- Email: [email protected]
CT imaging in response evaluation of patients with small cell lung cancer
Lung Cancer (2006) 54, 41—49
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/lungcan
PET/CT imaging in response evaluation of patients with small cell lung cancer Barbara M. Fischer a,∗, Jann Mortensen a, Seppo W. Langer b, Annika Loft a, Anne K. Berthelsen a, Gedske Daugaard b, Ulrik Lassen b, Heine H. Hansen b a
Department of Nuclear Medicine and PET, Centre of Diagnostic Imaging, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark b Department of Oncology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark Received 12 April 2006; received in revised form 4 June 2006; accepted 11 June 2006
KEYWORDS Response evaluation; Carcinoma; Small cell; Positron-emission tomography; PET/CT
Summary There is an increasing amount of evidence on the usability of PET in response evaluation of non-small cell lung cancer. However, data on SCLC is scarce and mainly retrospective. This prospective study assesses the use of PET (positron emission tomography) and PET/CT in response evaluation of patients with small cell lung cancer (SCLC). Methods: Assignment of early and final response was compared between PET, PET/CT, and CT in 20 patients with SCLC. Final response as assigned by CT (RECIST) served as reference. Results: At response evaluation after one cycle of chemotherapy major disagreement (responder versus non-responder) between PET and CT in predicting final response was seen in 1 of 12 patients. At final response evaluation major disagreement between PET, PET/CT and CT was seen in 2 of 19 patients (11%). All measurements of FDG-uptake were significantly correlated to size and changes in size as measured by CT. A significant difference in relative change in tumour FDG-uptake and volume was found between responding and non-responding patients. No significant difference was found between a visual and semi-quantitative analysis of PET. Conclusion: Response evaluation of SCLC by PET/CT is feasible, but it is uncertain whether it adds further information to evaluation by RECIST, thus further studies and standardization of methods are needed. © 2006 Elsevier Ireland Ltd. All rights reserved.
Abbreviations: SCLC, small cell lung cancer; PET, positron emission tomography; FDG, 18 F-fluorodeoxyglucose; RECIST, response evaluation criteria in solid tumours; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; ROI, region of interest; SUV, standardized uptake value: Q[ROI MBq/l] /(q[injected MBq] × normalization); SUVmax , Q equals maximum pixel value in ROI; SUVave , Q equals average pixel value in ROI; SUVbsa , SUV normalized to patient body surface area; SUVbw , SUV normalized to patient body weight; MTV, metabolic tumour volume: mean SUV times tumour volume using 50% (MTV50 ) or 75% (MTV75 ) isocontour; EORTC, European organization for research and treatment of cancer ∗ Corresponding author at: Department of Nuclear Medicine and PET, Copenhagen University Hospital, 3982 Blegdamsvej 9, DK-2100 Copenhagen, Denmark. Tel.: +45 39 563 550; fax: +45 35 453 898. E-mail address: malene.fi[email protected] (B.M. Fischer). 0169-5002/$ — see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2006.06.012
42
1. Introduction Small cell lung cancer (SCLC) is a rapidly disseminating cancer with an overall poor prognosis. SCLC accounts for approximately 15% of all new cases of lung cancer, equal to more than 75,000 cases in Europe and the USA in 2004 [1]. The initial response rate to chemotherapy is high (60—70%), but the majority of the patients relapses shortly after end of therapy [2]. Tumour response is presently evaluated by comparison of tumour size on CT scan before and after chemo- or radiotherapy. Previously this was measured in two-dimensions; WHO criteria [3] and more recently in one dimension; RECIST [4]. Interestingly the definition of tumour response using these criteria is based upon an experiment performed 30-years ago determining the accuracy with which 16 experienced oncologists could measure tumour size by palpation [5]. Structural imaging techniques such as chest radiographs and CT provide excellent anatomical details and are essential tools in the care of patients with lung cancer. However, whether structural imaging is the most valid measure of tumour response is uncertain [6]. Recently data on several solid tumours, including non-small cell lung cancer, has indicated a possible advantage of response evaluation by positron emission tomography (PET) with the radioactive glucose-analogue FDG [7]. Since FDG preferentially accumulates in viable tumour cells and not in fibrotic or necrotic tissue [8], a change in FDG-uptake on PET scan might be a better parameter for monitoring the response and it might be able to assess response before structural changes occur. Assessing early sub-clinical response could enable the clinician to optimise treatment of the single patient as well as shortening clinical trials evaluating new anti-cancer agents [6,7]. The current evidence shows that FDG-PET response has a variable correlation with CT response, and is probably more accurate than CT response [9]. The amount of evidence on the value of PET in therapy assessment of patients with non-small cell lung cancer (NSCLC) is growing [6], but the amount of data on SCLC is very scarce and consists primarily of retrospective data [10—13]. The purpose of this study was to prospectively evaluate the feasibility of PET/CT in response evaluation of patients with SCLC.
2. Materials and methods The local ethical committee approved the study and written, informed consent was obtained from all participants. Eligible patients had histological or cytological proven SCLC and were referred to our institution for final staging and treatment. The following patients were ineligible: patients with type-1 diabetes, known former or present malignant disease apart from SCLC, claustrophobia, pregnancy and age below 18 years. Apart from standard staging and evaluation examinations participating patients were offered PET/CT before, after one or two cycles of chemotherapy and at the end of therapy. Chemotherapy regimen at our institution consists of a platinium compound (carboplatin or cis-platin) and etoposide, sometimes combined with vincristine and/or topetecan administered every 3 weeks for six cycles. Patients with limited disease are also offered radiotherapy. Standard evaluation examinations consist of chest X-ray prior to each cycle of chemotherapy and CT of
B.M. Fischer et al. the chest and upper abdomen 3 weeks after the last cycle of chemotherapy. All staging examinations were performed 1—5 days before initiation of therapy, final evaluation examinations were performed 3 weeks after the last cycle and before commencement of further therapy. Patients were followed for at least 1 year after inclusion or until death. Follow-up data was retrieved from the hospital files.
2.1. PET/CT PET/CT scans were performed after a fasting period of 6 h; 400 MBq 18-F-FDG was given i.v. and the patient rested for at least 1 h (mean time from injection to scan 92 min, median 84 min). Plasma glucose was measured by glucometer prior to scanning (mean 4.7 mmol/l, median 4.6 mmol/l). The patient was scanned from the head to the upper thigh on an integrated PET/CT system (GE Discovery LS, General Electric Medical Systems, Milwaukee, WI). A diagnostic CT protocol (80—120 mA, 140 kV, tube-rotation time of 0.5 s per rotation, pitch 6 and a slice thickness of 5 mm, with intravenous contrast) was applied followed by PET-scan (3 or 5 min emission scan per table position, depending on whether the patient could keep the arms above the head during scanning or not). Data from the CT-scan was used for attenuation correction. The PET scan was reconstructed by OSEM (ordered subset expectation minimization).
2.2. Classification of response Response was assessed according to the RECIST criteria [4]. PET-scans were analysed visually according to the criteria described by MacManus et al. [9]: complete metabolic response (CMR), partial metabolic response (PMR), stable metabolic disease (SMD) and progressive metabolic disease (PMD). The response was evaluated on PET/CT in consensus between radiologist and nuclear medicine physician. All results were reported on predefined report forms. A semi quantitative analysis of the PET scans was performed. Maximum and average standardized uptake value (SUVmax and SUVave ) corrected for bodyweight (bw) and body surface area (bsa) was assigned for each primary tumour, mediastinal masses and distant metastasis (the largest/hottest metastasis in each affected organ). Region of interest (ROI) was defined semi automatically using the isoconturing tool provided with the software (Leonardo VD30A, Syngo VX49B, ©Siemens AG, Berlin, Germany). Two metabolic tumour volumes, MTV, were calculated as average SUV times tumour volume defined by the isocontouring line of 50% and 75% of maximum SUV, respectively. On the basis of changes in SUVmax the response criteria as suggested by EORTC [14] was applied.
2.3. Data analysis Following variables were derived from the PET/CT scan of each patient: tumour response on CT, visual assessment of tumour response on PET and PET/CT, semi quantitative measures of tumour response on PET: maximum SUV and MTV, each reported as mean with standard deviation. The
Response evaluation in SCLC by PET/CT
43
qualitative assessment of response between CT, PET and PET/CT were compared by McNemar-Bowker (4 × 4 table) and McNemars test (2 × 2 table, responders (CR/CMR and PR/PMR) versus non-responders (SD/SMD and PD/PMD)) for paired data and agreement was measured by kappa () [15—17]. Differences between groups of scale variables with equal variances were compared by t-test, comparison of non-parametric variables were performed using the Mann—Whitney U-test, level of significance p < 0.05 (twotailed). A correlation analysis (Pearson) was performed between patient demographics (height, weight, age and sex) and SUV. Correlations between other parameters were examined by Spearman’s . Inter- and intra-observer variability was calculated as intra-class correlation coefficient (ICC) and r2 . Statistical analysis was performed using SPSS version 14.0.
3. Results 3.1. Patient data and follow-up Thirty patients received baseline PET/CT (prior to initiation of therapy), 20 patients received both baseline PET/CT, early evaluation and/or final PET/CT (after one and six cycles of chemotherapy respectively), resulting in a total of 62 scans (Fig. 1). Of the 20 (5 males and 15 females) patients receiving at least one evaluation scan the mean age was 64 years (range 51—77 years). Fifteen patients were diagnosed with ED (extensive disease) and five patients with LD (limited disease). All patients received combination chemotherapy (median six series, range 1—15). Six patients received radiotherapy as a part of first line therapy. The clinical course of the individual patients who received at least one PET/CT during or after therapy is summarized in Fig. 2. At the time of data analysis 13/20 (65%) patients had died of SCLC, median survival in this group being 10.7 months (6.2—20.2). The remaining seven patients were followed for a median of 16.5 months (12.3—27.7).
3.2. Visual assessment of response In the following sections incongruence in distinction between responding (CR/CMR or PR/PMR) and nonresponding patients (SD/SMD or PD/PMD) will be termed major disagreement, whereas, incongruence between e.g. CR versus PMR will be termed minor disagreement. PET/CT scan was performed after the first cycle of chemotherapy in 12 patients (early response assessment, Table 1). Major disagreement between early PET and early CT was seen in one patient: patient 11 (Fig. 2) achieved PR on CT and PMD on PET, PET/CT was equivocal; the patient progressed during the last cycle of chemotherapy and final PET/CT was never performed. Minor disagreement was observed in patient 6 who achieved PR on CT but CMR on PET, the patient relapsed and died after 10 months. Major disagreement between early PET/CT and final CT was seen in patient 15 who after one cycle of chemotherapy had SD on CT and PET/CT and SMD on PET, final evaluation showed PR and CMR on PET/CT and PET; the patient died of SCLC 3 months after ended therapy. The 1-year survival rate for patients responding on early PET was 70% (7 of the 10
Fig. 1 Flow chart summarizing the inclusion of patients, number of scans and distribution in response categories after final CT. (*) No PET/CT performed due to decreasing performance status (n = 1), disease progression (n = 1), no show (n = 1) and in one patient PET/CT was not scheduled without further explanation stated in the hospital files. One patient received separate CT and PET and is not included in the final analysis. (†) CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease.
responding patients); on PET/CT and CT it was 64% (7 of 11 patients). Two patients on PET and 1 patient on PET/CT and CT did not respond, both died of SCLC 6 and 8 months after inclusion. PET/CT scan was performed after six cycles of chemotherapy in 19 patients (final response assessment). Table 2 presents cross-tabulation of the results of final response evaluation by CT, PET and PET/CT. Seventeen patients were categorized as responders (CR or PR) on the final CT evaluation and two patients as non-responders (SD and PD). Overall disagreement was found in eight patients (42%), but major disagreement (responder versus non-responder) was only found in two (11%) patients. PET was more likely than CT and PET/CT to assign the patients as having complete response (eight patients versus four and five, respectively). However, no significant difference was found between CT, PET, and PET/CT in categorizing each patient as responder or non-responder (McNemar, p > 0.5), agreement between CT and PET/CT as well as CT and PET was moderate ( = 0.44, p = 0.054) and agreement between PET and PET/CT was excellent ( = 1.00, p = 0.000). Looking at all four response categories the level of agreement between CT and PET/CT was substantial ( = 0.69, p = 0.000), agreement between PET/CT and PET was still very good
44
B.M. Fischer et al.
Fig. 2 Clinical course of all patients receiving baseline PET/CT and PET/CT during and/or after chemotherapy. ( ) PET/CT; (ED) extensive disease; (LD) limited disease; ( ) radiotherapy as a part of first line treatment. (CR/CMR) complete response (CT)/complete metabolic response (PET); (PR/PMR) partial response (CT)/partial metabolic response (PET); (SD/SMD) stable disease (CT)/stable metabolic disease (PET); (PD/PMD) progressive disease (CT)/progressive metabolic disease (PET); ( ) relapse/progression of disease; (䊉) death of disease; () alive at follow-up; ( ) chemotherapy, number of completed series.
( = 1.00, p = 0.000), whereas, agreement between CT and PET was only fair ( = 0.28, p = 0.081). Survival rate after 1 year was similar for PET, CT and PET/CT: 65% (11/17 patients) for responding patients, and 50% (one of two patients) for non-responding patients.
3.3. Quantitative assessment of response FDG-uptake before, during and after therapy was measured for primary tumour, mediastinal metastases (if discernible from primary tumour) and the most intense focus in each affected organ. Forty-eight foci were evaluated at baseline,
26 at early evaluation and 44 at final evaluation. Inter- and intra-observer variation was estimated as r2 and ICC (intraclass correlation coefficient) based on data from the first 10 consecutive patients (30 foci) that completed evaluation scans (Fig. 3). Correlation was high both between two independent readers as well as repeated measurements by the same reader. The mean percent changes in SUVmax and MTV50 on a patient-by-patient basis (primary tumour) are illustrated in Table 3. A significant difference in percentage change from baseline until final evaluation in tumour FDG-uptake and size was found between responders and non-responders: SUVmax decreased 47% more in responders than in non-
Response evaluation in SCLC by PET/CT
45
Table 1 Cross-tabulation of: (A) early response measured by CT (RECIST)a and PET (visual analysis)b ; (B) early response measured by PET and final response measured by CT; (C) early response and final response measured by CT PET early CMR
PMR
SMD
PMD
Total
(A) Cross-tabulation of early response measured by CT (RECIST)a and PET (visual analysis) CT early CR PR 1 9 1 SD 1 PD Total
1
9
1
1
12
(B) Cross-tabulation of early response measured by PET and final response measured by CT CT final CR 3 PR 1 6 1 SD PD 1c Total
1
9
1
1
PR
SD
PD
0 11 1 0
3 8 0 1 12
CT early CR
Total
(C) Cross-tabulation of early response and final response measured by CT CT final CR 3 PR 7 1 SD PD 1c Total
0
11
1
0
3 8 0 1 12
a
CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. CMR: complete metabolic response, PMR: partiel metabolic response, SMD: stable metabolic response, PMD: progressive metabolic disease. c This patient had PMD on PET and PR on CT, relapsed during the last course of chemotherapy and never received final PET/CT. b
responders (95% C.I. [−84%, −10%], p = 0.04). Neither of the non-responding patients had an early evaluation scan, thus a similar analysis at this time point could not be performed. However, in the early evaluation scans from 12 responding patients an average decline in SUVmax and MTV to approximately 37% and 8% of baseline values could be observed after the first cycle of chemotherapy. At the same time a decline in size at CT to 49% percent of baseline could be observed. A significant correlation was found between results from PET (SUVmax , SUVave and MTV) and results from CT (size in mm) (Spearman’s 0.65, 0.65 and 0.79, respectively, p < 0.001). The performance of visual assessment of PET and a quantitative approach using the EORTC-criteria was compared using kappa and McNemars test. No significant difference in assignment of response categories was found between visual assessment of PET and EORTC, and CT and EORTC respectively. Agreement between visual assessment of PET and EORTC was good ( = 0.73, p = 0.000), between CT and EORTC poor ( = 0.15, p = 0.363). The results were identical whether SUV was corrected for body weight or body surface area.
4. Discussion The purpose of this prospective study was to assess the value of PET/CT in response evaluation after treatment of SCLC. We found that PET and PET/CT can be used for response evaluation after chemotherapy in SCLC: changes in FDGuptake by primary tumour and metastases are significantly correlated to changes in size as measured by CT. Since SCLC is a rapidly proliferating but also, in contrast to NSCLC, rapidly responding cancer this could be expected. PET/CT scan was performed after the first cycle of chemotherapy in 12 patients, major disagreement between early PET and early CT was observed in one patient (8%). Assessment of response after one or two cycles of chemotherapy could lead to an early change in treatment strategy in case of treatment failure. This could potentially improve the treatment of the single patient, but the numbers of effective salvage regimens for SCLC are still scarce. After definitive chemotherapy (six series) incongruence between final response evaluated by PET and CT was seen in eight (42%) of the patients. Major disagreement was observed in two patients (11%), this could have influenced
46
B.M. Fischer et al.
Table 2
Crosstabulation of final response measured by CT (RECIST)a , PET (visual analysis)b and PET/CT PET CMR
PMR
SMD
PMD
Total
(A) Cross-tabulation of final response measured by CT (RECIST)c CT CR 3 1 PR 5 7 1 SD 1 PD 1 0 Total
8
4 13 1 1
9
1
1
19
PR
SD
PD
PET/CT CR
(B) Cross-tabulation of final response measured by PET (visual analysis) CT CR 4 PR 1 11 1 SD 1 PD 1 0 Total
5
12
1
PMR
SMD
Total d
4 13 1 1
1
19
PET CMR
(C) Cross-tabulation of final response measured by PET/CT PET/CT CR 4 1 PR 5 7 SD 1 PD Total
9
8
PMD
Total
e
1
1
5 12 1 1
1
19
a
CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. b CMR: Complete metabolic response, PMR: Partiel metabolic response, SMD: Stable metabolic response, PMD: Progressive metabolic disease. c McNemar-Bowker test (4 × 4 table): 2.667, p = 0.264. Mcnemar test (2 × 2 table): p = 1.00. d McNemar-Bowker test (4 × 4 table): 1.000, p = 0.607. Mcnemar test (2 × 2 table): p = 1.00. e McNemar-Bowker test (4 × 4 table): 2.667, p = 0.102. Mcnemar test (2 × 2 table): p = 1.00.
patient management (responder versus non-responder) e.g. changing decision regarding prophylactic cranial irradiation. When comparing PET/CT and CT overall incongruence was seen only in three patients (16%) causing possible change in patient management in two patients. In spite of these findings no significant difference in response (RECIST) as assigned by CT and response assigned by PET and PET/CT during and after chemotherapy was found. Previous published studies on response evaluation and restaging of patients with PET in SCLC are retrospective [10—12]. Kamel et al. [10] found incongruence in 6/20 patients (30%); with impact on patient management in 3/20 (15%). Blum et al. [11] reported even higher numbers; incongruent results was reported in 63% of the 25 patients (affecting management in 10 patients (40%)). Neither paper reported any level of significance. Both studies were retrospective; thus the majority of the PET scans was
performed on a specific clinical indication. Since PET was not a part of standard staging and response evaluation at either institution only a minority of the patients receiving PET for response evaluation had a baseline PET for comparison: Blum et al. reported that the number of PET-scans with impact on patient management decreased from 40% to 25% when looking only on patients with both baseline and restaging PET. Pandit et al. has suggested an independent prognostic value of PET imaging in SCLC [18]. Our study was too small to perform survival analysis, but 1-year survival rate in the patients responding on early PET and CT was 0.70 and 0.64, respectively. However, regarding cases of discordance between CT and PET at final response assessment (Fig. 2), a simple analysis of the correlation between response category and survival ≥12 months reveals five errors by PET and three errors by CT. At early response evaluation one error
Response evaluation in SCLC by PET/CT
47
Fig. 3
Inter- and intra-observer variation.
by PET and one by CT is observed. Studies on other solid tumours have indicated that PET [19—22] after one or two cycles of chemotherapy can accurately predict response. However, none of the studies reported results from simultaneous response evaluation by CT for comparison. The design of this study was planned in accordance with the recommendations by EORTC from 1999 [14]. On ROI definition we applied a semi automatically threshold method using the maximum pixel value and 50% and 75% isocontour. The threshold method has previously been reported to have high test—retest reproducibility [23] and to be relatively independent of tumour size and geometry [24]. We found excellent inter- and intra-observer repeatability concerning both FDG-uptake and tumour volume using the threshold method. SUV has to be normalized to the body weight (SUVbw ), body surface area (SUVbsa ) or lean body mass (SUVlbm ): The EORTC PET study group recommended that SUV corrected for body surface area should be the minimum standard of measurement. We found that the mean relative change in FDG-uptake from before to after therapy was significantly smaller when measured as SUVbsa compared to SUVbw . This difference of an estimated two percent is small but should however be kept in mind when comparing relative changes in FDG-uptake using different means of normalization. In an attempt to ease comparison between response evaluation by CT and PET; four response categories corresponding to the RECIST criteria has been suggested by EORTC based on relative changes in tumour SUV [14] and by MacManus et al. based on visual analysis [9]. We applied both methods on our data and found no significant difference.
The literature on PET in response evaluation of SCLC is scarce and on PET/CT next to nothing. The sum of our and previous findings indicates that response evaluation by PET and PET/CT is feasible both during and after chemotherapy. Whether this will change patient management or merely serve as another prognostic factor remains to be settled [18,25], and larger prospective studies on PET/CT in SCLC is needed. However, it is our experience that performing a prospective study on a SCLC population is challenging: only 20 of the initially included 30 patients received a response evaluation scan. One explanation is the poor performance status of many patients with SCLC resulting in frequent changes in treatment schedules. This combined with a fully booked PET/CT system makes the logistics in a response evaluation study like this difficult. Furthermore, in order to include a larger number of patients a multicenter study would probably be necessary. Since 1985 the 5-year survival rate for patients with limited disease SCLC in the United States has only increased from 10.5% to 13.3% and an unchanged 14% of the patients do not receive any treatment, thus new and improved treatment strategies are certainly warranted [26]. Current evidence on NSCLC indicates that assessing early sub-clinical response on PET can help optimise treatment of the single patient as well as shortening clinical trials evaluating new anti-cancer agents [6,7]. Even though SCLC is a clinical and histological distinct entity different from NSCLC in many ways, it might be sound to regard the data on PET and SCLC as a reassuring proof of principle and adapt the growing amount of evidence from NSCLC in the planning of SCLC treatment protocols. On the other hand: the lack of salvage therapies and the fact that SCLC very rapidly
48
B.M. Fischer et al.
Table 3 SUVa , MTVb and size on CT of primary tumour as percent of baseline findings after first cycle of chemotherapy (early) and after ended therapy (final)
Earlyc All CR PR
d
n
SUV %baseline
SD
MTV %baseline
SD
CT %baseline
SD
12 3 9
36.63 39.35 35.73
17.46 19.50 17.90
7.84 3.51 9.29
8.94 1.94 9.98
48.80 51.31 47.96
22.39 4.51 26.10
Finale
n
SUV %baseline
SD
MTV %baseline
SD
CT %baseline
SD
All Responder CR PR Non-responder SD PD
19 17 4 13 2 1 1
26.17 21.23f 11.00 24.37 68.20f 69.24 67.15
26.95 23.88 22.01 24.37 1.48
11.55 3.11g 2.79 3.21 79.08g 48.15 110.00
27.28 5.62 5.59 5.88 43.73
34.92 25.56 0.00 33.42 114.52 91.53 137.50
36.91 24.12 0.00 22.15 32.51
a
SUV: standardized uptake value, normalized to body weight. MTV: metabolic tumour volume, mean SUV times tumour volume using 50% isocontour, normalized to body weight. c Result after final PET/CT scan, categorized according to RECIST (CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. d The two patients categorized as non-responders on the final PET/CT scan missed PET/CT after the first cycle of chemotherapy. e Result after final PET/CT scan, categorized according to RECIST (CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. f Difference in mean between responders and non-responders: −46.97, 95% C.I [−83.51, −10,42], p = 0.04. g Difference in mean between responders and non-responders: −75.97, 95% C.I [−95.39, −56.55], p = 0.02. b
response to chemotherapy indicates that the therapeutic efficacy [27,28] of PET in SCLC currently is less promising than in NSCLC.
give any definitive conclusion regarding the value of PET/CT in response evaluation of SCLC, thus further studies are needed.
5. Conclusion
References
PET and PET/CT can be used for response evaluation after chemotherapy in SCLC; changes in FDG-uptake by primary tumour and metastases are significantly correlated to changes in size as measured by CT. No significant difference was found between a visual and semi-quantitative analysis of the PET scans. Our results indicate that it is uncertain whether response assessment, early or final, by PET and PET/CT add further benefit to the treatment of patients with SCLC as compared to assessment by CT alone. Giving the low number of participants in this study we cannot
[1] Jackman DM, Johnson BE. Small-cell lung cancer. Lancet 2005;366:1385—96. [2] Simon GR, Wagner H. Small cell lung cancer. Chest 2003; 123:259S—71S. [3] Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer 1981;47:207—14. [4] Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS, et al. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 2000;92:205—16. [5] Weber WA. Use of PET for monitoring cancer therapy and for predicting outcome. J Nucl Med 2005;46:983—95.
Response evaluation in SCLC by PET/CT [6] Vansteenkiste J, Fischer BM, Dooms C, Mortensen J. Positronemission tomography in prognostic and therapeutic assessment of lung cancer: systematic review. Lancet Oncol 2004;5: 531—40. [7] Juweid ME, Cheson BD. Positron-emission tomography and assessment of cancer therapy. N Engl J Med 2006;345(5):496— 507. [8] Higashi K, Clavo AC, Wahl RL. Does FDG uptake measure proliferative activity of human cancer cells? In vitro comparison with DNA flow cytometry and tritiated thymidine uptake. J Nucl Med 1993;34:414—9. [9] MacManus MP, Hicks RJ, Matthews JP, McKenzie A, Rischin D, et al. Positron emission tomography is superior to computed tomography scanning for response-assessment after radical radiotherapy or chemoradiotherapy in patients with non-small-cell lung cancer. J Clin Oncol 2003;21(7): 1285—92. [10] 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(12):1911—7. [11] Blum R, MacManus MP, Rischin D, Michael M, Hicks R. Impact of positron emission tmography on the management of patients with small-cell lung cancer. Am J Clin Oncol 2004;27(2): 164—71. [12] Zhao DS, Valdivia AY, Li Y, Blaufox DM. 18 F-fluorodeoxyglucose positron emission tomography in small-cell lung cancer. Semin Nucl Med 2002;XXXII(4):272—5. [13] Schumacher T, Brink I, Mix M, Reinhardt M, Herget G, Digel W, et al. FDG-PET imaging for the staging and followup of small cell lung cancer. Eur J Nucl Med 2001;28(4): 483—8. [14] Young H, Baum R, Cremerius U, Herholz K, Hoekstra OS, Lammertsma AA, et al. Measurement of clinical and subclinical tumour response using [18F]-fluorodeoxyglucose and positron emission tomography: review and 1999 EORTC recommendations. Eur J Cancer 1999;35(13):1773—82. [15] Altman DG. Some common problems in medical research. In: Practical statistics for medical research. 1st ed. London: Chapman & Hall/CRC; 1991. p. 396—439. [16] Ludbrook J. Statistical techniques for comparing measurers and methods of measurement: a critical review. Clin Exp Pharmacol Physiol 2002;29:527—36.
49 [17] G¨ onen M, Panageas KS, Larson SM. Statistical issues in analysis of diagnostic imaging experiments with multiple observations per patient. Radiology 2001;221:763—7. [18] Pandit N, Gonen M, Krug L, Larson SM. Prognostic value of [18 F]FDG-PET imaging in small cell lung cancer. Eur J Nucl Med 2003;30(1):78—84. [19] Weber WA, Petersen V, Schmidt B, Tyndale-Hines L, Link T, et al. Positron emission tomography in non-small-cell lung cancer: prediction of response to chemotherapy by quantitative asessment of glucose use. J Clin Oncol 2003;21:2651—7. [20] Schelling M, Avril N, N¨ ahrig J, Kuhn W, R¨ omer WSD, Werner M, et al. Positron emission tomography using [18 F]Fluorodeoxyglucose for monitoring primary chemotherapy in breast cancer. J Clin Oncol 2000;18:1689—95. [21] Schwarz JD, Bader M, Jenicke L, Hemminger G, J¨ anicke F, Avril N. early prediction of response to chemotherapy in metastatic breast cancer using sequential 18 F-FDG PET. J Nucl Med 2005;46:1144—50. [22] Avril N, Sassen S, Schmalfeldt B, Naehrig J, Rutke S, Weber WA, et al. Prediction of response to neoadjuvant chemotherapy by sequential F-18-fluorodeoxyglucose positron emission tomography in patients with advanced-stage ovarian cancer. J Clin Oncol 2005;23(30):7445—53. [23] Krak NC, Boellaard R, Hoekstra OS, Twisk JWR, Hoekstra CJ, Lammertsma AA. Effects of ROI definition an reconstruction method on quantitative outcome and applicability in a response monitoring trial. Eur J Nucl Med Mol Imag 2005;32:294—301. [24] Lee JR, Madsen MT, Bushnel D, Menda Y. A threshold method to improve standardized uptake value reproducibility. Nucl Med Commun 2000;21(7):685—90. [25] Markman M. Use of positron emission tomography scans in ovarian cancer: a diagnostic technique in search of an indication. J Clin Oncol 2005;23(30):7385—7. [26] Lee CB, Morris DE, Fried DB, Socinski MA. Current and evolving treatment options for limited stage small cell lung cancer. Curr Opin Oncol 2006;18:162—72. [27] Price P, Laking G. How should we introduce clinical PET in the UK? The oncologists need to have a view. Clin Oncol 2004;16:172—5. [28] Markman M. Use of positron emission tomography scans in ovarian cancer: a diagnostic technique in search of an indication. J Clin Oncol 2005;23(30):7385—6.
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/lungcan
PET/CT imaging in response evaluation of patients with small cell lung cancer Barbara M. Fischer a,∗, Jann Mortensen a, Seppo W. Langer b, Annika Loft a, Anne K. Berthelsen a, Gedske Daugaard b, Ulrik Lassen b, Heine H. Hansen b a
Department of Nuclear Medicine and PET, Centre of Diagnostic Imaging, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark b Department of Oncology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark Received 12 April 2006; received in revised form 4 June 2006; accepted 11 June 2006
KEYWORDS Response evaluation; Carcinoma; Small cell; Positron-emission tomography; PET/CT
Summary There is an increasing amount of evidence on the usability of PET in response evaluation of non-small cell lung cancer. However, data on SCLC is scarce and mainly retrospective. This prospective study assesses the use of PET (positron emission tomography) and PET/CT in response evaluation of patients with small cell lung cancer (SCLC). Methods: Assignment of early and final response was compared between PET, PET/CT, and CT in 20 patients with SCLC. Final response as assigned by CT (RECIST) served as reference. Results: At response evaluation after one cycle of chemotherapy major disagreement (responder versus non-responder) between PET and CT in predicting final response was seen in 1 of 12 patients. At final response evaluation major disagreement between PET, PET/CT and CT was seen in 2 of 19 patients (11%). All measurements of FDG-uptake were significantly correlated to size and changes in size as measured by CT. A significant difference in relative change in tumour FDG-uptake and volume was found between responding and non-responding patients. No significant difference was found between a visual and semi-quantitative analysis of PET. Conclusion: Response evaluation of SCLC by PET/CT is feasible, but it is uncertain whether it adds further information to evaluation by RECIST, thus further studies and standardization of methods are needed. © 2006 Elsevier Ireland Ltd. All rights reserved.
Abbreviations: SCLC, small cell lung cancer; PET, positron emission tomography; FDG, 18 F-fluorodeoxyglucose; RECIST, response evaluation criteria in solid tumours; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; ROI, region of interest; SUV, standardized uptake value: Q[ROI MBq/l] /(q[injected MBq] × normalization); SUVmax , Q equals maximum pixel value in ROI; SUVave , Q equals average pixel value in ROI; SUVbsa , SUV normalized to patient body surface area; SUVbw , SUV normalized to patient body weight; MTV, metabolic tumour volume: mean SUV times tumour volume using 50% (MTV50 ) or 75% (MTV75 ) isocontour; EORTC, European organization for research and treatment of cancer ∗ Corresponding author at: Department of Nuclear Medicine and PET, Copenhagen University Hospital, 3982 Blegdamsvej 9, DK-2100 Copenhagen, Denmark. Tel.: +45 39 563 550; fax: +45 35 453 898. E-mail address: malene.fi[email protected] (B.M. Fischer). 0169-5002/$ — see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2006.06.012
42
1. Introduction Small cell lung cancer (SCLC) is a rapidly disseminating cancer with an overall poor prognosis. SCLC accounts for approximately 15% of all new cases of lung cancer, equal to more than 75,000 cases in Europe and the USA in 2004 [1]. The initial response rate to chemotherapy is high (60—70%), but the majority of the patients relapses shortly after end of therapy [2]. Tumour response is presently evaluated by comparison of tumour size on CT scan before and after chemo- or radiotherapy. Previously this was measured in two-dimensions; WHO criteria [3] and more recently in one dimension; RECIST [4]. Interestingly the definition of tumour response using these criteria is based upon an experiment performed 30-years ago determining the accuracy with which 16 experienced oncologists could measure tumour size by palpation [5]. Structural imaging techniques such as chest radiographs and CT provide excellent anatomical details and are essential tools in the care of patients with lung cancer. However, whether structural imaging is the most valid measure of tumour response is uncertain [6]. Recently data on several solid tumours, including non-small cell lung cancer, has indicated a possible advantage of response evaluation by positron emission tomography (PET) with the radioactive glucose-analogue FDG [7]. Since FDG preferentially accumulates in viable tumour cells and not in fibrotic or necrotic tissue [8], a change in FDG-uptake on PET scan might be a better parameter for monitoring the response and it might be able to assess response before structural changes occur. Assessing early sub-clinical response could enable the clinician to optimise treatment of the single patient as well as shortening clinical trials evaluating new anti-cancer agents [6,7]. The current evidence shows that FDG-PET response has a variable correlation with CT response, and is probably more accurate than CT response [9]. The amount of evidence on the value of PET in therapy assessment of patients with non-small cell lung cancer (NSCLC) is growing [6], but the amount of data on SCLC is very scarce and consists primarily of retrospective data [10—13]. The purpose of this study was to prospectively evaluate the feasibility of PET/CT in response evaluation of patients with SCLC.
2. Materials and methods The local ethical committee approved the study and written, informed consent was obtained from all participants. Eligible patients had histological or cytological proven SCLC and were referred to our institution for final staging and treatment. The following patients were ineligible: patients with type-1 diabetes, known former or present malignant disease apart from SCLC, claustrophobia, pregnancy and age below 18 years. Apart from standard staging and evaluation examinations participating patients were offered PET/CT before, after one or two cycles of chemotherapy and at the end of therapy. Chemotherapy regimen at our institution consists of a platinium compound (carboplatin or cis-platin) and etoposide, sometimes combined with vincristine and/or topetecan administered every 3 weeks for six cycles. Patients with limited disease are also offered radiotherapy. Standard evaluation examinations consist of chest X-ray prior to each cycle of chemotherapy and CT of
B.M. Fischer et al. the chest and upper abdomen 3 weeks after the last cycle of chemotherapy. All staging examinations were performed 1—5 days before initiation of therapy, final evaluation examinations were performed 3 weeks after the last cycle and before commencement of further therapy. Patients were followed for at least 1 year after inclusion or until death. Follow-up data was retrieved from the hospital files.
2.1. PET/CT PET/CT scans were performed after a fasting period of 6 h; 400 MBq 18-F-FDG was given i.v. and the patient rested for at least 1 h (mean time from injection to scan 92 min, median 84 min). Plasma glucose was measured by glucometer prior to scanning (mean 4.7 mmol/l, median 4.6 mmol/l). The patient was scanned from the head to the upper thigh on an integrated PET/CT system (GE Discovery LS, General Electric Medical Systems, Milwaukee, WI). A diagnostic CT protocol (80—120 mA, 140 kV, tube-rotation time of 0.5 s per rotation, pitch 6 and a slice thickness of 5 mm, with intravenous contrast) was applied followed by PET-scan (3 or 5 min emission scan per table position, depending on whether the patient could keep the arms above the head during scanning or not). Data from the CT-scan was used for attenuation correction. The PET scan was reconstructed by OSEM (ordered subset expectation minimization).
2.2. Classification of response Response was assessed according to the RECIST criteria [4]. PET-scans were analysed visually according to the criteria described by MacManus et al. [9]: complete metabolic response (CMR), partial metabolic response (PMR), stable metabolic disease (SMD) and progressive metabolic disease (PMD). The response was evaluated on PET/CT in consensus between radiologist and nuclear medicine physician. All results were reported on predefined report forms. A semi quantitative analysis of the PET scans was performed. Maximum and average standardized uptake value (SUVmax and SUVave ) corrected for bodyweight (bw) and body surface area (bsa) was assigned for each primary tumour, mediastinal masses and distant metastasis (the largest/hottest metastasis in each affected organ). Region of interest (ROI) was defined semi automatically using the isoconturing tool provided with the software (Leonardo VD30A, Syngo VX49B, ©Siemens AG, Berlin, Germany). Two metabolic tumour volumes, MTV, were calculated as average SUV times tumour volume defined by the isocontouring line of 50% and 75% of maximum SUV, respectively. On the basis of changes in SUVmax the response criteria as suggested by EORTC [14] was applied.
2.3. Data analysis Following variables were derived from the PET/CT scan of each patient: tumour response on CT, visual assessment of tumour response on PET and PET/CT, semi quantitative measures of tumour response on PET: maximum SUV and MTV, each reported as mean with standard deviation. The
Response evaluation in SCLC by PET/CT
43
qualitative assessment of response between CT, PET and PET/CT were compared by McNemar-Bowker (4 × 4 table) and McNemars test (2 × 2 table, responders (CR/CMR and PR/PMR) versus non-responders (SD/SMD and PD/PMD)) for paired data and agreement was measured by kappa () [15—17]. Differences between groups of scale variables with equal variances were compared by t-test, comparison of non-parametric variables were performed using the Mann—Whitney U-test, level of significance p < 0.05 (twotailed). A correlation analysis (Pearson) was performed between patient demographics (height, weight, age and sex) and SUV. Correlations between other parameters were examined by Spearman’s . Inter- and intra-observer variability was calculated as intra-class correlation coefficient (ICC) and r2 . Statistical analysis was performed using SPSS version 14.0.
3. Results 3.1. Patient data and follow-up Thirty patients received baseline PET/CT (prior to initiation of therapy), 20 patients received both baseline PET/CT, early evaluation and/or final PET/CT (after one and six cycles of chemotherapy respectively), resulting in a total of 62 scans (Fig. 1). Of the 20 (5 males and 15 females) patients receiving at least one evaluation scan the mean age was 64 years (range 51—77 years). Fifteen patients were diagnosed with ED (extensive disease) and five patients with LD (limited disease). All patients received combination chemotherapy (median six series, range 1—15). Six patients received radiotherapy as a part of first line therapy. The clinical course of the individual patients who received at least one PET/CT during or after therapy is summarized in Fig. 2. At the time of data analysis 13/20 (65%) patients had died of SCLC, median survival in this group being 10.7 months (6.2—20.2). The remaining seven patients were followed for a median of 16.5 months (12.3—27.7).
3.2. Visual assessment of response In the following sections incongruence in distinction between responding (CR/CMR or PR/PMR) and nonresponding patients (SD/SMD or PD/PMD) will be termed major disagreement, whereas, incongruence between e.g. CR versus PMR will be termed minor disagreement. PET/CT scan was performed after the first cycle of chemotherapy in 12 patients (early response assessment, Table 1). Major disagreement between early PET and early CT was seen in one patient: patient 11 (Fig. 2) achieved PR on CT and PMD on PET, PET/CT was equivocal; the patient progressed during the last cycle of chemotherapy and final PET/CT was never performed. Minor disagreement was observed in patient 6 who achieved PR on CT but CMR on PET, the patient relapsed and died after 10 months. Major disagreement between early PET/CT and final CT was seen in patient 15 who after one cycle of chemotherapy had SD on CT and PET/CT and SMD on PET, final evaluation showed PR and CMR on PET/CT and PET; the patient died of SCLC 3 months after ended therapy. The 1-year survival rate for patients responding on early PET was 70% (7 of the 10
Fig. 1 Flow chart summarizing the inclusion of patients, number of scans and distribution in response categories after final CT. (*) No PET/CT performed due to decreasing performance status (n = 1), disease progression (n = 1), no show (n = 1) and in one patient PET/CT was not scheduled without further explanation stated in the hospital files. One patient received separate CT and PET and is not included in the final analysis. (†) CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease.
responding patients); on PET/CT and CT it was 64% (7 of 11 patients). Two patients on PET and 1 patient on PET/CT and CT did not respond, both died of SCLC 6 and 8 months after inclusion. PET/CT scan was performed after six cycles of chemotherapy in 19 patients (final response assessment). Table 2 presents cross-tabulation of the results of final response evaluation by CT, PET and PET/CT. Seventeen patients were categorized as responders (CR or PR) on the final CT evaluation and two patients as non-responders (SD and PD). Overall disagreement was found in eight patients (42%), but major disagreement (responder versus non-responder) was only found in two (11%) patients. PET was more likely than CT and PET/CT to assign the patients as having complete response (eight patients versus four and five, respectively). However, no significant difference was found between CT, PET, and PET/CT in categorizing each patient as responder or non-responder (McNemar, p > 0.5), agreement between CT and PET/CT as well as CT and PET was moderate ( = 0.44, p = 0.054) and agreement between PET and PET/CT was excellent ( = 1.00, p = 0.000). Looking at all four response categories the level of agreement between CT and PET/CT was substantial ( = 0.69, p = 0.000), agreement between PET/CT and PET was still very good
44
B.M. Fischer et al.
Fig. 2 Clinical course of all patients receiving baseline PET/CT and PET/CT during and/or after chemotherapy. ( ) PET/CT; (ED) extensive disease; (LD) limited disease; ( ) radiotherapy as a part of first line treatment. (CR/CMR) complete response (CT)/complete metabolic response (PET); (PR/PMR) partial response (CT)/partial metabolic response (PET); (SD/SMD) stable disease (CT)/stable metabolic disease (PET); (PD/PMD) progressive disease (CT)/progressive metabolic disease (PET); ( ) relapse/progression of disease; (䊉) death of disease; () alive at follow-up; ( ) chemotherapy, number of completed series.
( = 1.00, p = 0.000), whereas, agreement between CT and PET was only fair ( = 0.28, p = 0.081). Survival rate after 1 year was similar for PET, CT and PET/CT: 65% (11/17 patients) for responding patients, and 50% (one of two patients) for non-responding patients.
3.3. Quantitative assessment of response FDG-uptake before, during and after therapy was measured for primary tumour, mediastinal metastases (if discernible from primary tumour) and the most intense focus in each affected organ. Forty-eight foci were evaluated at baseline,
26 at early evaluation and 44 at final evaluation. Inter- and intra-observer variation was estimated as r2 and ICC (intraclass correlation coefficient) based on data from the first 10 consecutive patients (30 foci) that completed evaluation scans (Fig. 3). Correlation was high both between two independent readers as well as repeated measurements by the same reader. The mean percent changes in SUVmax and MTV50 on a patient-by-patient basis (primary tumour) are illustrated in Table 3. A significant difference in percentage change from baseline until final evaluation in tumour FDG-uptake and size was found between responders and non-responders: SUVmax decreased 47% more in responders than in non-
Response evaluation in SCLC by PET/CT
45
Table 1 Cross-tabulation of: (A) early response measured by CT (RECIST)a and PET (visual analysis)b ; (B) early response measured by PET and final response measured by CT; (C) early response and final response measured by CT PET early CMR
PMR
SMD
PMD
Total
(A) Cross-tabulation of early response measured by CT (RECIST)a and PET (visual analysis) CT early CR PR 1 9 1 SD 1 PD Total
1
9
1
1
12
(B) Cross-tabulation of early response measured by PET and final response measured by CT CT final CR 3 PR 1 6 1 SD PD 1c Total
1
9
1
1
PR
SD
PD
0 11 1 0
3 8 0 1 12
CT early CR
Total
(C) Cross-tabulation of early response and final response measured by CT CT final CR 3 PR 7 1 SD PD 1c Total
0
11
1
0
3 8 0 1 12
a
CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. CMR: complete metabolic response, PMR: partiel metabolic response, SMD: stable metabolic response, PMD: progressive metabolic disease. c This patient had PMD on PET and PR on CT, relapsed during the last course of chemotherapy and never received final PET/CT. b
responders (95% C.I. [−84%, −10%], p = 0.04). Neither of the non-responding patients had an early evaluation scan, thus a similar analysis at this time point could not be performed. However, in the early evaluation scans from 12 responding patients an average decline in SUVmax and MTV to approximately 37% and 8% of baseline values could be observed after the first cycle of chemotherapy. At the same time a decline in size at CT to 49% percent of baseline could be observed. A significant correlation was found between results from PET (SUVmax , SUVave and MTV) and results from CT (size in mm) (Spearman’s 0.65, 0.65 and 0.79, respectively, p < 0.001). The performance of visual assessment of PET and a quantitative approach using the EORTC-criteria was compared using kappa and McNemars test. No significant difference in assignment of response categories was found between visual assessment of PET and EORTC, and CT and EORTC respectively. Agreement between visual assessment of PET and EORTC was good ( = 0.73, p = 0.000), between CT and EORTC poor ( = 0.15, p = 0.363). The results were identical whether SUV was corrected for body weight or body surface area.
4. Discussion The purpose of this prospective study was to assess the value of PET/CT in response evaluation after treatment of SCLC. We found that PET and PET/CT can be used for response evaluation after chemotherapy in SCLC: changes in FDGuptake by primary tumour and metastases are significantly correlated to changes in size as measured by CT. Since SCLC is a rapidly proliferating but also, in contrast to NSCLC, rapidly responding cancer this could be expected. PET/CT scan was performed after the first cycle of chemotherapy in 12 patients, major disagreement between early PET and early CT was observed in one patient (8%). Assessment of response after one or two cycles of chemotherapy could lead to an early change in treatment strategy in case of treatment failure. This could potentially improve the treatment of the single patient, but the numbers of effective salvage regimens for SCLC are still scarce. After definitive chemotherapy (six series) incongruence between final response evaluated by PET and CT was seen in eight (42%) of the patients. Major disagreement was observed in two patients (11%), this could have influenced
46
B.M. Fischer et al.
Table 2
Crosstabulation of final response measured by CT (RECIST)a , PET (visual analysis)b and PET/CT PET CMR
PMR
SMD
PMD
Total
(A) Cross-tabulation of final response measured by CT (RECIST)c CT CR 3 1 PR 5 7 1 SD 1 PD 1 0 Total
8
4 13 1 1
9
1
1
19
PR
SD
PD
PET/CT CR
(B) Cross-tabulation of final response measured by PET (visual analysis) CT CR 4 PR 1 11 1 SD 1 PD 1 0 Total
5
12
1
PMR
SMD
Total d
4 13 1 1
1
19
PET CMR
(C) Cross-tabulation of final response measured by PET/CT PET/CT CR 4 1 PR 5 7 SD 1 PD Total
9
8
PMD
Total
e
1
1
5 12 1 1
1
19
a
CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. b CMR: Complete metabolic response, PMR: Partiel metabolic response, SMD: Stable metabolic response, PMD: Progressive metabolic disease. c McNemar-Bowker test (4 × 4 table): 2.667, p = 0.264. Mcnemar test (2 × 2 table): p = 1.00. d McNemar-Bowker test (4 × 4 table): 1.000, p = 0.607. Mcnemar test (2 × 2 table): p = 1.00. e McNemar-Bowker test (4 × 4 table): 2.667, p = 0.102. Mcnemar test (2 × 2 table): p = 1.00.
patient management (responder versus non-responder) e.g. changing decision regarding prophylactic cranial irradiation. When comparing PET/CT and CT overall incongruence was seen only in three patients (16%) causing possible change in patient management in two patients. In spite of these findings no significant difference in response (RECIST) as assigned by CT and response assigned by PET and PET/CT during and after chemotherapy was found. Previous published studies on response evaluation and restaging of patients with PET in SCLC are retrospective [10—12]. Kamel et al. [10] found incongruence in 6/20 patients (30%); with impact on patient management in 3/20 (15%). Blum et al. [11] reported even higher numbers; incongruent results was reported in 63% of the 25 patients (affecting management in 10 patients (40%)). Neither paper reported any level of significance. Both studies were retrospective; thus the majority of the PET scans was
performed on a specific clinical indication. Since PET was not a part of standard staging and response evaluation at either institution only a minority of the patients receiving PET for response evaluation had a baseline PET for comparison: Blum et al. reported that the number of PET-scans with impact on patient management decreased from 40% to 25% when looking only on patients with both baseline and restaging PET. Pandit et al. has suggested an independent prognostic value of PET imaging in SCLC [18]. Our study was too small to perform survival analysis, but 1-year survival rate in the patients responding on early PET and CT was 0.70 and 0.64, respectively. However, regarding cases of discordance between CT and PET at final response assessment (Fig. 2), a simple analysis of the correlation between response category and survival ≥12 months reveals five errors by PET and three errors by CT. At early response evaluation one error
Response evaluation in SCLC by PET/CT
47
Fig. 3
Inter- and intra-observer variation.
by PET and one by CT is observed. Studies on other solid tumours have indicated that PET [19—22] after one or two cycles of chemotherapy can accurately predict response. However, none of the studies reported results from simultaneous response evaluation by CT for comparison. The design of this study was planned in accordance with the recommendations by EORTC from 1999 [14]. On ROI definition we applied a semi automatically threshold method using the maximum pixel value and 50% and 75% isocontour. The threshold method has previously been reported to have high test—retest reproducibility [23] and to be relatively independent of tumour size and geometry [24]. We found excellent inter- and intra-observer repeatability concerning both FDG-uptake and tumour volume using the threshold method. SUV has to be normalized to the body weight (SUVbw ), body surface area (SUVbsa ) or lean body mass (SUVlbm ): The EORTC PET study group recommended that SUV corrected for body surface area should be the minimum standard of measurement. We found that the mean relative change in FDG-uptake from before to after therapy was significantly smaller when measured as SUVbsa compared to SUVbw . This difference of an estimated two percent is small but should however be kept in mind when comparing relative changes in FDG-uptake using different means of normalization. In an attempt to ease comparison between response evaluation by CT and PET; four response categories corresponding to the RECIST criteria has been suggested by EORTC based on relative changes in tumour SUV [14] and by MacManus et al. based on visual analysis [9]. We applied both methods on our data and found no significant difference.
The literature on PET in response evaluation of SCLC is scarce and on PET/CT next to nothing. The sum of our and previous findings indicates that response evaluation by PET and PET/CT is feasible both during and after chemotherapy. Whether this will change patient management or merely serve as another prognostic factor remains to be settled [18,25], and larger prospective studies on PET/CT in SCLC is needed. However, it is our experience that performing a prospective study on a SCLC population is challenging: only 20 of the initially included 30 patients received a response evaluation scan. One explanation is the poor performance status of many patients with SCLC resulting in frequent changes in treatment schedules. This combined with a fully booked PET/CT system makes the logistics in a response evaluation study like this difficult. Furthermore, in order to include a larger number of patients a multicenter study would probably be necessary. Since 1985 the 5-year survival rate for patients with limited disease SCLC in the United States has only increased from 10.5% to 13.3% and an unchanged 14% of the patients do not receive any treatment, thus new and improved treatment strategies are certainly warranted [26]. Current evidence on NSCLC indicates that assessing early sub-clinical response on PET can help optimise treatment of the single patient as well as shortening clinical trials evaluating new anti-cancer agents [6,7]. Even though SCLC is a clinical and histological distinct entity different from NSCLC in many ways, it might be sound to regard the data on PET and SCLC as a reassuring proof of principle and adapt the growing amount of evidence from NSCLC in the planning of SCLC treatment protocols. On the other hand: the lack of salvage therapies and the fact that SCLC very rapidly
48
B.M. Fischer et al.
Table 3 SUVa , MTVb and size on CT of primary tumour as percent of baseline findings after first cycle of chemotherapy (early) and after ended therapy (final)
Earlyc All CR PR
d
n
SUV %baseline
SD
MTV %baseline
SD
CT %baseline
SD
12 3 9
36.63 39.35 35.73
17.46 19.50 17.90
7.84 3.51 9.29
8.94 1.94 9.98
48.80 51.31 47.96
22.39 4.51 26.10
Finale
n
SUV %baseline
SD
MTV %baseline
SD
CT %baseline
SD
All Responder CR PR Non-responder SD PD
19 17 4 13 2 1 1
26.17 21.23f 11.00 24.37 68.20f 69.24 67.15
26.95 23.88 22.01 24.37 1.48
11.55 3.11g 2.79 3.21 79.08g 48.15 110.00
27.28 5.62 5.59 5.88 43.73
34.92 25.56 0.00 33.42 114.52 91.53 137.50
36.91 24.12 0.00 22.15 32.51
a
SUV: standardized uptake value, normalized to body weight. MTV: metabolic tumour volume, mean SUV times tumour volume using 50% isocontour, normalized to body weight. c Result after final PET/CT scan, categorized according to RECIST (CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. d The two patients categorized as non-responders on the final PET/CT scan missed PET/CT after the first cycle of chemotherapy. e Result after final PET/CT scan, categorized according to RECIST (CR: complete response, PR: partiel response, SD: stable disease, PD: progressive disease. f Difference in mean between responders and non-responders: −46.97, 95% C.I [−83.51, −10,42], p = 0.04. g Difference in mean between responders and non-responders: −75.97, 95% C.I [−95.39, −56.55], p = 0.02. b
response to chemotherapy indicates that the therapeutic efficacy [27,28] of PET in SCLC currently is less promising than in NSCLC.
give any definitive conclusion regarding the value of PET/CT in response evaluation of SCLC, thus further studies are needed.
5. Conclusion
References
PET and PET/CT can be used for response evaluation after chemotherapy in SCLC; changes in FDG-uptake by primary tumour and metastases are significantly correlated to changes in size as measured by CT. No significant difference was found between a visual and semi-quantitative analysis of the PET scans. Our results indicate that it is uncertain whether response assessment, early or final, by PET and PET/CT add further benefit to the treatment of patients with SCLC as compared to assessment by CT alone. Giving the low number of participants in this study we cannot
[1] Jackman DM, Johnson BE. Small-cell lung cancer. Lancet 2005;366:1385—96. [2] Simon GR, Wagner H. Small cell lung cancer. Chest 2003; 123:259S—71S. [3] Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer 1981;47:207—14. [4] Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS, et al. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 2000;92:205—16. [5] Weber WA. Use of PET for monitoring cancer therapy and for predicting outcome. J Nucl Med 2005;46:983—95.
Response evaluation in SCLC by PET/CT [6] Vansteenkiste J, Fischer BM, Dooms C, Mortensen J. Positronemission tomography in prognostic and therapeutic assessment of lung cancer: systematic review. Lancet Oncol 2004;5: 531—40. [7] Juweid ME, Cheson BD. Positron-emission tomography and assessment of cancer therapy. N Engl J Med 2006;345(5):496— 507. [8] Higashi K, Clavo AC, Wahl RL. Does FDG uptake measure proliferative activity of human cancer cells? In vitro comparison with DNA flow cytometry and tritiated thymidine uptake. J Nucl Med 1993;34:414—9. [9] MacManus MP, Hicks RJ, Matthews JP, McKenzie A, Rischin D, et al. Positron emission tomography is superior to computed tomography scanning for response-assessment after radical radiotherapy or chemoradiotherapy in patients with non-small-cell lung cancer. J Clin Oncol 2003;21(7): 1285—92. [10] 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(12):1911—7. [11] Blum R, MacManus MP, Rischin D, Michael M, Hicks R. Impact of positron emission tmography on the management of patients with small-cell lung cancer. Am J Clin Oncol 2004;27(2): 164—71. [12] Zhao DS, Valdivia AY, Li Y, Blaufox DM. 18 F-fluorodeoxyglucose positron emission tomography in small-cell lung cancer. Semin Nucl Med 2002;XXXII(4):272—5. [13] Schumacher T, Brink I, Mix M, Reinhardt M, Herget G, Digel W, et al. FDG-PET imaging for the staging and followup of small cell lung cancer. Eur J Nucl Med 2001;28(4): 483—8. [14] Young H, Baum R, Cremerius U, Herholz K, Hoekstra OS, Lammertsma AA, et al. Measurement of clinical and subclinical tumour response using [18F]-fluorodeoxyglucose and positron emission tomography: review and 1999 EORTC recommendations. Eur J Cancer 1999;35(13):1773—82. [15] Altman DG. Some common problems in medical research. In: Practical statistics for medical research. 1st ed. London: Chapman & Hall/CRC; 1991. p. 396—439. [16] Ludbrook J. Statistical techniques for comparing measurers and methods of measurement: a critical review. Clin Exp Pharmacol Physiol 2002;29:527—36.
49 [17] G¨ onen M, Panageas KS, Larson SM. Statistical issues in analysis of diagnostic imaging experiments with multiple observations per patient. Radiology 2001;221:763—7. [18] Pandit N, Gonen M, Krug L, Larson SM. Prognostic value of [18 F]FDG-PET imaging in small cell lung cancer. Eur J Nucl Med 2003;30(1):78—84. [19] Weber WA, Petersen V, Schmidt B, Tyndale-Hines L, Link T, et al. Positron emission tomography in non-small-cell lung cancer: prediction of response to chemotherapy by quantitative asessment of glucose use. J Clin Oncol 2003;21:2651—7. [20] Schelling M, Avril N, N¨ ahrig J, Kuhn W, R¨ omer WSD, Werner M, et al. Positron emission tomography using [18 F]Fluorodeoxyglucose for monitoring primary chemotherapy in breast cancer. J Clin Oncol 2000;18:1689—95. [21] Schwarz JD, Bader M, Jenicke L, Hemminger G, J¨ anicke F, Avril N. early prediction of response to chemotherapy in metastatic breast cancer using sequential 18 F-FDG PET. J Nucl Med 2005;46:1144—50. [22] Avril N, Sassen S, Schmalfeldt B, Naehrig J, Rutke S, Weber WA, et al. Prediction of response to neoadjuvant chemotherapy by sequential F-18-fluorodeoxyglucose positron emission tomography in patients with advanced-stage ovarian cancer. J Clin Oncol 2005;23(30):7445—53. [23] Krak NC, Boellaard R, Hoekstra OS, Twisk JWR, Hoekstra CJ, Lammertsma AA. Effects of ROI definition an reconstruction method on quantitative outcome and applicability in a response monitoring trial. Eur J Nucl Med Mol Imag 2005;32:294—301. [24] Lee JR, Madsen MT, Bushnel D, Menda Y. A threshold method to improve standardized uptake value reproducibility. Nucl Med Commun 2000;21(7):685—90. [25] Markman M. Use of positron emission tomography scans in ovarian cancer: a diagnostic technique in search of an indication. J Clin Oncol 2005;23(30):7385—7. [26] Lee CB, Morris DE, Fried DB, Socinski MA. Current and evolving treatment options for limited stage small cell lung cancer. Curr Opin Oncol 2006;18:162—72. [27] Price P, Laking G. How should we introduce clinical PET in the UK? The oncologists need to have a view. Clin Oncol 2004;16:172—5. [28] Markman M. Use of positron emission tomography scans in ovarian cancer: a diagnostic technique in search of an indication. J Clin Oncol 2005;23(30):7385—6.