Lymph node metastases of cutaneous melanoma: Diagnosis by B-scan and color Doppler sonography

Lymph node metastases of cutaneous melanoma: Diagnosis by B-scan and color Doppler sonography

Lymph node metastases of cutaneous melanoma: Diagnosis by B-scan and color Doppler sonography Matthias Moehrle, MD, Andreas Blum, MD, Gernot Rassner, ...

160KB Sizes 0 Downloads 33 Views

Lymph node metastases of cutaneous melanoma: Diagnosis by B-scan and color Doppler sonography Matthias Moehrle, MD, Andreas Blum, MD, Gernot Rassner, MD, and Michael Juenger, MD Tuebingen, Germany Background: Sonography is a sensitive, noninvasive method that can be used to detect regional lymph node metastases. Color Doppler sonography (CDS) can supply further information on lymph node perfusion. Objective: We evaluated the usefulness of CDS for differentiating between benign lymphadenopathies and lymph node metastases of cutaneous melanoma. Methods: In a prospective study, reactive inflammatory lymph nodes (rLN) and lymph node metastases of cutaneous melanoma (mLN) were examined by sonography and CDS. Lymph node echogenicity and shape (length/depth ratio) were determined by sonography. The vascularization pattern of the lymph nodes was established with CDS. We recorded the Doppler frequency spectra at the hili of the lymph nodes and then calculated the resistance and pulsatility indices (RI, PI). Results: The echogenicity of the lymph node centers had a sensitivity of 96% and a specificity of 100%. The shape differed highly significantly between the two groups (P < .001). The criterion length/depth ratio < 2 had a sensitivity of 85% and specificity of 86%. Hilus vessels could be detected in 14 of 22 rLN (64%). These vessels, however, were not present in any of the metastases. The RI and the PI in detectable lymph node vessels differed between the two groups (RI: P < .05; PI: not significant), but because of the overlap between the two groups, these indices were of no diagnostic value. In the presence of 2 or more of the following 3 criteria: length/depth ratio < 2, hypoechoic center, and the absence of hilus vessels, diagnosis of metastasis of malignant melanoma had a sensitivity of 100% and a specificity of 96%. Conclusion: CDS improves the diagnostic accuracy of conventional sonography. The measurement of Doppler curves in lymph node vessels and the calculation of pulsatility and resistance indices, on the other hand, is time-consuming and seems to be of no diagnostic value. (J Am Acad Dermatol 1999;41:703-9.)

B

- scan sonography (BSS) is a sensitive diagnos-

tic procedure that can be used to detect regional lymph node metastases in patients with cutaneous melanoma.1 Using a number of (sono)morphologic criteria for evaluation there is an accurate way to differentiate between benign and malignant lymphadenopathies.2-4 The detection of tumor vessels is a new application of color Doppler sonography (CDS). Tumor perFrom the Department of Dermatology, University of Tuebingen. The study was financed by funds from the Department of Dermatology. Accepted for publication May 27, 1999. Reprint requests: Matthias Moehrle, MD, Department of Dermatology, University of Tuebingen, Liebermeisterstrasse 25, 72076 Tuebingen, Germany. Copyright © 1999 by the American Academy of Dermatology, Inc. 0190-9622/99/$8.00 + 0 16/1/100414

fusion by vessels larger than 0.1 mm in diameter, that is, arterioles and venules, can be examined by means of commercially available instruments without contrast medium. Flow in multiple vessels is indicated by color pixels in a B-scan. Spectral analysis, however, is only possible in individual, arbitrarily chosen vascular sections. Perfusion patterns are detectable in a subjective, semiquantitative manner. CDS studies of primary tumors, especially in the area of gynecology,5-9 and of lymph nodes have not been entirely consistent in their selection criteria for malignancy.10 Among the few publications available on the perfusion of metastases,11-22 there are no specific data on lymph node metastases of cutaneous melanoma.23 With CDS we may be able to differentiate between benign and malignant (or even melanoma-typical) lymphadenopathies qualitatively, by analyzing perfu703

704 Moehrle et al

J AM ACAD DERMATOL NOVEMBER 1999

Table I. Formulas for defining the pulsatility index (PI) and resistance index (RI) Pulsatility index by Gosling and King24 PI = Systolic peak flow velocity–end-diastolic flow velocity Time averaged maximum velocity Resistance index by Pourcelot25 RI = Systolic peak flow velocity–end-diastolic flow velocity Systolic peak flow velocity

A

sion patterns, or quantitatively, by analyzing hemodynamics. This prospective study was designed to evaluate the usefulness of examining lymph node pathologies with CDS. Therefore we assessed the sensitivity and specificity of various criteria for differentiating lymphadenopathies found in follow-up examinations of melanoma patients.

METHODS

B

C Fig 1. A, B-scan sonography of reactive and metastatic lymph nodes. Shape (length/depth ratio). B, Color Doppler sonography of reactive and metastatic lymph nodes. Pulsatility index (PI). C, Color Doppler sonography of reactive and metastatic lymph nodes. Resistance index (RI).

We examined 22 reactively enlarged lymph nodes (18 patients) and 26 lymph node metastases of cutaneous melanoma (19 patients) with BSS and CDS. Reactively enlarged lymph nodes were localized in the groin (n = 20), in the axillae (n = 1), and in the neck (n = 1); lymph node metastases of cutaneous melanoma were localized in the groin (n = 10), in the axillae (n = 12), in the neck (n = 1), infraclavicularly (n = 1), and on the abdomen (n = 2). The diagnosis of reactively altered lymph node was made on the basis of clinical findings in conjunction with an inflammatory disease (abscess, erysipelas, erythroderma, atopic dermatitis, or psoriasis), in absence of any previous and present malignancy. Patients from our melanoma follow-up clinic with palpable or sonographically (BSS) suspect lymph nodes were included. The diagnosis of lymph node metastasis of cutaneous melanoma was confirmed histologically. In 24 of 26 lymph nodes we found tumor infiltration with perinodal tumor growth. Patients with any previous chemotherapy, radiation, or regional surgery were excluded. All 48 lymph nodes were examined with an ATL Ultramark 9 with a 7.5 MHz linear-array transducer in both the B-scan and the slow-flow mode (–5 to +5 cm/s) on at least two planes. To avoid compression artifacts the transducer was applied with as little pressure as possible. The examination procedures were performed by trained physicians (M. M. and A. B.), and there was minimal interobserver and intraobserver variability. The lymph nodes were visualized by conventional BSS in transversal and sagittal sections (length × width × depth). The sonomorphologic features that were used included shape, border to the surrounding tissues, and echogenicity of the lymph node center and periphery. CDS revealed vessels in 21 of the 22 reactively altered lymph nodes, but in only 19 of the 26 melanoma metastases. Here the sample volume cursor was positioned on the vessel lumen, and the size and angle were adjusted to

Moehrle et al 705

J AM ACAD DERMATOL VOLUME 41, NUMBER 5, PART 1

Table II. Results (number [N], median, minimum, maximum, standard deviation (SD), and P value [MannWhitney test]) of reactive lymph nodes (LN) and metastatic lymph nodes indicating length, width, depth (in millimeters), length/depth ratio, pulsatility index (PI), and resistance index (RI) Length

No. Median Minimum Maximum SD P value*

Width

Depth

Length/Depth

PI

RI

rLN

mLN

rLN

mLN

rLN

mLN

rLN

mLN

rLN

mLN

rLN

mLN

22 20.35 10.60 32.50 6.02 —

24 21.15 3.60 42.70 10.15 —

22 15.25 6.30 23.90 5.22 —

25 17.70 1.60 39.20 8.16

22 7.25 3.40 18.60 3.12

25 15.90 1.70 40.70 8.44

22 2.82 1.20 5.97 0.95

24 1.27 1.00 2.00 0.41 < .001

19 1.15 0.86 1.60 0.23

15 1.50 0.75 2.45 0.49 NS

19 0.66 0.58 0.77 0.06

15 0.78 0.51 0.97 0.12 < .05

NS, Not statistically significant. *Mann-Whitney test.

the vessels. Systolic peak flow velocity and end-diastolic flow velocity were determined from the Doppler spectrum, and then the pulsatility index (PI) and the resistance index (RI) were calculated. PI and RI, which are independent from the Doppler angle, can reveal information about peripheral perfusion resistance (Table I).24,25 The highest PI and RI per lymph node were considered for statistical evaluation. Vascular lymph node architecture was also evaluated with respect to perfusion type (central, peripheral, or mixed), hilus vessels, and focal or total loss of perfusion.

Statistical analysis Statistical significance was calculated with the MannWhitney U test for independent parameters. The relatively small size of the groups has to be considered.

RESULTS B-mode sonography The length/depth ratio is a parameter describing lymph node shape. In reactive lymph nodes, which have an oval shape, this ratio had a median of 2.82 (± 0.95); and in the metastases of cutaneous melanoma, which have a rather round shape, the length/depth ratio’s median was 1.27 (± 0.41) (P < .001) (Table II, Fig 1, A). To distinguish inflammatory lymph nodes from lymph node metastases with BSS, a value of less than 2 for the length/depth ratio had a sensitivity of 85% and a specificity of 86%. The positive predictive value was 88% and the negative predictive value was 83%. All of the inflammatory lymph nodes had an echogenic center surrounded by a hypoechoic seam (a “cockade”) (Fig 2). Of all the metastases, 96% lacked such a cockade sign (Fig 3). Thus the echogenicity of the lymph node centers allows us to differentiate between reactive lymph nodes and metastases.

CDS In the lymphadenitis group, perfusion was highest in the center in 10 of the lymph nodes (46%), whereas 11 lymph nodes (50%) had a mixed perfusion pattern. Among the melanoma metastases only 2 lymph nodes (8%) revealed a central perfusion pattern; in 8 lymph nodes (31%) it was mixed and in another 8 lymph nodes (31%) perfusion was primarily peripheral. Hilus vessels (Fig 4) could be demonstrated in 14 lymphadenitides (64%), but in none of the metastases (Fig 5). A total of 15 metastases (58%) had a partial, another 7 (27%) a total lack of perfusion. Pulsatility index and resistance index A total of 44 Doppler spectra were obtained from 19 of the 22 lymphadenitides and a total of 28 Doppler spectra from 15 of the 26 lymph node metastases. The highest RI and PI per lymph node were taken for statistical evaluation. RI and PI differed between reactive lymph nodes and metastases (RI: P < .05; PI: not significant) (Table II, Fig 1, B and C). Combination of BSS and CDS Given one of the criteria “length/depth ratio < 2” or “missing cockade sign,” the diagnosis of metastasis was never false negative, and 14% of the reactive lymph node diagnoses were false positive. Given the 2 criteria “length/depth ratio < 2” and “hypoechoic lymph node center” the positive predictive value was 100%; the combination of “length/depth ratio > 2” and “positive cockade sign (hyperechoic center)” had a negative predictive value of 100%. Given 2 of the 3 criteria “length/depth ratio < 2,” “hypoechoic lymph node center,” and “absence of

706 Moehrle et al

J AM ACAD DERMATOL NOVEMBER 1999

Fig 2. B-scan sonography. Transverse section of inflammatory inguinal lymph node of patient with erysipelas. Oval shape and typical echogenic center surrounded by hypoechoic seam (cockade). Lymph node is next to lumen of greater saphenous vein (“VSM”).

Fig 3. B-scan sonography. Cervical lymph node metastasis of cutaneous melanoma. Global shape and low central echogenicity.

Table III. Evaluation of criteria of B-mode and CDS for differentiating between reactive lymph nodes (n = 22) and metastases of melanoma (n = 26)

Sensitivity Specificity Positive predictive value Negative predictive value Accuracy

1 Length/ depth ≤ 2

2 No echogenic center

3 No pos hilum vessels

1 and 2 Both pos

1 or 2 At least one pos

1, 2, or 3 At least two pos

85% (22) 86% (19) 88% 83% 85% (41)

96% (25) 100% (22) 100% 96% 98% (47)

100% (26) 64% (14) 76% 100% 83% (40)

73% (19) 86% (19) 100% 100% 79% (38)

100% (26) 86% (19) 90% 100% 94% (45)

100% (26) 95% (21) 96% 100% 98% (47)

Number of lymph nodes is given in parentheses.

hilum vessels,” the sensitivity for the diagnosis of metastasis of melanoma was 100%, with only one false positive (positive predictive value 96%, specificity 95%) (Table III).

DISCUSSION BSS of enlarged lymph nodes is highly accurate in identifying lymph node metastases.26-28 Sonographicpathologic correlation showed that histologic characteristics determine the sonographic appearance. Tumor infiltration results in a round shape, loss of central echogenicity, and cortical widening. In contrast, reactively enlarged lymph nodes preserve the nodal architecture.21,29,30 BSS criteria for malignancy are size, shape, hypoechoic center (absence of a

“cockade”), and similar changes of parameters during follow-up examinations.4,31-33 BSS is a highly valuable tool as a noninvasive diagnostic procedure in melanoma clinics.23 There are few indications for computed tomographic scans or magnetic resonance imaging examinations.1-3,34 Positron emission tomography has no advantage over BSS in diagnosing lymph node metastases.35 Fine needle aspiration cytology has been described as a low-risk, accurate method to detect metastases of squamous cell carcinoma36 and of cutaneous melanoma.37,38 This study showed a sensitivity of 85% in evaluating lymph node shape (length/depth ratio < 2) with BSS. In addition, echogenicity of the lymph node border and center raises this value to

J AM ACAD DERMATOL VOLUME 41, NUMBER 5, PART 1

Moehrle et al 707

Fig 4. Color Doppler sonography. Same reactive lymph node (“LK”) as in Fig 2 in a sagittal section. Central perfusion pattern from hilum vessels of lymph node. Pronounced blood flow in femoral vein (“V FEM”).

Fig 5. Color Doppler sonography. Same lymph node metastasis of cutaneous melanoma as in Fig 3. Lack of visible perfusion within metastasis. At right border of lymph node there is a vessel that may be adjacent to lymph node.

100% (Table III). In only 69% of the lymph node metastases were we able to detect intranodal blood vessels with CDS. CDS revealed decreased intranodal perfusion in 83% of these lymph nodes. In total, 85% of the metastases showed a complete or partial loss of perfusion. The absence of hilus vessels and reduced perfusion support the diagnosis of a metastasis. In acute inflammation enlarged nodes maintain their architecture. Normal flow to the hilus of the node is preserved, and blood flow is increased because of the inflammation. Malignancy may invade, destroy, or compress the nodal architecture, resulting in a loss of the normal central blood flow. Doppler shift is dependent on Doppler angles, which cannot be determined exactly within the nodes. The PI and RI, however, are independent of such Doppler angles. PI is a function of systolic peak flow velocity, end-diastolic flow velocity, and also time-averaged maximum velocity, which is calculated from the envelope of the Doppler spectrum. This envelope is rather susceptible to artifacts.24,25 CDS visualizes vessels of a diameter of 0.1 mm and greater. Blood vessels cannot be detected in micrometastases or in weakly perfused, partially necrotic metastatic lesions. CDS improves the accuracy of differential diagnoses between reactively and metastatically enlarged lymph nodes. Specifically metastases of squamous cell carcinoma (SCC) of the head and neck have

been accurately detected by CDS. Higher PI and RI in SCC metastases than in reactive lymph nodes have been attributed to increased perfusion resistance. This may be caused by the smaller diameter of vessels in the tumor in conjunction with mechanical compression by tumor masses.13,14,16,18 Recently the assessment of vascular pattern and vascular density by CDS and power Doppler sonography has been valuable in the differentiation between metastatic and benign lymphadenopathies.11,18,19,21,23 The PI and RI in reactive lymph nodes that we report herein are consistent with the results of other studies.13,16,22 Metastases of cutaneous melanoma apparently have lower indices than those reported for lymph nodes with metastases of SCC, histiocytosis X, myeloproliferative syndrome, or malignant lymphomas in the literature.13,14,16,17,20-22 In this study RI (significantly) and PI (not significantly) differ in lymph node metastases of cutaneous melanoma and reactive lymphadenopathies. Because of the wide range of values and overlap between the two groups, metastasis cannot be excluded by RI and PI (Fig 1, B and C). Combining BSS criteria (shape, echogenicity of lymph node center) with a qualitative CDS criterion (absence of hilum vessels) we could enhance the sensitivity of noninvasive lymph node sonography to up to 100% in this limited series (Table III).

708 Moehrle et al

BSS and CDS are promising techniques for examining lymph nodes suspected of harboring macrometastatic but not micrometastatic melanoma. We suggest the following practical strategy based on the results of this study: 1. Estimation of lymph node size and shape with conventional BSS 2. Qualitative check of lymph node border and particularly of the center to obtain further information 3. Qualitative CDS criteria such as hilum vessels and a focal or total loss of perfusion to verify the diagnosis The recording of Doppler spectra in lymph node vessels and the calculation of PI and RI is time-consuming and appears to have no diagnostic value. REFERENCES 1. Prayer L,Winkelbauer H, Gritzmann N,Winkelbauer F, Helmer M, Pehamberger H. Sonography versus palpation in the detection of regional lymph-node metastases in patients with malignant melanoma. Eur J Cancer 1990;26:827-30. 2. Stutte H, Erbe S, Rassner G. Lymph node sonography in the after care of malignant melanoma. Hautarzt 1989;40:344-9. 3. Lohnert JD, Bongartz G, Wernecke K, Peters PE, Macher E, Brocker EB. Sensitivity and specificity of sonographic diagnosis of the lymph nodes in malignant melanoma. Radiologe 1988;28:317-9. 4. Tregnaghi A, DeCandia A, Calderone M, Cellini L, Rossi CR, Talenti E, et al. Ultrasonographic evaluation of superficial lymph node metastases in melanoma. Eur J Radiol 1997;24:216-21. 5. Blohmer JU, Chaoui R, Schmalisch G, Bollmann R, Lau HU. Differential breast tumor diagnosis by comparing blood circulation of the tumor with the contralateral breast using color coded, pulsed Doppler ultrasound. Geburtshilfe Frauenheilkd 1995;55:1-6. 6. Hamper UM, Sheth S, Abbas FM, Rosenshein NB, Aronson D, Kurman RJ. Transvaginal color Doppler sonography of adnexal masses: differences in blood flow impedance in benign and malignant lesions. Am J Roentgenol 1993;160:1225-8. 7. Hata K, Hata T, Manabe A, Kitao M. Ovarian tumors of low malignant potential: transvaginal Doppler ultrasound features. Gynecol Oncol 1992;45:259-64. 8. Kurjak A, Predanic M, Kupesic Urek S, Jukic S. Transvaginal color and pulsed Doppler assessment of adnexal tumor vascularity. Gynecol Oncol 1993;50:3-9. 9. Villena Heinsen C, Ertan AK, Tossounidis I, Hollander M, Konig J, Schmidt W. Diagnostic value of color Doppler ultrasound in breast tumors. Geburtshilfe Frauenheilkd 1995;55:541-7. 10. Swischuk LE, Desai PB, John SD. Exuberant blood flow in enlarged lymph nodes: findings on color flow Doppler. Pediatr Radiol 1992;22:419-21. 11. Tschammler A, Wirkner H, Ott G, Hahn D. Vascular patterns in reactive and malignant lymphadenopathy. Eur Radiol 1996;6: 473-80. 12. Giovagnorio F, Caiazzo R, Avitto A. Evaluation of vascular patterns of cervical lymph nodes with power Doppler sonography. J Clin Ultrasound 1997;25:71-6. 13. Steinkamp HJ, Maurer J, Cornehl M, Knobber D, Hettwer H, Felix R. Recurrent cervical lymphadenopathy: differential diagnosis with color-duplex sonography. Eur Arch Otorhinolaryngol 1994;251:404-9.

J AM ACAD DERMATOL NOVEMBER 1999

14. Tschammler A, Gunzer U, Reinhart E, Hohmann D, Feller AC, Muller W, et al. The diagnostic assessment of enlarged lymph nodes by the qualitative and semiquantitative evaluation of lymph node perfusion with color-coded duplex sonography. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1991; 154:414-8. 15. Tschammler A, Knitter J, Wittenberg G, Krahe T, Hahn D. Quantification of lymph node perfusion using color duplex ultrasonography. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1995;163:203-9. 16. Schreiber J, Mann W, Lieb W. Color duplex ultrasound measurement of lymph node perfusion: a contribution to diagnosis of cervical metastasis. Laryngorhinootologie 1993;72:187-92. 17. Steinkamp HJ, Rausch M, Maurer J, Hosten N, Schedel H, Langer R, et al. Color-coded duplex sonography in the differential diagnosis of cervical lymph node enlargements. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1994;161:226-32. 18. Wu CH, Chang YL, Hsu WC, Ko JY, Sheen TS, Hsieh FJ. Usefulness of Doppler spectral analysis and power Doppler sonography in the differentiation of cervical lymphadenopathies. AJR Am J Roentgenol 1998;171:503-9. 19. Wu CH, Shih JC, Chang YL, Lee SY, Hsieh FJ. Two-dimensional and three-dimensional power Doppler sonographic classification of vascular patterns in cervical lymphadenopathies. J Ultrasound Med 1998;17:459-64. 20. Chang DB,Yuan A,Yu CJ, Luh KT, Kuo SH,Yang PC. Differentiation of benign and malignant cervical lymph nodes with color Doppler sonography. AJR Am J Roentgenol 1994;162:965-8. 21. Na DG, Lim HK, Byun HS, Kim HD, Ko YH, Baek JH. Differential diagnosis of cervical lymphadenopathy: usefulness of color Doppler sonography. AJR Am J Roentgenol 1997;168:1311-6. 22. Choi MY, Lee JW, Jang KJ. Distinction between benign and malignant causes of cervical, axillary, and inguinal lymphadenopathy: value of Doppler spectral waveform analysis. AJR Am J Roentgenol 1995;165:981-4. 23. Omlor G, DillMuller D, Gross G, Kautz G, Schuder G, Zaun H, et al. Elective lymph node dissection (ELND) in malignant melanomas—the role of ultrasound. ZBL Chir 1996;121:469-73. 24. Gosling RG, King DH. Arterial assessment by Doppler-shift ultrasound. Proc R Soc Med 1974;67:447-9. 25. Pourcelot L. Applications cliniques de l’examen Doppler transcutané. Les Colloques de l’Institut National de la Santé et de la Recherche Médicale INSERM 1974;34:213-40. 26. Bruneton JN, Roux P, Caramella E, Demard F, Vallicioni J, Chauvel P. Ear, nose, and throat cancer: ultrasound diagnosis of metastasis to cervical lymph nodes. Radiology 1984;152:771-3. 27. Steinkamp HJ, Knobber D, Schedel H, Maurer J, Felix R. Palpation and sonography in after-care of head-neck tumor patients: comparison of ultrasound tumor entity parameters. Laryngorhinootologie 1993;72:431-8. 28. Leicher Duber A, Bleier R, Duber C, Thelen M. Cervical lymph node metastases: a histologically controlled comparison of palpation, sonography, and computed tomography. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1990;153:575-9. 29. Vassallo P, Edel G, Roos N, Naguib A, Peters PE. In-vitro high-resolution ultrasonography of benign and malignant lymph nodes: a sonographic-pathologic correlation. Invest Radiol 1993;28:698-705. 30. Maurer J, Willam C, Steinkamp HJ, Knollmann FD, Felix R. Keratinization and necrosis: morphologic aspects of lymphatic metastases in ultrasound. Invest Radiol 1996;31:545-9. 31. Sakai F, Kiyono K, Sone S, Kondo Y, Oguchi M, Watanabe T, et al. Ultrasonic evaluation of cervical metastatic lymphadenopathy. J Ultrasound Med 1988;7:305-10. 32. Steinkamp HJ, Cornehl M, Hosten N, Pegios W, Vogl T, Felix R.

J AM ACAD DERMATOL VOLUME 41, NUMBER 5, PART 1

Cervical lymphadenopathy: ratio of long- to short-axis diameter as a predictor of malignancy. Br J Radiol 1995;68:266-70. 33. Rainer T, Ofner G, Marckhgott E. Ultrasound diagnosis of regional lymph node metastasis of the neck in patients with headneck neoplasms: sono-morphologic criteria and diagnostic accuracy. Laryngorhinootologie 1993;72:73-7. 34. Prayer L, Winkelbauer F, Gritzmann N, Weislein H, Helmer M, Pehamberger H. Primary lymph node metastases in malignant melanoma studied via high-resolution real-time sonography— its value and indications. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1989;151:294-7. 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

Moehrle et al 709

lymph node metastases of malignant melanoma. Arch Dermatol 1995;131:1394-8. 36. van den Brekel MW, Stel HV, Castelijns JA, Croll GJ, Snow GB. Lymph node staging in patients with clinically negative neck examinations by ultrasound and ultrasound-guided aspiration cytology. Am J Surg 1991;162:362-6. 37. Schoengen A, Binder T, Faiss S, Weber L, Zeelen U. Fine needle aspiration cytology of metastatic malignant melanoma: improvement of results with ultrasound control. Hautarzt 1993; 44:703-7. 38. Fornage BD, Lorigan JG. Sonographic detection and fine-needle aspiration biopsy of nonpalpable recurrent or metastatic melanoma in subcutaneous tissues. J Ultrasound Med 1989; 8:421-4.