cell lung cancer (NSCLC) patients is heterogeneous. Extent of disease at diagnosis is one of the most useful prognostic markers [1]. All patients are staged by the tumornode-metastases (TNM) classification [2] and treatment strategy is based on the stage of disease. Patients with metastatic, stage IV disease are incurable with short life expectancy and choice of treatment in these patients must be carefully balanced between the possible benefit of treatment and side effects. A precise evaluation of the individual patient’s prognosis is therefore of the greatest importance. However, evaluation of survival time is challenging for clinicians and they have a tendency to overestimate the prognosis
∗ Corresponding author at: Department of Clinical Biochemistry, Aarhus University Hospital, Noerrebrogade 44, 8000 Aarhus C, Denmark. E-mail address: [email protected] (A. Winther-Larsen). http://dx.doi.org/10.1016/j.lungcan.2016.01.024 0169-5002/© 2016 Elsevier Ireland Ltd. All rights reserved.
[3]. This leads to a growing tendency to overtreatment with patients receiving therapy near end of life [4]. New tools to guide prognosis estimation and choice of treatment are therefore highly warranted. In the recent years, growing interest has emerged for the use of 2 -deoxy-2 -[18F] fluoro-d-glucose (F-18-FDG) positron emission tomography/computed tomography (PET/CT) scans as an imaging biomarker of prognosis and treatment effect. The most widely used PET indicator of metabolic activity is the maximal standardized uptake value (SUVmax ). However, the usefulness of SUVmax may be limited since it only represents a single voxel value and not sufficiently reflects the entire tumor burden. Therefore, interest has therefore emerged for the use of volume-based PET measurements since they are capable of measuring the metabolic tumor burden by incorporating both metabolic activity and volumetric data [5]. Metabolic tumor volume (MTV) is solely a volumetric parameter delineated by a specific metabolic threshold while tumor lesion glycolysis (TLG) in addition incorporates the metabolic activity of the
82
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
tumor. In previous studies, MTV and TLG evaluated before first-line of treatment have shown promising results as markers of outcome in stage IV NSCLC patients [6–9]. However, the value of metabolic tumor burden assessment prior to following lines of treatment has not been evaluated. Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR TKIs) are a treatment option for advanced EGFR wild-type (wt) patients in the second or third-line palliative setting [10,11]. Though, response rates are only approximately 8% [12] and the only recommended factor to guide patient selection is performance status (PS) [10–12]. Moreover, these patients have a short life expectancy and better markers are highly needed to prevent use of unnecessary and potentially harmful treatment at the end of life. Thus, the purpose of this study was to evaluate if MTV and TLG could be a markers of outcome in advanced EGFR wt NSCLC patients deemed to start second or third-line treatment with erlotinib.
2. Materials and methods 2.1. Patients and treatment In this prospective, single-centre study, we enrolled advanced NSCLC patients starting treatment with erlotinib in a palliative setting from April 2013 until August 2015 at the Department of Oncology, Aarhus University Hospital, Denmark. Patients were eligible for enrolment if they fulfilled the following criteria for erlotinib treatment: histologically or cytologically proven NSCLC, age ≥18 years, performance status ≤2 and no prior TKI treatment. If patients had uncontrolled diabetes mellitus or severe respiratory problems, they were excluded from enrolment. All patients gave informed written consent before inclusion in the study and the study was approved by the Central Denmark Region Committees on Biomedical Research Ethics (no. 1-10-72-19-12). For the purpose of this study, we included only patients from the total cohort who were EGFR wt and received erlotinib as second or third-line of treatment. All enrolled patients received an initial dose of 150 mg erlotinib daily. If side effects appeared, erlotinib was continued together with symptomatic management of toxicity before any dose reduction. However, if the patient had continuous side effects of grade 2 or more according to the CTC criteria version 4.0, dose-reduction was performed. The patients were treated until radiologic or clinical progression, unacceptable toxicity, or death. All patients had a pre-treatment CT scan of the chest and abdomen performed prior to start of erlotinib. Follow-up was done by CT scans every 12 weeks during the treatment period combined with clinical and biochemical evaluation every fourth week in the first 12 weeks and subsequently every sixth week as standard of care in our institution. At the end of treatment, the patients were evaluated regarding either subsequent treatment with pemetrexed or docetaxel or best supportive care according to standard of care at our institution. Treatment response evaluated on CT scans was defined according to Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1) criteria [13]. Neuroimaging was performed on clinical indication. In patients with adenocarcinoma, EGFR mutation testing was performed as part of the routine diagnostic work-up by use of the Therascreen EGFR RGQ PCR kit (QIAGEN, Manchester, UK) according to the manufacturer’s protocol. Testing for EML4-ALK fusions were incorporated as part of the routine diagnostic work-up in adenocarcinoma patients during the inclusion period and performed by fluorescence in situ hybridization. TNM stage was evaluated at time of inclusion according to the 7th edition of the American Joint Committee on Cancer [2]. Data on baseline characteristics and response was collected from medical files.
Table 1 Basic characteristics for all patients (N = 51). Characteristics
All patients N (%)
Age, years Median (range)
67 (49–83)
Gender Female Male
23 (45) 28 (55)
PS, ECOG 0 1 2
4 (8) 37 (72) 10 (20)
Smoking status Never Formera Current
1 (2) 37 (73) 13 (25)
Stage IIIa IIIb IV
2 (4) 2 (4) 47(92)
Brain metastases Yes No
8 (16) 43 (84)
Histology Adenocarcinoma Squamous cell
43 (84) 8 (16)
EML4-ALK gene fusionb Positive Negative Unknown
0 28 (65) 15 (35)
Erlotinib treatment 2nd line 3rd line
42 (82) 9 (18)
Prior treatment 1st line Carboplatin/vinorelbinec Carboplatin/vinorelbine/bevacizumabd 2nd linee Pemetrexed Docetaxel Palliative radiotherapyf
27 (53) 24 (47) 5 (56) 4 (44) 3 (6)
PS, performance status; ECOG, Eastern Cooperative Oncology Group; EML4-ALK, echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase. a Former smoker was defined as having stopped smoking at time of diagnosis. b Only patients with adenocarcinoma were tested. c Carboplatin day 1 (AUC 5) and vinorelbine day 1 and day 8 (60–80 mg/m2 (p.o.)) every 3 weeks for a maximum of four cycles. d Bevacizumab (7.5 mg/m2 i.v. day 1) was given in combination with chemotherapy. Patients with disease control received subsequent maintenance therapy every 3 weeks until progression or toxicity. e Only including patients treated with erlotinib in 3rd line. f Only including radiotherapy delivered less than four weeks prior to the positron emission tomography/computed tomography scan.
2.2. FDG-PET/CT imaging F-18-FDG-PET/CT scans were performed prior to start of second or third-line erlotinib on a combined PET/CT scanner (Siemens Biograph TruePoint 40) at the Department of Nuclear Medicine and PET-Centre Aarhus University Hospital, Denmark. After a fasting period of at least 6 h, the patients were injected intravenously with mean 350 ± 92 MBq of F-18-FDG (5 MBq per kg +/−10%, with a minimum of 200 MBq and a maximum of 600 MBq). Bedside plasma glucose concentrations were measured prior to injection of F-18FDG and a mean of 6.3 ± 0.8 mmol/l (range; 4.6–8.8 mmol/l) was measured. Scans were obtained at a mean of 59 ± 5 min (range; 51–74 min) after 18 F FDG-injection. The acquisition time was 3 min per bed position. A whole-body low-dose CT scan (50 mAS, 120kVp) was performed for attenuation correction purposes and to deter-
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
83
106 paents screened for enrollment 1 paent failed inclusion criteria 38 paents declined to parcipate 67 paents enrolled 4 paents withdrew content prior to PET scan 10 paents were EGFR mutaon-posive 2 EGFR wild type paents received erlonib as 1st line of treatment 51 paents analyzed
1.00 0.00
0 1
20
0
5
10 Time (months)
15
20
0 0
Number at risk High TLG 25 Low TLG 24
2 4
1 1
1 0
0 0
D
High MTV Low MTV
0.75
1.00
0 2
1.00
0 6
15
P=0.002
0.00
0.25
0.50
Overall survival
P<0.001
High TLG Low TLG
0.75
10 Time (months)
0.50
5
0.00
Overall survival
P=0.027
0.25
0 Number at risk High MTV 24 Low MTV 25
C
0.75
Progression−free survival
0.75 0.25
0.50
P<0.001
High TLG Low TLG
0.50
B High MTV Low MTV
0.00
Progression−free survival
A
0.25
1.00
Fig. 1. Flow diagram of patient selection.
0 Number at risk High MTV 24 Low MTV 25
5
6 18
10 15 Time (months) 2 8
0 4
20
0 2
25
0
5
10 15 Time (months)
20
25
0 0
Number at risk High TLG 25 Low TLG 24
8 16
4 6
0 2
0 0
1 3
Fig. 2. Kaplan–Meier survival curves for progression free survival (PFS) and overall survival (OS) according to MTV and TLG above (high) and below (low) median values (A) PFS according to high and low MTV (B) PFS according to high and low TLG (C) OS according to high and low MTV (D) OS according to high and low TLG. Differences between groups were calculated using the log-rank test.
mine the anatomical location. Following the scan, the images were reconstructed using TrueX algorithm from Siemens (21 subsets and 3 iterations) in a matrix of 168 × 168 (4.0 mm/pixel) and postfiltered with a 3.0 mm FWHM Gaussian filter. All SUV values were normalized to lean body mass (SUL). Same scanner model, protocol for acquisition and reconstruction software was used for all patients.
2.3. Evaluation of FDG-PET/CT imaging All PET scans were analysed using Siemens Syngo.via software by an experienced nuclear medicine physician blinded to the patient outcome. All extra cerebral tumor lesions were evaluated according to the recommendations in the Positron Emission tomography Response Criteria in Solid Tumors (PERCIST 1.0) guide-
84
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
Table 2 Distribution of patient characteristics between groups with low and high MTV and TLG (N = 49). Characteristics
Low MTV (N = 25)
High MTV (N = 24)
P valuea
Low TLG (N = 24)
High TLG (N = 25)
P valuea
Age <67 years >67 years
12 (52) 13 (50)
11 (48) 13 (50)
0.879
10 (43) 14 (54)
13 (57) 12 (46)
0.469
Gender Female Male
14 (64) 11 (41)
8 (36) 16 (59)
0.111
12 (55) 12 (44)
10 (45) 15 (56)
0.482
PS, ECOG 0−1 2
19 (49) 6 (60)
20 (51) 4 (40)
0.524
18 (46) 6 (60)
21 (54) 4 (40)
0.435
Smoking statusb Former or never Current
18 (50) 7 (54)
18 (50) 6 (46)
0.812
17 (47) 7 (54)
19 (53) 6 (46)
0.682
Stage III IV
3 (75) 22 (49)
1 (25) 23 (51)
0.317
3 (75) 21 (47)
1 (25) 24 (53)
0.289
Brain metastases Yes No
2 (25) 23 (56)
6 (75) 18 (44)
0.110
2 (25) 22 (54)
6 (75) 19 (46)
0.136
Histology Adenocarcinoma Squamous cell
20 (49) 5 (62)
21 (51) 3 (38)
0.478
19 (46) 5 (62)
22 (54) 3 (38)
0.403
EML4-ALK gene fusionc Negative Unknown
13 (48) 12 (54)
14 (52) 10 (46)
0.656
11 (41) 13 (59)
16 (59) 9 (41)
0.201
Erlotinib treatment 2nd line 3rd line
21 (53) 4 (44)
19 (47) 5 (56)
0.473
20 (50) 4 (44)
20 (50) 5 (56)
0.527
12 (46) 13 (56)
14 (54) 10 (44)
0.469
11 (42) 13 (56)
15 (58) 10 (44)
0.321
1 (25) 3 (75)
3 (75) 1 (25)
0.486
1 (25) 3 (75)
3 (75) 1 (25)
0.486
Prior treatment 1st line Carboplatin/vinorelbined Carboplatin/vinorelbine/bevacizumabe 2nd linef Pemetrexed Docetaxel
MTV, metabolic tumor volume; TLG, tumor lesion glycolysis; PS, performance status; ECOG, Eastern Cooperative Oncology Group; EML4-ALK, echinoderm microtubuleassociated protein-like 4-anaplastic lymphoma kinase. Low MTV ≤ 100.42 cm3; High MTV > 100.42 cm3; Low TLG ≤ 395.27; High TLG > 395.27. a P value was calculated using 2 Test or Fishers exact test where appropriate. b Former smoker was defined as having stopped smoking at time of diagnosis. c Only patients with adenocarcinoma were tested. d Carboplatin day 1 (AUC 5) and vinorelbine day 1 and day 8 (60–80 mg/m2 (p.o.)) every 3 weeks for a maximum of four cycles. e Bevacizumab (7.5 mg/m2 i.v. day 1) was given in combination with chemotherapy. Patients with disease control received subsequent maintenance therapy every 3 weeks until progression or toxicity. f Only including patients treated with erlotinib in 3rd line.
line [14]; a lesion was considered measurable if SULpeak was at least 1.5 times the mean liver SUL (in a standard 3 cm spherical ROI in normal right lobe of the liver). All measurable lesions were three dimensionally delineated using a threshold approach at liver mean + 2SD using a semi-automated contouring tool. Each tumor’s MTV in cubic centimeters (cm3 ) was automatically calculated by the software. Whole-body MTV was calculated by adding up all the volumes of the malignant lesions in each patient. TLG was calculated as the MTV multiplied with the mean SUV of each delineated lesion. Finally, whole-body TLG was calculated by adding up all the TLGs of all measurable lesions in each patient. 2.4. Statistical analysis For the statistical analyses, clinical variables were grouped into two categories except for age (continuous variable). Median values were used as cut-offs for dichotomisation of MTV and TLG. The association between PET parameters and clinical characteristics were calculated using the 2 test or the Fishers exact test where appropriate. OS was determined as the time from start of erlotinib treatment until death of any course or last follow-up date
(October 2nd, 2015). Progression-free survival (PFS) was defined as the time from first administration of erlotinib to first documentation of either radiological or clinical progression or death. Patients ending erlotinib treatment without progression or death were censored at the time of discontinuation. Patients still being treated with erlotinib on the date of the last follow-up were censored on that day. Estimates of median PFS and OS were calculated using the Kaplan–Meier method and compared by the log-rank test. Univariate and multivariate hazard ratios (HR) were determined using the Cox proportional hazards model. MTV and TLG were analysed in two separate multivariate models (model 1 and model 2) since they were highly correlated (spearman correlation = 0.948, p < 0.001) and multicollinearity may exist. Due to the low number of events, only four variables were included in the multivariate analysis. All tests were two-sided, and p-values less than 0.05 were considered to be statistically significant. Statistical analyses were performed using SPSS statistics version 20.0 for windows (IBM SPSS Statistics, Chicago, IL, USA). STATA version 13 (Stata Corporation, Texas, USA) was used for Kaplan-Meier survival analysis.
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
85
Table 3 Univariate and multivariate Cox regression analyses of progression-free survival (N = 49). Adjusted HR (95% CI) Model 1a
P value
Adjusted HR (95% CI) Model 2b
P value
0.562
0.67 (0.32–1.40) 1.00
0.287
0.68 (0.33–1.40) 1.00
0.300
1.05 (0.44–2.50) 1.00
0.918
1.04 (0.43–2.53) 1.00
0.928
0.88 (0.36–2.13) 1.00
0.772
Stage IV III
0.43 (0.12–1.46) 1.00
0.174
Brain metastases Yes No
3.74 (1.59–8.82) 1.00
0.003
3.14 (1.18–8.37) 1.00
0.022
3.62 (1.44–9.11) 1.00
0.006
Erlotinib treatment 2nd line 3rd line
0.87 (0.41–1.85) 1.00
0.723
MTV, cm3 >100.42 ≤100.42
5.00 (2.36–10.57) 1.00
<0.001
5.44 (2.46–12.02) 1.00
<0.001
TLG >395.27 ≤395.27
2.00 (1.06–3.76) 1.00
0.032
2.17 (1.11–4.26) 1.00
0.024
HR (95% CI)
P value
Age
1.02 (0.98–1.06)
0.433
Gender Female Male
0.64 (0.34–1.19) 1.00
0.158
Histology Adenocarcinoma Squamous cell
0.97 (0.45–2.12) 1.00
0.943
Smoking Never or former Current
0.82 (0.42–1.60) 1.00
PS, ECOG 0−1 2
HR, hazard ratio; CI; confidence interval; MTV, metabolic tumor volume; TLG, tumor lesion glycolysis; PS, performance status; ECOG, Eastern Cooperative Oncology Group. a Model 1 containing the four variables: MTV, Smoking status, PS and brain metastases. b Model 2 containing the four variables: TLG, Smoking status, PS and brain metastases.
3. Results 3.1. Patients and FDG-PET scans Fifty-one patients were included in this study. A flow diagram of inclusion is shown in Fig. 1. Patient characteristics are summarized in Table 1. No patients were lost to follow-up. An F-18-FDG-PET/CT scan was performed in all patients and conducted median 1 day (range; 0–21) before start of erlotinib. Tumor volume assessment was not possible in two patients due to carcinomatosis of the lung. Hence, MTV and TLG values were assessable in 49 patients. Median values of MTV and TLG, used for dichotomisation, were 100.42 cm3 (range; 16–1826) and 395.27 (range; 66–7034), respectively. Clinicopathological factors were well balanced between the two MTV groups and TLG groups (Table 2). 3.2. FDG-PET/CT parameters and survival The median PFS for all patients was 2.7 months (95% CI; 2.5–2.9) and median OS was 5.9 months (95% CI; 4.4–7.4). At last follow-up date, one patient was still undergoing treatment and 10 patients were still alive. Reasons for ending erlotinib treatment were radiological or clinical progression of disease (N = 42), toxicity (N = 7) or death (N = 1). Both MTV and TLG values were significantly correlated to PFS. As shown in Fig. 2A, patients with low MTV had a median PFS of 2.8 months (95% CI; 2.6–3.0) compared with 2.2 months (95% CI; 1.7–2.7) in patients with high MTV (p < 0.001). The same was seen for TLG as shown in Fig. 2B (2.8 months (95% CI; 2.7–2.9) versus 2.3 months (95% CI; 1.8–2.7); p = 0.027). Furthermore, both variables were significantly associated with OS (Fig. 2C and 2D). The
median OS for patients with low MTV was 12.6 months (95% CI; 7.6–17.7) compared with 2.6 months (95% CI; 1.6–3.6) in patients with high MTV (p < 0.001). Patients with low TLG had a median OS of 8.6 months (95% CI; 8.1–9.1) while patients with high TLG had a median OS of 3.1 months (95% CI; 1.6–4.6) (p = 0.002). The association between MTV, TLG and survival was consistently found when tertiles were used as cut-offs (Supplementary Table 1). In univariate analysis, MTV, TLG, and cerebral metastases were found to significantly correlate with both PFS and OS (Tables 3 and 4). To evaluate the independent value of the two PET parameters, a multivariate analysis was conducted. Both high MTV (HR = 5.44 (95% CI; 2.46–12.02)) and high TLG (HR = 2.17 (95% CI 1.11–4.26)) remained independent predictors of shorter PFS (Table 3) and shorter OS (MTV: HR = 4.80 (95% CI 2.08–11.06); TLG: HR = 2.76 (95% CI 1.33–5.71)) (Table 4). Presence of cerebral metastases was also found to be an independent marker for both shorter PFS and shorter OS, although the latter was only observed for model 2 containing TLG (Table 4). 4. Discussion The purpose of this study was to evaluate the association between metabolic tumor burden, assessed as MTV or TLG, and outcome in 51 EGFR wt NSCLC patients treated with erlotinib in second or third-line. We found both PET parameters to be highly significant markers of both PFS and OS independent of other known prognostic factors. Of special interest, we observe a negative impact on outcome of high MTV; all patients with high MTV had non-responding disease with a PFS below 2.9 months, while several patients in the low MTV group had PFS longer than 5 months. Furthermore, we found a remarkable difference in OS between the groups. Patients
86
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
Table 4 Univariate and multivariate Cox regression analyses of overall survival (N = 49). Adjusted HR (95% CI) Model 1a
P value
Adjusted HR (95% CI) Model 2b
P value
0.44 (0.19–1.03) 1.00
0.058
0.45 (0.20–1.06) 1.00
0.068
0.001
2.26 (0.93–5.48) 1.00
0.072
3.10 (1.32–7.31) 1.00
0.010
1.21 (0.53–2.76) 1.00
0.652
1.24 (0.53–2.92) 1.00
0.627
1.00 (0.43–2.34) 1.00
0.999
MTV, cm3 >100.42 ≤100.42
4.94 (2.30–10.64) 1.00
<0.001
4.80 (2.08–11.06) 1.00
<0.001
TLG >395.27 ≤395.27
2.80 (1.43–5.48) 1.00
0.003
2.76 (1.33–5.71) 1.00
0.006
HR (95% CI)
P value
Age
1.01 (0.97–1.05)
0.798
Gender Female Male
0.66 (0.34–1.28) 1.00
0.217
Histology Adenocarcinoma Squamous cell
1.25 (0.52–3.01) 1.00
0.622
Smoking Never or former Current
1.43 (0.68–3.01) 1.00
0.348
PS, ECOG 0−1 2
0.72 (0.33–1.58) 1.00
0.412
Stage IV III
0.58 (0.20–1.66) 1.00
0.306
Brain metastases Yes No
3.75 (1.67–8.43) 1.00
Erlotinib treatment 2nd line 3rd line
HR, hazard ratio; CI; confidence interval; PS, performance status; ECOG, Eastern Cooperative Oncology Group; MTV, metabolic tumor volume; TLG, tumor lesion glycolysis. a Model 1 containing the four variables: MTV, PS, brain metastases and line of erlotinib treatment. b Model 2 containing the four variables: TLG, PS, brain metastases and line of erlotinib treatment.
with low values of MTV had a median OS of more than 12 months which is remarkably long for EGFR wt patients treated in second or third-line. Opposite, patients with high MTV and/or TLG had a very short median OS of approximately the same lengths as the PFS found in these patients (high MTV: 2.2 and 2.6 months; high TLG: 2.3 and 3.1 months). Response rates to second and third-line treatments are poor in EGFR wt patients and the only marker stated in the ESMO [10] and ASCO recommendations [11] to guide treatment decision is PS. Yet, even though our patients were selected for treatment based on their PS, the overall PFS and OS remained short; although comparable with previous studies in this patient group [15,16]. This highlights the need for new potential markers to identify patients who will benefit from treatment. Our data indicates that metabolic tumor burden correlates to tumor aggressiveness and non-responsive disease and is a highly promising clinical tool that may allow better patient selection for palliative treatment in the future. To our knowledge, this is the first study to evaluate MTV and TLG as markers of outcome in NSCLC patients treated in second or thirdline. Two previous studies have evaluated metabolic tumor burden as a marker of outcome after first-line TKI treatment; however, with conflicting results. Keam et al. [17] evaluated a retrospective cohort of 75 EGFR mutation-positive NSCLC patients with stage IIIb or IV and found, in line with us, TLG to be independently predictive of longer PFS. Opposite, Kahraman et al. [18] could not show any significant correlation between either MTV nor TLG and PFS in a prospective study of 30 advanced NSCLC patients treated with erlotinib (EGFR mutation-positive (N = 5), EGFR wt (N = 18), and untested patients (N = 7)). The relative low number of patients and, thereby, lack of power in their study, could explain the contradicting results. Furthermore, metabolic tumor burden has been
evaluated as a prognostic factor. Several studies have correlated metabolic tumor burden measured on the diagnostic PET scan in untreated advanced NSCLC patients with OS [6–9,19,20], and all studies except for one have found a positive correlation between high metabolic tumor burden and short OS. Therefore, it is likely that volumetric PET parameters hold a prognostic value. It is well known that tumor burden when classified according to the TNM classification is a highly useful prognostic marker, and it is likely that metabolic tumor burden can refine the classification in the future. Though, we are not able to determine from our study if metabolic tumor burden is a true predictive marker of effect from second or third-line TKI treatment or a prognostic marker. A study including a control group of non-TKI treated patients would be required to clarify this. Moreover, the usefulness of pre-treatment MTV or TLG before other second and third-line treatment options as docetaxel, pemetrexed and immunotherapy would be important to address in future studies. The majority of the previously published studies have been of retrospective nature with the limitations associated with this. Our study was performed prospectively, which allowed us to perform the F-18-FDG-PET scan immediately at the beginning of the treatment and not several days to weeks before as in the case in the majority of the retrospective studies. Furthermore, we used the same scanner model, protocol for acquisition and reconstruction software in all patients. Thereby, we could reduce the possible inter-individual variability of the scans. Additionally, we had complete clinical data on all our patients, including data on brain metastasis, and could adjust for these variables. On the other hand, our study had some limitations to consider. Firstly, the number of patients was limited; yet, we found highly significant differences in PFS and OS. Secondly, our study was a single-centre study and
A. Winther-Larsen et al. / Lung Cancer 94 (2016) 81–87
the results will have to be validated in a multicenter study to show the reproducibility of our findings. 5. Conclusion We demonstrated high pre-treatment MTV and TLG to be independent factors for shorter PFS and OS in advanced EGFR wt NSCLC patients treated with erlotinib in second or third-line. If our result can be validated, this study identifies a promising new tool for improved stratification of this patient cohort, which hopefully can refine the process of optimal patient selection in the future. Conflict of interest The authors have none to declare. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.lungcan.2016.01. 024. References [1] P. Goldstraw, D. Ball, J.R. Jett, T. Le Chevalier, E. Lim, A.G. Nicholson, et al., Non-small-cell lung cance, Lancet (London, England) 378 (2011) 1727–1740. [2] R. Rami-Porta, J.J. Crowley, P. Goldstraw, The revised TNM staging system for lung cancer, Ann. Thorac. Cardiovasc. Surg. 15 (2009) 4–9. [3] P. Glare, K. Virik, M. Jones, M. Hudson, S. Eychmuller, J. Simes, et al., A systematic review of physicians’ survival predictions in terminally ill cancer patients, BMJ 327 (2003) 195–198. [4] P. Pacetti, G. Paganini, M. Orlandi, A. Mambrini, M.C. Pennucci, A. Del Freo, et al., Chemotherapy in the last 30 days of life of advanced cancer patients, Support. Care Cancer 23 (2015) 3277–3280. [5] S.M. Larson, Y. Erdi, T. Akhurst, M. Mazumdar, H.A. Macapinlac, R.D. Finn, et al., Tumor treatment response based on visual and quantitative changes in global tumor glycolysis using PET-FDG imaging. The visual response score and the change in total lesion glycolysis, Clin Positron Imaging 2 (1999) 159–171. [6] S.W. Yoo, J. Kim, A. Chong, S.-Y. Kwon, J.-J. Min, H.-C. Song, et al., Metabolic tumor volume measured by F-18 FDG PET/CT can further stratify the prognosis of patients with stage IV non-small cell lung cancer, Nucl. Med. Mol. Imaging 46 (2012) 286–293 (2010).
87
[7] T.-Y. Ho, P.-C. Chou, C.-T. Yang, N.-M. Tsang, T.-C. Yen, Total lesion glycolysis determined per RECIST 1.1 criteria predicts survival in EGFR mutation-negative patients with advanced lung adenocarcinoma, Clin. Nucl. Med. 40 (2015) e295–9. [8] H. Zhang, K. Wroblewski, D. Appelbaum, Y. Pu, Independent prognostic value of whole-body metabolic tumor burden from FDG-PET in non-small cell lung cancer, Int. J. Comput. Assist. Radiol. Surg. 8 (2013) 181–191. [9] H.W. Chung, K.Y. Lee, H.J. Kim, W.S. Kim, Y. So, FDG PET/CT metabolic tumor volume and total lesion glycolysis predict prognosis in patients with advanced lung adenocarcinoma, J. Cancer Res. Clin. Oncol. 140 (2014) 89–98. [10] M. Reck, S. Popat, N. Reinmuth, D. De Ruysscher, K.M. Kerr, S. Peters, Metastatic non-small-cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up, Ann. Oncol. 25 (Suppl 3) (2014) iii27–39. [11] G.A. Masters, S. Temin, C.G. Azzoli, G. Giaccone, S. Baker, J.R. Brahmer, et al., Systemic therapy for stage IV non-small-cell lung cancer: American Society of Clinical Oncology Clinical Practice Guideline Update, J. Clin. Oncol. 33 (2015) 3488–3515. [12] E. Arriola, Á. Taus, D. Casadevall, Is there a role for epidermal growth factor receptor tyrosine kinase inhibitors in epidermal growth factor receptor wild-type non-small cell lung cancer? World J. Clin. Oncol. 6 (2015) 45–56. [13] E.A. Eisenhauer, P. Therasse, J. Bogaerts, L.H. Schwartz, D. Sargent, R. Ford, et al., New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1), Eur. J. Cancer 45 (2009) 228–247. [14] R.L. Wahl, H. Jacene, Y. Kasamon, M.A. Lodge, From RECIST to PERCIST: evolving considerations for PET response criteria in solid tumors, J. Nucl. Med. 50 (Suppl. 1) (2009) 122S–150S. [15] M.C. Garassino, O. Martelli, M. Broggini, G. Farina, S. Veronese, E. Rulli, et al., Erlotinib versus docetaxel as second-line treatment of patients with advanced non-small-cell lung cancer and wild-type EGFR tumours (TAILOR): a randomised controlled trial, Lancet Oncol. 14 (2013) 981–988. [16] T. Ciuleanu, L. Stelmakh, S. Cicenas, S. Miliauskas, A.C. Grigorescu, C. Hillenbach, et al., Efficacy and safety of erlotinib versus chemotherapy in second-line treatment of patients with advanced, non-small-cell lung cancer with poor prognosis (TITAN): a randomised multicentre, open-label, phase 3 study, Lancet Oncol. 13 (2012) 300–308. [17] B. Keam, S.J. Lee, T.M. Kim, J.C. Paeng, S.-H. Lee, D.-W. Kim, et al., Total lesion glycolysis in positron emission tomography can predict gefitinib outcomes in non-small-cell lung cancer with activating EGFR mutation, J. Thorac. Oncol. 10 (2015) 1189–1194. [18] D. Kahraman, A. Holstein, M. Scheffler, T. Zander, L. Nogova, A.A. Lammertsma, et al., Tumor lesion glycolysis and tumor lesion proliferation for response prediction and prognostic differentiation in patients with advanced non-small cell lung cancer treated with erlotinib, Clin. Nucl. Med. 37 (2012) 1058–1064. [19] S. Liao, B.C. Penney, H. Zhang, K. Suzuki, Y. Pu, Prognostic value of the quantitative metabolic volumetric measurement on 18F-FDG PET/CT in stage IV nonsurgical small-cell lung cancer, Acad. Radiol. 19 (2012) 69–77. [20] A.D. Nygaard, P.C. Holdgaard, K.-L.G. Spindler, N. Pallisgaard, A. Jakobsen, The correlation between cell-free DNA and tumour burden was estimated by PET/CT in patients with advanced NSCLC, Br. J. Cancer 110 (2014) 363–368.