Compu ted tomography (CT) is an integral part of the

Transcription

1 Positron Emission Tomography of Lung Tumors and Mediastinal Lymph Nodes Using [ 1 8 FlFluorodeoxyglucose Walter J. Scott, MO, Jane L. Schwabe, MO, Naresh C. Gupta, MO, Naresh A. Dewan, MBBS, Steve D. Reeb, MO, Jeffrey T. Sugimoto, MO, and the Members of the PET-Lung Tumor Study Group* Departments of Surgery, Pulmonary Medicine, and Radiology, Creighton University Medical Center and Omaha Veterans Affairs Medical Center, Omaha, Nebraska Positron emission tomography detects increased glucose uptake in malignant tissue using the glucose analogue [2-18Flfluoro-2-deoxy-o-glucose. We reviewed the scans obtained in 62 patients with lung tumors. All had undergone computed tomography and had tissue-based diagnoses: 22 had adenocarcinomas, 12 had squamous cell carcinomas, 13 had other malignancies, 1 had organizing pneumonia, 1 had a hamartoma, and 13 had granulomas. Positron emission tomography with [2_ 18F]fluoro-2 deoxy-n-glucose identified 44 of 47 malignancies. Two of three false-negative findings were tumors that were 1 em" or less and the other was a bronchioloalveolar carcinoma. All three false-positive findings were granulomas. The sensitivity and specificity of the technique were 93.6% and 80%, respectively, and the positive and negative predictive values were 93.6% and 80%, respectively. The differential uptake ratio was determined in all 62 pa- tients. The mean differential uptake ratio (± the standard error of the mean) for malignant tumors was 6.4 ± 0.56 and that for benign tumors was 1.14 ± 0.26 (p < , t test>. Twenty-five of the patients had N2 lymph nodes evaluated pathologically. Positron emission tomography with [2-18Flfluoro-2-deoxy-o-glucose identified negative N2 nodes in 19 of 22 patients (86%) with negative nodes and positive N2 nodes in 2 of 3 patients (66%) with positive nodes, including one instance missed by computed tomography. Positron emission tomography with [2J 8Flfluoro-2-deoxy-o-glucose is a highly accurate and noninvasive method for identifying malignant lung tumors. The clinical application of this imaging technique in the evaluation of N2 lymph nodes awaits the results of ongoing prospective studies. (Ann Thorae Surg ) Compu ted tomography (CT) is an integral part of the noninvasive evaluation of thoracic tumors and the monitoring of the therapeutic response. The spatial resolution of CT provides morphologic detail superior to that provided by other imaging techniques. However, CT has a number of limitations, including a limited ability to distinguish benign from malignant tumors [1]. A high rate of glycolysis is a biochemical feature of malignant tissue, a fact first noted by Warburg [2,3] and confirmed by others [4]. In contrast to CT, positron emission tomography (PET) with [2-18F]fluoro-2-deoxy-o-glucose (FDG) provides images based on the differences in glucose metabolism between tissues [5]. [2-18F]fluoro-2-deoxy-D-glucose is a glucose analogue that is labeled with a positron-emitting isotope, fluorine 18, and FDG competes with glucose for uptake into the cell. [2-18F]fluoro-2-deoxy-D-glucose is phosphorylated and then accumulates in the cell because it cannot be metabolized further and cannot diffuse out of the Presented at the Thirtieth Anniversary Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 31-Feb 2, * The Members of the PET-Lung Tumor Study Group are listed in Appendix 1. Address reprint requests to Dr Scott, Department of Surgery, Creighton University Medical Center, 601 N 30th St, Omaha, NE by The Society of Thoracic Surgeons cell. Therefore, the amount of FDG that accumulates in the cell over a fixed period is proportional to the rate of cellular glucose metabolism. The ability of PET-FDG imaging techniques to detect differences in glucose metabolism between malignant and benign tissues is the basis for its use in the practice of clinical oncology. In early reports of the use of PET-FDG imaging in patients with lung cancer, it was described as a promising technique for differentiating malignant from benign tissue and for detecting mediastinal lymph node metastases [1, 5-9]. This report describes the findings from a retrospective analysis of our initial experience with PET-FDG imaging in patients with various lung tumors. Material and Methods We retrospectively analyzed the case histories of 62 patients who underwent PET-FDG imaging of lung tumors from 1992 through August All patients underwent CT scanning of the chest before PET-FDG imaging. The CT scans were obtained on a General Electric 9800-Hilite-HTD scanner (Brookfield, WI). Routine images were obtained to the level of the adrenal glands, with 10-mm collimation at 10-mm table increments, using 120 k.v, 120-mA, 2-second scans reconstructed with standard /94/$7.00

2 Ann Thorac Surg SCOTT ET AL 699 algorithms. (Thin-section images for the determination of lesion composition used 3-mm collimation, 3-mm table increments, 120-kV, 120-mA, 2-second scans reconstructed with standard algorithms.) Isovue 300 (iopamidol 61%) was used for intravenous contrast enhancement and was infused at a rate of 0.5 ml/min for 60 seconds, then increased to 1.5 ml/min to a complete infusion of 140 ml for the total volume. Mediastinal lymph nodes were considered abnormal if they exceeded 1 ern in the short-axis diameter. Radiopharmaceutical Agents The FDG was synthesized according to Hamacher and colleagues' [10] method using the Siemens/CTI (CTI, Knoxville, TN) chemical processing control unit. The fluorine 18 was obtained from the Siemens/CTI radioisotope delivery system 112, an 11-meV proton cyclotron, and sent directly to the chemical processing control unit for FDG synthesis. Pyrogenicity, thin-layer chromatography, and sterility tests were performed. The purity of FDG exceeded 95% for all doses. The average dose was 10.5 mci and ranged from 9.8 to 11.0 mci. Positron Emission Tomography A Siemens/CTI ECAT 931 positron emission tomograph (CTI, Knoxville, TN) with a 12-cm axial field of view was used for all patients. The ECAT 931 has eight detector rings with 512 BGO crystals per ring. It uses direct and crossplanes coincidence detection to generate 15 slices of 8-mm thickness each per bed position. For the thorax, two-bed positions were obtained, covering 24 cm axially. Reconstruction in a 128 X 128 matrix with a Hann filter (0.5 cutoff) yields 8-mm resolution in all three dimensions. Scatter correction is not available with this unit. The acquisition protocol was as follows. The patient took nothing orally for at least 4 hours. Informed consent was obtained. A rectilinear transmission scan was acquired to aid in positioning, after inspection of a chest x-ray study or CT scan to ascertain the location of the lung mass. A transmission scan was acquired for 10 to 20 minutes for each bed position, depending on the specific activity of the ring sources at the time of the study, for at least 12 million counts per plane. During the transmission scan, marks were made on the patient's skin to aid in repositioning for the emission scan. One-half milliliter of blood was drawn for determination of the baseline blood glucose level (no manipulation of the blood glucose level was done) and the data recorded. The FDG was administered intravenously. After a 60-minute uptake period, the patient was repositioned in the scanner. An emission scan was acquired at 20 minutes for each bed position for a total of 40 minutes. Image reconstruction using measured attenuation correction was performed using a Hann filter with 0.5-pixel cutoff. The transaxial images were reconstructed into sagittal and coronal planes. Positron Emission Tomographic Scan Interpretation and Differential Uptake Ratio Calculation Qualitative images were displayed on an SUN microsysterns SPARKstation 1+ and interpreted by a boardcertified nuclear medicine physician experienced in PET. In some cases, the interpreter had access to chest x-ray studies, CT scans, or results. No coregistration of the CT and PET scans was performed. Scan reports from the time of the clinical study (qualitative, subjective interpretation) were used to classify the nodule as benign or malignant and to determine the presence or absence of metastasis to the mediastinum. Subsequently, a blinded, board-certified nuclear medicine physician experienced in PET calculated the differential uptake ratios (DUR). Images were displayed in Imagetool (CTI) on a SUN microsystems SPARKstation 1+ using axial planes. A small region of interest was drawn over the portion of the lung mass with the greatest concentration of FDG. This yielded the number of microcuries per milliliter within the outlined region. The DUR was calculated from the following formula: (/-LCi/mL) * (patient weight in kilograms)/(dose in millicuries). The DUR, then, corrects for the weight of the patient (larger patients have a larger volume of distribution of FDG) and the dose of FDG administered. No recovery coefficient correction was applied. For images without a clear focus of increased FDG in the lung parenchyma, a small region of interest of 50 pixels was drawn in the general area of the lung mass, as determined from the CT scan. Differential uptake ratios were not calculated for the mediastinum. Tissue diagnoses were obtained for all primary tumors on the basis of the findings yielded by percutaneous needle biopsy, bronchoscopy, mediastinoscopy, video-assisted techniques, or open thoracotomy. Mediastinal lymph node staging was performed according to the mapping system devised by the American Thoracic Society in patients with non-small cell lung cancer who underwent mediastinoscopy or lung resection. The mean DURs (± the standard error of the mean [SEM]) for primary tumors that were determined by histologic studies to be benign were compared with those for the malignant tumors using an unpaired, two-sided t test. A p value of 0.05 or less indicated a significant difference. Both the CT and the PET-FDG scans were evaluated to determine their accuracy in imaging the hilar and mediastinal lymph nodes by comparing each of them with the histologic findings in the lymph node tissue obtained. Results Positron emission tomography of lung tumors using [2-1HF]fluoro-2-deoxy-o-glucose was performed in 62 patients (15 women, 47 men). No adverse effects attributable to PET-FDG imaging were identified in any patient. Tissue-based diagnoses were obtained in each patient. Of the 62 tumors, 47 were malignant: there were 22 adenocarcinomas (primary and metastatic), 12 squamous cell carcinomas, 5 non-small cell carcinomas, 3 small cell carcinomas, 1 large cell carcinoma, 2 melanomas, 1 carcinoid, and 1 Hodgkin's lymphoma. There were 15 benign tumors, and these consisted of 13 granulomas, 1 hamartoma, and 1 organizing pneumonia. The size of the primary tumors was determined from the CT scans or the resected specimens, and ranged from 0.7 to

3 700 SCOTT ET AL Ann Thorae Surg 6 cm. Nineteen (31%) of the lung tumors were greater than 3 em in diameter. Of the remaining 43 tumors, 5 were 3 ern in diameter, 16 were 2.9 to 2.0 em, 20 were 1.9 em to 1.0 em, and 2 were less than 1 cm. Serum glucose levels in specimens drawn just before the FDG was injected were less than 120 mg/dl in 57 of the 62 patients. The levels in the remaining 5 patients ranged from 125 to 150 mg/dl. No false-negative PET-FDG scan findings occurred in patients with an elevated glucose level. Based on visual inspection of the images, the PET-FDG scans correctly identified 44 of the 47 malignancies. Of the three false-negative findings, two were for tumors less than 1 em" and the other was for a bronchioloalveolar carcinoma. All three false-positive findings were for granulomas. The sensitivity and specificity of PET-FDG imaging for differentiating benign from malignant primary lung tumors were 93.6% and 80%, respectively. The positive and negative predictive values were 93.6% and 80%, respectively. The accuracy was 90%. The DUR was calculated for all 62 patients. The mean ± SEM DUR for malignant tumors was 6.4 ± 0.56 and that for benign tumors was 1.14 ± 0.26 (p < ). The DURs for histologically benign tumors ranged from 0 to The DURs for histologically malignant tumors ranged from 0.9 to The DURs for the 3 patients with histologically malignant tumors that were judged to be benign based on the observations from visual inspection of the PET scan images (false-negative PET scan) were 1.64 and 1.86 for the 2 patients with small (1.0 em or less) tumors and 0.9 for the patient with a 3.0-cm-diameter bronchioloalveolar carcinoma. The DURs for 3 patients with histologically benign tumors judged to be malignant on the basis of the visual inspection of the PET scan images (false-positive PET scan) were 1.92, 2.97, and The distribution of DURs for histologically malignant and histologically benign tumors is shown in Figure 1. In retrospect, a DUR of 2.0, chosen to distinguish benign from malignant lung tumors, was found to yield the fewest false-negative (3/47) and falsepositive (2/15) results in this group of data. The results of histologic analysis of mediastinal lymph 21 Malignant u o o Oaf oe$ ctp n=15 g"ot'j o o ~ Benign Fig 2. Positron emission tomographic image demonstrating [2-18Flfluoro-2-deoxY-D-glucose uptake in both the primary tumor and subcarinal lymph nodes. nodes were available from 25 of the 62 patients (40%). The PET-FDG scans correctly identified the absence of mediastinallymph node metastases in 19 of 22 patients (86%) with histologically negative lymph nodes. Positive mediastinal lymph nodes were correctly identified in 2 of 3 patients (66%) with histologically positive nodes. One of the falsepositive mediastinal node results occurred in a patient with active granulomatous disease. The other two instances of false-positive mediastinal nodes based on PET FDG imaging occurred in patients with extensive ipsilateral metastases in the hilar lymph nodes. In retrospect, PET-FDG imaging was actually found to have detected these hilar lymph node metastases, but they were misc1assified as mediastinal lymph node metastases because of their proximity to the mediastinum. The one false-negative mediastinal finding according to PET-FDG imaging occurred in a patient who had numerous subcarinai lymph node metastases with relatively low levels of radiotracer accumulation. The higher-background FDG uptake in the mediastinum, characteristic of the normal mediastinum but not normal lung, made these lymph node metastases difficult to detect. The mediastinal lymph nodes in this instance were also not abnormal according to CT criteria. The mediastinum was free of lymph node metastases in 20 of 22 patients (91%) with histologically negative lymph nodes according to the CT scan appearances. Mediastinal metastases were incorrectly diagnosed on the basis of CT scan findings in two instances. In one instance, an enlarged mediastinal lymph node identified on CT scans contained no tumor. In the second instance, an anterior mediastinal mass seen preoperatively on CT scans and deemed suspicious for metastasis was not appreciated at thoracotomy. Both of these patients had a normal mediastinum based on the findings from PET-FDG imaging. Only 1 of the 3 patients with mediastinal lymph node metastases was identified by CT scanning. Positron emission tomography with [2_ 18p] fluoro-2-deoxy-d-glucose correctly identified subcarinal lymph node metastases in 1 of the 2 patients with proven mediastinal metastases and a normal mediastinum on CT scans, as shown in Figures 2 and ~ ~ o \ ~ ~. o ~ ~ ~ j -1 L -' Fig 1. Distribution of the differential uptake ratios for histologically malignant and histologicallybenign tumors.

4 Ann Thorac Surg SCOTT ET AL 701 Fig 3. Normal mediastinum on a computed tomographicscan in the patient with positive subcarinal nodes identified by positron emission tomography using 12-1RFlfluoro-2-deoxY-D-glucose. Comment The spatial resolution of CT scans provides images with unsurpassed anatomic detail In the evaluation of suspected intrathoracic malignancies, CT scans characterize the primary tumor and its relationship to other structures, detect asymptomatic metastases, and identify enlarged lymph nodes. Additional advantages of CT include its ability to visualize small amounts of calcium, which may suggest a benign diagnosis [11]. Although they can provide unsurpassed morphologic detail, CT scans cannot definitively determine whether a tumor is malignant or benign. Computed tomography is currently the preferred imaging technique for evaluating mediastinal adenopathy. The diagnosis of lymph node metastases relies exclusively on the demonstration of lymph node enlargement. Mediastinal lymph nodes that are not enlarged may contain small foci of tumor, but many enlarged lymph nodes will be hyperplastic or inflammatory. The sensitivity and specificity of CT criteria for the detection of mediastinal nodal metastases observed in recent series are summarized in Table 1 [12-15]. The relatively low sensitivity and specificity of CT in detecting mediastinal lymph node metastases generally requires that the metastases be confirmed by biopsy findings, especially if positive findings would rule out surgical therapy. Computed tomographic scans are also used to determine the response of lung cancer to radiotherapy or chemotherapy. However, the radiographically observed response does not always correlate with the pathologically observed response. This has been most convincingly shown in trials of neoadjuvant therapy for advanced-stage lung cancers [16]. Thus, despite its ability to provide highly detailed images of anatomic structures, CT has definite limitations in the evaluation and follow-up of patients with lung cancer. Positron emission tomography using [2- IHF]fluoro-2 deoxy-o-glucose addresses some of these limitations. Positron emission tomography is an in vivo method of quantitating metabolism in normal and diseased tissues [2]. Warburg [3] was the first to note that a high rate of glycolysis is a biochemical feature of malignant tissue. Positron emission tomography can detect the uptake of the radiolabeled glucose analogue FOG. This compound enters cells in a competitive fashion with other hexoses, and the rate of uptake is proportional to the rate of the hexokinase-mediated reaction that transforms FOG into the phosphorylated form. The rate of the hexokinase reaction, in turn, is proportional to the rate of glycolysis in tissue under steady-state conditions. The phosphorylated FOG becomes trapped in the cell because it is not a substrate for glucose phosphate isomerase, the next enzyme in the glycolytic pathway [17]. Positron emission tomography can detect increased uptake of FOG in abnormal cells, and, because of this, could be used for the diagnosis of cancers and the monitoring of the effects of antitumor treatment. Our initial experience with PET-FOG imaging of lung tumors and mediastinal nodes confirms the experience of others [5-9]. Recently, Patz and associates [18] reported that PET-FOG imaging detected malignancy in 51 patients with focal pulmonary abnormalities, with a sensitivity of 89% and a specificity of 100%. These authors used a OUR threshold of 2.5 for diagnosing malignancy. We found that PET-FOG imaging detected malignant lung tumors with a sensitivity of 93.6%, a specificity of 80%, and an overall accuracy of 90%. False-negative results occurred in 2 patients with very small tumors «1 em") and in 1 patient with a bronchioloalveolar carcinoma, a tumor which, in many instances, has a low-grade malignant potential The rate of glycolysis in that tumor may have resulted in a low rate of FOG uptake. Studies are under way to determine whether FOG uptake is a measure of the proliferative activity of tumor cells [19]. The two small tumors in our patients were peripheral "scar" carcinomas possessing relatively few malignant cells that were interspersed in a large amount of fibrous stroma. The actual amount of malignant tissue taking up FOG may have been relatively small, which resulted in low overall FOG uptake. Falsenegative results are also possible in small tumors because of partial-volume effects. The metabolic activity of tumors below a certain size (1.6 cm for our scanner) is underestimated because the scanner averages the activity of a smaller «1.6-cm) tumor with the activity of the rim of normal tissue surrounding it, out to a diameter of 1.6 ern. For this reason, OURs for smaller tumors may be falsely low. In our study, false-positive diagnoses were made in 3 patients, in all of whom granuloma or inflammatory disease was ultimately diagnosed. In at least 1 of these Table 1. Sensitivity and Specificity of Computed Tomographic Criteria for Detecting Mediastinal Nodal Metastases Reference Sensitivity Specificity McKenna et al [13], % 60% Ferguson et al [12], % 83% Staples et al [15], % 65% Webb et al [1], % 69% McLoud et ai [14], ';;) 62%

5 702 SCOTT ET AL Ann Thorac Surg patients, review of the chest x-ray studies clearly revealed that the nodule seen on PET-FOG images had grown at a rate faster than that of lung cancers, suggesting the presence of a relatively acute inflammatory process. Others have found that FOG accumulates in a site of active inflammation [18]. It has been shown to localize in tumor cells and in the macrophages and granulation tissue that may surround the tumor [20, 21]. The uptake in macrophages and granulation tissues may explain the FOG uptake in nonmalignant inflammatory lesions. The OURs calculated in our 62 patients suggest that this may be a valuable method for differentiating benign from malignant tumors. On the basis of our limited data, it is too early to formulate definitive OUR criteria for malignancies. In retrospect, a OUR threshold for malignancy of 2.0 yielded the greatest number of true-positive and truenegative PET scan findings in our study. However, studies of medical decision-making techniques have revealed that the threshold value for a specific test is influenced by the pretest probability of a condition in a certain population. Furthermore, the relative importance of not missing a finding (limiting the false-negative results) weighed against the consequences of treating those patients who do not actually have a certain finding (accepting false-positive results) will influence the choice of a diagnostic threshold for the OUR value. Additionally, the basis of the PET image is the relative uptake of FOG by the tumor (a specific region of interest) versus that taken up by surrounding tissue. Therefore, the background OUR value is important. The average OUR of normal tissue varies for different regions in the chest (Gobar L5, personal communication, 1994). Therefore, a higher OUR threshold might be required in order to diagnose a malignancy against the background of a region with normal but more metabolically active tissue. The number of patients who underwent pathologic staging of their mediastinal lymph nodes in our early experience is limited. Therefore, few conclusions can be drawn from these data regarding the ability of PET-FOG imaging to detect mediastinal lymph node metastases. Of the three false-positive mediastinal studies, two were in patients with large hilar lymph node masses adjacent to the mediastinum. These two false-positive readings might have been correctly localized if the spatial resolution of the PET-FOG image had been better. The lack of precise anatomic detail is an inherent weakness of all radionuclide studies. We are working to develop computer algorithms that superimpose (coregister) axial CT images on PET-FOG images to improve the localization of areas of uptake. These preliminary results await confirmation from larger series and from other centers where PET technology is available. Currently, PET seems most useful as an adjunct to CT in directing evaluation of patients with lung tumors. With improved technology, PET-FOG may help to eliminate some of the invasive staging procedures now performed. Finally, PET-FOG should be evaluated as a method capable of noninvasively measuring the response of patients to the treatment of lung cancer. We sincerely appreciate and acknowledge the assistance of Sandra K. Nichols in the preparation and editing of the manuscript. References 1. Webb WR, Gatsonis C, Zerhouni EA, et al. CT and MR imaging in non-small cell bronchogenic carcinoma: report of the Radiologic Diagnostic Oncology Group. Radiology 1991; 178: Warburg O. The metabolism of tumors. London: Constable, Warburg O. On the origin of cancer cells. Science 1956;123: Weber G. Enzymology of cancer cells. Part I. N Engl J Med 1977;296: Strauss L, Conti P. The applications of PET in clinical oncology. J Nucl Med 1991;32: Gupta N, Frank A, Dewan N, et al. Solitary pulmonary nodules: detection of malignancy with PET with 2-[F-18] fluoro-2-deoxy-d-glucose. Radiology 1992;184: Knopp MV, Strauss LG, Haberdorn U, et al. Positron emission tomography (PET) with F-18-deoxy-glucose in the imaging and staging of bronchogenic carcinoma. Eur J Nucl Med 1991;32: Kubota K, Matsuzawa T, Ito M, et al. Differential diagnosis of lung tumor with positron emission tomography: a prospective study. J Nucl Med 1990;31: Patz EF, Lowe VI, Hoffman JM, et al. Focal pulmonary abnormalities: evaluation with F-18 fluorodeoxyglucose PET scanning. Radiology 1993;188: Hamacher K, Coenen HH, Stocklin G. The efficient stereospecific synthesis of no-carrier-added 2-['HF]-fluoro-2-deoxy-Dglucose using aminopolyether supported nucleophilic substitution. J Nucl Med 1986;27: White C, Templeton P, Belani C. Imaging in lung cancer. Semin Oncol 1993;20: Ferguson MK, MacMahon H, Little AG, Golomb HM, Hoffman PC, Skinner DB. Regional accuracy of computed tomography of the mediastinum in staging of lung cancer. J Thorac Cardiovasc Surg 1986;91: McKenna R [r, Libshitz H, Mountain CE, et al. Roentgenographic evaluation of mediastinal nodes for preoperative assessment in lung cancer. Chest 1985;88: McLoud T, Bourgouin P, Greenberg R, et al. Bronchogenic carcinoma: analysis of staging in the mediastinum with CT by correlative lymph node mapping and sampling. Radiology 1992;182: Staples C, Muller N, Miller R, et al. Mediastinal nodes in bronchogenic cancer: comparison between CT and mediastinoscopy. Radiology 1988;167: Rusch VW, Albain KS, Crowley JJ, et al. Surgical resection of stage IlIA and stage I1IB non-small-cell lung cancer after concurrent induction chemoradiotherapy. A Southwest Oncology Group trial. J Thorac Cardiovasc Surg 1993;105: Barrio JR. Biochemical principles in radiopharmaceuticals. In: Phelps ME, Mazziotta JC, Shelbert H, eds. Positron emission tomography and autoradiography: principles and applications for the brain and heart. New York: Raven, 1986: Patz EF, Lowe VI, Hoffman JM, et al. Focal pulmonary abnormalities: evaluation with F-18 fluorodeoxyglucose PET scanning. Radiology 1993;188: 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: Brown RS, Fisher SJ, Wahl RL. Autoradiographic evaluation of the intratumoral distribution of 2-deoxy-D-glucose and monoclonal antibodies in xenografts of human ovarian adenocarcinoma. J Nucl Med 1993;34: Kubota R, Yamada S, Kubota K, Ishiwata K, Tamahashi N, Ido T. Intratumoral distribution of fluorine-18-fluoro-deoxyglucose in

6 Ann Thorac Surg SCOTT ET AL 703 vivo: high accumulation in macrophages and granulation tissues studied by microautoradiography. J Nucl Med 1992;33: Appendix 1. Members of the PET-Lung Tumor Study Group The members of the PET-Lung Tumor Study Group at Creighton University Medical Center are Naresh A. Dewan, MBBS, pulmonary medicine; Albert R. Frank, MD, radiation oncology; Naresh C. Gupta, MD, nuclear medicine; Lisa S. Gobar, MD, nuclear medicine; James A. Mailliard, MD, oncology; Bangaruswamy Mouli, MD, nuclear medicine; Steve D. Reeb, MD, pulmonary medicine; Jane L. Schwabe, MD, surgery; Walter J. Scott, MD, surgery; John J. Sunderland, PhD, nuclear medicine; John D. Terry, MD, radiology; and Franc Wallace, MD, oncology. DISCUSSION DR DONALD M. HOPKINS (Sacramento, CAl: This is an exciting new technology that Dr Scott has reported on. He has given us some very important information pertinent to the use of this new tool that we can employ for the diagnosis and management of intrathoracic malignancies. We need a lot more data to guide us in how to use positron emission tomography (PET) appropriately, but I think this will be forthcoming as more thoracic surgeons become acquainted and experienced with its use. It is a very expensive study, and in the present climate of cost containment, we can anticipate a struggle in obtaining reimbursement for its use. Our experience in Sacramento has been very similar to that Dr Scott has reported. In addition, we have identified unsuspected metastatic lesions on our PET scans in several of our patients with bronchogenic carcinoma. We have been able to differentiate benign from malignant adrenal gland masses in 4 patients, and we have been able to distinguish benign from malignant lesions picked up on our staging bone scans. It is too early to know how helpful PET scan imaging will be in the management of intrathoracic malignancies, but the potential looks very exciting. DR SCOTT: Thank you for your comments, Dr Hopkins. We, too, are concerned about the cost of PET technology. As with any new application of technology, we need to identify the clinical situations in which PET imaging using [2-' 8F]fluoro-2-deoxy-D-glucose provides important information. Our early data suggest that this imaging technique may play an important role in the noninvasive evaluation of patients with solitary pulmonary nodules and in the evaluation of patients for the presence of regionally advanced disease. However, more data are necessary before the clinical utility-of the method can be determined.