- © 2009 by American Society of Clinical Oncology
Two-Center Evaluation of Dynamic Sentinel Node Biopsy for Squamous Cell Carcinoma of the Penis
- Joost A.P. Leijte,
- Ben Hughes,
- Niels M. Graafland,
- Bin K. Kroon,
- Renato A. Valdés Olmos,
- Omgo E. Nieweg,
- Cathy Corbishley,
- Sue Heenan,
- Nick Watkin and
- Simon Horenblas
- From the Departments of Urology, Nuclear Medicine, and Surgical Oncology, the Netherlands Cancer Institute, Antoni van Leeuwenhoek Hospital, Amsterdam, the Netherlands; and the Departments of Urology, Pathology, and Radiology, St George's Hospital, London, United Kingdom.
- Corresponding author: Simon Horenblas, MD, PhD, the Netherlands Cancer Institute, Antoni van Leeuwenhoek Hospital, Plesmanlaan 121, Amsterdam, the Netherlands 1066 CX; e-mail: s.horenblas{at}nki.nl.
-
Presented at the 23rd Annual European Association Urology Congress, March 26-29, 2008, Milan, Italy.
Abstract
Purpose Sentinel node biopsy is used to evaluate the nodal status of patients with clinically node-negative penile carcinoma. Its use is not widespread, and the majority of patients with clinically node-negative disease undergo an elective inguinal lymph node dissection. Reservations about the use of sentinel node biopsy include the fact that most current results come from one institution and the supposedly long learning curve associated with the procedure. The purpose of this study was to address these issues by analyzing results from two centers and by evaluating the learning curve.
Patients and Methods All patients undergoing sentinel node biopsy for penile carcinoma at two centers were included. The sentinel node identification rate, false-negative rate, and morbidity of the procedure were calculated. Results from the first 30 procedures were assessed for a potential learning curve.
Results A total of 323 patients with penile squamous cell carcinoma, which included 611 clinically node-negative groins, were scheduled for sentinel node biopsy. A sentinel node was found in 572 of the 592 groins (97%) that proceeded to sentinel node biopsy. In 79 groins, a sentinel node was positive for tumor. Six inguinal node recurrences occurred after a negative sentinel node procedure, all within 15 months after sentinel node biopsy. The combined false-negative rate was 7%. Complications occurred in 4.7% of explored groins. None of the false-negative procedures occurred in the initial 30 procedures.
Conclusion Sentinel node biopsy is a suitable procedure to stage clinically node-negative penile cancer, and it has a low complication rate. No learning curve was demonstrated in this study.
INTRODUCTION
There is no consensus on the management of patients with clinically node-negative (cN0) penile carcinoma.1 Approximately 20% to 25% of these patients have occult inguinal metastases, and there is evidence that an elective inguinal lymph node dissection leads to better survival compared with surgical removal at the time that metastases become clinically apparent.2,3 However, elective inguinal lymph node dissection in all patients with cN0 disease is unnecessary in 75% to 80%, and this procedure is associated with substantial morbidity.4,5 Current imaging techniques, such as ultrasound, computed tomography, and positron emission tomography, as well as risk prediction on the basis of primary tumor characteristics, have been unable so far to reliably identify patients with cN0 disease and occult metastasis.6,7 Similar clinical challenges in breast cancer and melanoma have been addressed with the use of sentinel node biopsy. This technique has become a routine procedure in the management of patients with cN0 breast cancer and has an increasingly important role in melanoma. For both diseases, good reliability and acceptable complication rates have been reported, and numerous patients have been spared unnecessary lymph node dissections.8,9
It seems contradictory that, although penile carcinoma was one of the first malignancies in which the concept of sentinel node biopsy was described, its use for this disease is still under debate and is not widespread.10
Sentinel node biopsy for penile carcinoma, which used a modern technique of lymphatic mapping with blue dye and a radioactive tracer, was introduced in 1994, and its reliability and low morbidity have been reported by various authors.11–13 Reservations about the use of sentinel node biopsy include the fact that most of the current results are from one institution and the supposedly long learning curve associated with the procedure.14 The purpose of this study was to address these issues by analyzing sentinel node biopsy results from two different centers with regard to sentinel node identification rate, false-negative rate, and complication rate of the procedure and by evaluating the learning curve.
PATIENTS AND METHODS
The data regarding sentinel node biopsy from two referral centers for penile carcinoma were combined. The database from the Netherlands Cancer Institute (NKI) in Amsterdam contains 297 patients, with 550 cN0 groins, who were treated in the period between January 1994 and November 2007. The database from St George's Hospital (SGH) in London contains 134 patients, with 263 cN0 groins, who were treated between January 2004 and November 2007. Both databases are prospectively kept and include detailed information on preoperative examinations, surgical procedures, histopathologic findings, and follow-up.
General Overview of Primary Tumor Treatment
The penile tumor was treated surgically, in most instances in the same operation session as the sentinel node biopsy. The primary tumor treatment has been described in detail previously.15,16 In general, T1 and T2 tumors smaller than 2 cm were treated with penile-preserving methods. T2 tumors larger than 2 cm, T3 tumors, and T4 tumors were treated with glans amputation and partial or total amputation. Re-excision was carried out in case of a positive margin.
Indications for Sentinel Node Biopsy and Inclusion Criteria for Analysis
In both institutes, sentinel node biopsy is performed in cN0 groins of patients with a penile squamous cell carcinoma staged as T1G2 or greater.17,18 More than 95% of penile malignancies are squamous cell carcinomas. Patients with other types of penile malignancies were excluded from analysis.
Sentinel node biopsy was introduced at NKI in 1994. This procedure is used for patients with bilateral cN0 groins and also for staging the cN0 groin in patients with unilateral, clinically node-positive disease.19 Since 2001, several modifications have been made in the procedure to decrease the false-negative rate, including the addition of preoperative ultrasound of cN0 groins and immediate fine-needle aspiration cytology if a suspicious node is seen during ultrasound investigation.20 Sentinel node biopsy was introduced at SGH in 2004 on the basis of the modified protocol from the NKI. Only patients treated after the protocol was modified were included in this analysis (ie, 189 patients, with 348 cN0 groins, who were treated after 2001 at NKI and 134 patients, with 263 cN0 groins, who were treated at SGH). Thus, a total of 323 patients, with 611 cN0 groins, were available for analysis.
Patients with clinically node-positive disease underwent an inguinal lymph node dissection on this side after fine-needle aspiration cytology. A pelvic lymph node dissection was done if histopathology revealed two or more tumor-positive inguinal nodes or extracapsular growth.21
Protocols for Sentinel Node Biopsy
Both institutes use essentially the same protocol, except for some minor differences. Firstly, the preoperative ultrasound at SGH is performed immediately before sentinel node surgery for logistical reasons; at NKI there are 2 weeks between ultrasound and surgery. Furthermore, at SGH, preoperative scintigraphy is performed the same day as sentinel node surgery. At NKI, scintigraphy is performed the previous day. The protocols from both institutes have been described in detail previously.12,22
Technique of Lymphatic Mapping
Lymphoscintigraphic images were obtained from all patients before surgery after intradermal, peritumoral injection of a 99mtechnetium-labeled nanocolloid (Nanocoll; GE Healthcare, Eindhoven, the Netherlands). Immediate dynamic imaging was performed and was followed by static imaging at 30 minutes, 90 minutes (SGH), and 2 hours (NKI) postinjection by using a dual-head gamma camera. Anterior and lateral images were obtained routinely with simultaneous transmission scanning by using a 57cobalt flood source to outline the body contour for orientation. Other views were made if indicated. The locations of sentinel nodes were marked on the skin after analysis of the lymphoscintigraphic images was performed by the nuclear medicine physician. The sentinel node was defined as a node on a direct lymphatic drainage pathway from the primary tumor.
Sentinel Node Surgery and Pathology Evaluation
Shortly before surgery, 1 mL of patent blue dye was injected intradermally around the tumor (NKI) or circumferentially around the penile shaft (SGH). Sentinel nodes were identified intraoperatively with the aid of the preoperative lymphoscintigraphic images, a gamma-ray detection probe, and the patent blue dye. After removal of the sentinel nodes, the groin was palpated for suspicious nodes that failed to pick up any tracer. All harvested sentinel nodes were histopathologically analyzed for metastasis with serial sectioning (2 mm at SGH, 150 μm at NKI) and immunohistochemical staining by using CAM5.2 (Becton Dickinson, San Jose, CA) and pankeratin. An ipsilateral completion inguinal lymph node dissection was performed in patients who had tumor-positive sentinel nodes. A pelvic lymph node dissection was done in case histopathologic evaluation of the inguinal lymph node dissection specimen revealed two or more tumor-positive nodes or extracapsular growth.
Follow-Up
All complications of sentinel node biopsy in the immediate postoperative period were recorded. The standard follow-up schedule at NKI consisted of control visits every 2 months during the first 2 years after sentinel node biopsy, as 3-month intervals in the 3 year, and as 6-month intervals thereafter. At SGH, control visits were scheduled every other month in the first year, every 4 months in the second year, and every 6 months in the third year after sentinel node biopsy. In both institutes, follow-up visits consisted of physical examination, and the focus was on a potential inguinal recurrence. On indication, ultrasound and computed tomography were used. The median follow-up duration of the entire group was 17.9 months (range, 1 to 69 months).
Definition of False-Negative Procedure
We classified the sentinel node biopsy as a false-negative procedure if a regional nodal recurrence developed after a negative sentinel node procedure. In patients with a new primary tumor or penile recurrence that predated the regional nodal recurrence, the sentinel node procedure was not considered a false negative.
Statistical Analysis
The false-negative rate of sentinel node biopsy was calculated according to standard statistical methods (ie, false-negative procedures ÷ true-positive procedures + false-negative procedures) for both institutes separately and by using the combined data.
To evaluate the presence of a learning curve, we analyzed the results of the first 30 procedures at SGH for the false-negative rate and for complications.
RESULTS
A total of 323 patients, with 611 cN0 groins, were included in this study. A detailed overview of patient characteristics is listed in Table 1.
Patient Demographic and Clinical Tumor Characteristics
Preoperative Ultrasound-Guided Fine-Needle Aspiration Cytology
Preoperative fine-needle aspiration cytology was tumor positive in 26 clinically node-negative groins (NKI, n = 15; SGH, n = 11). At NKI, these patients underwent an immediate inguinal lymph node dissection on the affected side. Because the fine-needle aspiration cytology was performed the same day as sentinel node surgery at SGH, the results were not available at the time of surgery, and all patients proceeded to sentinel node surgery. Thus, lymphatic mapping was performed in 596 cN0 groins.
Visualization and Identification Rate
See Table 2. All nonvisualized groins were explored for radioactivity by using the gamma probe, were explored for blue dye, and were intraoperatively palpated for suspicious nodes, except for four cN0 groins at SGH in which an immediate inguinal lymph node dissection was performed because of nonvisualization.
Visualization and Identification Rate
Histopathologic Analysis
Seventy-nine (13%) of the 592 explored cN0 groins contained a tumor-positive sentinel node (42 of 333 cN0 groins at NKI and 37 of 259 cN0 groins at SGH).
Fourteen (18%) of the 79 completion inguinal lymph node dissections after removal of a positive sentinel node revealed additional inguinal metastasis.
False-Negative Procedures and False-Negative Rate
A total of six false-negative procedures (NKI, n = 3; SGH, n = 3) were identified during follow-up, all within 15 months. These patients, who had inguinal recurrences, underwent inguinal lymph node dissection, except for one patient who underwent radiotherapy of the groin at NKI.
Of the three patients with false-negative procedures at NKI, two died at 4 and 10 months after the inguinal recurrence. The remaining patient was alive at 13 months without evidence of disease. The false-negative rate at NKI was 6.7%. At SGH, all three patients who underwent false-negative procedures were still alive after chemotherapy for lung metastases (n = 2) and adjuvant radiotherapy to the groin (n = 1) at 3, 6, and 3 months after the recurrence, respectively. The false-negative rate at SGH was 7.5%.
The combined false-negative rate was 7% (ie, six false-negative groins ÷ 79 sentinel node tumor-positive groins + six false-negative groins).
Complications
A complication occurred in 28 of the 592 explored groins, which led to a complication rate of 4.7% in operated groins (Table 3).
Overview of Complications
Evaluation of Learning Curve
None of the false-negative procedures at SGH occurred during the first 30 procedures of SGH. The first false-negative procedure that occurred in the series of SGH was the forty-first consecutive patient. No complications occurred in the first 30 patients.
DISCUSSION
This study shows that sentinel node biopsy is a suitable procedure to identify patients with occult inguinal metastasis. In this large series of patients treated according to a standardized protocol, a sensitivity of 93% was found. The complication rate was low, and almost all complications were transient and could be managed conservatively. Furthermore, this study demonstrated that the encouraging results from an experienced center can also be achieved in a center introducing the technique.
The initial results of sentinel node biopsy at NKI were plagued by a relatively high false-negative rate of 22%.19 A detailed analysis of the false-negative procedures was performed, and several modifications to the procedure were made. Preoperative ultrasound was added as a standard examination to decrease the risk of missing a sizeable metastatic lymph node that was not detected during physical examination but that could lead to tumor blockage, rerouting of lymphatic fluid and, thus, a potential false-negative procedure. Groins without a visualized sentinel node are now routinely explored with the gamma probe for radioactivity and for blue dye, and intraoperative palpation of the wound is performed in all patients. Also, several modifications were made in the histopathologic analysis. All sentinel nodes are serially sectioned and immunohistochemically stained to avoid missing micrometastasis. Since the modifications were instituted, the false negative rate decreased to 4.8%.12
The results from this study contrast with several other studies that were published recently on the topic of sentinel node biopsy for penile cancer.23–25 Spiess et al25 reported on their experience with sentinel node biopsy in 31 patients, who all underwent completion inguinal node dissection after the sentinel node biopsy. They found a false-negative rate of 29%. In their series of 27 patients, Gonzaga-Silva et al23 reported a false-negative rate as high as 75% when using an isolated gamma probe technique to identify the sentinel node. The possible reasons for the latter poor results have been discussed before; these results could be caused by suboptimal technique—for instance, lacking the preoperative ultrasound evaluation of the groins or preoperative lymphoscintigraphy.26–28
Another explanation of the high false-negative rates in these prior studies, combined with the initial false-negative rate of 22% that was found at NKI,19,23–25 may represent the presence of a learning curve to reach an acceptable accuracy of the procedure.
A false-negative sentinel node biopsy may have an impact on the chances of survival. Surgical resection of occult metastatic disease offers a survival benefit compared with inguinal node dissection when occult disease becomes clinically apparent.2,3 Patients who underwent a false-negative procedure, therefore, should be considered undertreated compared with those who underwent elective inguinal lymph node dissection. Therefore, many advocate a learning phase when introducing the technique, in which a completion lymph node dissection is performed regardless of the tumor status of the sentinel node. In two of the largest multicenter trials that evaluated the value of sentinel node biopsy in melanoma (MSLT-I) and breast cancer (ALMANAC), learning phases of 30 and 40 procedures, respectively, were required in the participating centers.29,30
However, in the ALMANAC study that involved 14 breast units in the United Kingdom, no learning curve could be demonstrated in an analysis of the initial procedures.30 From a statistical point of view, there are also some remarks to be made regarding the length and accuracy of a learning phase. Tanis et al31 calculated that a learning phase of at least 300 node-positive procedures is necessary to determine with 95% certainty that a surgeon can find the sentinel node in 95% of the patients, which provides a false-negative rate of no more than 5%.31 If 13% of cN0 groins contain involved nodes, this means that a statistically sound learning phase requires 2,307 groins. This number is unrealistic, especially in a rare disease such as penile cancer.
Interestingly, in a recent multicenter trial that prospectively evaluated the role of sentinel node biopsy in vulvar carcinoma, a learning phase of only 10 procedures was required for participating centers. The rationales for this shorter learning phase were the smaller variability of lymphatic drainage compared with melanoma and breast cancer and the assertion that the draining nodes often are superficially located in the inguinal region and, therefore, are easier to harvest.32
In the limited number of patients in this study, we could not demonstrate the presence of a learning curve at SGH after introduction of sentinel node biopsy. No completion lymph node dissection was performed in the initial procedures. Although this, for statistical reasons, could be caused by chance, it could also be that a substantial part of the learning curve can be skipped when an optimized protocol is used.
The multidisciplinary nature of sentinel node biopsy, which relies also on the experience of the nuclear medicine physicians and pathologists, should not be overlooked. Currently, the sentinel node procedure is performed in many institutes for breast cancer and melanoma, and a considerable amount of knowledge regarding the visualization, identification, and histopathologic analysis of the sentinel node is available from which urologists can draw.
A weak point of this study is the median follow-up time of 17.9 months. A longer follow-up period may reveal more false-negative procedures and may decrease the sensitivity of sentinel node biopsy. It is generally assumed that a false-negative sentinel node procedure becomes clinically apparent within 2 years.12
Should sentinel node biopsy for penile carcinoma be performed by every urologist? We want to stress that, in a rare disease such as penile carcinoma, centralization of management into specialized centers may be required for the involved physician to gain and maintain the required amount of experience.
This study was performed in two tertiary referral hospitals for penile cancer. It seems sensible to perform the procedure in such centers. On the basis of a lack of learning curve in this study, one might advocate that the procedure also can be performed in hospitals with a smaller volume of patients, which assumes that experience with the sentinel node procedure in other malignancies, such as breast cancer and melanoma, has been obtained. The risk of a false-negative procedure should be weighed against the morbidity of an elective inguinal lymph node dissection.
In conclusion, with the use of an up-to-date protocol, sentinel node biopsy for penile carcinoma is a suitable procedure to identify patients with clinically node-negative disease who have occult metastasis. The sentinel node identification rate was 97%, the false-negative rate was 7%, and the complication rate was 4.7%. In this study, no learning curve could be demonstrated.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The author(s) indicated no potential conflicts of interest.
AUTHOR CONTRIBUTIONS
Conception and design: Joost A.P. Leijte, Ben Hughes, Bin K. Kroon, Renato A. Valdés Olmos, Omgo E. Nieweg, Nick Watkin, Simon Horenblas
Administrative support: Joost A.P. Leijte, Ben Hughes, Niels M. Graafland
Provision of study materials or patients: Renato A. Valdés Olmos, Cathy Corbishley, Sue Heenan, Nick Watkin, Simon Horenblas
Collection and assembly of data: Joost A.P. Leijte, Ben Hughes, Niels M. Graafland, Bin K. Kroon, Renato A. Valdés Olmos, Cathy Corbishley, Sue Heenan, Nick Watkin, Simon Horenblas
Data analysis and interpretation: Joost A.P. Leijte, Ben Hughes, Niels M. Graafland
Manuscript writing: Joost A.P. Leijte, Ben Hughes, Niels M. Graafland, Omgo E. Nieweg, Nick Watkin, Simon Horenblas
Final approval of manuscript: Joost A.P. Leijte, Ben Hughes, Niels M. Graafland, Bin K. Kroon, Renato A. Valdés Olmos, Omgo E. Nieweg, Cathy Corbishley, Sue Heenan, Nick Watkin, Simon Horenblas
Footnotes
-
Published in part for the Netherlands Cancer Institute (Eur Urol 2007;52:170-177) and for St George's Hospital (BJU Int 2007;100:561-565).
-
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
- Received October 21, 2008.
- Accepted February 13, 2009.
