Inguinal Node Status by Ultrasound in Vulva Cancer

Gynecologic Oncology 77, 93–96 (2000) doi:10.1006/gyno.1999.5702, available online at http://www.idealibrary.com on

Inguinal Node Status by Ultrasound in Vulva Cancer 1 D. K. Abang Mohammed, M.B., Ch.B., 2 R. Uberoi, FRCR, A. de B. Lopes, MRCOG, and J. M. Monaghan, FRCOG/FRCS (Ed) Gynaecological Oncology Centre/Radiology Department, Queen Elizabeth Hospital, Gateshead NE9 6SX, United Kingdom Received August 11, 1999

were 97 and 85% for FIGO stage I and II disease, respectively [2]. Groin dissections in these early cancers are carried out for diagnostic purposes with no known therapeutic benefit and are associated with significant morbidity [3, 4]. Wound morbidity is reported to be worse in en bloc (68%) versus triple incision (37%) and in bilateral (37%) versus unilateral dissection (0%). Conservative and unilateral groin surgery for vulva cancer appears safe in cases selected based on the primary tumor characteristics and location [5, 6]. Preoperative assessment of node status has the potential advantage of being a precise and direct method of selecting cases for groin dissection. The isolation of the sentinel node appears promising with the use of patent blue dye and radioisotope injection but requires an excisional procedure [7–9]. Less invasively, ultrasound has been used to assess nodal status in head and neck tumors, breast cancers, upper aerodigestive tract malignancy, and cervical cancers [10 –14]. The accuracy of ultrasound ranges between 67 and 95%, varying according to the different parameters examined, which included short axis diameter, long axis diameter, long axis to short axis (L/S) ratio, shape, nodal vascular pattern, echogenic pattern, and regularity of nodal outline. Fine-needle aspiration (FNA) has been used to supplement the ultrasonic node status in some of these studies [10, 12]. These studies share the common theme of staging a tumor preoperatively to decide on the optimum treatment providing the best outcome with minimum morbidity. This study used ultrasound to predict inguinal node metastases in women undergoing groin dissection for vulva cancer. Two sonographic criteria were selected to predict nodal metastases: short axis diameter greater than 8 mm (⬎8 mm) and a long axis-to-short axis ratio less than or equal to 2 (L/S ⱕ 2). Cutoff levels of between 5 and 10 mm for the short axis diameter have been described by previous authors [10, 12, 15] and we elected an intermediate value of 8 mm for our study. The use of an L/S ratio ⬍2 appears a more consistent parameter in previous studies [11, 15]. The aim of our study was to test the accuracy of these two sonographic criteria in predicting nodal status and, therefore, the diagnostic role of ultrasound in the preoperative assessment of groins in vulva cancer.

Objective. The objective of this study was to determine the value of ultrasound in preoperative assessment of groin node status in vulva cancer. Materials and methods. Women with clinically uninvolved groins who were undergoing groin node dissection for vulva cancer in our department over an 18-month period were recruited into the study. A preoperative scan of each groin to be dissected was performed to identify any suspicious lymph nodes containing metastases. Suspicious nodes were defined by two sonographic criteria: short axis diameter (>8 mm) and a long axis/short axis ratio (L/S < 2). Each suspicious node was sampled by ultrasoundguided fine-needle aspiration (FNA). Results. Twenty women, with an average age of 70 years, consented to the study. Seventeen had bilateral groin node dissection and three had unilateral groin node dissection. Six (16%) of the seventeen dissected groins contained metastases. Short axis had a better overall accuracy (89%) but failed to detect a singular micrometastasis. The L/S ratio identified all positive groins but had a high false-positive rate (62%) and an overall accuracy of 67%. The combination of both criteria did not improve the overall accuracy when compared with the individual criterion. FNA was not diagnostic in three, representative in two, and falsely negative in one. Conclusion. Although L/S ratio has a lower overall accuracy, it correctly identified all groins with metastases. This has a great impact on treatment and prognosis. Its high false-positive rate may be improved by more diagnostic FNA. These sonographic criteria show good potential for segregating those with groin metastases requiring surgical treatment from those with uninvolved nodes. This experience has to be expanded to prove its clinical effectiveness. © 2000 Academic Press Key Words: vulva cancer; inguinal lymph node; ultrasound.

INTRODUCTION More than two thirds of vulva cancers present early and have a good prognosis in the absence of nodal involvement [1–3]. Corrected 5-year survival rates published from our department 1

Poster presentation 7– 8 May 1999, British Gynaecological Cancer Society (BGCS), Liverpool, United Kingdom. Poster presentation 26 –30 September 1999, 7th Biennial IGCS Meeting, Rome, Italy. 2 To whom correspondence should be addressed. Fax: 0191-482 5604. 93

0090-8258/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

94

MOHAMMED ET AL.

TABLE 1 FIGO Stages of Study Population FIGO stage

Number

Ib II III IV Recurrence

9 5 3 2 1

Total

20

MATERIALS AND METHODS Women referred with vulva cancer to the Gynaecological Oncology Centre, Gateshead, during a 20-month period in 1997–1999 were recruited into the study. The study had approval from the Local Research Ethics Committee (LREC). Women with grossly involved (fixed or ulcerating) groin nodes were excluded. A preoperative ultrasound assessment of the groin(s) was performed by a single radiologist (R.U.). With a 7.5-mHz linear array (Acuson XP10) the groin adjacent to the femoral vessels was scanned from just below the profunda femoris bifurcation to just above the inguinal ligament in longitudinal and transverse planes. All lymph nodes were assessed for size, looking at the maximum length and taking the short axis tangential to this across the maximum diameter of the node. This gave the L/S ratio and the maximum short axis for the lymph node. Lymph nodes that were suspicious on both criteria were then aspirated under ultrasound guidance by the radiologist, using a 21-gauge Sheba needle for cytological reporting. The cytological categories were: suspicious for malignancy, no evidence of malignancy, and nondiagnostic specimen. The histological status of the lymph nodes was correlated to the two sonographic criteria used, separately as well as in combination. The cytological findings of the FNA were also compared with the nodal histology. RESULTS Of the 26 women recruited, 20 gave informed consent to the study. Of these, 17 had bilateral and 3 had unilateral groin dissection. Average age of the study group was 70 years (range: 42– 88). The tumor histology was squamous cell in 18 women, adenoid cystic carcinoma of Bartholin’s gland in one woman, and malignant melanoma in one woman. One woman who had unilateral surgery performed 4 years previously presented with a contralateral tumor which was staged as a recurrence with impalpable inguinal nodes. The FIGO staging is given in Table 1. There were 13 T1 tumors (ⱕ2 cm in largest diameter) and seven T2 (⬎2 cm in largest diameter). Two of six groins with

palpable, but not obviously involved, lymph nodes had metastasis on histology. In seven groins, the lymph nodes were too small to be visualized by ultrasound and all had negative histology. For the analysis, their short axis diameters were taken as less than 8 mm and the L/S ratios could not be determined. This accounted for the discrepancy in the total number of groins when comparing the two sonographic criteria. Of the total of 37 groins, 6 (16.2%) were positive for metastasis on histology. One of these contained a single intracapsular metastasis measuring 8 mm on histology. Its short axis diameter measured less than 8 mm on ultrasound with a rounded configuration (L/S ratio ⬍ 2). Table 2 compares the histological node status of each groin with the short axis diameter, the L/S ratio, and the two criteria in combination. When both criteria were used, a lymph node was considered to be suspicious when positive on both criteria. There were 8 groins in which both criteria were positive, 8 groins in which only the L/S ratio was positive, and 14 groins in which both criteria were negative. The sensitivity, specificity, positive predictive value, negative predictive value, falsepositive rate, false-negative rate, and accuracy of the two criteria separately as well as in combination are also given.

TABLE 2 Sonographic Criteria, Nodal Histology, and Summary of Statistics Nodal histology

Short axis ⬎8 mm ⱕ8 mm N ⫽ 37

Positive

Negative

5 1 6

3 28 31

Sensitivity ⫽ 83.3, specificity ⫽ 90.3, PPV ⫽ 62.5, NPV ⫽ 96.6, false positive ⫽ 37.5, false negative ⫽ 3.4, accuracy ⫽ 89.2 L/S ratio ⱕ2 ⬎2 N ⫽ 30

6 0 6

10 14 24

Sensitivity ⫽ 100, specificity ⫽ 58.3, PPV ⫽ 37.5, NPV ⫽ 100, false positive ⫽ 62.5, false negative ⫽ 0, accuracy ⫽ 66.7 Short axis and L/S ratio Both positive One/both negative N ⫽ 30

5 1 6

3 21 24

Sensitivity ⫽ 83.3, specificity ⫽ 87.5, PPV ⫽ 62.5, NPV ⫽ 95.5, false positive ⫽ 37.5, false negative ⫽ 4.5, accuracy ⫽ 86.7

INGUINAL NODE STATUS BY ULTRASOUND IN VULVA CANCER

The detection of the occult metastasis by the L/S ratio resulted in a 100% sensitivity when compared with that by the short axis diameter. However, the L/S ratio had a poorer specificity, false-positive rate, and overall accuracy. In 7 of the 10 false-positive groins, L/S ratios less than 2 were found in lymph nodes measuring 8 mm or less. In comparison, only 1 of the 6 positive groins with an L/S ratio less than 2, measured less than 8 mm. Conversely, the short axis diameter demonstrated better specificity, positive predictive value, and overall accuracy. However, the sensitivity and false negative value were affected by the occult metastasis. When both criteria were imposed, the accuracy was comparable to that of the short axis diameter alone. The occult metastasis was missed as only one of the criteria identified the node as suspicious. Of the eight groins suspicious on both criteria, six had FNA and one woman with bilateral suspicious groins refused the procedure. The FNA proved to be of poor yield in three of the samples. Two of the three diagnostic cytology results were representative. The remaining cytology sample which was falsely negative had pus aspirated from the lymph node. DISCUSSION Groin node dissection is of therapeutic value in the presence of metastases. In selected early stages of vulva cancer, groin dissections have become less radical with significant reduction in morbidity and hospital stay [3, 4]. The preoperative identification of women without nodal metastasis resulting in groinsparing surgery would prevent the frequently described morbidity of wound breakdown, infection, and lymphocyst formation, as well as secondary lower limb lymphedema. The gold standard for determining groin nodal status is histology. A nonsurgical method has to match the surgical histology as a measure of its accuracy, ideally with no false negatives and minimum false positives. The overall accuracy of ultrasound in predicting metastasis in our study was best achieved at 89% by the short axis diameter criterion. This was higher when compared with work relating to squamous carcinoma of the head and neck which reported an accuracy of 83% [15]. Using a 5-mm cutoff in axillary nodes, the accuracy was 68%, and with a 10-mm cutoff in upper aerodigestive tract, the accuracy was 81% [10, 12]. These comparisons may not be justified as our numbers are small in comparison and the normal lymph node sizes may also vary in different anatomical sites. In our study group the short axis diameter appears to be best at detecting negative nodes but performed less well in identifying all positive groins. In clinical terms, as in our case, a single occult metastasis requires surgical excision to achieve a good prognosis. The fact that lymph node metastasis has been described in lymph nodes as small as 4 mm [16] is liable to make this an unreliable criterion if used on its own. In a study involving head and neck tumors, nodal width was found to be

95

the best predictor of metastases, and the authors recommended that in conjunction with poor echogenicity, lymph nodes as small as 3 mm be aspirated [17]. In comparison, the L/S ratio in our study had a poorer overall accuracy of 67% but an excellent sensitivity and negative predictive value. This is in contrast to the work done on cervical node assessment which had an accuracy of 88 –95% [15, 18]. The occult metastasis, falsely negative on its short axis diameter, was detected by the L/S ratio. It suggests that the shape of a lymph node may be the first to be altered in early tumor infiltration. The rounded shape, however, was also seen in benign nodes, resulting in a high false-positive rate (62%) for the L/S ratio. All of the seven groins with no visualized nodes and therefore indeterminate L/S ratios had reassuringly negative histology. It will be interesting to see if this remains a consistent feature with an increasing number of cases and with different radiologists doing the scans. When a suspicious node was defined on both criteria being positive, the occult metastasis was not included. However, if a suspicious node was identified on either criterion being positive, the findings would be identical to those of the L/S ratio. Although the L/S ratio has a high false-positive rate, the false-negative rate is eliminated. This false-positive rate may be reduced by more diagnostic FNA. Several studies [10, 19] reported that FNA had positively contributed to their pickup rates. In our study, three of six FNAs performed were unsuitable for diagnosis. The diagnostic rate of FNA may be improved by a fresh specimen interpretation instead of an air-dried specimen. This also has the benefit of allowing for a nondiagnostic sample to be repeated. In lymphoma, midsized-needle biopsy (MNB) as a diagnostic technique has been described to be more representative when compared with FNA [20]. Further work is required to determine the contribution of FNA to groin node assessment. Our preliminary data would suggest that patients with ultrasound-undetectable inguinal lymph nodes (19%) and those with an L/S ratio greater than 2 (38%) can be safely omitted from having groin node surgery. In these 57% of cases follow-up by ultrasound would be appropriate. With improved FNA technique one may confirm the benign or malignant status of lymph nodes with one or both identified sonographic criteria. If not, then diagnostic groin dissection will have to be the standard in this subgroup of patients, until other preoperative investigations such as the sentinel node biopsy proves successful. Ultrasound is safe, noninvasive, and highly acceptable to patients. The two sonographic criteria used in this study are reproducible and display a promising accuracy for use in preoperative groin assessment. This potential has to be developed and tested further before clinical implementation can be justified and made with confidence.

96

MOHAMMED ET AL.

REFERENCES 1. Creasman WT, Phillips JL, Menck HR: The National Cancer Data Base Report on Early Stage Invasive Vulvar Carcinoma. Cancer 80:505–513, 1997 2. Grimshaw RN, Murdoch JB, Monaghan JM: Radical vulvectomy and bilateral inguinal–femoral lymphadenectomy through separate incisions: Experience with 100 cases. Int J Gynecol Cancer 3:18 –23, 1993 3. Lin JY, DuBeshter B, Angel C, Dvoretsky PM: Morbidity and recurrence with modifications of radical vulvectomy and groin dissection. Gynecol Oncol 47:80 – 86, 1992 4. Hopkins MP, Reid GC, Morley GW: Radical vulvectomy: The decision for the incision. Cancer 72:799 – 803, 1993 5. Hoffman MS, Cavanagh D: The modern managemnt of invasive squamous cell carcinoma of the vulva. Contemp Rev Obstet Gynecol June:137–141, 1997 6. Stehman FB, Bundy BN, Dvoretsky PM, Creasman WT: Early stage 1 carcinoma of the vulva treated with ipsilateral superficial inguinal lymphadenectomy and modified radical hemivulvectomy: A prospective study of the Gynaecologic Oncology Group. Obstet Gynecol 79:490 – 497, 1992 7. DeCesare SL, Fiorica JV, Roberts WS, Reintgen D, Arango H, Hoffman MS, Puleo C, Cavanagh D: A pilot study utilizing intraoperative lymphoscintigraphy for identification of the sentinel lymph nodes in vulva cancer. Gynecol Oncol 66:425– 428, 1997 8. Levenback C, Burke TW, Morris M, Malpica A, Lucas KR, Gerhenson DM: Potential applications of intraoperative lymphatic mapping in vulva cancer. Gynecol Oncol 59:216 –220, 1995 9. Terada KY, Coel MN, Ko P, Wong JH: Combined use of intraoperative lymphatic mapping and lymphoscintigraphy in the management of squamous cell cancer of the vulva. Gynecol Oncol 70:65– 69, 1998 10. Bonnema J, van Geel AN, van Ooijen B, Mali SPM, Tijam SL, HenzenLongmans SC, Schmitz PIM, Wiggers T: Ultrasound-guided aspiration biopsy for detection of nonpalpable axillary node metastases in breast cancer patients: New diagnostic method. World J Surg 21:270 –274, 1997 11. Cheung TH, Yang WT, Yu MY, Lo WK, Ho S: New development of laparoscopic ultrasound and laparoscopic pelvic lymphadenectomy in the management of patients with cervical carcinoma. Gynecol Oncol 71:87– 93, 1998

12. Takes RP, Righi P, Meeuwis CA, Manni JHJ, Knegt P, Marres HAM, Spoelstra HAA, De Boer MF, van der Mey AGL, Bruaset I, Ball V, Weisberger E, Radpour S, Kruyt RH, Joosten FBM, Lameris JS, van Oostayen JA, Kopecky K, Caldemeyer K, Henzen-Logmans SC, Wiersma-van Tilburg JM, Bosman FT, van Krieken JHJM, Hermans J, Baatenburg de Jong RJ: The value of ultrasound with ultrasound-guided fine-needle aspiration biopsy compared to computed tomography in the detection of regional metastases in the clinically negative neck. Int J Radiat Oncol Biol Phys 40:1027–1032, 1998 13. Wu CH, Hsu MM, Chang YL, Hsieh FJ: Vascular pathology of malignant cervical lymphadenopathy. Cancer 83:1189 –1196, 1998 14. Giancarlo T, Palmierie A, Giacomarra V, Russolo M: Pre-operative evaluation of cervical adenopathies in tumours of the upper aeredogestive tract. Anticancer Res 18:2805–2810, 1998 15. Steinkemp HJ, Cornehl M, Hosten N, Pegios W, Vogl T, Felix R: Cervical lymphadenopathy: Ratio of long-to short-axis diameter as a predictor of malignancy. Br J Radiol 68:266 –270, 1995 16. Tschammler A, Ott G, Schang T, Seelbach-Goebel B, Schwager K, Hahn D: Lymphadenopathy: Differentiation of benign from malignant disease— Color doppler US assessmant of intranodal angioarchitecture. Radiology 208(1):117–123, 1998 17. Kruyt RH, van Putten WL, Levendag PC, de Boer MF, Oudkerk M: Biopsy of nonpalpable cervical lymph nodes: Selection criteria for ultrasound-guided biopsy in patients with head and neck squamous cell carcinoma. Ultrasound Med Biol 22(4):413– 419, 1996 18. Takashima S, Sone S, Nomura N, Kobayashi T, Nakamura H: Nonpalpable lymph nodes of the neck: Assessment with US and US-guided fine-needle aspiration biopsy. J Clin Ultrasound 25(6):283–292, 1997 19. Atula TS, Grenman R, Varpula MJ, Kurki TJ, Klemi PJ: Assessment of cervical lymph node status in head and neck cancer patients: Palpation, computed tomography and low field magnetic resonance imaging compared with ultrasound-guided fine-needle aspiration cytology. Head Neck 25(2):152–161, 1997 20. Elvin A, Sundstrom C, Larsson SG, Lindgren PG: Ultrasound-guided 1.2 mm cutting-needle biopsies of head and neck tumours. Acta Radiol 38(3):376 –380, 1997