ultrasound-guided fine needle aspiration cytology prior to sentinel lymph node biopsy in melanoma...

Ultrasound-guided Fine Needle Aspiration Cytology prior toSentinel Lymph Node Biopsy in Melanoma Patients

Christiane Voit, MD,1 Martina Kron, PhD,2 Gregor Schafer, MD,1 Alfred Schoengen, MD,3

Heike Audring,1 Ansgar Lukowsky,1 Markus Schwurzer-Voit, MD,1 Wolfram Sterry, MD,1

Helmut Winter, MD,1 and Jurgen Rademaker, MD4

1Department of Dermatology of the Charite, Humboldt University, Berlin, Germany2Department of Biometry and Medical Documentation, University of Ulm, Ulm, Germany

3Department of Medical Oncology, Armed Forces Hospital, University of Ulm, Ulm, Germany4Department of Radiology, Memorial Sloan-Kettering Cancer Center, New York, New York, USA

Background: Sentinel lymph node biopsy (SLNB) allows early detection of metastases,thereby enabling early treatment in melanoma patients likely to benefit from adjuvant ther-apies. This prospective study analyzes the possible benefits of additional ultrasound (US) andfine needle aspiration cytology (FNAC) of sentinel nodes (SN) prior to SLNB.Method: Over a 2-year period 127 melanoma patients with 151 SN were scheduled for

SLNB. All SN were initially identified with lymphoscintigraphy, then identified and evaluatedby US and the cells aspirated for cytology (FNAC). US findings and FNAC results werecompared to surgical findings.Results: Of 127 patients, 114 had one SN each, 12 had two, and one had three. In vivo US

achieved a sensitivity of 79% (95% CI: 62–91%) and a specificity of 72% (95% CI: 62–81%).FNAC showed a sensitivity of 59% (95% CI: 41–76%) and a specificity of 100% (95% CI: 95–100%). The combination of these two in vivo methods achieved an overall sensitivity of 82%(95% CI: 65–93%) and an overall specificity of 72% [95% CI: 62–81%].Conclusion: Combined US and FNAC provides important information prior to SLNB in

that both procedures identify metastases in the lymph nodes (sensitivity > 80%). Patients withpositive FNAC may proceed directly to complete lymph node dissection (cLND) instead ofhaving initial SLNB. Thus, combined US and FNAC may prevent unnecessary anesthesia andsurgical management as well reduce costs. In our study 16% (19/121) fewer SLNB procedureswere carried out, subsequently replaced by cLND. For patients with a negative combination ofin vivo US and FNAC, SLNB remains the best diagnostic option.Key Words: Sentinel lymph node biopsy—Fine needle aspiration cytology—Ultrasound—

Melanoma.

In recent years, both the incidence of and themortality rate from melanoma have increased despite

great efforts to achieve early diagnosis.1 Theinvolvement of loco-regional lymph nodes is the mostimportant predictive factor for overall survival.2 A‘‘diagnostic’’ sentinel lymph node biopsy (SLNB)provides early identification of metastases in the ex-cised lymph node;3 this in turn justifies the ‘‘thera-peutic,’’ and more aggressive, complete lymph nodedissection (cLND).A detection rate of approximately 97% for sentinel

nodes (SN) is achieved by combining techniques ofradio-colloidal technetium-99m application, blue-

Received June 8, 2005; accepted January 23, 2006; publishedonline October 25, 2006Part of this work was presented in May 2003 at the Melanoma

session of the 39th Annual Meeting of the American Society ofClinical Oncology in Chicago, Illinois, and at the Oral MelanomaPoster Discussion session of the 40th Annual Meeting of theAmerican Society of Clinical Oncology in New Orleans, Louisiana.Address correspondence and reprint requests to: Christiane Voit,

MD; E-mail: [email protected]

Published by Springer Science+Business Media, Inc. � 2006 The Society ofSurgical Oncology, Inc.

Annals of Surgical Oncology, 13(12):1682–1689

DOI: 10.1245/s10434-006-9046-4

1682

dye-mapping, and intraoperative gamma-probe.4–6

Ultrasound (US) and fine needle aspiration cytology(FNAC) has been shown to identify local regionalmetastases earlier than physical examination with ahigh sensitivity and specificity during the follow-up ofmelanoma patients.7,8 FNAC has been proven to be apotent diagnostic tool with high sensitivity and evenhigher specificity,7,9–12 but it is only applied in spe-cialized medical centers.13

The purpose of this prospective study was toanalyze the possible advantages of US and FNACprior to SLNB for patients with melanoma. Morespecifically, our aim was to assess the sensitivity andreliability of identifying sentinel lymph nodes withUS and the ability to identify positive–involved–sentinel lymph nodes with US and FNAC. As partof this process, we investigated whether the SNidentified by lymphoscintigraphy was the same oneas that identified by US and whether the site of theSN indicated by US was identical to the siteapparent following lymphoscintigraphy. Sentinellymph nodes were examined with US and FNACand the findings were compared with the surgicalfindings from SLNB. A high diagnostic accuracy ofUS and FNAC would justify the use of both pro-cedures before SLNB and thereby allow somepatients to proceed directly to the therapeutic cLNDwithout undergoing SNLD.

PATIENTS AND METHODS

Patients

Our prospective study included 127 consecutivepatients who were scheduled for SLNB after theexcision of suspected melanoma between July 2001and August 2003. The institutional ethical reviewboard approved the study, and informed consent wasobtained from all patients enrolled. At the time ofenrolment, the new version of the American JointCommittee on Cancer (AJCC) staging system formelanoma was applied to our patient population.Patients� tumor data were not known in all casesprior to the in vivo examination, but in cases of pri-mary melanoma with a tumor depth of less than 1mm and without ulceration and/or regression, pa-tients were excluded from the study. Figure 1 sum-marizes the order of the examinations andinterventions in our study. Table 1 summarizes thepatients� clinical data.

Methods

All patients were scheduled for SLNB and exam-ined by US in the B-Mode and Power Doppler invivo before and also after lymphoscintigraphy in or-der to identify the location that had been proposedfor the SN by lymphoscintigraphy. Our aim in usingUS was to clearly depict the location of the suspectedSN prior to surgical treatment. Since patients withtruncal, head, or neck melanoma commonly drain tomore than one basin and marking with lymphoscin-tigraphy proved to be helpful, FNAC material wasonly aspirated during the second sonographic exam-ination after lymphoscintigraphy. The US findingswere passed on to the surgeon. Repeated FNAC tookplace to aspirate material from the suspected in vivo

Excision of PT (cutaneous melanoma)

Ultrasound of lymphatic basin and

contralateral side

Identification of SN by sonographically criteria,

depending on time schedule without the skin

markings of the lymphoscintigraphy

Lymphoscintigraphy

Ultrasound of lymphatic basin and

contralateral side

Identification of SN by sonographical criteria

after lymphoscintigraphy (with the marks of

the scintigraphy on patient´s skin) but without

hand held gamma probe.

FNAC for in vivo cytology

Histological data from the melanoma are

available

SLNB

Ultrasound of the excised sentinel ex vivo

FNAC for ex vivo

↓

↓

↓

↓

↓

↓

↓

FIG. 1. Flowchart of the procedures and interventions carriedout in this study. US, Ultrasonography; SN, sentinel node;

3FNAC, fine needle aspiration cytology; SNLB, sentinel lymphnode biopsy.

ULTRASOUND-GUIDED FINE NEEDLE ASPIRATION CYTOLOGY 1683

Ann. Surg. Oncol. Vol. 13, No. 12, 2006

SN prior to SLNB. In addition, the excised SN–de-tected intraoperatively by a combination of lympho-scintigraphy and blue-dye mapping as standardprocedure–was sonographically examined betweentwo gel stand-off pads following surgery and beforehistology. This was carried out in order to comparethe sonographic shape and measurements of the exvivo SN with the sonographic shape and measure-ments of the in vivo SN. See Fig. 1.

Classification via Ultrasound

Pre-operatively we performed a high-resolution USexamination of the lymphatic basin and the lym-phatic drainage of the tumor. All US examinationswere performed using the high-end device Technos(ESAOTE, Italy) equipped with three transducers ofbetween 3.5 and 14 MHz (B-mode, 30 pictures persecond, color Doppler, Power Mode). The lymphnode was measured, classified as benign (b), suspectfor benignity (sfb), suspect for malignancy (sfm) ormalignant (m). The region was always examined incomparison to the contralateral side.

Fine Needle Aspiration in Detail

FNAC was performed with a hand-held ‘‘Binder’’-valve, which provides an especially short distancebetween the button for initiation of aspiration andthe region of interest. This makes it possible toaspirate even very small targets without losing con-tact with the lesion in the process. US-guided FNACuses an alcoholic fluid as a conductor medium, thusminimizing the danger of infection. The fine needlefor superficial lymph nodes has a diameter ofapproximately 0.4 mm (26G). For deeper lymphnodes (depth > 25 mm) a 22 G lumbar punctureneedle is used. The negative pressure for aspiration isperformed with a 20-ml syringe by fixing the plungerat the 10 ml position, thereby creating an approxi-mate negative pressure of about )300 cm H2O. Weperformed numerous aspirations under sonographicguidance to receive multiple smears for cytodiag-nostic evaluation. A smear was considered to betechnically efficient if it contained approximately 100cells. FNAC procedures performed in small targets,such as intranodal areas within a sentinel node, witha needle diameter of only 0.4 mm often obtain asmaller number of cells and thus tend to give�unrepresentative� results. In order to deliver repre-sentative results, multiple and repeated FNAC pro-cedures must be performed.

Postsurgical Evaluation (Post-SLNB)

Directly after SLNB the excised node was studiedby US between two gel-pads of 10-mm thickness toachieve a high focus within the lymph node (Fig. 1).Again we documented the diameter and volume andcompared the structure of the parenchyma to thatshown in the in vivo pictures. The malignancy wasassessed and classified again in vitro as b, sfb, sfm, orm (see above). As a result we were able to determine

TABLE 1. Summary of the clinical data

Patients

GenderFemale 66 (52%)Male 61 (48%)

Age (years)Median (min–max) 60 (20–88)Median (min–max) 2.1 (0.4–18.0)

Tumor Breslow depth mm)< 1 mm 29 (23%)‡ 1 to < 2 mm 30 (24%)‡ 2 to < 4 mm 32 (25%)‡ 4 mm 34 (27%)Missing 2 (1%)

Primary tumor siteHead, neck 14 (11%)Limbs 61 (48%)Trunk 51 (40%)Unknown 1 (1%)

Type of histologya

SSM 82 (65%)NM 30 (24%)LMM 2 (2%)ALM 5 (4%)Others 8 (6%)

UlcerationNo 72 (57%)Yes 38 (30%)Threatening 15 (12%)Missing 2 (1%)

RegressionNo 75 (59%)Yes 47 (37%)Threatening 3 (2%)Missing 2 (1%)

Clark LevelII 1 (1%)III 49 (39%)IV 65 (51%)V 10 (8%)Missing 2 (1%)

AJCC stagingIa 26 (20%)Ib 20 (16%)IIa 18 (14%)IIb 21 (17%)IIc 6 (5%)IIIa 12 (9%)IIIb 11 (9%)IIIc 13 (10%)

a SSM, Superficial spreading melanoma; NM, nodular mela-noma; LMM, Lentigo maligna melanoma; ALM, acrolentiginousmelanoma.

C. VOIT ET AL.1684


whether the excised SN coincided with the one seenand punctured in vivo (Fig. 2). The SN was thensliced into 1-mm thick sections and evaluated histo-logically. Standard techniques, including those forHMB-45 and Melan-A, were performed.

Patterns in Ultrasound

To estimate criteria for the SN we evaluated thenode�s size and shape in the expected lymphatic basinin comparison to circumjacent nodes. A more sym-metrical broadening of the parenchymal periphery, acap-like structure where the afferent lymphatic vesselsenter, and a certain pattern of vascularization (rathercentral) as described elsewhere have been found to berelatively typical patterns for the SN.14,15 Our expe-rience suggests that involved SN show a moreasymmetrical broadening of the periphery (oftenhump-like) with a localized hypervascularization inthe periphery and/or echo-free islands within thenode and/or a slight increase in vascularization. Thenon-sentinel lymph nodes in the depth of the subcu-taneous tissue usually present an iso-echoic structure,so that they only rarely can be discriminated from thesurrounding fatty tissue. In most cases a non-malig-nant lymph node has a Vasallo-index > 2.16,17

Statistics

To assess diagnostic value separately for US andFNAC, sensitivity, specificity, positive and negativepredictive values with corresponding 95% confidence

intervals (95% CI) were calculated. Additionally,sensitivity, specificity, and predictive values werecalculated for the combination of US and FNAC bydefining a positive test as a test that is positive in atleast one of the two methods. The statistical analyseswere performed with the Statistical Analysis SystemVersion 8.2 (SAS Institute, Cary, N.C.).

RESULTS

Median tumor Breslow depth was 2.1 mm, andmost patients had Clark level III (39%) or IV (51%)(Table 1). One hundred forty-one SN were evaluatedin 127 patients; of these, 114 patients each had onesentinel lymph node evaluated, 12 patients had two,and 1 patient had three. Three patients presentedwith palpable regional lymph nodes. In 121 of the 141sentinel nodes evaluated ex vivo we found that the invivo and ex vivo sonographic examinations matched.In one case, neither the US nor surgery was able todetect any lymph node, leading to a unique matchingresult. Five SN were not seen on the in vivo US, orwere post-sentinels, i.e. second-echelon lymph nodeswith no signs of malignancy and therefore were notsubject to further evaluation. For an additional 14SN, the in vivo and ex vivo sonographic examina-tions did not match, but none of these SN was his-tologically malignant.Certain patterns on US, such as asymmetrical

broadening of the parenchymal periphery (oftenhump-like), a cap-like structure, and a certain patternof vascularization, are considered to be typical formetastasis in sentinel lymph nodes14,18 (Fig. 3).When we compared the in vivo US evaluation withthe histological one (Table 2), 34 of the 121 in vivoand ex vivo matching nodes were histologicallymalignant or suspicious for malignancy (30 m, 4 sfm,respectively). For those SN classified as merely sus-picious for malignancy based on histology, very fewcells (< 10) could be detected by histopathology. Ofthese 34 nodes, 27 had already been evaluated asmalignant or suspicious for malignancy based on thein vivo US examination, indicating that in vivo UShad a sensitivity of 79% (95% CI: 62–91%). Six of the34 nodes were classified as ‘‘possibly benign’’ (sfb)based on the US, and four were histologically clas-sified as being merely suspicious for malignancy,meaning that very few cells (< 10) could be detectedby histopathology. Of the 121 lymph nodes withmatching in vivo and ex vivo US results, 87 werehistologically negative (87 benign and 0 sfb). In 63 of87 cases, the US examinations had already classified

FIG. 2. Technical arrangement of the ex vivo ultrasound of thesentinel node. The excised SN and adjacent tissue are shownbetween stand-off gel pads directly after excision and before ex vivoUS examination.



the lymph nodes as uninvolved (37 benign and 26sfb), resulting in a specificity of 72% (95% CI: 62–81).The positive predictive value was 53% (95% CI: 38–67%), and the negative predictive values was 90%(95% CI:80–96%).The results of in vivo cytology were found to be

valuable, i.e. an adequate specimen could be obtainedin 105 of the 121 nodes (Table 2). The remainingcytological smears were not valuable due to a limitednumber of cells or non-representative cells. Histologyof the SN was negative in 73 of 105 cases (73 benign

and 0 sfb), of which 73 had already been proven to benegative by the cytological analysis, thereby reflectinga specificity of 100% (95% CI: 95–100%) for in vivocytology when an adequate specimen was obtained.Sensitivity was only 59% (95% CI: 41–76%) for invivo cytology. Of the 32 SN found to be positive byhistological evaluation (28 malignant and 4 sfm), 19were also identified as being positive by in vivocytology (FNAC). The positive predictive value was100% (95% CI: 82–100%), and the negative predictivevalue was 85% (95% CI: 76–92%).

FIG. 3. Comparison between the patterns of the in vivo and ex vivo SN. (Two cases, surgically confirmed).

C. VOIT ET AL.1686


The combination of US and FNAC of in vivo SNcorrectly classified 28 of the 34 histologically provenmalignant/sfm cases as being positive, thus achievinga sensitivity of 82% (95% CI: 65–93%) (Table 2). Thiscombination of US and FNAC also correctly iden-tified 63 of 87 histologically proven benign SN as notinvolved, reflecting a specificity of 72% (95% CI: 62–81%). The positive predictive value was 54% (95% CI:39–68%), and the negative predictive value was 91%(95% CI: 82–97%).By restricting our analysis to all histologically

clearly malignant SN and taking all suspect US

findings (sfb, sfm, and the malignant findings)together as one group, we would achieve a sensi-tivity of 100% (30 nodes; 95% CI: 88–100%) butonly gain a worse specificity of 41% (37 of 90; 95%CI: 31–52%). Only one histologically sfm SN wasclassified as sonographically benign. We did not seebleeding or infection or any other adverse side ef-fects from the FNAC, and the procedure was welltolerated by the patients. To date, we have not ob-served any spreading of tumor cells along the fineneedle tract and have not found any recurrencesrelated to the FNAC procedure. The median volumeof SN in vivo was 842 mm3, while that of SN exvivo was 714 mm3. Volume measurements of lymphnodes tend to be greater when taken in vivo becausethe surrounding fatty tissue is more often includedin the measurement, thus artificially augmenting thenumbers.

DISCUSSION

To our knowledge, the study reported here is thefirst prospective study to include in vivo examinationsof the SN in melanoma patients. Malignant mela-noma of the skin primarily metastasizes to the re-gional lymph nodes.19 It is therefore important tohave accurate tools for detecting relapse as early aspossible. Once a lymph node is involved, the 5- or 10-year survival rate drops dramatically.2,20

It has been shown that the status of the SN afterSLNB is the most important single prognostic fac-tor.21

For physicians unfamiliar with ultrasound, it maybe difficult to imagine that the SN could show adifferent sonographic-morphologic pattern, as wasdemonstrated in another study.14 During immuno-logic processes such as infection or peripheral tumorgrowth, a lymph node loses its inactive structure andproceeds to a reactive state. This transformationtends to be reversible as long as no neo-vasculariza-tion or (micro-) metastases exist within the node. Theshape of the SN tends to be oval, and it lays down anecho-poor peripheral band, much like the pulparound the stone of a cherry. The hilum reveals in-creased echoes. An increase in blood perfusionthroughout the whole node and partial hyper-vascu-larization within the periphery in the power modemay be identified, and should be taken signaling thelocation of a process in the parenchyma that is sus-picious for a malignancy. A balloon-shaped paren-chyma is always highly suspicious for malignancy,especially if it contains the above-mentioned local

FIG. 4. Typical patterns of the lymph node suspected to be aninvolved SN (two different cases, surgically confirmed). The nodesreveal asymmetrical broadening of the parenchyma, asymmetricalhypervascularization, and partial loss of the hilum.



hyper-perfusion. The maximum asymmetrical cortexthickness, as shown in a recent breast cancer study,22

combined with irregular hyper-vascularization hasturned out to be the most important feature forpredicting sentinel node involvement. There is nocorrelation between size and assessment of malig-nancy of a lymph node.8

Due to the fact that melanoma cells are scattered ina disconnected fashion within the lymph node, thesecells are best suited for the aspiration process, and anamazingly high specificity has been achieved so far.The specificity of FNAC is 100%, which indicatesthat the probability of producing false positive find-ings is nearly zero. False positive results were notproduced, even when nests of nevus cells within thelymph node were found (in 9/141 excised SN).7,23–27

The relatively low sensitivity of FNAC can beexplained by the small size of our targets, whichsometimes measure only about 3 · 3 mm. Not onlythe entire sentinel but sometimes a small area withinthe node is the �region of interest.� In order to be ableto examine the cells within such small areas, we per-formed FNAC repeatedly.In four cases histology revealed only a very small

number of melanoma cells, or even just one mela-noma cell. We documented these cases under thehistologically positive sentinels, and the patients werescheduled for cLND.US is very useful in treatment planning as it pro-

vides the surgeon with information about surround-ing structures, blood vessels, and suspicious findingsin additional lymph nodes. It is best if the surgeonschedules the US examination directly prior to SLNBas this will provide information on the in vivo con-ditions and the extent of agreement with the skinmarking of the scintigraphy. If the SN reveals a po-sitive cytology, the SLNB can be replaced by a LND,thus saving a surgical procedure as well as time.

US cannot and has never been intended to replacethe established method of SN identification with acombination of lymphoscintigraphy and blue-dyemapping, since it is not able to achieve as high assensitivity.28 Furthermore, it was not our intention tosuggest replacing SLNB with FNAC. The importantpoint we wish to make–based on the results of thisstudy–is the potential of high-resolution US andFNA for identifying a subgroup of patients who donot need to undergo surgical SN biopsy but can in-stead undergo high-resolution US and fine needleaspiration. Our data only show that in a selectednumber of patients it is justified to proceed directlyfrom FNAC to cLND. In our study, 16% (19/121) ofthe SLNB procedures could be omitted; these wereimmediately replaced by LNDs. At our institution(Charite, Berlin), US and FNAC are now performedas regular staging procedures before SLNB, andpatients with positive FNAC directly undergo ther-apeutic LND. Using the gamma probe during in vivoUS could be helpful and will be tested in furtherstudies.

ACKNOWLEDGMENTS

Supported by Deutsche Krebshilfe (grant: 70-2791-Vo I).

REFERENCES

1. Beddingfield FC III. The melanoma epidemic: res ipsa loqui-tur. Oncologist 2003; 8:459–65.

2. Buzzell RA, Zitelli JA. Favorable prognostic factors in recur-rent and metastatic melanoma. J Am Acad Dermatol 1996;34:798–803.

3. Morton DL, Wen DR, Wong JH, Economou JS, Cagle LA,Storm FK, et al. Technical details of intraoperative lymphaticmapping for early stage melanoma. Arch Surg 1992; 127:392–9.

TABLE 2. Sensitivity and specificity of US and FNAC in combination

Histology:malignant Histology: benign

PositiveaNega-tivea Positive

Nega-tive

Sensitivity Specificity PPVc NPVd

M sfm b sfb M sfm b sfb 95% CIb 95% CI 95% CI 95% CI

US 22 5 1 6 3 21 37 26 79% (62;91) 72% (62;81) 53% (38;67) 90% (80;96)FNAC 17 2 9 4 0 0 49 24 59% (41;76) 100% (95;100) 100% (82;100) 85% (76;92)US and FNAC combined 28 6 24 63 82% (65;93) 72% (62;81) 54% (39;68) 91% (82;97)

a b, benign; sfb, suspect for benignity; sfm, suspect for malignancy; m, malignant.b CI, Confidence interval.c PPV, Positive predictive value.d NPV, Negative predictive value.

C. VOIT ET AL.1688


4. Krag DN, Meijer SJ, Weaver DL, Loggie BW, Harlow SP,Tanabe KK, et al. Minimal-access surgery for staging ofmalignant melanoma. Arch Surg 1995; 130:654–8.

5. Albertini JJ, Cruse CW, Rapaport D, Wells K, Ross M, De-Conti R, et al. Intraoperative radio-lympho-scintigraphy im-proves sentinel lymph node identification for patients withmelanoma. Ann Surg 1996; 223:217–24.

6. Balch CM, Buzaid AC, Atkins MB, Cascinelli N, Coit DG,Fleming ID, et al. A new American Joint Committee onCancer staging system for cutaneous melanoma. Cancer 2000;88:1484–91.

7. Rossi CR, Seno A, Vecchiato A, Foletto M, Tregnaghi A, DeCandia A, et al. The impact of ultrasound scanning in thestaging and follow-up of patients with clinical stage I cutane-ous melanoma. Eur J Cancer 1997; 33:200–3.

8. Voit C, Mayer T, Kron M, Schoengen A, Sterry W, Weber L,et al. Efficacy of ultrasound B-scan compared with physicalexamination in follow-up of melanoma patients. Cancer 2001;91:2409–16.

9. Perry MD, Seigler HF, Johnston WW. Diagnosis of metastaticmalignant melanoma by fine needle aspiration biopsy: a clinicaland pathologic correlation of 298 cases. J Natl Cancer Inst1986; 77:1013–21.

10. Fornage BD, Lorigan JG. Sonographic detection and fine-needle aspiration biopsy of nonpalpable recurrent or meta-static melanoma in subcutaneous tissues. J Ultrasound Med1989; 8:421–4.

11. Layfield LJ, Ostrzega N. Fine needle aspirate smear mor-phology in metastatic melanoma. Acta Cytol 1989; 33:606–12.

12. Schoengen A., Binder T, Schiffelholz W, Schulz PC, Zeelen U.Fine-needle aspiration guided by ultrasound in suspectedcancer. Onkologie 1994; 17:420–6.

13. Voit C, Schoengen A, Schwurzer-Voit M, Weber L, Ulrich J,Sterry W, et al. The role of ultrasound in detection and man-agement of regional disease in melanoma patients. Semin Oncol2002; 29:353–60.

14. Kahle B, Hoffend J, Wacker J, Hartschuh W. Preoperativeultrasonographic identification of the sentinel lymph node inpatients with malignant melanoma. Cancer 2003; 97:1947–54.

15. Voit C, Schaefer G, Schoengen A, Schwurzer-Voit M, MayerT, Audring H, Lukowsky A, Winter H, Hasert R, Sterry W,Kron M. Ultrasound, in vivo/in vitro, cytology and tyrosinaseRT-PCR of the sentinel node in melanoma. J Clin Oncol ProcAm Soc Clin Oncol 2003; 22:16a.

16. Vassallo P, Wernecke K, Roos N, Peters PE. Differentiation ofbenign from malignant superficial lymphadenopathy: the roleof high-resolution US. Radiology 1992; 183:215–20.

17. Vassallo P, Edel G, Roos N, Naguib A, Peters PE. In-vitrohigh-resolution ultrasonography of benign and malignant

lymph nodes. A sonographic-pathologic correlation. InvestRadiol 1993; 28:698–705.

18. Voit C, Kron M, Schoengen A, Audring H, Lukowsky A,Sterry W, Winter H, Hasert R, Mayer T, Schaefer G. In vivoand ex vivo ultrasound, fine needle aspiration and molecularbiology of Sentinel nodes (SN) in melanoma patients. J ClinOncol 2004; 22:712a.

19. Ghossein RA, Carusone L, Bhattacharya S. Molecular detec-tion of micrometastases and circulating tumor cells in mela-noma prostatic and breast carcinomas. In Vivo 2000; 14:237–50.

20. Balch CM, Cascinelli N, Sim FH, Soong S-J, Taylor WF,Bufalino R. Elective lymph node dissection: Results ofprospective randomized surgical trials. In: Balch CM,Houghton AN, Sober AJ, Soong S-J, eds. Cutaneous Mela-noma, Third Edition. St. Louis: Quality Medical Publishing,Inc., 1998:209–25.

21. Gershenwald JE, Colome MI, Lee JE, Mansfield PF, Tseng C,Lee JJ, et al. Patterns of recurrence following a negative sen-tinel lymph node biopsy in 243 patients with stage I or IImelanoma. J Clin Oncol 1998; 16:2253–60.

22. Deurloo EE, Tanis PJ, Gilhuijs KG, Muller SH, Kroger R,Peterse JL, et al. Reduction in the number of sentinel lymphnode procedures by preoperative ultrasonography of the axillain breast cancer. Eur J Cancer 2003; 39:1068–73.

23. Voit C, Mayer T, Proebstle TM, Weber L, Kron M, Krup-ienski M, et al. Ultrasound-guided fine-needle aspirationcytology in the early detection of melanoma metastases. Cancer2000; 90:186–93.

24. Engzell U, Esposti PL, Rubio C, Sigurdson A, Zajicek J.Investigation on tumour spread in connection with aspirationbiopsy. Acta Radiol Ther Phys Biol 1971; 10:385–98.

25. Fornari F, Civardi G, Cavanna L, Di Stasi M, Rossi S, SbolliG, et al. Complications of ultrasonically guided fine-needleabdominal biopsy. Results of a multicenter Italian study andreview of the literature. The Cooperative Italian Study Group.Scand J Gastroenterol 1989; 24:949–55.

26. von Schreeb T, Arner O, Skovsted G, Wikstad N. Renal ade-nocarcinoma. Is there a risk of spreading tumour cells indiagnostic puncture?. Scand J Urol Nephrol 1967; 1:270–6.

27. Voit C, Kron M, Mayer T, Proebstle TM, Schwurzer-Voit M,Weber L, Mooser G, Zeelen U, Sterry W, Schoengen A.Ultrasound guided fine needle aspiration cytology (FNAC) forearly verification of melanoma. Proc Am Soc Clin Oncol J ClinOncol 2001; 21:360a.

28. Hocevar M, Bracko M, Pogacnik A, Vidergar-Kralj B, BesicN, Zgajnar J, et al. The role of preoperative ultrasonography inreducing the number of sentinel lymph node procedures inmelanoma. Melanoma Res 2004; 14:533–6.



ultrasound-guided fine needle aspiration cytology prior to sentinel lymph node biopsy in melanoma...

Documents