Radio! Oncol 2000; 34(1): 1-9.
Correlation of 18F-FDG-PET and histopathology in patients with
malignant melanoma
Karl H. Bohuslavizki1, Susanne Klutmann1, Karsten Neuber2, Juliane Wedler2, Jiirgen Altenhoff2, Sabine Kroger1, Ralph Buchert1, Christian Bleckmann1
and Malte Clausen1
Departments of ^Nuclear Medicine and 2Dermatology, University Hospital Eppendorf Hamburg,
Germany
Background. Preliminary reports suggest that PET using 1SF-FDG may be a valuable diagnostic tool in patients with advanced malignant melanoma. Therefore, the aim of this study was to correlate PET lesions with histological findings in staging of malignant melanoma.
Patients and methods. A total of82 patients with malignant melanoma underwent 107 PET examinations for primary staging or therapy monitoring. After an intravenous injection of 370 MBq 1SF-FDG whole-body images were acquired on an ECAT EXACT 47 (921) with an axial field-of-view of 16.2 cm (Siemens, CTI). Tumor-suspicious PET findings, results of physical examination, and tumor-suspicious lesions by conventional imaging, i.e. ultrasound, CT, and MRI, were evaluated histologically and correlated with each other on a lesion-by-lesion basis.
Results. PET detected 124 lesions with an increased focal bracer uptake. In addition, physical examination and conventional imaging revealed 65 tumor-suspicious lesions. In total, 189 tumor-suspicious lesions were evaluated histologically. 7SF-FDG-PET was true-positive in 115/189 lesions and false-positive in 9/189 lesions. In 21 out of 65 PET-negative lesions, biopsy could not confirm melanoma tissue. In confrast, in 44 out of 65 PET-negative lesions further biopsy revealed malignant melanoma tissue. Sensitivity and specificity of 1SF-FDG-PET for the detection of malignant melanoma tissue were 72% and 70%, respectively. Negative and positive predictive values of 1SF-FDG-PET imaging were 32% and 93 %, respectively. With respect to anatomical localization, two thirds of false negative/positive PET lesions were located in the skin or mucosal area.
Conclusions. 1SF-FDG-PET is a valuable diagnostic tool in order to prove tumor-suspicious lesions in malignant melanoma. However, for exclusion of skin metastases an accurate physical examination of patients with malignant melanoma by a dermatologist is indispensable. With respect to very aggressive treahnent modalities of advanced malignant melanoma 1SF-FDG-PET may help to select the appropriate treatment protocol for each individual patient.
Key words: melanoma-diagnosis-pathology; tomography, emission-computed; 1SF-FDG-PET; neoplasm staging; histology; treatment strategy
2
Bohuslavizki KH et al. / 1SFFDG-PET and histopathology in patients with melanoma
Introduction
Cutaneous malignant melanoma is one of the most common malignancies with a twofold to threefold increasing incidence over the last 40 years.1 The most important prognostic factor is tumor staging at the time of diagnosis.2 According to the recommendations of the American Joint Commission on Cancer (AJCC) the clinical stage is divided into four groups. Clinical stages I and II are defined for primary malignant melanomas limited to the site of the origin without any evidence of a tumor spread elsewhere. In case of palpable local lymph node involvement or disseminated disease, the patients are classified as clinical stage III and IV, respectively. At the time of the first presentation, in nearly 80 % of all patients with the clinical stage I or II with, a mean 5-year survival rate of 85 % was noted.2 However, one third of the latter patients will have clinically undetectable lymph node metastases which, if left untreated, will significantly worsen the survival rate.3,4 Thus, an accurate tumor staging is a prerequisite for selecting the adequate treatment protocol.
Conventional imaging, i.e. computed tomography, magnet resonance imaging, and ultrasound are valuable and well-established diagnostic tools in pretherapeutic staging.5-8 However, these imaging modalities allow an identification of morphologic changes only, whereas tumor tissue in normal-sized lymph nodes can not be detected by definition.9 Moreover, morphologically orientated imaging permits a screening of a pre-selected body area only. Since malignant melanomas are
Received 29 December 1999 Accepted 10 January 2000
Correspondence to: Dr. Karl H. Bohuslavizki, Department of Nuclear Medicine, University Hospital Eppendorf, Martinistr. 52, D-20246 Hamburg, Germany: Phone: +49 40 42803 4047; Fax: +49 40 42803 6775; E-mail: bohu@uke.uni-hamburg.de
known for their aggressive lymphatic and hematogenic spread potency3'7, one single non-invasive imaging modality with simultaneous imaging of the whole-body would significantly facilitate pre-therapeutic management in these patients. A number of radiotracers have therefore been suggested, i.e. 67Ga-citrate,10 1231-benzamide, 123I-a-methyl-tyrosine,9 and 99mTc-labelled antimelanoma-antibodies,11 though a great number of false-negative findings were reported for all of these radiotracers.9,12 In contrast, initial experiences demonstrated the clinical potency of positron emission tomography (PET) using 2-[18F]-fluoro-2-deoxy-D-glucose (18F-FDG) for the detection of both local and systemic spread of metastatic malignant melanoma.1-13"22 Thus, within its geometric resolution of about 4-6 mm (FWHM), PET is able to detect tumor tissue independent of morphological changes due to an increased rate of glycolysis in malignant transformed cells. Since the early detection of malignant melanoma metastases increases patients' survival rate,23,24 PET imaging might be a valuable diagnostic tool in detecting melanoma metastases.
Therefore, the aim of this study was to determine the clinical value of 18F-FDG-PET in patients with malignant melanoma by comparing PET to findings of both clinical examination and morphological imaging, and to correlate these findings with histological results on a lesion-to-lesion basis.
Patients and methods
Patients
A total of 82 patients (37 female, 45 male) aged from 28 to 80 years with histologically proven malignant melanoma were investigated. The primary tumors were located in the skin area of the head and neck region in 17 patients, in the upper extremities in 6, in the lower extremities in 18, on the chest wall in 3, on the back in 29, and on the abdominal wall
Radiol Oncol 2000; 34(1): 1-9.
Bohuslnvizki KH et al. / 1SF-FDG-PET and histopathology in patients with melanoma
3
1
7. I
/ /
fc*. u ii & J&

«P» - .f,
U
/
t* 9

S



» M
V
VLL D
LAO
RVL
RAO
Figure l. Maximum intensity projections (MIPs) from the left lateral (V LL D), anterior (R V L), right anterior oblique (RAO), and left anterior oblique (LAO) views of a patient (58/f) with a primary malignant melanoma of the right lower extremity. The patient was clinically suspected for inguinal lymph node metastases. Note increased 18F-FDG uptake of both the primary tumor and the lymph node metastasis (arrow). Since no distant metastases were detected by 18F-FDG-PET the patient was classified stage IIIB.
in 2 patients. Unusual sites of the primary malignant melanoma were the vulva and the retina in one and two patients, respectively. The anatomic site of the primary tumor was unknown in the remaining 4 patients.
A description of the Clark level25 was available for 57 patients with the following distribution: level I, no patient; level II, 2 patients; level III, 13 patients; level IV, 34 patients; level V, 2 patients. Two patients were classified Clark level II/III, level III/IV and level IV/V, respectively.
Moreover, the classification results of the primary lesion thickness according to the Breslow scheme26 were available for 59 patients. There were 10 patients with thin
lesions (0,75 mm or less), 38 with intermediate lesions (0,76-3,99 mm), and 11 patients with thick lesions of 4 mm or greater.
PET scanning
The patients fasted for at least 12 hours prior to PET-scanning in order to minimize blood insulin levels and glucose utilization of normal tissue.27 Whole-body emission images were acquired without attenuation correction 60 min after i.v. injection of 370 MBq 18F-FDG using an ECAT EXACT 47 (921) scanner (Sie-mens/CTI) with an axial field-of-view of 16.2 cm.
The patients were placed in the PET gantry feet first with both arms folded over the
Radiol Oncol 2000; 34(1): 1-9.
4
Bohuslavizki KH et al. / lsF-FDG-PET and histopathology in patients with melanoma
V LL D	LAQ	RVL	-RAO
Figure 2. MIPs in same views as in figure 1 of a patient (69/f) after surgical treatment of a malignant melanoma located at the right leg. Note multiple metastases in 1SF-FDG-PET (arrows). However, a total of six metastases located within the skin area could not be identified by 18F-FDG-PET imaging but by an accurate clinical examination.
abdomen. Images were acquired for 4 min per bed position covering the feet up to the middle of the femurs. Then, the patients were repositioned in the gantry head first, and the second set of images was acquired from the brain down to the waist. Prior to the third acquisition set from the waist down to the lower extremities, the patients were asked to empty the bladder in order to decrease urine activity. Emission data were reconstructed by filtered back projection using a Hanning filter with a cut-off frequency of 0.4 of Nyquist frequency. Thus, transaxial spatial resolution was approximately 12 mm. PET-images were printed on transparency film (Helios 810, Sterling) using a linear gray scale with highest activity displayed in black. Images were displayed with an upper threshold of five times
Radiol Oncol 2000; 34(1): 1-9.
of the mean activity in the lung. Standardized documentation included both 20 transversal and 20 coronal slices with a slice thickness of 13.5 mm each, and maximum-intensity projections (MIPs) in the anterior, left lateral, right-anterior-oblique, and left-anterior-oblique view as published previously.28
Evaluation
Two independent nuclear medicine physicians interpreted PET images visually. All tumor-suspicious findings were evaluated histologically. Moreover, ali patients underwent physical examinations as well as morphological imaging, i.e. chest X-ray, CT scans of the chest, brain and abdomen or MRI. Additional tumor-suspicious lesions of conventional
Bohuslavizki KH et al. / lSF FDG-PET and histopathology in patients with melanoma
5
<
t
/
1
/ I
\ m

V LL D
LA0
RVL
RAO
Figure 3. MlPs in same views as in figure 1 of a patient (56/rn) after the excision of a malignant melanoma of the right leg. Patient showed up with a tumor-suspicious lesion in the right lung in postsurgical X-ray of the chest. I8F-FDG-PET revealed an increased glucose metabolism within the lesion, and histology confirmed malignant melanoma metastasis. Thus, the patient was classified as stage IV.
imaging or physical examination were evaluated histologically, too. The results of histology and tumor-suspicious lesions were correlated on a lesion-by-lesion basis with special respect to anatomical localization.
Results
In a total of 82 patients PET detected 124 lesions with an increased focal tracer uptake. Following physical examination and conventional imaging, 65 additional lesions were recorded to be suspicious for malignant melanoma metastases. Thus, a total of 189 lesions were evaluated histologically.
PET was true-positive in 115/124 lesions (Figure 1) and false-positive in 9/124 lesions. In 21/65 PET-negative lesions, biopsy did not confirm melanoma tissue, whereas histology was positive for melanoma metastases in the remaining 44 lesions (Figure 2). Thus, for the detection of malignant melanoma tissue PET had an overall sensitivity of 72% and an over-
all specificity of 70%. The negative predictive value of 18F-FDG PET was 32%. In contrast, the positive predictive value was 93%.
PET findings were either false-positive or false-negative in a total of 53 lesions. With regard to anatomical localization, 33 out of these lesions were located in the skin or within mucosal areas. Two lesions were situated in the lungs and three in the mesenterium. Five lesions were located in the axillary lymph nodes. The remaining 10 lesions were located in different regions of the body, e.g. the lung, the mediastinum. Thus, with respect to anatomical localization, almost two thirds of either false-positive or false-negative PET findings were located in the skin or within mucosal areas. If false-positive or false-negative skin or mucosal lesions were not taken into account, PET remained false-negative in 11 lesions with histologically proven melanoma tissue. Thus, in this subgroup, the sensitivity of PET scanning in melanoma patients increased to 91%.
Radiol Oncol 2000; 34(1): 1-9.
6
Bohuslavizki H et al. / 1SF-FDG-PET and histopathology in patients with melanoma
Discussion
Initial studies assessed the clinical utility of 18F-FDG-PET for the detection of metastatic malignant melanoma. Gritters and coworkers3 studied 12 patients with a total of 52 biopsy- or CT-diagnosed melanoma lesions. All patients underwent additional 18F-FDG-PET. Their initial data demonstrated the potential role of 18F-FDG-PET for the detection of metastatic malignant melanoma, especially in untreated extrathoracic lesions. Steinert and coworkers22 examined 33 patients with primary diagnosis or known relapse of malignant melanoma. In their patients, 18F-FDG-PET showed a sensitivity of 92% for the detection of malignant melanoma lesions. Moreover, the specificity was 77% without further clinical information and 100% with clinical information. Corresponding findings were demonstrated by Holder and coworkers16 who recommended 18F-FDG-PET as a primary strategy imaging modality in the staging of melanoma patients.
In this study, a total of 82 patients with malignant melanoma underwent 107 18F-FDG-PET examinations. In 115 out of 124 lesions with pathological focal increased !8F-FDG uptake, PET was true-positive. Moreover, in 9 PET-positive lesions, histology could not confirm malignant melanoma tissue. In contrast, in 44 out of 65 lesions detected by clinical examination or morphological imaging, only histological evaluation revealed malignant melanoma tissue. However, 21 out of 65 PET-negative lesions were true-negative. In accordance with the findings of other investigations the results of this study show that whole-body ^F-FDG-PET is an accurate imaging modality in patients with malignant melanoma in order to screen the whole-body for the presence of metastases.
However, in this study, ^F-FDG-PET had a sensitivity of 72% and a specificity of 70% only. It is remarkable that 53 lesions were either false-negative or false-positive in ^F-
FDG-PET. Two thirds of these lesions were located within the skin or mucosal areas. Thus, the skin and mucosal areas might be problematic regions for the detection of malignant melanoma metastases with 18F-FDG-PET. One possible cause of false-positive results is the fact that ^F-FDG is excreted via the urine. Thus, the contaminations of the skin, predominantly at the lower extremities might be interpreted as tumor-suspicious lesions of the skin. Moreover, it is known that patients treated with interferon alpha and interleukin-2 exhibit cutaneous inflammatory infiltrations at the injection site,24 which may cause false-positive results in 18F-FDG-PET. However, the majority of false PET findings were false-negative. The limited impact of tóF-FDG-PET for the detection of skin and mucosal metastases might be due to physiological and technological reasons. First, tumor lesions located within the regions of high physiological 18F-FDG uptake, i.e. the brain or the kidneys, might not be identified by 18F-FDG-PET imaging. Second, the detection of small skin/mucosal lesions with diameters of less than 5 mm might be limited by geometrical resolution of ^F-FDG-PET. Moreover, PET-images in this study were reconstructed by filtered backprojection. As a consequence, melanoma metastases in borderline areas, i.e. the skin, can hardly be differentiated from non-malignant transformed tissue. This problem might be solved by time-consuming iterative reconstruction algorithms. Moreover, the high number of false-negative PET lesions located within the skin and mucosal area underlines the necessity of an accurate and careful physical examination of the patient in daily clinical patient management. With these limitations in mind, whole-body 18F-FDG-PET is a suitable imaging modality in order to prove tumor-suspicious lesions in malignant melanoma. However, for exclusion of skin metastases an accurate physical examination by a dermatologist is still indispensable.
Radiol Oncol 2000; 34(1): 1-9.
Bohuslavizki KH et al. / 1SF-FDG-PET and histopathology in patients with melanoma
7
Any diagnostic test should, in principle, not only be judged with respect to its statistic data, e.g. sensitivity, specificity, negative and positive predictive value, and accuracy, but rather in the light of its effect on treatment strategy. The therapeutic approach in malignant melanoma mainly depends on the extent of the disease. In clinical stages I (pT1/T2 NO MO) and II (pT3/T4 NO MO), excision of the primary malignancy is the golden standard. In the last few years elective lymphadenecto-my was abandoned since its additional value for the patients' survival rate was demonstrated in retrospective patient studies only,29'30 but not in randomized prospective patient studies.30 If patients show up with regional lymph node metastases or in-transit-metastases but no distant metastases (stage IIIb), the therapeutic approach includes therapeutic lymphadenectomy. However, 10-year survival-rate decreases from 97% in patients staged pTlNOMO to 19% in patients staged Nl or N2 and MO melanoma.31 The primary treatment goal in patients with Ml malignant melanoma (clinical stage IV) is the reduction of tumoral masses in order to prolong patients' life expectancy as well as to improve the quality of life.32 In principle, there are three therapeutic options: surgery, external radiotherapy, and chemotherapy. In case of isolated metastases, an operative treatment has proved to be helpful in the prolongation of patients' life expectancy. Most studies demonstrated life prolongation in case of total resection of all tumoral masses only.33 Thus, 10-year survival-rate was expected to be as low as 3% in these patients with advanced malignant melanoma.31 However, there is no well-established, standardized systemic treatment protocol for patients with distant metastases. Thus, the treatment strategy itself is still under clinical investigation and is the focus of several patient studies. There is now evidence that patients with stage IV malignant melanoma benefit from an aggressive chemotherapy with interleukin-2 and inter-
feron alpha. These authors report of 5-year survival-rate of up to 10%.24'34
Thus, in addition to sensitivity and specificity of high-resolution ultrasonography of 70% and 90%,35 respectively, even patients with advanced malignant melanoma may benefit from the detection of metastases by 18F-FDG-PET due to several reasons. First, patients' survival rate decreases with an increasing number of involved lymph node regions.36 Second, prognosis of patients is better with an early detection of metastases and with small tumor masses at the time of detection.36 Third, ^F-FDG-PET has been proved superior in the detection of lung metastases (Figure 3) as compared to conventional, well-established computed tomogra-phy.37'38 Last, 18F-FDG-PET offers the advantage to image the whole body in one single procedure which is especially important because in malignant melanoma often unexpected, aberrant metastatic spread is found. Thus, ^F-FDG-PET has already been suggested for the staging of malignant melanoma.22
Conclusions
1SF-FDG-PET is a valuable diagnostic tool in order to prove tumor-suspicious lesions in malignant melanoma. However, for exclusion of metastases physical examination by dermatologist and conventional imaging are indispensable. With respect to very aggressive treatment modalities of advanced malignant melanoma ^F-FDG-PET may help to select the appropriate treatment protocol for the individual patient.
References
l. Macfarlane DJ, Sondak V, Johnson T, Wahl RL. Prospective evaluation of 2-[18F]-2-deoxy-D-glucose positron emission tomography in staging of regional lymph nodes in patients with cutaneous malignant melanoma. J Clin Oncol 1998; 16: 1770-6.
Radiol Oncol 2000; 34(1): 1-9.
8
Bohuslavizki KH et al. / 1SF-FDG-PET and histopathology in patients with melanoma
2.	Sober AJ, Koh HK. Melanoma and other pigmented skin lesions. In: Isselbacher KJ, Braunwald E, Wilson JD, Martin JB, Fauci, A. S., Kasper DL, editors. Harrison's principles ofintemal medicine. 13 ed. New Yoork: McGraw-Hill, Inc; 1994. p. 1867-71.
3.	Gritters LS, Francis IR, Zasadny KR, Wahl RL. Initial assessment of positron emission tomography using 2-fluorine-18-fluoro-2-deoxy-D-glucose in the imaging of malignant melanoma. J Nucl Med 1993; 34: 1420-7.
4.	Zartman GM, Thomas MR, Robinson WA. Metastatic disease in patients with newly diagnosed malignant melanoma. N Nucl Med 1987; 35: 163-4.
5.	Ginaldi S, Wallace S, Shalen P, Luna M, Handel S. Cranial computed tomography of malignant melanoma. AJR 1981; 136: 145-9.
6.	Shirkhoda A, Albin J. Malignant melanoma: correlating abdominal and pelvic CT with clinical staging. Radiology 1987; 165: 75-8.
7.	Silverman PM, Heaston DK, Korobkin M, Seigler HF. Computed tomography in the detection of abdominal metastases from malignant melanoma. Invest Radiol 1984; 19: 309-12.
8.	Patten RM, Shuman WP, Teefey S. Metastases from malignant melanoma to the axial skeleton: a CT study of frequency and appearance. AJR 1990; 155: 109-12.
9.	Steinert H, Boni R, Huch-Boni RA, Capaul R, von Schulthess GK, Westera G. 1231-alpha-methyltyro-sine scintigraphy in malignant melanoma. Nuklearmedizin 1997; 36: 36-41.
10.	Bekerman C, Hoffer PB, Bitran JD. The role of Gallium-67 in the clinical evaluation of cancer. Sem Nucl Med 1985; 15: 72-103.
11.	Bockisch A, Oehr P, Biltz H, Hotze A, Kreysel HW, Biersack HJ. The clinical value of radioimmunode-tection (RID) using antimelanoma antibodies. Tumordiag und Ther 1991; 12: 112-6.
12.	Boni R, Steinert H, Huch Boni R, Von Schulthess GK, Meyer J, Dummer R, et al. Radioiodine-labelled alpha-methyl-tyrosine in malignant melanoma: cell culture studies and results in patients. Br J Dermatol 1997; 137: 96-100.
13.	Damian DL, Fulham MJ, Thompson E, Thompson JF. Positron emission tomography in the detection and management of metastatic melanoma. Melanoma Res 1996; 6: 325-9.
14.	Boni R. Whole-body positron emission tomography: an accurate staging modality for metastatic melanoma. Arch Dermatol 1996; 132: 833-4.
15.	Hsueh EC, Gupta RK, Glass EC, Yee R, Qi K, Morton DL. Positron emission tomography plus serum TA90 immune complex assay for detection of occult metastatic melanoma. J Am Coll Surg 1998; 187: 191-7.
16.	Holder WD Jr, White RL Jr, Zuger JH, Easton EJ Jr, Greene FL. Effectiveness of positron emission tomography for the detection of melanoma metastases. Ann Surg 1998; 227: 764-71.
17.	Engel H, Steinert H, Buck A, Berthold T, Huch Boni RA, von Schulthess GK. Whole-body PET: physiological and artifactual fluorodeoxyglucose accumulations. J Nucl Med 1996; 37: 441-6.
18.	Kern KA. [14C]deoxyglucose uptake and imaging in malignant melanoma. J Surg Res 1991; 50: 643-7.
19.	Reske SN, Bares R, Bull U, Guhlmann A, Moser E, Wannenmacher MF. Clinical value of positron emission tomography (PET) in oncologic questions: results of an interdisciplinary consensus conference. Nuklearmedizin 1996; 35: 42-52.
20.	Rinne D, Baum RP, Hor G, Kaufmann R. Primary staging and follow-up of high risk melanoma patients with whole- body lsF-fluorodeoxyglucose positron emission tomography: results of a prospective study of 100 patients. Cancer 1998; 82: 1664-71.
21.	Paquet P, Hustinx R, Rigo P, Pierard GE. Malignant melanoma staging using whole-body positron emission tomography. Melanoma Res 1998; 8: 59-62.
22.	Steinert HC, Huch Boni RA, Buck A, Boni R, Berthold T, Marincek B, et al. Malignant melanoma: staging with whole-body positron emission tomography and 2-[F-18]-fluoro-2-deoxy-D-glucose. Radiology 1995; 195: 705-9.
23.	Reintgen DR, Balch CM, Kirkwood J, Ross M. Recent advances in the care of the patient with malignant melanoma. Ann Surg 1997; 225: 1-14.
24.	Richards JM, Gale D, Mehta N, Lestingi T. Combination of chemotherapy with interleukin-2 and interferon alfa for the treatment of metastatic melanoma. J Clin Oncol 1999; 17: 651-7.
25.	Clark WHJ, From L, Bernandino E, Mihm M. The histogenesis and biologic behavior of primary malignant melanomas of the skin. Cancer 1969; 29: 705-26.
26.	Breslow A. Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970; 172: 902-8.
Radiol Oncol 2000; 34(1): 1-9.
Bohuslavizki KH el al. / 1SF-FDG-PETand histopathology in palients with melanoma
9
27.	Minn H, Leskinen-Kallio S, Lindholm P, Bergman J, Ruotsalainen U, Terss M, et al. (18-F)fluo-rodeoxyglucose uptake in tumors: kinetic vs. stady-state methods with reference to plasma insulin. J Comput Assist Tomogr 1993; 17: 115-23.
28.	Bleckmann C, Buchert R, Schulte U, Lorenzen J, Bohuslavizki KH, Mester J, et al Onko-PET: Bildanalyse mittels Monitor versus standardisierter Filmdokumentation. Nuklearmedizin 1999; 38: 56-60.
29.	Kaufmann R. Operative Therapie des primsren Melanoms. Onkologe 1996; 2: 449-52.
30.	Balch CM, Milton GW, Cascinelli NFSH. Elective lymph node dissection: pros and cons. In: Balch CM, Houghton AN, Milton GW, Sober AJ, Soong SJ, editors. Cutaneous melanoma. Philadelphia: Lippincott, J. B.; 1992. p. 345-66.
31.	Hsffner AC, Garbe C, Biittner P, Orfanos CE, Rassner G, Burg G. The prognosis of primary and metastasizing melanoma. An evaluation of the TNM classification in 2495 patients and proposals for their revision. Br] Cancer 1992; 66: 856-61.
32.	Garbe C. Malignes Melanom. In: Seeber S, Schütte J, editors. Tiernpiekonzeple Onkologie. 3 ed. Berlin: Springer; 1998. p. 799-829.
33.	Göhl J, Meyer T, Haas C, Altendorf Hofman A, Hohenberger W. Ist die chirurgsche Therapie von Fernmetastasen maligner Melanome sinnvoll? Langenbecks Arch Chir 1996; 113(Suppl): 122-6.
34.	Hoffmann R, Müller I, Neuber K, Lassmann S, Buer J, Probst M, et al. Risk and outcome in metastatic malignant melanoma patients receiving DTIC, cisplatin, BCNU and tamoxifen followed by immuntherapy with interleukin 2 and interferon alpha2a. Br J Cancer 1998; 78: 1076-80.
35.	Blessing C, Feine U, Geiger L, Carl M, Rassner G, Fierlbeck G. Positron emission tomography and ultrasonography. A comparative retrospective study assessing the diagnostic validity in lymph node metastases of malignant melanoma. Arch Dermatol 1995; 131: 1394-98.
36.	Stachelmann WK, Rapaport DP, Soong SJ. Prognostic clinical and pathologic features. In: Balch CM, Houghton AN, Sober AJ, Soong SJ, editors. Cutaneous melanoma. St. Louis: Quality Medical Publishing, Inc.; 1998. p. 26-42.
37.	Hoh CK, Hawkins RA, Glaspy JA, Dahlbom M, Tse NY, Hoffman EJ et al Cancer detection with whole-body PET using 2-[18F]fluoro-2-deoxy-D-glu-cose. J Compul Assist Tomogr 1993; 17: 582-9.
38.	Hoh CK, Schiepers C, Seltzer MA, Gambhir SS, Silverman DH, Czernin J, et al PET in oncology: will it replace the other modalities? Semin Nucl Med 1997; 27: 94-106.
Radiol Oncol 2000; 34(1): 1-9.