What's known on the subject? and What does the study add?
According to the current European Association of Urology Guidelines, dynamic sentinel node biopsy is the recommended approach to assess lymph node status in men with cN0 intermediate and high risk penile cancer. Nevertheless, most encouraging results derive from a limited number of studies.
The present study shows a false-negative rate of 15%, comparable with or better than several previous studies. Nevertheless, the aim should be a false-negative rate of no more than 5%. We conclude that increased overall experience and the use of the complete modern dynamic sentinel node biopsy protocol are paramount to improve results.
• To evaluate the false-negative rate and complication rate of dynamic sentinel node biopsy (DSNB) in penile cancer.
PATIENTS AND METHODS
• In this retrospective study, 58 unilaterally or bilaterally clinically lymph node negative (cN0) patients with penile cancer (57 squamous cell carcinomas and one malignant melanoma), scheduled for DSNB at the Örebro University Hospital, Sweden, between 1999 and 2011, were analysed.
• Preoperative ultrasonography and fine-needle aspiration cytology of suspicious nodes were not introduced until 2008.
• Patients were assessed by lymphoscintigraphy using 99mtechnetium nanocolloid on the day before surgery and the dissection of sentinel nodes was aided by the lymphoscintigraphic images and intraoperative detection of radiotracer and patent blue dye.
• The false-negative rate and complication rate were calculated per groin.
• Of the 58 patients, 32 (55%) underwent preoperative ultrasonography.
• Two patients had positive fine-needle aspiration cytology and discontinued further DSNB protocol. Of the remaining 56 patients, all but one were bilaterally cN0 and hence 111 cN0 groins were assessed by lymphoscintigraphy.
• In the 55 bilaterally cN0 patients, lymphoscintigraphy visualized a bilateral sentinel node in 34 (62%).
• At surgery, all excised sentinel nodes were radioactive while 43% were additionally blue. In total, at least one sentinel node was harvested in 96 (86%) of the DSNB staged groins.
• A positive sentinel node was found in 11 groins (bilaterally in three patients). During a median follow-up of 21 months, two false-negative cases emerged, producing a false-negative rate of 15%. Both false-negative cases occurred during the first half of the study. The complication rate was 10%. The majority of complications were minor and transient.
• DSNB is a minimally invasive staging tool in men with cN0 penile cancer, enabling early detection of metastatic disease and thus optimal care.
• Our false-negative rate of 15% is comparable or even favourable in comparison with several previous studies, but far from the 5% or less that we aim for. The complication rate found is somewhat higher than previously reported.
• With increased overall experience and the continued use of the complete DSNB protocol, we believe our results will improve and the complication rate will decrease.
Penile cancer is a rare disease in the western world. In Sweden the annual incidence is 2.2 per 100 000 men . The strongest prognostic factor for survival in these men is the presence and extent of lymph node metastases . In lymph node positive patients cure is still possible by means of lymph node dissection (LND) with 5-year survival rates of roughly 20–75%, depending on the N stage .
The main focus of recent debate has been who should be referred for LND and when. On the one hand, in patients with actual nodal metastases the disease-specific 3-year survival was 84% when subjected to early LND compared with only 35% after delayed LND following surveillance . On the other hand, LND is associated with high complication rates and of no clinical benefit in patients with no lymph node metastases .
In patients who present with palpable inguinal lymph nodes, up to approximately 70% actually have lymph node metastases . Hence, the current European Association of Urology (EAU) Guidelines recommend LND in these men . The approach in patients who are clinically lymph node negative (cN0) is more controversial. The EAU Guidelines (2004) suggested a risk-adapted strategy based on histological characteristics of the primary tumour, proposing LND in high and possibly also intermediate risk tumours . However, the accuracy of this strategy has been questioned repeatedly. In one study, 82% of the cN0 patients managed according to these guidelines underwent unnecessary LND . In two additional large series of cN0 patients, 77% with high risk and up to 91% with intermediate risk penile cancer would have been subjected to unnecessary LND if managed according to the outlined guidelines [8,9].
The sentinel node concept was introduced in penile cancer by Cabanas  more than 30 years ago. However, the initial concept was based on a static location of the sentinel node and associated with a high false-negative rate (FNR) and eventually fell out of favour . The concept of sentinel node was reintroduced in penile cancer in 1994 by Horenblas and colleagues . The method, labelled dynamic sentinel node biopsy (DSNB), includes preoperative lymphoscintigraphy using 99mtechnetium nanocolloid, preoperative patent blue dye injection and intraoperative guidance with a gamma ray detection probe to visualize the individual lymphatic drainage pattern . In a study by Leijte et al. , patients staged by DSNB between 1994 and 2001 had an unsatisfactory FNR of 19%, while it was reduced to 5% only in patients treated between 2001 and 2004. The majority of false-negative cases were assumed to be caused by tumour blockage with or without rerouting to a ‘neo’-sentinel node . To avoid this phenomenon, preoperative ultrasound-guided fine-needle aspiration cytology (FNAC) of suspicious inguinal lymph nodes was added to the protocol in the latter group and, together with some additional minor modifications and an increased overall experience, improved the results significantly . These results have been reproduced by Hadway et al. . In a combined analysis by the two above-mentioned groups, the FNR was 7% and the complication rate less than 5% .
In the current EAU Guidelines DSNB is the recommended approach in cN0 patients with intermediate and high risk penile cancer . Since most of the encouraging results derive from a limited number of studies, additional validation is desirable. The present study is the first report on DSNB in penile cancer from Sweden.
PATIENTS AND METHODS
Unilaterally or bilaterally cN0 patients with assumed or confirmed intermediate or high risk squamous cell carcinoma (SCC) or malignant melanoma of the penis, scheduled for DSNB at the Örebro University Hospital, Sweden, were retrospectively included in this study.
The local ethics committee approved the study. Up to 2008, ultrasonography was not included in our routine evaluation. Since then, patients scheduled for DSNB were assessed preoperatively by ultrasonography of the groins and FNAC was performed in the case of suspicious nodes. In the case of positive FNAC further DSNB protocol was discontinued in view of the positive cytology and the patient was subjected to an ipsilateral LND.
Following application of local anaesthetic spray (10% xylocaine), approximately 80–90 MBq 99mtechnetium nanocolloid (Nanocoll®, GE Healthcare, Amersham, UK) was injected intradermally on four peritumoural sites or, when the primary tumour had already been removed, around the resection site. Immediately after the tracer injection a dual-head gamma camera (E.cam®, Siemens Healthcare, Stockholm, Sweden) was used to visualize the lymphatic drainage. During the initial dynamic sequence 20 anterior images, each of 30 s duration, were acquired and directly followed by a 5 min static anterior and lateral image, which was repeated 120 min post injection. Additional imaging was performed if indicated. A sentinel node was defined as an inguinal node with direct drainage from the site of the injection. The approximate locations of the sentinel nodes were marked with permanent ink on the skin of the patient. Patent blue dye (Patent Blue V sodium 2.5%, Guerbet, Roissy, France) was injected intradermally on the same sites as the radiotracer a few minutes prior to surgery, performed the day after lymphoscintigraphy. Two experienced surgeons performed all procedures. The sentinel nodes were identified by the intraoperative use of a gamma ray detection probe (neoprobe®neo2000®gamma detection system, Neoprobe Corporation, Dublin, OH, USA) in conjunction with the lymphoscintigraphic images and the staining of the lymphatic vessels and nodes by patent blue dye. The sentinel nodes were formalin fixed, cut into 2 mm slices and paraffin embedded. Three levels per block, with 200 µm separation, were stained using haematoxylin and eosin and immunohistochemistry was carried out. In malignant melanoma immunohistochemistry was done with S-100, supplemented by HMB45 and Melan A if indicated, and in SCC with AE1/AE3 (pan-CK). The DAKO EnVisionTM system was used for visualization. A complementary ipsilateral LND was performed in the case of a positive sentinel node a few weeks after DSNB.
Referred patients generally had their first postoperative control at our institute, while the remaining follow-up was carried out at the referring hospital. Patients were recommended follow-up according to the EAU Guidelines [6,7].
A false-negative groin was defined as a regional recurrence in a groin previously staged as negative by DSNB. During the initial learning phase a few patients underwent immediate LND following DSNB regardless of lymph node status found, in order to assess the accuracy of the DSNB procedure. In these cases the LND histopathology report was used as reference. In the remaining cases the status at the last follow-up was used as reference. The follow-up time was calculated from DSNB to a regional recurrence, LND because of a local recurrence or last follow-up. Follow-up time was not assessed in patients who underwent bilateral LND to evaluate DSNB accuracy or because of DSNB findings.
The FNR was calculated per groin according to the standard definition (i.e. the false-negative cases divided by the true-positive plus false-negative cases). The complication rate was analysed per DSNB-operated groin. Groins undergoing immediate LND were excluded in this analysis.
Between April 1999 and April 2011 (12 years) 56 patients met the inclusion criteria. Two patients had local penile recurrences 5 and 12 years respectively after initial combined primary tumour surgery and DSNB. These recurrences were assessed as new cases and, consequently, 58 patients/cases were included in the study. Fifty-three (91%) of the patients were referred to us from 20 different hospitals in Sweden. DSNB was performed only sporadically during the first half of the study period, with 11 (19%, two per year) patients scheduled for DSNB compared with 47 (81%, eight per year) during the last half. This was because expertise and equipment were not always available during the first half of the study period. The inclusion criteria remained unchanged.
The mean age among the included patients was 60 years (range 37–84 years). One patient had a malignant melanoma, while the remaining patients all had SCC. The histological characteristics of the tumours are summarized in Table 1. One patient was unilaterally clinically lymph node positive. The remaining 57 patients were all bilaterally cN0.
Table 1. Histological characteristics of the primary tumours and distribution of positive sentinel nodes, false-negative cases and positive FNAC
No. of patients
No. of patients
The numerals in parentheses indicate the number of groins with a positive sentinel node (numerator) and the corresponding number of patients concerned (denominator) as determined by DSNB. *One false-negative groin. †One groin (††two groins) with positive FNAC on preoperative ultrasonography.
Thirty-two patients (55%) were assessed by preoperative ultrasonography with FNAC of suspicious inguinal lymph nodes, while 26 patients (45%) underwent DSNB before preoperative ultrasonography was introduced at our institute. Two bilaterally cN0 patients had a positive FNAC, unilaterally in the first case and bilaterally in the second (Table 1). Both these patients underwent immediate bilateral LND and both were found to have bilateral inguinal metastases. Consequently, 56 patients (55 bilaterally cN0 and one unilaterally cN0, equalling 111 cN0 groins) continued the DSNB protocol (Fig. 1).
The 56 remaining patients (with 111 cN0 groins) all underwent lymphoscintigraphy. In the 55 bilaterally cN0 patients, lymphoscintigraphy identified bilateral sentinel nodes in 30 (55%), unilateral in 24 (44%), while one patient (2%) had no radiotracer uptake in either groin. In the patient who was unilaterally cN0, a sentinel node was identified in the groin of interest. In seven bilaterally cN0 patients with a unilateral uptake on the initial lymphoscintigraphy, a repeat DSNB focusing exclusively on the contralateral side was performed within a short period of time. Among these, four patients (57%) had a sentinel node identified in the previously radiotracer-silent groin. Including the repeat DSNB procedures, 34 (62%) of the bilaterally cN0 patients also had bilateral uptake on lymphoscintigraphy. In all, lymphoscintigraphy identified at least one sentinel node in 89 cN0 groins (80%), while 22 (20%) were radiotracer silent (Fig. 1).
In the 22 radiotracer-silent groins, nine were subjected to surveillance. No metastases became evident in these groins during the course of follow-up. Another eight radiotracer-silent groins (including both groins in the patient with bilateral absent uptake) were surgically explored and sentinel nodes were identified based on palpation, intraoperative radiotracer detection and blue dye staining. One of these groins harboured a positive sentinel node. The remaining five radiotracer-silent groins were subjected to LND. One of these groins contained metastases (Fig. 1).
During surgery, in one patient in whom lymphoscintigraphy was bilateral, we were not able to identify a sentinel node in one of the groins. Instead a LND was performed but no metastases were found. Hence, at surgery, a sentinel node was harvested in 96 (86%) of the DSNB staged groins (88 aided by lymphoscintigraphic visualization and eight groins by exploration). A total of 180 sentinel nodes were removed from the 96 DSNB staged groins (1.9 nodes per groin, range 1–4). All nodes (100%) were radioactive and 77 (43%) were additionally blue. A positive sentinel node was encountered in 11 groins, including the one found in a radiotracer-silent groin described above. Three patients had bilateral positive sentinel nodes (Fig. 1).
Eight patients (with 15 DSNB staged groins), all among the initial cases, underwent immediate LND, regardless of DSNB results, to assess the accuracy of the procedure. There was no discrepancy between DSNB and LND results in these patients.
The median follow-up time was 21 months (range 4–143 months). During the follow-up, two false-negative groins were encountered. These occurred in patients 22 and 28, 4 and 11 months respectively after a negative DSNB. Both patients died of penile cancer, 20 and 25 months respectively after the regional recurrence. These data produce an FNR of 15% (two false-negative groins divided by 11 true-positive plus two false-negative groins). When comparing patients who did and did not undergo preoperative ultrasonography the results were similar, with an FNR of 17% (1/(5 + 1)) in the first group compared with 14% (1/(6 + 1)) in the second group. The primary tumour characteristics of the patients with a false-negative DSNB are shown in Table 1.
All DSNB positive groins were subjected to a complementary ipsilateral LND. Only two groins (18%) contained further metastases on LND histopathology, while in the remaining nine groins (82%) all additional nodes were tumour-free. The patient with malignant melanoma had a unilateral positive sentinel node but no additional metastases on LND and no regional recurrences during follow-up (11 months).
Eighty-one groins (96 DSNB staged groins minus the 15 groins that underwent immediate LND) were available for the complication analysis. A complication was noted in eight groins (10%). Four groins experienced mild haematomas that resolved spontaneously. One groin became infected. One patient had a mild unilateral lymphoedema of the leg, while one patient had a bilateral scrotal lymphoedema. In the latter case the patient was bilaterally DSNB positive and hence subjected to a complementary bilateral LND within a short period of time. No groin suffered multiple complications. Types and frequencies of the complications are summarized in Table 2.
Table 2. Complications following DSNB
No. of groins
Early resection of lymph node metastases provide a survival advantage , but LND in all cN0 patients constitutes over-treatment in the vast majority [5,8,9,11–13], putting these patients on a high risk of complications without any clinical benefit . According to the recently updated EAU Guidelines, DSNB is recommended in cN0 patients with intermediate and high risk penile cancer . However, the reported accuracy of DSNB varies considerably, with an FNR ranging from 5% to 75% .
We report on 58 patients scheduled for DSNB. Five of these patients (Table 1) had tumours not necessitating lymph node staging according to the EAU Guidelines. The reason these patients were scheduled for DSNB was clinical over-staging. Two patients with a locally recurrent penile cancer, 5 and 12 years after initial combined primary tumour surgery and DSNB, were included in the study a second time, undergoing a repeat DSNB. This strategy was proved feasible by Graafland et al. .
We were able to identify a bilateral sentinel node on lymphoscintigraphy in 62% of the bilaterally cN0 patients. This is comparable with Perdona et al. , who reported on a bilateral sentinel node visualization in 14 out of 22 patients (64%), but is in contrast to, among others, Hadway et al. , who identified bilateral sentinel nodes in 68 out of 75 patients (91%). A unilateral visualization on lymphoscintigraphy constitutes a difficult clinical problem in the management of patients with penile cancer by DSNB. In our retrospective study this situation was handled in different ways: LND, surveillance, repeat DSNB or exploration for sentinel nodes based on palpation, intraoperative radiotracer detection and/or blue dye staining (Fig. 1). Our present strategy is primarily to repeat DSNB and if, again, no sentinel node is visualized, to perform a modified LND. In our opinion this is a reliable course of action. Interestingly, a clear trend towards more frequent bilateral tracer uptake was noticeable in the latter half of our series (data not shown). This might be due to increased overall experience.
Two patients in our study discontinued further DSNB protocol because of positive FNAC on preoperative ultrasonography. In the remaining 56 DSNB procedures, 11 true-positive and two false-negative groins were found, producing an FNR of 15%. While these results are comparable or even favourable to the majority of previous studies , they are still unsatisfactory and far from our ambition of an FNR of 5% or less. Two European high volume referral centres for penile cancer have reported such results [11,12]. In their combined analysis, including 323 patients, an FNR of 7% was established .
Two patients in our study suffered a regional recurrence after a negative DSNB. In the first case no preoperative ultrasonography was performed, and we hypothesize that this is a case of tumour blockage and rerouting as described by Kroon et al. . In the second case, preoperative ultrasonography was performed and actually showed a suspicious node, but the FNAC was negative as was the histopathological examination of the harvested sentinel node. One possibility is that tumour cells in the true sentinel node were missed by the FNAC, while tumour blockage and rerouting caused the incorrect node to be excised at surgery. Re-evaluation of the DSNB histopathology in these two cases confirmed normal lymphatic tissue and cytology.
One patient in our study had absent radiotracer uptake in both groins. This patient was not assessed preoperatively by ultrasonography. Both groins were explored for sentinel nodes based on palpation, intraoperative radiotracer detection and/or blue dye staining. In his left groin, two out of four harvested radioactive but non-blue sentinel nodes contained metastases. We hypothesize that this is an additional case of tumour blockage where preoperative ultrasonography would have identified the metastases. If so, the patient would have discontinued further DSNB protocol and been referred immediately for LND, hence saving time and money. This further emphasizes the importance of using preoperative ultrasonography.
Although the number of patients with less than T1G2 tumours in our series is small, none of these patients had a positive sentinel node or a false-negative DSNB (Table 1). These findings support the DSNB indications in the current EAU Guidelines . Three patients with T1G2 tumours in our study were found to have lymph node metastases (Table 1). In two of these patients lymph node metastases were found by DSNB, while the third patient had a false-negative DSNB procedure. The risk of lymph node metastases in patients with T1G2 tumours in the literature varies between zero and 50% . In a study by Hughes et al.  including 105 cN0 T1G2 patients, 9% were found to have lymph node metastases. Our finding supports Hughes et al. , who state that T1G2 tumours ‘warrant surgical and potentially curative staging’.
In our study, all sentinel nodes were radioactive but only 43% were additionally blue. Our results contrast with those of Hadway et al. , Perdona et al.  and Kroon et al.  who report on 70–87% of the excised nodes being both blue and radioactive. No doubt, blue staining of lymph vessels and nodes is helpful and facilitates correct identification during surgery. The above and also the relatively low bilateral radiotracer uptake in our study might be due to a suboptimal injection technique on our part. We are currently considering changing the injection site of both tracer and patent blue dye from peritumoural to the distal shaft as described by Hadway et al. .
Our study covers a 12-year period. In order to increase sensitivity, the DSNB methodology has been modified during this period, most importantly by the introduction of preoperative ultrasonography . In our opinion, modern DSNB should include preoperative ultrasonography, lymphoscintigraphy, patent blue dye injection, intraoperative radiotracer detection and palpation and a thorough histopathological analysis of the harvested sentinel nodes. Considering the excellent results in a few but large series using this methodology [11–13] it is highly questionable to use anything less. To exemplify, Gonzaga-Silva et al.  reported on an FNR of 75% in a study of 27 patients using an isolated gamma probe technique. Comparing modern DSNB with its previous methodology shows considerable improvements with regard to sensitivity , even though we were not able to confirm this in our study. However, it is difficult to determine to what extent the increased overall experience contributes to the improved results. Interestingly, no learning curve was identified in a study evaluating the introduction of modern DSNB . Nevertheless, in our analysis, both false-negative cases emerged in the first half of the study. This might be due to a shorter follow-up in the last half, but might also demonstrate an actual learning curve.
The complication rates after LND have decreased somewhat in more recent studies . Nevertheless, up to an 87% overall risk of complications and up to a 57% risk of severe lymphoedema  proves that LND is high risk surgery. In our analysis, 10% of the DSNB staged groins had complications. The majority were minor and transient. Nevertheless, the complication rate following DSNB in the present study is slightly higher than reported by others [11–13,19].
Even if, in the right hands, DSNB is both safe and accurate in patients with cN0 penile cancer, it is not necessarily the ultimate solution. Several alternatives have been evaluated. Neither 18F-fluorodeoxyglucose positron emission tomography/CT nor isolated ultrasound-guided FNAC is sensitive enough in cN0 groins [20,21]. Tabatabaei et al.  evaluated lymphotropic nanoparticle enhanced MRI with ferumoxtran-10 for lymph node staging in seven patients and found complete sensitivity and a specificity of 97%, but the study is small and needs to be confirmed. Ficarra et al.  have proposed a nomogram to predict lymph node involvement and report on a good concordance index. However, according to the nomogram, moderately differentiated and superficially growing tumours are at higher risk of metastases than poorly differentiated and vertically growing tumours respectively . These findings are inconsistent with previous studies, necessitating further validation . In our opinion, there is currently no reliable non-invasive or minimally invasive alternative to DSNB.
Our study has several limitations. First, this is a retrospective study with all potential drawbacks the study design brings. For instance, a patient selection bias cannot be excluded in the patients staged by DSNB before it became a routine procedure and the analysis of complications is dependent on the intensity of follow-up and quality of the medical charts. Second, regional recurrences will usually appear within 2 years after the primary treatment . Hence, a minimum of 24 months' follow-up is desirable. Our median follow-up was 21 months but with a wide range (4–143 months). Since further regional recurrences can appear in the future, our true FNR might be higher than we report now. Finally, since our DSNB protocol has changed during the study period, a just comparison with other series is difficult.
We owe a debt of gratitude to urologist Swen-Olof Andersson (MD, PhD) for insightful input on the manuscript, and nuclear physicist Håkan Geijer (MD, PhD) with staff and pathologist Mats Karlsson (MD, PhD) for their dedication and collaboration.