Accurate tumor staging is essential for choosing the

Transcription

1 Detection of Extrathoracic Metastases by Positron Emission Tomography in Lung Cancer Walter Weder, MD, Ralph A. Schmid, MD, Helke Bruchhaus, MD, Sven Hillinger, MD, Gustav K. von Schulthess, MD, and Hans C. Steinert, MD Department of Surgery and Division of Nuclear Medicine, Department of Radiology, University Hospital, Zürich, Switzerland Background. Accurate staging of non small cell lung cancer is essential for treatment planning. We evaluated in a prospective study the role of whole-body 2-[ 18 F]fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET) in mediastinal nodal staging with a positive predictive value of 96%. The study was continued to further evaluate the value of whole-body FDG PET in detecting unexpected extrathoracic metastases (ETMs) in patients qualifying for surgical treatment by conventional staging. Methods. One hundred patients underwent clinical evaluation, chest and upper abdominal computed tomography scan, mediastinoscopy (lymph nodes greater than 1 cm on computed tomography), and routine laboratory tests. In 94 patients with stage IIIa or less and 6 with suspected N3 a whole-body FDG PET was performed. If clinical signs of ETMs were present additional diagnostic methods were applied. All findings in the FDG PET were confirmed histologically or radiologically. Results. Unexpected ETMs were detected in 13 (14%) of 94 patients (stage IIIa or less) at 14 sites. In addition 6 of 94 patients were restaged up to N3 after PET. The suspected N3 disease (stage IIIb) on computed tomography was confirmed by PET in all 6 patients. There was no false positive finding of ETM. Weight loss was correlated with the occurrence of ETM: more than 5 kg, 5 of 13 patients (38%); more than 10 kg, 4 of 6 patients (67%). Pathologic laboratory findings were not predictive for ETM. Conclusions. Whole-body FDG PET improves detection of ETMs in patients with non small cell lung cancer otherwise elegible for operation. In 14% of patients (stage IIIa or less), ETMs were detected, and in total, 20% of the patients were understaged. (Ann Thorac Surg 1998;66:886 93) 1998 by The Society of Thoracic Surgeons Accurate tumor staging is essential for choosing the appropriate treatment strategy of cancer in general and of bronchogenic carcinoma in particular. In non small cell lung cancer (NSCLC) the treatment regimen depends on preoperative staging. Curable surgical resection is possible for early stages of bronchial carcinoma (stage IIIa or less, stage T3 N2 M0 or less). Despite radical surgical treatment the overall 5-year survival rate remains low (20% to 40%). One reason is undetected extrathoracic metastases (ETMs), which cause underestimation of the tumor stage. Current modalities used in preoperative staging of NSCLC include noninvasive imaging techniques, such as chest roentgenography, computed tomography (CT), and if clinically indicated, magnetic resonance imaging (MRI), bone scintigraphy, and ultrasonography, as well as bronchoscopy and mediastinoscopy. Positron emission tomography (PET) is a wellestablished imaging technique, which does not depict the anatomic morphology as well as CT and MRI, but detects local differences in tissue metabolism. This principle is used to examine tissue types with high metabolism, eg, brain, heart, and malignant tumors. Presented at the Thirty-fourth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 26 28, Address reprint requests to Dr Weder, Department of Surgery, University Hospital, Rämistr 100, CH-8091 Zürich, Switzerland. Warburg [1] described in 1930 the high glucose metabolism of malignant tumors. With the radioactive labeled glucose analogue 2-[ 18 F]fluoro-2-deoxy-d-glucose (FDG) it is possible to visualize glucose metabolism in vivo. Therefore, FDG PET can be used effectively for differentiation of malignant and nonmalignant tissue. Especially in lung cancer, excellent FDG uptake has been described previously [2, 3]. In our previous study N staging in NSCLC with FDG PET proved to be highly accurate in 96% of cases with a sensivity of 89% and a specifity of 99% [4]. The positive predictive value was 96%, and the negative predictive value was 97%. We reviewed the data and expanded our prospective study to determine the value of FDG PET in preoperative M staging in NSCLC. Furthermore, we evaluated whether a positive correlation between patient data (history, physical examination, tumor characteristics, histology, and N staging) and metastatic disease could be found to facilitate selection of candidates for operation by FDG PET. Material and Methods Patients Between February 1994 and September 1997 we performed FDG PET in 107 patients with NSCLC diagnosed by cytology or histology. Only patients who qualified for 1998 by The Society of Thoracic Surgeons /98/$19.00 Published by Elsevier Science Inc PII S (98)

2 Ann Thorac Surg WEDER ET AL 1998;66: POSITRON EMISSION TOMOGRAPHY IN LUNG CANCER 887 surgical therapy were included in the study. All patients who underwent neoadjuvant therapy were excluded. In 2 patients definitive histologic diagnosis revealed small cell lung cancer, and 5 patients were excluded because of incomplete data. Preoperative staging included history, physical examination, and blood tests as well as chest x-ray films in two planes and CT scan of the thorax and upper abdomen. Clinical signs for metastases were further investigated with bone scintigraphy, MRI of the brain, or other diagnostic modalities. Enlarged mediastinal lymph nodes ( 1 cm) on CT scan were further evaluated by mediastinoscopy. Therefore, data from 100 patients were used for analysis. Ninety-four patients were in stage IIIa or less and 6 were in suspected stage IIIb before the whole-body FDG PET. These included 80 men aged 39 to 80 years (mean, 61 years) and 20 women aged 41 to 79 years (mean, 58 years). Positron Emission Tomography Imaging Whole-body FDG PET scanning was performed with a commercially available instrument (Advance; GE Medical Systems, Milwaukee, WI) with the whole-body mode implemented as standard software. The evaluation extended from head to pelvis. Emission and transmission scans were obtained with an acquisition time of 6 minutes per field of view. The total examination time per patient was about 1 hour. To suppress myocardial glucose utilization, patients were asked to fast for at least 4 hours before undergoing whole-body FDG PET. The glucose analogue FDG initially was produced at an outlying facility and later in our own radiopharmaceutical laboratory by using techniques described before [5]. Transmission scanning began immediately after administration of 300 to 400 MBq of FDG. Emission scanning followed at 40 minutes after administration of FDG. The acquired data were reconstructed by using standard back-projection techniques in axial sections and then were reformatted into coronal and sagittal views. The scans were documented using a color laser printer with a digital interface (Agfa Multimedia 315; A&S Kopiersysteme, Münster, Germany). Because the primary substrate for brain metabolism is glucose, which supplies more than 95% of brain requirements, we used FDG uptake in the brain as the reference uptake value [6]. A lesion was defined as a focus of increased FDG uptake greater than the intensity of the surrounding activity, excluding the renal pelvis, urinary bladder, and myocardium, if present. A lesion with intense FDG uptake comparable to the physiologically high FDG uptake in the brain and additionally demonstrating nodular appearance was defined as malignant. Each PET finding fulfilling the previously mentioned properties was confirmed as malignant either histologically, by another imaging method (ultrasonography, bone scan, CT, or MRI), or by follow-up. The PET scans were interpreted independently and prospectively by an experienced nuclear physician without knowledge of any histopathologic or radiologic data. For all patients exact histopathologic diagnosis was assessed. Correlation of Positron Emission Tomography With Clinical Findings and Laboratory Data Case histories of all patients were reviewed for indicators of metastases (weight loss, pain, enlarged lymph nodes, or neurologic deficiencies). Radiologic data (preoperative chest roentgenogram and CT scan of the chest and upper abdomen) were examined for bone, liver, and pararenal metastases. In addition all pathologic laboratory parameters were registered. The staging was done according to the American Thoracic Society TMN system [7] and the staging classifications of the American Joint Committee [8]. Results Histologic Examination of the Primary Tumor Of the 100 patients with NSCLC, 53 had squamous cell carcinoma, 38 had adenocarcinoma, and 9 had large cell carcinoma. In women the distribution was as follows: 5 had squamous cell carcinoma (25% of female patients), 14 had adenocarcinoma (70%), and 1 had large cell carcinoma (5%). In men 48 squamous cell carcinomas (60% of male patients), 24 adenocarcinomas (30%), and 8 large cell carcinomas (10%) were found. Extrathoracic Metastases Unexpected ETM were detected at 14 sites in 13 (14%) of the patients, who were stage IIIa or less before PET (Figs 1 4). In all 6 patients with suspected N3 (stage IIIb) by conventional staging, metastatic disease was confirmed by PET. Thus, with respect to the entire study population, in 19 patients ETMs were suspected with the whole-body PET in 24 locations (13 bone, 3 liver, 4 adrenal gland, 3 supraclavicular lymph nodes, and 1 central nervous system). In 15 patients metastatic disease was confirmed by other methods in at least one site. In 7 patients it was confirmed histologically and in 8 patients with conventional imaging methods (Table 1). In 4 patients positive extrathoracic findings were suspected with whole-body PET but not confirmed with conventional radiologic or clinical methods at the time of first clinical workup. In 2 patients bone metastases at the site of the PET finding were confirmed within 6 months. In 1 patient with N3 disease a suspected metastasis in the adrenal gland found with PET was not further investigated, and in 1 patient no follow-up data were available. There was no false positive finding for ETM. No correlation of age with the occurrence of ETM was found (with ETM, 58 years; range, 39 to 79 years; no ETM, 62 years; range, 41 to 80 years). Correlation of Positron Emission Tomography With Histologic Findings and N and M Staging The correlation of histologic type and metastatic disease demonstrated that patients with primary adenocarcinoma of the lung had the highest frequency of ETMs (Table 2). Of the 100 patients, 69 had an N0/N1 stage, 19 had an

3 888 WEDER ET AL Ann Thorac Surg POSITRON EMISSION TOMOGRAPHY IN LUNG CANCER 1998;66: Fig 1. (A) A 73-year-old patient with squamous cell carcinoma. The 2-[ 18 F]fluoro-2-deoxy-d-glucose positron emission tomography scan on sagittal view demonstrates the metastasis as an enhancement dorsal of the first lumbar vertebra. (B) Corresponding skeletal scintigraphy. An enhancement is noted on the right side but with the characteristics of degenerative alterations. The metastasis can not be detected. (C) Corresponding computed tomography scan of the first lumbar vertebra, performed after positron emission tomography demonstrated a metastasis in the base of the right processus spinosus. The metastasis was confirmed histologically. N2 stage, and 12 had an N3 stage. In 6 patients the N3 disease was suspected on CT before the PET scan and confirmed by mediastinoscopy. N staging was histologically confirmed in 92 patients. In 8 patients no further mediastinal investigation was performed because of metastatic disease. None of these patients had evidence of inflammatory lymphadenopathy. Of the 69 patients conventionally staged as N0/N1, 6 (9%) had a positive ETM finding in PET. Seven (28%) of the 25 patients with N2 stage and 6 (100%) of the 6 with stage N3 were positive for ETMs. Of importance is the finding that 13 (14%) patients of the 94 with stage IIIa or less (diagnosed according to the described workup) had unexpected ETMs detected by PET. In addition in these 94 patients, previously unknown N3 disease was detected in 6 patients (6.4%). All of them had mediastinal lymph nodes less than 1 cm in the chest CT scan. In conclusion, of 94 patients who qualified for radical surgical treatment (stage IIIa or less) with the applied preoperative staging methods, on the basis of the new findings in whole-body PET imaging 19 (20%) patients were upstaged to stage IIIb or IV (accordingly N3 Mx or Nx M1) (Table 3). Correlation of Positron Emission Tomography With Clinical Findings and Laboratory Data Eight (10%) of 81 patients in M0 stage and 5 (33%) of 19 patients in M1 stage had weight loss more than 5 kg. Two (2%) of 81 patients with M0 had a weight loss of more than 10 kg; in contrast, 4 (21%) of 19 patients with M1 stage lost more than 10 kg.

4 Ann Thorac Surg WEDER ET AL 1998;66: POSITRON EMISSION TOMOGRAPHY IN LUNG CANCER 889 Fig 2. (A) A 51-year-old patient with squamous cell carcinoma. Conventional preoperative staging (chest x-ray film in two planes, chest computed tomography scan) was negative for metastatic disease. 2-[ 18 F]fluoro-2-deoxy-d-glucose positron emission tomography shows a metastasis of the lower thoracic vertebral column. (B) Corresponding computed tomographic scan of the thoracic vertebral column. The metastasis was confirmed histologically. Analysis of pathologic laboratory findings were not predictive for ETM (hemoglobin, hematocrit, glutamicoxaloacetic transaminase, glutamic-pyruvate transaminase, -glutamyl transferase, alkaline phosphatase, C-reactive protein). Forty-four (54%) patients with M0 and 12 (63%) of those with M1 showed pathologic laboratory findings. Thirty-two (73%) of those with M0 had an elevated C-reactive protein level, as did 9 (75%) with M1. Hemoglobin and hematocrit were reduced in 22 (50%) of the M0 patients, whereas they were reduced in 9 (75%) with M1. Elevated liver enzymes were found in 6 (14%) patients without metastatic disease, but only in 1 (8%) who was staged M1. The frequency of an increased alkaline phosphatase level was similar in both groups (7 [16%] in M0 and 2 [17%] in M1) (Table 4). Comment Our prospective study in patients with NSCLC was originally designed to evaluate the role of FDG PET for mediastinal N staging. Use of FDG PET for mediastinal N staging is limited with respect to the exact anatomic Fig 3. (A) A 43-year-old patient with adenocarcinoma in the right upper lobe. 2-[ 18 F]fluoro-2-deoxy-d-glucose positron emission tomography demonstrates a rib metastasis on the right side lateral to the liver, which was interpreted as a rib fracture in the chest x-ray film, skeletal scintigraphy, and computed tomographic scan. The coronal view shows the primary tumor and the rib metastasis. (B) Corresponding computed tomographic scan with enhancement in the rib, interpreted as rib fracture. The metastasis was confirmed histologically.

5 890 WEDER ET AL Ann Thorac Surg POSITRON EMISSION TOMOGRAPHY IN LUNG CANCER 1998;66: Table 2. Comparison of Histology With M Staging History n M0 M1 Proved M1 Suspected Squamous (89%) 6 (11%) 0 (0%) Adeno (71%) 8 (21%) 3 (8%) Large cell 9 7 (78%) 1 (11%) 1 (11%) Total Adeno adenocarcinoma; Large cell large cell carcinoma; Squamous squamous cell carcinoma. Fig 4. A 55-year-old patient with squamous cell carcinoma. The positron emission tomography imaging (coronal view) shows detection of a retroclavicular metastasis. The diagnosis was confirmed histologically. Physiologic enhancement of 2-[ 18 F]fluoro-2-deoxy-dglucose uptake in brain and myocardium is observed. differentiation of N1/N2 in the hilar region. However, N3 stage is detected reliably. In this previously published study a high accuracy in detecting malignant lesions in the mediastinum could be achieved. This might be related to our highly qualified team of nuclear physicians and a low rate of inflammatory disease in our study population. However, mediastinal staging by mediastinoscopy is still the method of choice with the highest accuracy. In addition it provides histologic confirmation of lymph node metastases and allows differentiation from nodal enlargement caused by reactive inflammatory hyperplasia (eg, poststenotic pneumonia) and other nonmalignant disease (eg, sarcoidosis, histiocytosis). Moreover, it is difficult to distinguish hilar (N1) from mediastinal (N2) disease at the tracheobronchial angle by PET. However, the continuous evaluation of the data made clear that the role of FDG PET in detecting ETMs is at least equally important. Of the 94 patients (stage IIIa or less) with NSCLC enrolled in this study, 13 patients (14%) demonstrated metastatic findings in the whole-body FDG PET. This is in accordance with the study of Valk and associates [9], which described a detection of previously unsuspected distant metastasis in 11 (11%) of 99 patients with NSCLC. In 15 patients ETMs were confirmed histologically or radiologically in at least one location. In 4 patients positive findings by PET were not confirmed at the time of diagnosis. But in 2 of these patients ETMs became clinically manifest during follow-up. These results indicate that whole-body FDG PET is a highly sensitive method for detection of unknown ETMs in NSCLC. Similar to previous studies [10, 11], 6 patients without enlarged mediastinal lymph nodes were upstaged to N3 after whole-body PET and confirmed histologically. Table 1. Data From 15 Patients With Proved Extrathoracic Metastases Patient No. Age (y) Sex Tumor Size (cm) Histology N Stage M1 in PET Reference Method Weight Loss 1 71 M 1.1 Adeno N3 Cervical lymph nodes Ultrasonography 10 kg in 12 mo 2 51 M 4.0 Squamous N2 Thoracic vertebra Histology NAD 3 64 M 6.0 Squamous N3 Liver and adrenal gland Liver histology, adrenal 6kgin5mo gland suspected 4 58 F 3.0 Adeno N3 Adrenal gland Ultrasonography NAD 5 55 M 3.2 Squamous N2 Cervical lymph node Histology NAD 6 57 M 4.0 Squamous N0 Rib Spotfilm radiograph NAD 7 43 M 3.5 Adeno N0 Rib Histology NAD 8 39 M 2.0 Large cell N3 Liver and pelvis Liver histology, bone scan 14 kg in 6 mo 9 46 F 4.5 Adeno N2 Cervical vertebra and rib Bone scan NAD M 2.5 Adeno N2 Thoracic vertebra MRI and bone scan NAD M 4.0 Squamous N1 Lumbar vertebra Histology 10 kg in 12 mo M 4.5 Adeno N0 Adrenal gland Histology NAD M 3.0 Adeno N3 CNS and skeletal CNS CT, skeletal suspected 14 kg in 6 mo F 3.0 Squamous N0 Liver CT and ultrasonography NAD F 6.3 Adeno N3 Cervical lymph node and rib Clinical findings NAD Adeno adenocarcinoma; CT computed tomography; F female; Large cell large cell carcinoma; M male; MRI magnetic resonance imaging; NAD nothing abnormal detected; PET positron emission tomography; Squamous squamous cell carcinoma.

6 Ann Thorac Surg WEDER ET AL 1998;66: POSITRON EMISSION TOMOGRAPHY IN LUNG CANCER 891 Conventional imaging methods for tumor staging include CT scan and MRI, ultrasonography, and bone scintigraphy. Ultrasonography, MRI, and CT are applied for the examination of a selected anatomic region. Bone scan is a sensitive but rather unspecific method and furthermore is limited to the specific organ system. In contrast, whole-body PET imaging with FDG allows scanning the entire body including all organ systems. Although the incidence of metastatic disease in NSCLC is high, routine staging of all patients with skeletal scintigraphy and CT scan of the head and the abdomen is not routinely applied, because the likelihood of a true positive finding in bone or CT scan in an asymptomatic patient is small [10]. Using skeletal scintigraphy as a routine diagnostic method, only 14% of the positive solitary findings in patients with cancer were caused by a metastasis of the known tumor [12]. Also, the metaanalysis by Silvestri and coworkers [13] demonstrated that the negative predictive value for detecting metastatic disease in the CT scan in patients with NSCLC is not higher than the clinical evaluation. Based on this experience, in current clinical practice preoperative staging in bronchogenic carcinoma is symptom based. However, in our study skeletal metastases were detected with the whole-body FDG PET in 13 patients. In 10 of these patients skeletal scintigraphy was not performed previously because of lack of clinical symptoms. In contrast to conventional imaging modalities the rate of positive findings for M1 in our series is surprisingly high. Moreover, the patients underwent FDG PET without clearly defined clinical symptoms of metastases. Whole-body FDG PET allows screening of the entire body for metastatic disease with a single examination and can detect even small, asymptomatic metastases (4 to 5 mm). In 3 of our patients a metastasis in cervical or supraclavicular lymph nodes was detected by PET imaging, of whom 2 had a completely inconspicuous physical examination even with the knowledge of the PET finding. This illustrates the high resolution power of a modern PET scanner of about 4 to 5 mm. Also in the ability to distinguish solitary benign pulmonary nodules from malignant tumors FDG PET has proved to be highly sensitive (93%) and specific (88%) [14, Table 3. Staging of Patients Before and After PET Imaging a Staging Before PET After PET N0/N1 M N2 M N3 M0 6 b 6 N0/N1 M N2 M N3 M a N and M staging of all 100 patients enrolled in the study before and after PET imaging. Before whole-body PET imaging only 6 patients were not eligible for surgical treatment, whereas after PET scan 13 patients were not resectable because of newly detected stage M1. b Suspected on computed tomographic scan, proved by mediastinoscopy. PET positron emission tomography. Table 4. Laboratory Data Laboratory data M0 M1 CRP increase 73% 75% Hb/Hct decrease 50% 75% GOT/GPT increase 14% 8% ALP increase 17% 17% ALP alkaline phosphatase; CRP C-reactive protein; Hb hemoglobin; Hct hematocrit; GOT/GPT glutamic-oxaloacetic transaminase/glutamic-pyruvate transaminase. 15]. However, increased FDG uptake is of course not specific for malignant tissue and is seen in infectious and inflammatory processes [15, 16]. In our series only patients with confirmed diagnosis of NSCLC were enrolled in the study. All additional infectious sites detected in our series could be differentiated from malignant lesions. Previous studies comparing the specifity of FDG PET and CT scan to detect metastases in the liver and the adrenal gland demonstrated that the FDG PET has a higher sensitivity and specifity than CT [9, 17]. Our results indicate that patients with M1 disease can be diagnosed effectively with the whole-body FDG PET. Therefore, a negative PET finding for M and N3 stage indicates a high positive predictive value for a curable resection. This could lead to the conclusion that PET would be the ideal screening method to detect metastatic disease in NSCLC or other thoracic tumors [18]; however, in current clinical practice the availability and cost of the equipment limits the application of this method. The use of PET for the evaluation of all patients with NSCLC is currently not feasible for most centers. Based on the experience in our studies, we currently use FDG PET as an additional tool to exclude patients with M1 disease in patients with borderline indications from the physiologic standpoint (limited pulmonary function, coronary artery disease), and in patients with suspected advanced disease such as severe weight loss, locally advanced disease, solitary brain metastases, and local recurrence that might qualify for reoperation. The aim is to avoid extensive or futile surgical resections in these high-risk groups. How these guidelines influence the treatment strategy and costs, or whether whole-body PET should be applied as a screening tool for metastases in all lung cancer patients, needs further investigation. In conclusion, our results indicate that whole-body FDG PET is an excellent method to detect unknown ETMs in NSCLC, as 13 (14%) of 94 patients selected to undergo operation with conventional methods were not eligible for radical surgical treatment after PET imaging because of newly detected stage M1 disease. References 1. Warburg O. The metabolism of tumors. New York: Smith, 1931; Nolop KB, Rhodes CG, Brudin LH, et al. Glucose utilization in vivo by human pulmonary neoplasms. Cancer 1987;60:

7 892 WEDER ET AL Ann Thorac Surg POSITRON EMISSION TOMOGRAPHY IN LUNG CANCER 1998;66: Lowe VJ, DeLong DM, Hoffman JM, Coleman RE. Optimum scanning protocol for FDG-PET evaluation of pulmonary malignancy. J Nucl Med 1995;36: Steinert HC, Hauser M, Allemann F, et al. Non small cell lung cancer: nodal staging with FDG PET versus CT with correlative lymph node mapping and sampling. Radiology 1997;202: Hamacher K, Coenen HH, Stocklin G. Efficient stereospecific synthesis of no-carrier-added 2-[ 18 F]-fluoro-2-deoxy-dglucose using amino-polyether supported nucleophilic substitution. J Nucl Med 1986;27: Engel H, Steinert H, Buck A, Berthold T, Huch Böni RA, von Schulthess GK. Whole-body PET: physiological and artifactual fluorodeoxyglucose accumulations. J Nucl Med 1996;37: Tisi GM, Friedman PJ, Peters EM, et al. American Thoracic Society: clinical staging of primary lung cancer. Am Rev Respir Dis 1983;127: Friedman PJ. Lung cancer: update on staging classifications. Am J Radiol 1988;150: Valk PE, Pounds TR, Hopkins DM, et al. Staging non small cell lung cancer by whole-body positron emission tomographic imaging. Ann Thorac Surg 1995;60: Chin R, Ward R, Keyes JW, et al. Mediastinal staging of non small cell lung cancer with positron emission tomography. Am J Respir Crit Care Med 1995;152: Rigo P, Paulus P, Kaschten BJ, et al. Oncological applications of positron emission tomography with fluorine-18 fluorodeoxyglucose. Eur J Nucl Med 1996;23: Jacobson AF, Cronin EB, Stomper PC, Kaplan WD. Bone scans with one or two new abnormalities in cancer patients with no known metastases: frequency and serial scintigraphic behavior of benign and malignant lesions. Radiology 1990;175: Silvestri GA, Littenberg B, Colice GL. The clinical evaluation for detecting metastatic lung cancer. A meta-analysis. Am J Respir Crit Care Med 1995;152: Dehdashti F, Siegel BA, Griffeth LK, et al. Benign versus malignant intraosseus lesions: discrimination by means of PET with 2-[F-18]-fluoro-2-deoxy-d-glucose. Radiology 1996; 200: Gupta NC, Maloof J, Gunel E. Probability of malignancy in solitary pulmonary nodules using fluorine-18-fdg and PET. J Nucl Med 1996;37: Patz EF, Lowe VJ, Hoffmann JM, et al. Focal pulmonary abnormalities: evaluation with F-18 fluorodeoxyglucose PET scanning. Radiology 1993;188: Erasmus JJ, Patz EF, McAdams HP, et al. Evaluation of adrenal masses in patients with bronchogenic carcinoma using 18F-fluorodeoxyglucose positron emission tomography. Am J Radiol 1997;168: Block MI, Patterson GA, Sundaresan RS, et al. Improvement in staging of esophageal cancer with the addition of positron emission tomography. Ann Thorac Surg 1997;64: DISCUSSION DR SUDHIR R. SUNDARESAN (St. Louis, MO): I congratulate Dr Weder on a very fine study and an excellent presentation. There are a considerable number of published reports now dealing with PET for evaluating solitary pulmonary nodules and for staging the mediastinum in lung cancer. And at least for staging the mediastinal lymph node status, the available information suggests that PET is consistently superior to CT in evaluating nodal stage. And if you average the data from roughly a half dozen available reports, you find that the average sensitivity of PET approaches that of mediastinoscopy, and the average specificity of PET is more than 90%. When you consider these favorable data and also consider that PET studies are noninvasive and provide whole-body imaging in about an hour with a very low radiation dose, the potential application for PET to detect extrathoracic spread of lung cancer becomes very relevant. I am aware of only three publications reporting the prevalence of unexpected extrathoracic metastases in lung cancer patients who are deemed operable by conventional staging methods. In an earlier report that is a retrospective study of only 34 patients by Lewis and colleagues, the prevalence of unexpected distant metastases was 29%. There have been two more recent prospective studies published. The first of these, by Bury and colleagues looking at 61 patients, found a 10% prevalence of unexpected distant metastases; another report, by Valk and colleagues looking at 99 patients, had a prevalence of 11% and also no false-positives. Now, in the current study by Dr Weder and his colleagues, they have shown that in a group of 94 lung cancer patients that were staged at most as IIIa, and therefore considered operable, there was a 14% prevalence of unexpected thoracic metastases and no false-positives. So in this respect their findings are consistent with the prevalence available in the very few published reports. Also, their correlation of positive PET findings with histology and also the nodal stage seem entirely appropriate. Doctor Weder, I would like to ask three questions. First, although you found that PET turned up unexpected distant metastases missed by CT scan, have you encountered the opposite? In other words, PET suggests benign status of a distant CT abnormality. For example, this was an important finding in the study by Valk and colleagues. In their study, CT showed 19 distant abnormalities, usually in the liver, adrenal, or contralateral lung, and PET scanning showed these to be false-positives in 14 of those 19 cases. My second question has to do with the concept of PET scanning providing a sort of one-stop shopping for lung cancer staging. Our nuclear medicine physicians tell us that PET is unreliable for detecting brain metastases because of the obligatory high rate of glucose metabolism in the brain. So our approach, if we are trying to rule out a patient for surgery by searching for distant metastatic disease, and especially if we are looking for brain metastases, is to rely on CT or MR scanning and not a PET scan. In your study you did find one brain metastasis by PET scan and this is despite the fact that the reference value of FDG uptake you used to define a positive PET finding was the brain FDG uptake itself. Now, I am told that PET can be used to evaluate brain tumors, but that specialized scanning protocols need to be employed. So my question here is, have you used or are you considering implementing the use of these sorts of maneuvers to expand the capability of PET to detect brain metastases? Finally, because of the costs involved, we are all striving to find the most efficient and cost-effective means of staging the patient with lung cancer. Now that your group has completed a very nice study evaluating nodal status, and now on unexpected distant metastasis, can you please elaborate further on where you see PET fitting into the staging algorithm for lung cancer at your institution. DR PETER C. PAIROLERO (Rochester, MN): I would just like to point out that at the moment the Health Care Financing Administration does not pay for PET scanning in the United States. In the near future they will develop an algorithm describing under