FLUOR-18-FLUORODEOXYGLUCOSE POSITRON EMISSION TOMOGRAPHY (FDG-PET) IN MALIGNANT MELANOMA: Diagnostic comparison with conventional imaging methods

Transcription

1 Acta Radiologica ISSN: (Print) (Online) Journal homepage: FLUOR-18-FLUORODEOXYGLUCOSE POSITRON EMISSION TOMOGRAPHY (FDG-PET) IN MALIGNANT MELANOMA: Diagnostic comparison with conventional imaging methods B. Krug, M. Dietlein, W. Groth, H. Stützer, T. Psaras, A. Gossmann, K. Scheidhauer, H. Schicha & K. Lackner To cite this article: B. Krug, M. Dietlein, W. Groth, H. Stützer, T. Psaras, A. Gossmann, K. Scheidhauer, H. Schicha & K. Lackner (2000) FLUOR-18-FLUORODEOXYGLUCOSE POSITRON EMISSION TOMOGRAPHY (FDG-PET) IN MALIGNANT MELANOMA: Diagnostic comparison with conventional imaging methods, Acta Radiologica, 41:5, To link to this article: Published online: 09 Jul Submit your article to this journal Article views: 48 Full Terms & Conditions of access and use can be found at

2 Acta Radiologica 41 (2000) Copyright C Acta Radiologica 2000 Printed in Denmark All rights reserved ACTA RADIOLOGICA ISSN FLUOR-18-FLUORODEOXYGLUCOSE POSITRON EMISSION TOMOGRAPHY (FDG-PET) IN MALIGNANT MELANOMA Diagnostic comparison with conventional imaging methods B. KRUG 1,M.DIETLEIN 2,W.GROTH 3,H.ST TZER 4,T.PSARAS 1,A.GOSSMANN 1,K.SCHEIDHAUER 2, H. SCHICHA 2 and K. LACKNER 1 Departments of 1 Radiology, 2 Nuclear Medicine, 3 Dermatology and 4 Statistics, University of Cologne, Cologne, Germany. Abstract Purpose: To assess the diagnostic value of fluor-18-fluorodeoxyglucose posi- Key words: Melanoma, staging; tron emission tomography (FDG-PET) in screening for melanoma metastases. emission CT; ultrasonography; Material and Methods: The case records of 94 melanoma patients who had comparative studies. been examined by whole-body FDG-PET between 1995 and 1999 were evaluated retrospectively. Forty patients showed evidence of lymphogenous and 42 Correspondence: Barbara Krug, of hematogenous metastasis. The maximal interval between PET and the diag- Institut und Poliklinik für nostic procedure under comparison was 2 weeks. Confirmation of the findings Radiologische Diagnostik der was based on histology or the clinical or radiological course. Universität zu Köln, Results: In 24 patients, all diagnostic examinations including CT had been D Köln-Lindenthal, Germany. performed within 2 weeks from PET. In no case did PET change the staging. FAX π In 13 patients, PET agreed with morphological diagnosis in the number of metastatically invaded organs. This included 3 patients without metastases. The Accepted for publication 14 March estimated number of organs invaded by metastases was higher with PET in patients and higher with morphological imaging techniques in 6 patients. Among the with higher or equivocal counts of organs with metastases there were 2 confirmed false-positive findings. Conclusion: In a selected patient population, FDG-PET was found to be inferior to CT for diagnosing lung and liver metastases. The supplementary use of FDG-PET is not generally of value once metastasis has been established. The treatment and prognosis of malignant melanoma is dependent on the tumor stage. Half of the patients with hematogenous metastases die within 6 months (8, 11, 15). The therapeutic options at this stage of the disease are limited to palliative treatment. A 5-year survival rate of 40% has been reported for patients with satellite, in-transit and regional lymph node metastases (11, 14), which are usually surgically excised as a curative strategy. In most cases, presence or absence of lym- phogenous or hematogenous metastasis is established by conventional radiography, US, CT, MR imaging or radionuclide bone imaging. The diagnostic accuracy varies depending on the method of examination, the technique used, the organ and the size of the tumor: Whereas focal lesions of over 1 cm in diameter can be reliably detected, staging can be difficult, especially of small tumors. In the case of lymph nodes, anatomical imaging modalities have to rely on indirect staging criteria such as 446

3 FDG-PET IN MALIGNANT MELANOMA size, shape and count, which means that metastases in normal-sized lymph nodes go undetected. Since the early 1990s, positron emission tomography (PET) with 2-fluorine-18-fluoro-2-deoxy-Dglucose (FDG) has been available as an alternative imaging modality. It is based on the finding that proliferating tissue has elevated rates of glucose metabolism and allows the whole body to be examined in one scanning sequence (10). Initial trials with the whole-body technique have indicated that the method has a higher sensitivity for detection of melanoma metastases than conventional imaging procedures (2, 3, 6, 7, 10, 12). Accordingly, a new concept of metastasis screening with FDG-PET and validation of positive through specific morphological follow-up diagnosis has now been introduced. The discrepancy between reports in the literature and our clinical experience led us to carry out a retrospective analysis of PET diagnostic findings in melanoma patients. The following questions were posed: a) Is FDG-PET superior to all morphological examining techniques in all organs, and how reliable is FDG-PET diagnosis in comparison with the usual combination of morphological techniques used in the follow-up treatment of tumors? and b) Where does FDG-PET rank in comparison to other imaging procedures for the diagnosis of melanoma? Material and Methods All melanoma patients referred to the Department of Dermatology and Venerology between June 1995 and March 1999 for examination by FDG- PET were retrospectively included in the study. The inclusion criteria were the performance of a FDG- PET examination in any patient with malignant melanoma, who was seen by the Dermatological Department during a time interval of 4 years. The results of the all imaging procedures documented in the medical records were evaluated by a radiologist (B.K.), a nuclear medicine physician (M.D.) and a dermatologist (W.G.) using a data collection sheet covering the patient s details, details of the examination, the hospital, the findings tabulated according to organ and examining technique and clinical course. The maximum permissible interval between PET and other examinations being compared with PET was 2 weeks. If there was doubt of the correct anatomical allocation of a PET focus, the corresponding image documentations were re-evaluated in order to avoid false PET results because of incorrect topographical analysis. Reviews of the radiological images were not carried out since the images were unavailable in most cases. Confirmation of positive was based on histology or definite results of other imaging procedures carried out within 2 weeks. Negative were considered as confirmed by a clinical/radiological disease-free follow-up of at least 6 months. Within 34 months, 42 women and 52 men underwent 118 FDG-PET examinations for suspected metastatic melanoma. Sixteen patients had 2 and 4 patients had 3 PET examinations at an average interval of 7 months (minimum 3 months, maximum 15 months). The average patient age was 44 years (minimum 17 years, maximum 66 years). At the time of the examination, 28 patients had lymphogenous and 19 hematogenous metastasis as proven by histology or definite results of radiography, US, CT or MR imaging. In 12 patients, lymphogenous metastasis was suspected on clinical grounds, while 23 patients had clinical evidence of hematogenous metastasis. Thirty-one patients had died of their disease by April The interval between the last PET and death ranged from a minimum of 0 to a maximum of 12 months (average 6 months). One PET examination was excluded from the evaluations on account of elevated blood sugar levels. Eighteen correlative imaging procedures and examinations were performed within the permissible time period. Fifty PET examinations were done by the Department of Nuclear Medicine and 68 by two independent institutes of nuclear medicine. All clinics had run a 1- to 2-year training program between 1993 and False-positive and false-negative findings were evenly distributed between the institutions involved. All institutions used the same ECAT EXACT scanner (Siemens- CTI). Although all PET examinations involved transmission and emission measurements, the acquisition protocols varied: whole-body PET was acquired in two sessions and involved repositioning, whereas partial body PET was performed in one session and was therefore less susceptible to malignment artifacts. Activitites of 300 to 400 MBq 18 F-FDG were injected for each protocol. The brain was included in 36 whole-body examinations. Seventy-seven examinations covered the neck, trunk and the extremities. The areas neck, neck/thorax, skull/neck/thorax, neck/thorax/ abdomen and pelvis/groin were each examined once. The radiodiagnostic procedures were performed by private practices in two-thirds and by the Department of Radiology in one-third of cases. The Department of Radiology used a digital selen system (Thoravision, Philips Medical Systems) for chest and a standard Bucky systems for chest and 447

4 B. KRUG ET AL. Table 1 Counts of described and confirmed lung metastases diagnosed by chest radiography and PET Chest radiography findings, metastases, n Multiple S False-negative 1 False-positive 0 skeletal radiography. US (CS 9300, Ecoscan) was done with 3.5 MHz transducers for examinations of the abdomen and with 7.5 MHz transducers for the small parts. CT was performed with Somatom Plus S and Somatom Plus 4 units (Siemens). CT of the thorax and the abdomen was done with spiral technique with oral (abdomen) and i.v. (thorax, abdomen) administration of contrast agents using standard imaging protocols. A total of 72 sonographies were performed in 40 patients: 27 on the day of the PET and 45 after PET. The mean interval between US and PET examinations was 6 days (minimum 0, maximum 14 days). Of the thoracic CT, 45 were made before, 3 on the same day and 26 after the PET examination (mean interval 7, minimum 0, maximum 14 days). In the 65 patients examined with CT, 7 were made on the day of PET, 59 before and 59 after (mean interval 6, minimum 1, maximum 14 days). The 18 MR examinations were not systematically evaluated owing to the small number of cases involved. Results Lungs: In 10 of the 21 confirmed cases with lung metastases (47.6%), the chest radiographic findings agreed with the PET results with respect to the number of metastases detected (Table 1). In 1 confirmed case (4.8%), a higher number of lung metastases were revealed by PET than by radiography. In 10 confirmed comparisons (47.6%), chest radiography showed a higher number of pulmonary nodules. In 1 false-negative PET finding, chest radiography confirmed disseminated pulmonary metastases. In 1 correct negative PET, chest radiography proved to be false-positive. The confirmed PET and CT findings were in agreement in 50% of cases (9 comparisons) (Table 2). In no confirmed case did PET detect a higher number of lung metastases than CT. In this re- spect, CT proved to be superior to PET in 9 confirmed cases, whereas in 3 cases with normal PET findings, lung metastases were confirmed by CT in the subsequent course of disease. Liver: In 12 out of the 20 confirmed metastatic cases (60%), PET corresponded with US findings in the liver (Table 3). In 5 confirmed comparisons (25%), PET showed a higher number of metastasislike changes than US. In 1 of the cases in this category, US examination 13 days prior to PET showed normal findings, while a US control 18 days after PET corroborated the PET diagnosis of disseminated liver metastasis. Three times (15%) US revealed a higher metastases count than PET. In 2 of these cases the proved to be false-negative. In 9 out of 19 confirmed comparisons (47.4%), the PET and CT results for the liver were in agreement (Table 4). In 2 of the CT examinations evaluated as normal (10.5%), the indicated liver metastases. In 1 patient with disseminated liver metastases, CT performed 2 days before PET was interpreted as normal, while US made 18 days after PET confirmed the diagnosis. The PET finding in the other case turned out to be false-positive. In 42.1% of confirmed comparisons (8 patients), CT showed a higher number of liver metastases Table 2 Counts of confirmed lung metastases diagnosed by CT and PET CT findings, metastases, n Multiple S False-negative 3 False-positive 0 Table 3 Counts of confirmed liver metastases diagnosed by US and PET US findings, metastases, n Multiple S False-negative 2 False-positive 2 448

5 FDG-PET IN MALIGNANT MELANOMA Table 4 Counts of confirmed liver metastases diagnosed by CT and PET CT findings, metastases, n Multiple S False-negative 2 False-positive 1 Table 5 Comparison of counts of confirmed lymphogenous metastases detected in the soft tissues of the neck, the axillae and the groin by US and PET US findings, metastases, n Multiple S False-negative 2 False-positive 2 than PET. These included 2 that proved to be false-negative. Lymph nodes and subcutaneous soft tissues: In 18 patients with confirmed diagnoses, the PET examinations and sonographies of the soft tissues of the neck, axillae and groin were available for comparison. PET and US findings were concordant in 11 patients (61.1%) with respect to the number of lymphogenous and hematogenous soft-tissue metastases detected (Table 5). In 3 cases (16.7%), PET produced a higher count of metastasis-like foci, while in 4 cases (22.2%) US revealed a greater number of such foci. Among these were 2 falsepositive and 2 false-negative. In 9 patients with confirmation of normal status and in 1 patient with several confirmed mediastinal metastases, the PET and chest radiographic findings were concordant. In 2 cases with normal chest radiography, lymph node metastases were detected by PET: In 1 patient lymph node metastases were diagnosed by CT within 2 months. In another patient, FDG accumulation produced by a pulmonary lesion was mistaken for a lymph node metastasis. In 10 of the 16 confirmed comparisons (62.5%), PET and CT results agreed with regard to number of mediastinal and retroperitoneal lymph nodes detected (Table 6). A higher number of lymph node metastases was diagnosed four times (25.0%) by PET and twice by CT (12.5%). In 1 patient with positive PET and negative CT findings, CT performed 2 months later revealed mediastinal lymph node metastases. Three were falsepositive, and twice lung metastases were falsely interpreted as hilar lymph node metastases. One retroperitoneal FDG accumulation that seemed to indicate a lymph node metastasis could no longer be detected on a PET control 10 months later, even though there had been no intervening therapy, and was finally attributed to radionuclides in the urine passing through the patient s ureter. There was 1 case of false-negative. Skeleton: In 7 confirmed findings, PET was compared with radionuclide bone imaging. In 4 patients (1 patient with 1 skeletal metastasis, 2 patients with 3 5 skeletal metastases and 1 patient with disseminated skeletal metastases) the results of the two examining techniques agreed. Two PET findings were false-positive. In 1 confirmed case, more foci were detected by radionuclide bone imaging than by PET. In 14 patients, the skeletal findings from cranial, cervical, thoracic and abdominal CT were compared, without using special bone-window techniques, with the corresponding PET results. In 9 patients (64.3%), the findings agreed, in 8 cases they were normal, and in 1 case disseminated bone metastases were detected. In 4 patients with normal CT findings, skeletal metastases were suspected on the basis of PET. However, 3 of these proved to be false-positive. Synopsis of comparisons: Confirmation of the diagnosis was only possible in a minority of cases, the number of comparisons only being large enough to base single method comparisons on confirmed diagnoses (Tables 1 6). For that reason, Table 6 Counts of confirmed hilar, mediastinal and retroperitoneal lymph node metastases detected by CT and PET CT findings, metastases, n Multiple S False-negative 1 False-positive 3 449

6 B. KRUG ET AL. Table 7 Number of organs per patient diagnosed as sites of metastasis by PET in comparison with US (neck, axillae, groin, abdomen) and chest radiography Organs, n US findings in the neck, axillae, groin and abdomen and at chest radiography. Organs, n S S False-negative 0 False-positive 3 Table 8 Number of organs per patient diagnosed as sites of metastasis by PET in comparison with US (neck, axillae, groin) and CT (thorax, abdomen) Organs, n US findings of the neck, axillae and groin and CT findings of the thorax and abdomen. Organs, n Ø6 S S False-negative 0 False-positive 4 the synoptic comparisons of PET vs. two different combinations of radiological imaging modalities, which covered all organ systems and were used in daily routine, included both proven and unproven findings (Tables 7 and 8). In 29 patients, US of the soft tissues of the neck, the axillae, the groin and the abdomen, and chest radiography were performed within the permissible time interval (Table 7). In 16 patients (55.2%), the number of metastatically invaded organs detected by PET coincided with the correlative diagnostic methods. In 11 patients (37.9%), PET revealed a higher number of metastatically invaded organs than morphological diagnostic methods. However, 3 of these proved to be false-positive. Morphological diagnosis indicated a larger number of organs with metastases than PET in 2 cases (6.9%). In 13 of the 24 patients where US images of the soft tissues of the neck, the axillae and the groin, and CT images of the thorax and abdomen were available, the count of metastatically invaded organs agreed with the PET assessments (Table 8). In 5 patients (20.8%), PET showed a higher count of organs with metastasis-like alterations, while in 6 patients (25.0%) US and CT produced higher counts. In the group of PET reports with a higher count of metastatically invaded organs as well as in the group of PET reports with an equivocal count of metastatically invaded organs, there were 2 confirmed false-positive PET results when referring to single organ system comparisons, respectively. However, metastases suspected by PET in a confirmed healthy individuum or a normal PET results in a patient with proven metastases were not observed. 450

7 FDG-PET IN MALIGNANT MELANOMA Discussion A screening procedure that can be used for search for metastases needs, above all, to be highly sensitive. It should produce no false-negative findings and be able to detect metastases at an early, curable stage of development. However, high specificity, good biological tolerance and low costs are also important. Wide availability of equipment normally a valued criterion is not so critical in the case of melanoma owing to the relatively low incidence of the disease. According to studies published so far, FDG- PET is superior to morphological imaging techniques for the detection of melanoma metastases (2, 3, 7, 12). GRITTERS et al. (6) reported 100% accuracy for FDG-PET in detecting surface lymph node metastases in 12 patients, while STEINERT et al. (12) obtained a sensitivity of 92%, a specificity of 77% and an accuracy of 89% in the assessment of 33 melanoma patients with 40 metastases. BⁿNI et al. (2) described a sensitivity of 91%, a specificity of 67% and an accuracy of 92%. DAMIAN and coworkers (3) discovered 388 out of 415 metastases in 100 melanoma patients. Despite obtaining 67% specificity, BⁿNI et al. (2) approved the strategy of dispensing with confirmation of FDG accumulations by biopsy and regarded FDG-PET as superior to conventional imaging methods. In the opinion of DAMIAN et al. (3), it is possible to differentiate reliably between malignant and benign tissue on the basis of the degree of FDG uptake. Considering the high sensitivity reported, the availability of equipment and the costs involved, the Deutsche Gesellschaft für Nuklearmedizin (German Society of Nuclear Medicine) voted, in two meetings, to recommend the use of FDG-PET for lymph node staging and for examination of distant metastases in patients with high-risk melanomas, in cases where there was no morphological evidence of metastasis (4, 9, 10). In contrast to the publications quoted, we are primarily concerned with the diagnostic value of FDG-PET as a potential routine clinical procedure. The study compares results from PET, chest radiography, US, CT and skeletal scintigraphy examinations performed in various institutions by different investigators. The indications for examination did not always comply with the inclusion criteria of the Deutsche Gesellschaft für Nuklearmedizin (4) mentioned above. Thus, the patients selected for the study had tumors in a relatively advanced stage of development. The brain was included only in the first PET examinations, since clinical experience had shown that brain metastases are difficult to distinguish from background activity owing to the high rate of glucose metabolism in the normal parenchyma. The permissible time interval between the examination procedures under comparison was chosen shorter than that used by DAMIAN et al. (3) (10 weeks), GRITTERS et al. (6) (3 weeks) and STEINERT et al. (12) (3 weeks) in order to minimize the possibility of significant tumor progression between the studies. Although no reliable estimation of sensitivity, specificity and accuracy could be obtained owing to the small number of confirmed findings in the study, the confirmed false-negative lung, liver and lymph node findings suggest that the sensitivity of FDG-PET is unlikely to be high enough to exclude metastasis. The fact that morphological imaging revealed a higher number of metastasis-like lung and liver findings underlines this tendency, also in the unconfirmed cases. False-positive findings, present mostly in lymph nodes and the skeleton, could usually be traced to erroneous interpretations of physiological and pathophysiological foci of a non-metastatic origin, and should therefore not be regarded as shortcomings of the imaging method itself (5). The problem of low specificity appears not to have been sufficiently addressed (13): in the present investigation, FDG accumulations were usually interpreted as metastases without including the possibility of a benign origin in the differential diagnosis. In conclusion, the questions posed can be answered as follows: a) FDG-PET is inferior to CT for the detection of lung and liver metastases, and should therefore not be used alone to exclude metastasis. For this reason, the method offers no more diagnostic certainty than can already be obtained with the more economic combination of US (neck, axillae, groin) and CT (thorax, abdomen). b) Supplementary use of FDG-PET in addition to morphological diagnosis is not recommended in the case of lymphogenous and hematogenous metastasis. To what extent successful treatment based on false-negative morphological imaging results coinciding with correct positive justifies the supplementary use of FDG-PET for metastasis screening, is a question that for ethical and economic reasons can scarcely be answered in a clinical study. REFERENCES 1. BLESSING C., FEINE U., GEIGER L. et al.: Positron emission tomography and ultrasonography. Arch. Dermatol. 131 (1995), BⁿNI R., HUCH BⁿNI R. A., STEINERT H. et al.: Staging of melanoma by whole-body positron emission tomography 451

8 B. KRUG ET AL. using 2-fluorine-18-fluoro-2-deoxy-D-glucose. Br. J. Dermatol. 132 (1995), DAMIAN D. L., FULHAM M. J., THOMPSON E. et al.: Positron emission tomography in the detection and management of metastatic melanoma. Melanoma Res. 6 (1996), DEUTSCHE GESELLSCHAFT F R NUKLEARMEDIZIN: Nachrichten. Konsensus Onko-PET. Nuklearmedizin 36 (1997), ENGEL H., STEINERT H., BUCK A. et al.: Whole-body PET. Physiological and artifactual fluorodeoxyglucose accumulations. J. Nucl. Med. 37 (1996), GRITTERS L. S., FRANCIS I. R., ZASADNY K. R. et al.: Initial assessment of positron emission tomography using 2-fluorine-18-fluoro-2-deoxy-D-glucose in the imaging of malignant melanoma. J. Nucl. Med. 34 (1993), HOH C. K., HAWKINS R. A., GALSPY J. A. et al.: Cancer detection with whole-body PET using 2-[ 18 F]fluoro-2-deoxy-D-glucose. J. Comput. Assist. Tomogr. 17 (1993), LEDERMANN J. A.: Malignant melanoma. Medical and radiation therapy. Report of a meeting of physicians and scientists, University College London Medical School. Lancet 340 (1992), RESKE S. N.: Positronen-Emissions-Tomographie in der Onkologie. Dtsch. Ärzteblatt 95 (1998), C RESKE S. N., BARES R., B LL U. et al.: Klinische Wertigkeit der Positronen-Emissions-Tomographie (PET) bei onkologischen Fragestellungen. Ergebnisse einer interdisziplinären Konsensuskonferenz. Nuklearmedizin 35 (1996), SILVERBERG E., BORING C. C. & SQUIRES T. S.: Cancer statistics CA. Cancer J. Clin. 40 (1990), STEINERT H. C., HUCH BⁿNI R. A., BUCK A. et al.: Malignant melanoma. Staging with whole-body positron emission tomography and 2[F-18]-fluoro-2-deoxy-D-glucose. Radiology 195 (1995), TAHARA T., ICHIYA Y., KUWABARA Y. et al.: High [ 18 F]- fluorodeoxyglucose uptake in abdominal abscesses. A PET study. J. Comput. Assist. Tomogr. 13 (1989), TILLMANN D. M., AITCHISON T., WATT D. C. et al.: Stage II melanoma in the west of Scotland, Prognostic factors for survival. Eur. J. Cancer 27 (1991), WONG J. H., SKINNER K. A., KIM K. A. et al.: The role of surgery in the treatment of nonregionally recurrent melanoma. Surgery 113 (1993),