S. Horenblas, MD, PhD, FEBU, Head of the Department of Urology, Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital, and Professor of Urologic Oncology, Department of Urology, Free University Medical Center, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands.e-mail: email@example.com
There is no controversy about the need for surgery in patients with nodal invasion; about half of these patients can be cured with adequate surgery (Table 1) [1–5]. However, controversy remains about the extent of surgery. Should patients with unilateral involvement undergo bilateral lymphadenectomy? Should inguinal lymphadenectomy always be followed by pelvic lymphadenectomy? The treatment strategy in this group of patients should focus on measures that decrease the morbidity associated with lymph node resection. Up to 84% of the patients have short- or long-term complications (Table 2) [1,6–26]. With the advent of effective chemotherapy there is a further need to identify patients who could benefit from neoadjuvant or adjuvant treatment.
Surgery in patients with a low metastatic load gives excellent survival probabilities (Table 1). Hence the general advice is to undertake lymphadenectomy in all patients, preferably before the advent of clinical signs (apart from patients with carcinoma in situ or small non-infiltrating tumours). In almost 60% of the patients no single metastasis is found in the resected specimen [11,14,17–21,25,27–36]. Thus this advice leads to exposing all patients to the potential morbidity of lymphadenectomy, with benefit to only 40% at most. For many years there has been controversy between the proponents of lymphadenectomy at an early stage, before any clinical signs, and lymphadenectomy at the first sign of lymph node involvement, so-called ‘delayed lymphadenectomy’. The treatment strategy in this group of patients should focus on methods of timely identification of metastasis, to decrease the number of unnecessary operations and to identify patients who could benefit from lymphadenectomy. This review will focus on these controversial issues and the management strategy developed at the author's institute will be discussed.
The role of lymphadenectomy
Surgery plays a paramount role in the cure of patients with metastatic penile cancer, with cure achieved by surgery alone in ≈ 75% of patients presenting with one or two invaded nodes (Table 1 and Table 3) [14,16,18,19,21,35,37,38]. Even in patients with pelvic node involvement, cure by surgery alone can be obtained in almost 20% (Table 3). Surgery should strive for the complete removal of all lymph nodes associated with the drainage basin because of its therapeutic potential. All lymph nodes of the inguinal region, irrespective of their anatomical location, are considered to be first-line regional nodes. Similarly, all pelvic nodes are considered to be second-line regional nodes and should be removed on indication. The timing and extent of lymphadenectomy have been discussed at length. A scientifically sound answer on these issues can be given only after prospective randomized trials, but the rarity of this tumour precludes this type of clinical endeavour. The answers to date can only be obtained from large single-institution experience or from a form of meta-analysis. The issue of the timing of a lymph node dissection (LND) is of major interest in those patients who present with no clinical sign of metastatic disease. The critical issue is the unreliability of current clinical methods to detect occult lymph node metastasis (see Part 1). Because survival is clearly related to the relative amount of lymph node metastasis there is a clinical urgency to remove these metastases as soon as possible and preferably before clinical evidence of invasion has developed. Reviewing the literature, the issue of timing in this group of patients often seems a choice between ‘too early’ and ‘too late’. Proponents of an early or immediate lymphadenectomy point out that patients who do not undergo surgery but are observed often present at a stage where cure is no longer possible [1,21,30,39,40]. This situation usually occurs where patients do not meet rigid follow-up schedules. A meticulous follow-up is the key to success; the ‘wait-and-see’ strategy in this group of patients is similar to that used for patients with testicular cancer. Most importantly, observation in patients with clinically negative nodes can only be used if the follow-up is regular and at short intervals for ≈ 3 years; an interval of 2 months is advised. With this strict scheme the author's group have been able to salvage almost all patients who develop lymph node metastasis after treatment of the primary tumour . Nodal recurrences after 3 years are extremely rare [41,42].
Table 3. The 5-year survival by clinical and pathological N-stage
Even if lymphadenectomy is restricted to patients at high risk of developing metastasis, lymphadenectomy is beneficial to only 40%; in 60% the surgery is unnecessary, as no metastasis is found in the lymphadenectomy specimen . Ultrasound-guided fine-needle aspiration biopsy and the revival of sentinel node biopsy could end the discussion of early vs late LND by identifying patients with occult metastasis as soon as possible. Lymphadenectomy is not benign surgery but is prone to many complications, discussed later.
Patients with a confirmed metastasis must undergo surgery, but how intense the surgery should be is debatable. Various strategies have been developed. Some advise to always start with a pelvic node dissection and to continue with an inguinal dissection in those patients with no pelvic node involvement on frozen-section analysis [32,43]. The basic philosophy behind this strategy is the argument that surgery alone is futile in patients with pelvic node metastases. While this is true for gross metastatic involvement, it is not the case for microscopic invasion, where survival rates of 17–54% have been reported (Table 1).
In general, 20–30% of patients with positive inguinal nodes have positive pelvic nodes [6,12,14,33,35]. The likelihood of pelvic node involvement is related to the number of positive nodes in the inguinal specimen. Patients with one inguinal node metastasis and no involvement of the most proximal lymph node in the dissection specimen have a very low probability of pelvic node involvement . At the author's institute a pelvic LND is considered unnecessary in these patients. In all other patients with inguinal node involvement a unilateral pelvic node dissection of the affected site is recommended. This group of patients at high risk for pelvic metastasis could potentially benefit from pelvic CT, to detect pathological nodes suitable for fine-needle aspiration. The patients with macroscopic lymph node metastasis cannot be cured by surgery alone and are candidates for adjuvant or neoadjuvant treatment.
Should a bilateral inguinal dissection be undertaken in all patients with unilateral inguinal involvement? At the author's institute the timing of detection, the number of palpable nodes and the number of positive nodes found in the resection specimen are considered. Patients who develop a unilateral nodal recurrence during the follow-up are managed by unilateral dissection. The likelihood of bilateral involvement is related to the number of involved nodes in the resected specimen [14,35]. With two or more metastases the probability of occult contralateral involvement is 30% and warrants an early inguinal LND. Ultrasound-guided fine-needle aspiration biopsy and dynamic sentinel node biopsy could also solve the problem in those patients presenting initially with unilateral positive nodes, by identifying the occult metastasis in the contralateral inguinal region (Fig. 1) .
The role of radiation therapy for regional disease
Prophylactic radiation therapy in clinically node-negative patients is not advised, for the following reasons. First, most patients will not benefit; the argument is the same as for early LND. Second, all patients will be exposed to the complications of radiation therapy, e.g. short-term complications like epidermiolysis (Fig. 2) and long-term effects like lymphoedema and fibrosis. Third, the follow-up is more complicated because of the fibrotic changes, making a physical examination unreliable . Fourth, the data so far indicate a similar recurrence rate in series where radiation was compared with surveillance. Radiation failed to alter the course of the disease and in 17–42% of patients, metastasis in the regional nodes developed [7,37,46–50]. Comparing surgery with radiation is even less favourable for radiation. In a small series the 5-year survival after surgery was 67%, compared with 25% for the irradiated patients . In a series from India where surveillance, radiation therapy and LND were compared (not in a randomized trial but on sequential clinically node-negative patients) the relapse rate was 4% in the surgery group and 17% in the radiation group. As positive nodes were found in 15% of the operated patients this suggests that radiation did not prevent many of the expected relapses . Why irradiation failed is difficult to understand, especially in patients with presumed micrometastases only. As none of the studies was truly prospective or randomized, patient selection probably plays a significant role and explains the differences.
Some series have used preoperative radiation therapy in patients with large (> 4 cm) and/or fixed regional nodes [14,50,52–54]. In such cases inoperable nodes can be rendered operable but it is unknown if the fixation is caused by inflammatory reactions, that tend to resolve spontaneously, or by cancerous growth. In the series of Ravi et al., where preoperative radiation therapy was given to all patients with large nodes, extranodal infiltration was found in 8%, in contrast to 33% in 43 unirradiated patients. Subsequently, there was a statistically significant difference in groin recurrences, i.e. 3% vs 19% in the unirradiated group. From data of patients with squamous cell carcinoma (SCC) in the head and neck region, adjuvant postoperative radiation is advised in patients with extensive metastases and/or extranodal spread . More recent reports suggest that this strategy should possibly be replaced by adjuvant chemotherapy .
There seems to be no role for radiation therapy as a primary treatment in patients with confirmed lymph node metastasis. The 5-year survival is with radiation therapy as primary treatment (25%) is half that of surgery (50%) [30,51,57–60].
The prognosis of patients with lymph node metastasis
Generally, half the patients presenting with lymph node metastasis survive > 5 years after lymph node resection. Survival is related to the number of involved nodes or the metastatic load, defined as the relative number of metastases. Survival values from recent series are shown in Table 3.
Unresectable metastasis is a challenging issue; others and the present author have reported favourable results for preoperative radiation therapy, enabling subsequent surgical removal [14,50,52,53,56]. In the absence of randomized studies it is unclear whether survival is improved by adding radiation therapy. Combination chemotherapy appears to provide major responses in about two-thirds of the patients; the tumour is made resectable in half and with durable disease-free survival in 15–30% . These values are based on two combination chemotherapy regimens, i.e. the combination of cisplatin and 5-fluorouracil and that of vincristine, bleomycin and methotrexate [61,62].
Prognostic factors like the grade of tumour, metastatic load and site of metastases restrict the curative potential of surgery. Does adjuvant treatment improve survival in patients with poor prognostic factors? Again, a definitive answer cannot be given in the absence of prospective randomized trials, but experience with other locations for SCC can be helpful. Based on experience in head and neck surgery, the author's group have given adjuvant radiation therapy to all patients with two or more nodal metastases, extracapsular growth and pelvic metastases . The major goal of this treatment is increased local control. Another treatment is adjuvant combination chemotherapy; the combination of vincristine, bleomycin and methotrexate was used from 1979 to 1990 in 25 patients after LND . Most showed extranodal spread and seven had pelvic metastases. After a follow-up of 5–15 years, 16% relapsed. However, analysis of the data showed that those patients most in need of adjuvant treatment (bilateral metastases, pelvic involvement) fared worst. Neoadjuvant and adjuvant chemotherapy after node dissection is beneficial in head and neck surgery and in gynaecological oncology, and therefore it is also advised in penile cancer. Whether neoadjuvant and adjuvant combination therapy is of any benefit or even better than radiation therapy remains unanswered but warrants further clinical studies.
Technique of lymphadenectomy
The patient is placed supine with the legs abducted and externally rotated. A variety of incisions has been advised; for inguinal dissection they can be divided into horizontal and vertical but because of the vasculature of the inguinal skin, horizontal incisions are to be preferred. The key to minimal morbidity after lymphadenectomy is correct skin handling and correct dissection of the skin flaps . No lymph nodes are found in the layer between the skin and subcutaneous fascia. At the author's institute a para-inguinal incision, a few centimetres below the groin crease, is the preferred type of incision (Fig. 3). The skin should be incised until the subcutaneous fascia is identified. Then the proximal and distal skin flaps are developed. The boundaries of the dissection are as follows: proximally, the inguinal ligament; distally, the entrance of Hunter's canal, where the femoral vessels go under the muscles of the leg; the medial boundary is the adductor muscle; the lateral boundary is the sartorius muscle. The floor of the dissection consists of the fascia lata, the femoral vessels and the pectineus muscle (Fig. 4a,b). As the femoral nerve is located beneath the fascia lata, it is not seen during standard LND. The surgeon who seldom performs this operation has a tendency to carry the operation too far laterally and medially. After the dissection, transposition of the sartorius muscle according to Baronofsky is strongly advised, especially if wound complications are foreseen or if adjuvant radiation therapy is anticipated . The origin of the sartorius muscle at the anterior superior iliac spine is transected and the muscle sutured to the inguinal ligament. Excellent covering of the vessels is thus achieved with no long-term sequelae. The healthy muscle tissue promotes primary wound healing (Fig. 5).
The following modifications of the boundaries have been suggested . The saphenous vein is spared instead of ligated. The regions lateral to the femoral artery are excluded from resection. The dissection is focused on the so-called deep part of the dissection, the nodes around the fossa ovalis. Frozen-section analysis is recommended to change the modified dissection into a regular dissection in the presence of positive nodes. A long-term follow-up in two series showed that this method was reliable and with less morbidity than standard LND [64,65]. This modified dissection was criticized by Lopes et al., who found recurrences in two of 13 patients. As a staging procedure the exact role of this modified dissection, for reliability and morbidity, should be compared with the dynamic sentinel node procedure.
After dissection the wound edges are inspected; any area with doubtful vascularization should be removed. Some have suggested the use of fluorescein and a Wood's lamp, to better detect the viability of the skin edges . Extensive experience has not been reported and it seems not to have gained wide acceptance. There are no comparative studies on the use of antibiotics but it seems reasonable to give prophylactic antibiotics at the time of surgery, as this type of surgery should be considered a contaminated procedure, because of the often co-existing inflammatory reactions in the lymph nodes. Along with the general recommendations for prophylactic antibiotics, the author's group give one dose at the start of anaesthesia. Before closing the wound suction drains are placed, to prevent the initial formation of lymphocele and increase the chances of primary wound healing. After one week the vacuum is removed and spontaneous drainage observed. Drains are removed if the drainage is < 50 mL/day. Immediately after surgery ambulation is strongly advised and supported with individually fitted elastic stockings. Patients are advised to use the elastic stockings for at least a 6 months after surgery. After that period the need for elastic stockings depends on the individual tendency to develop lymphoedema. At the author's institution it is routine to give low molecular weight heparin as prophylaxis for thrombosis, starting the evening before surgery.
Pelvic lymphadenectomy can be undertaken simultaneously with the inguinal dissection or as a separate procedure. In the first case, removal of the nodes through one or two incisions has been described; most authors prefer two separate incisions. A comparison of the various types of incisions showed the lowest complication rate when two incisions were used [21,24]. The pelvic node dissection is undertaken either through a lower abdominal midline incision or a unilateral muscle-splitting incision (Fig. 6). The boundaries of the pelvic node dissection are more extended than is usual for staging prostate or bladder cancer, the rationale being that surgical removal of microscopic metastasis is of therapeutic benefit. Proximally, the boundaries are the common iliac vessels; distally, the passage of lymphatic vessels to the groin; laterally, the ilio-inguinal nerve; medially, the bladder and prostate; and the base is the deepest part of the obturator fossa. Care must be taken to completely remove the obturator fossa, especially the space behind the external iliac vessels, all the way to the exit site of the obturator nerve. A large node can usually be found there and if left is prone to recurrence with intractable pain, because of neural ingrowth. After the dissection suction drains are left in place and removed if the spontaneous drainage is < 50 mL. In the author's institution it is routine to use low molecular weight heparin as prophylaxis for thrombosis, starting the evening before surgery.
Is it necessary to perform an en-bloc dissection, to remove the primary tumour with the regional lymph nodes as one continuous specimen? This type of surgery was first introduced by Young in 1926, but was modified in 1934 by Graves, where the primary tumour was removed. There then followed a period in which the presumed infection resolved. While oncologically sound, clinical experience shows this to be unnecessary in most patients. The rationale is that lymphatic spread occurs through a process of embolism, not through continuous growth. On practical and theoretical grounds it is advisable to temporize the treatment of the primary tumour and the regional nodes. One reason is the elimination of the primary tumour as a focus of infection and the other is the deposition of ‘in-transit’ metastases in the regional nodes, thus eliminating the potential risk of metastatic outgrowth in the tract between the primary tumour and the groin. En-bloc dissection is only indicated in patients with extensive primary or recurrent disease with bilateral node involvement.
Laparoscopy has not had a major role in pelvic node dissection for penile SCC, probably because with the technique to date the same degree of completeness of dissection cannot be attained as with open dissection. In view of the therapeutic potential of node dissection, nothing but a complete dissection should be attempted .
Closing the skin defect after LND can be difficult in patients with a large metastasis with invasion of the skin. The skin involved should be removed, but this leaves large defects. Various methods can be used to manage this surgical problem. After a sartorius-plasty the wound can heal by secondary intention. Healing can be improved and hastened by a split-skin graft on the granulation tissue. Another method is the so-called ‘skin-stretch’ method ; by gradually increasing the pressure on the skin edges in a cyclic fashion, large skin defects are closed with no tension. Island flaps that can be useful for closure are the rectus abdominis pedicle, gracilis pedicle and the tensor fascia lata pedicle [70,71]. In cases of en-bloc removal, using scrotal skin and mobilizing the abdominal wall can often close the defect. The femoral vessels can be protected by transposition of the scrotal content and suturing the scrotal content to the inguinal ligament (Figs 7a–d).
Complications of LND
There are many potential complications after LND, even in the most experienced hands. This explains the reluctance to use node dissection in every patient presenting with penile cancer. The most common cited complications are wound infection (10–20%), skin necrosis (Fig. 8) with or without wound dehiscence (Fig. 9) (14–65%), lymphocele/seroma (Fig. 10) (19–45%), lymphoedema (Fig. 11a,b) (2–100%), and other complications like haemorrhage and thrombosis. A summary of the most relevant complications is given in Table 2. Several ways to prevent complications have already been mentioned but will be reviewed briefly. Understanding the course of the vasculature of the inguinal skin is important in planning the incision. In a series from Brazil, three types of incision were compared, i.e. a large bi-iliac incision, a transverse S-shaped incision and a skin-bridge technique with two separate incisions. The latter had the lowest percentage of skin necrosis and the least lymphoedema . Utmost care must be taken when handling the skin, to preserve its viability. Antibiotic prophylaxis and preventing thrombosis with low molecular weight heparin is strongly recommended. Individually fitted stockings should be available immediate after surgery. Extensive experience in nodal dissection is an important factor in preventing complications ; this is also shown in results from the author's institute. Thus it is reasonable to centralize this type of surgery and to refer patients to institutes with a large experience.
Lymphadenectomy plays a paramount role in treating metastatic penile cancer. The timing of lymphadenectomy in clinically node-negative patients is often a choice between ‘too early’ or ‘too late’; more accurate staging methods should solve this problem. In patients with metastasis, adjuvant and neoadjuvant treatment strategies should be developed and assessed. Because of the morbidity of lymph node resection and the rarity of the disease, centralization of expertise is strongly recommended.