Abstract
The aim of this study was to compare the utility of gamma camera using a coincidence detection system imaging (CDET) with 18-fluorodeoxyglucose to conventional imaging techniques in the detection of recurrence of non-small cell lung cancer. Sixty-nine patients were randomized into two groups for follow-up after surgery from October 2000 to December 2002. Each patient was evaluated every 6 months by conventional technique imaging in group A (n=33) or CDET imaging in group B (n=36) over two years. The direct costs of each procedure were evaluated. The major endpoint was the number of recurrences or new tumours detected. The two groups were similar. A total of 25 recurrences was detected (9 in group A and 16 in group B). Overall survival was similar in the two groups. CDET imaging was more expensive. CDET imaging provides earlier detection of recurrence, but does not modify survival outcome. Further studies are necessary to demonstrate the impact, if any, of 18-FDG imaging.
Surgery is the best treatment for early stage non-small cell lung cancer (NSCLC). One of the most important goals during follow-up after curative treatment of NSCLC is detection of recurrence (1). Follow-up of patients with completely resected NSCLC is always controversial (2). Most of the guidelines include recommendations for a post-treatment surveillance program. The recent guidelines incorporate clinical visits and chest computed tomography (CT) scanning during the first two years following curative intent therapy. Before dedicated PET cameras were available, gamma camera with a coincidence detection system was developed and used clinically (3). The detection of a pulmonary lesion with a diameter of at least 2 cm and evaluation of lymph nodes is similar between coincidence detection emission tomography (CDET) and positron-emission tomography (PET) imaging. The aim of this study was to compare the number of recurrences detected after curative treatment of NSCLC using conventional technique imaging versus CDET imaging.
Patients and Methods
Study patients. From October 2000 until December 2002, sixty-nine consecutive patients with curative-intent surgical resection for NSCLC were randomized into two groups. For all patients, preoperative staging included recent CT chest scan, liver ultrasound and bone scintigraphy. CDET initial staging was also performed in 18/33 patients in group A and in 20/36 patients in group B. For each patient, the initial decision for surgery was recommended by a consensus of the multidisciplinary board of lung cancer specialists of our institution. All the operations were performed in the same department by two experienced thoracic surgeons. Eligibility criteria included histologically confirmed resected NSCLC. The following exclusion criteria were applied: lung cancer other than NSLCC stages I, II or IIIA; history of previous malignancy; incomplete pulmonary resection; uncontrolled diabetes and major side-effect or death within 30 days of surgery. Oral and informed written consent were obtained for all patients. This study was approved by our local Ethics Committee. This study was also registered as clinical trial: NCT 00199615
Study design. The follow-up period was a minimum postoperative period of 2 years. On the first clinical visit one month after surgery, all the patients were blindly randomized between two follow-up procedures. In group A (n=33), chest CT scan with liver and adrenal gland section, abdominal ultrasonography, and bone scintigraphy (only if bone symptoms) were performed. In group B (n=36), CDET was performed. In the two groups, brain CT was performed at each imaging evaluation. These two surveillance programs were systematically performed at 6, 12, 18 and 24 months or earlier if recurrence was suspected. Recurrences were identified through one or more of the following: chest CT scan, bone scintigraphy, abdominal ultrasound, brain CT scan or magnetic resonance imaging (MRI). Confirmation of the diagnosis was made by pathology or cytology when clinically feasible, except for patients with multiple brain or bone metastases. Spiral CT examinations were performed using high-speed acquisition (Lightspeed, GE Medical Systems, Waukesha, WI, USA). CDET was performed using a dual-detector gamma camera (Axis; Phillips Medical Systems, Cleveland, OH, USA) equipped with a 19-mm sodium iodine crystal with septa operating in coincidence mode for acquisition. Serum glucose was measured before radiotracer injection. CDET was performed 60 minutes post 18F-fluorodeoxyglucose (18-FDG) injection, as previously reported (4). All coronal, transverse and sagittal slices were analyzed by visual interpretation after iterative reconstructions of CDET data (EMML, 20 iterations). CT and CDET data were reviewed by two experimented radiologist and nuclear physicians.
Costs of diagnostic procedures, one-day hospitalization and transportation between home and hospital.
Outcomes of study. The major end-point of the study was the number of recurrences or new tumours detected in symptomatic or asymptomatic patients. Local recurrence was defined by the reappearance of the same lung cancer in the hilar or mediastinal nodes and/or in the ipsilateral lung. Recurrence at any other site was noted as distant. A lesion of different histology was considered as a new primary tumour (1). The length of surgical follow-up, disease free survival and overall survival were calculated from the day of operation. The end of the disease-free interval was the day of clinical examination for symptoms, or the day of appearance of imaging findings consistent with cancer recurrence or new tumour.
Cost analysis. The perspective of the study was the French payer. The analysis included only direct medical costs for each patient included in both groups A and B. Reimbursement prices were determined for each procedure by the French Healthcare system using 2002 repayment tariffs. For each patient, imaging procedure, fixed hospital and patient transportation costs were collected. The cost of each diagnostic procedure is reported in Table I. For each patient, the transportation cost was calculated as a function of the distance between the patient's home and the hospital, according to specific area tariffs (Table I). Extra restaging, new hospitalization and new treatment costs were not considered in this study.
Statistical analysis. The length of the follow-up period was a minimum postoperative period of 2 years; follow-up also terminated at death. For this study, follow-up was closed on December 31, 2006. The quantitative variables were compared between the two groups using an unpaired Student's t-test. For the qualitative variables, the comparisons were made between the two groups using Chi-squared test or Fisher's test according to the sample size. The time between the date of operation and date of relapse was compared by a Mann-Whitney test. Overall survival was also defined as the time from operation to death or date of the last follow-up visit. A patient was considered disease-free until the detection of a recurrence or a new tumour. Survival was estimated by the Kaplan-Meier method. Cost evaluation of each follow-up strategy was analysed by the bootstrap method (5). Statistical significance was obtained when the p-value was less than 0.05. The software used was Statview 5.0 and SAS 9.13 (SAS Institute, Cary, USA) as well as Stata 8.0 (Stata Inc, Lakeway Drive, TX, USA).
Results
The study group included 69 patients surgically treated for NSCLC. For initial staging, CDET imaging was performed in only 18/33 patients in group A and 20/36 patients in group B. FDG uptake was detected in all the primary tumours imaged. Treatment with apparent curative results was recorded for 63 patients with stage I or II disease, and for 6 patients with stage IIIA disease (Table II). The sample included 33 patients in group A, and 36 patients in group B, with an overall median age of 62 years (range 42 to 82 years). The histological type and pathological stage of the tumours and surgical procedures are shown in Table II. Different surgical procedures were performed depending on the tumour location and comorbidity of patients. The median length of follow-up was 29 months (range 9 to 36 months) for group A and 25 months (range 5 to 36 months) for group B. Initial therapy was surgical resection in all patients, including neoadjuvant chemotherapy in 2 cases (stage IIIA, group B). For stage IIA or IIB patients, 10 had adjuvant chemotherapy, 4 in group A and 6 in group B. Radiotherapy was also performed in 4 patients (2 in each group). All the patients of stage IIIA had adjuvant chemotherapy and radiotherapy after surgery. A recurrence or new tumour was found in 9 patients from group A and 16 patients from group B. This difference was not significant (p=0.14). The diagnosis of recurrence was established by pathological biopsy (n=11), cytology (n=3), or subsequent imaging progression (n=10). Most of the recurrences occurred during the first two years following surgery (88%). The localization of recurrence is shown in Table III and the distribution by stage is noted in Table IV. In the conventional survey group, most of the recurrences were found in patients with clinical symptoms (Table V). In contrast, in the functional imaging group, recurrences were detected in asymptomatic patients (11 out of 16, p<0.01). Time to the detection of recurrence was 18±11.8 months in group A versus 12±9.9 months in group B. Two new malignancies were found: one head and neck tumour in group A and one second primary lung cancer with different histological diagnosis between the first and second tumour in group B (Figure 1). The patient with head and neck tumour had been treated by chemotherapy and radiotherapy. For the case with new lung tumour, chemotherapy following by erlonitib treatment was used for bronchoalveolar carcinoma. All the other patients received non-surgical treatment: chemotherapy and/or radiotherapy, except one patient in group B with adrenal gland recurrence (Figure 2). For this patient, treatment consisted of surgery followed by chemotherapy.
Study population characteristics.
Recurrences by site in the two groups.
Recurrences by stage in the two groups.
Mode of detection of recurrences.
At the end of this study of patients with recurrence or new tumour, 2 were alive and 7 dead in group A, and 3 alive and 13 dead in group B. During the follow-up period, death occurred as a result of non-pulmonary causes in 4 patients (3 in group A and 1 in group B): due to cardiac dysfunctions, pulmonary embolism. No significant statistical difference was observed for overall survival between the two groups: 29±17.1 months for group A, and 26.5±19.6 months for group B. In addition, comparing only patients with recurrence in the two groups, overall survival was again not significantly different (29.3±13.7 months for group A and 19.5±15.9 months for group B).
Cost analysis. The mean number of visits by each patient was 3.7±1.7 in group A and 4.2±1.9 in group B. The average cost (range) of follow-up visits and imaging was 755.47 € (640-864 €) per patient in group A, versus 1104.96 € (954-1240 €) in group B. The cost difference between each strategy was 349.48 € (p<0.001).
Discussion
The optimal follow-up of patients with resected NSCLC is controversial (6-9). Physical examination plus chest X-ray are recommended every 3 months in the first 2 years, then every 6 months for 3-5 years (10). The role of chest CT is a matter of controversy (11). Korst et al. have studied the role of CT for surveillance in patients with completely resected lung cancer (2). Thoracic surgeons detected 27 recurrences in 92 patients with CT abnormalities. But many negative investigations have been generated by CT follow-up. In a prospective study, Westeel et al. have reported that recurrences were more frequently detected by thoracic CT and fiberoptic bronchoscopy than by clinical examination and chest X-rays in patients after surgery for NSCLC (12). Such intensive follow-up improves survival in detecting recurrence in asymptomatic patients. A large prospective follow-up study in 1,073 patients have reported that treatment for recurrent NSCLC after curative surgical resection prolongs survival (13).
A: A 68-year-old man underwent a lobectomy of the right upper lobe: squamous carcinoma, pT2N0M0 (grey arrow). B: Two years later, large cell carcinoma of the left upper lobe was diagnosed: cT2N2M0, chemotherapy and radiotherapy (grey arrow).
CDET imaging revealed FDG fixation in the left adrenal gland (black arrow) of a 56-year-old man who had undergone a lobectomy of the right upper lobe one year previously of surgery confirmed metastatic disease. The grey arrow shows persistent apical pneumothorax following surgery.
Previous studies have reported that 18-FDG imaging appears to be more accurate than CT at detecting recurrent or new tumour in patients following curative treatment for NSCLC (14). Dual-head gamma camera modified for coincidence detection has been used for oncology 18-FDG imaging in some institutions (15). The sensitivity and specificity of CDET was only similar to dedicated-PET scanner for pulmonary lesions with a diameter of at least 2 cm (16).
In our institution, from 1999 to 2005, CDET imaging was usually performed for the diagnosis of thoracic malignancy (17); PET/CT replaced CDET in 2005. This study is the result of our experience concerning 18-FDG-CDET imaging in our institution, focusing on 2 years, of a prospective follow-up study. In our work, 18-FDG-CDET allowed early detection of recurrence in 11 asymptomatic patients. In contrast, CT and other conventional procedures detected recurrence only in one asymptomatic patient. In the present study, no patient had further surgery and only one patient had curative treatment for isolated adrenal gland metastasis.
A decade ago, Bury et al. (14) studied the usefulness of 18-FDG-PET for the detection of residual or recurrent disease in 126 patients from stage I to IIIB; for patients with apparent curative treatment (stage I to IIIA), 13 recurrent tumours were confirmed in 7 asymptomatic and 6 symptomatic patients. However, in 2009, recurrences are now detected at a smaller size. Latest generation PET/CT devices are more suitable than CDET for detecting lesions smaller than 1 cm. PET/CT improved the diagnosis of lung cancer recurrence in 24 out of 27 patients with suspected lesions by fusion of functional and anatomic data imaging, as reported by Keidar et al. (18). Hellwig et al. studied the accuracy of FDG-PET in detecting recurrent lung cancer in 62 patients. FDG-PET correctly identified 51 out of 55 relapses and helped in the selection of patients for second surgical treatment (19). Interestingly, Hicks et al. demonstrated that PET is an accurate technique for detecting recurrent disease in NSCLC after potentially curative treatment (20). In addition, Takenada et al. reported that integrated FDG-PET/CT is comparable to conventional imaging in detecting postoperative recurrence in NSCLC patients (21).
Follow-up is acceptable if it is cost-effective (6). Using a Markow model, surveillance was shown to be cost-effective in patients with previously resected stage IA NSCLC (22). In a large series of 563 patients, regular follow-up including clinical visit and X-rays was proposed for ten years in patients with curatively resected NSCLC (6). A recurrence or new tumour was detected in 239 patients with more than 70% of recurrences occuring during the first two years. Only 23 patients were surgically treated for metachronous NSCLC. As a consequence of high costs and limited benefits, a follow-up program every 6 months for the first 5 years was proposed.
PET imaging is essentially cost-effective in reducing futile surgery for NSCLC (23, 24). In our study, CDET imaging increased the follow-up cost. There were some limitations of our study. Firstly, not all patients had initial staging by FDG imaging before surgery, as occurs today. But recently, Gauger et al. studied the sites and times of recurrence in patients with clinical stage I NSCLC (25). All the patients had PET with conventional imaging before surgery; 40 out of 194 patients (20%) developed recurrent NSCLC. Initial PET did not detect minimal node involvement. Secondly, the sample size of the studied population is small and does not allow us to conclude that the use of CDET does not modify survival. Lack of detection of improved survival could easily be explained by the small cohort size here.
Our initial study had been stopped by the acquisition of a dedicated PET/CT system. The cost of the new strategy was acceptable using CDET technology, but PET or PET/CT imaging significantly increased the costs of the procedure.
In conclusion, 18-FDG imaging allows the earlier detection of recurrent or new NSLC in patients with curative treatment but without modifying survival significantly. On the other hand, 18-FDG imaging increases the cost of follow-up but does allow active treatment of some cases of recurrence. Prospective studies are necessary to select patients at high risk of developing recurrence or new tumour after NSCLC curative treatment and to select patients who will most likely benefit from second-line treatment.
Acknowledgements
The Authors gratefully acknowledge the assistance of E. Champagne in collecting data. This work was supported by a grant from the University Hospital of Limoges (local clinical research program, 2000).
Footnotes
-
Conflict of Interest Statement
None declared.
- Received November 18, 2009.
- Revision received May 18, 2010.
- Accepted May 24, 2010.
- Copyright© 2010 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
