- Email: [email protected]
102 POSTER Interval nodes in sentinel lymph node (SLN) mapping and dissection for melanoma patients
Poster session: Melanoma metastases (median time to recurrence 17 months); 9% presented with synchronous lymph node metastases. Most removed nodes were found in the neck (median 24.5) as compared to the axilla (median 18.0) and inguinal (median 13.0). Most positive nodes were found in the axilla (median 3.0) and inguinal (median 3.0); in the neck a median of 2.0. The tumour size (cm) in the axilla (4.08±1.8), inguinal (3.09±1.43) and neck (1.92±0.88) was significantly different (p=0.04). Mean SUVmax was 9.56±7.03 and was significantly higher for age>58 years (p=0.006), lymph nodes in the axilla (p=0.021) and tumour size ≥3 cm (p=0.010). In multivariate logistic regression (model with sex, age, localisation, removed nodes, positive nodes en tumour size) only tumour size was significantly correlated with the SUVmax (p=0.04). Conclusion: Most removed nodes were found in the neck; most positive and largest tumour size was found in the axilla. SUVmax was correlated with age, localisation and tumour size; however only tumour size was found to be significantly correlated in multivariate analysis. In future survival analysis an interaction term between tumour size and SUVmax should be taking into account. The impact of a high SUVmax in melanoma lymph node metastases should be further explored with respect to local and systemic adjuvant treatment strategies and outcome with respect to disease free and overall survival. 100
POSTER
Does a slightly elevated S100b in melanoma patients without demonstrable disease have clinical significance? T. van der Bom 1 , O.E. Nieweg 1 , C.M. Korse 2 , J.M.G. Bonfrer 2 , B.B.R. Kroon 1 . 1 The Netherlands Cancer Institute / Antoni van Leeuwenhoek Hospital, Department of Surgery, Amsterdam, The Netherlands; 2 The Netherlands Cancer Institute / Antoni van Leeuwenhoek Hospital, Departement of Clinical Chemistry, Amsterdam, The Netherlands Introduction. Some melanoma patients have a slightly elevated tumour marker S100B without demonstrable disease. The aims of this study were to examine the incidence of this unnerving problem, to find possible explanations and to consider clinical implications. Patients and methods. Subjects were identified from a database of all patients who were tested for S100B in our hospital between February 1999 and May 2006. S100B levels were assumed to be slightly elevated between 0.11 μg/l and 0.14 μg/l. When a slightly increased S100B level was found, the test was repeated. Patients were included in the study if both tests showed values within the range described above, if the imaging studies prompted by the elevated S100B did not reveal recurrence, and when there were no signs of tumour progression in the subsequent two months. The charts were reviewed for potential explanations and to document any recurrent disease down the line. Suspected sites of recurrence were biopsied. Results. Thirty-one patients out of 2120 in whom this marker was measured (1.5%) had slightly elevated S100B levels. Five of these 31 patients (16%) developed a recurrence during follow up. The recurrences were detected 10 to 65 months later, median 12 months. Pathological proof was obtained in each patient. Two of them were treated with curative intent, one by inguinal node dissection and the other by local excision. Both of them are alive and free of disease. The other three patients were treated palliatively with chemotherapy. One of them died from brain metastases. Six of the 31 patients (19%) were found to have increased tumour marker levels concurrent with another active disease process that has been linked to S100B elevation in the literature, without signs of tumour progression. The S100B elevation could not be explained in the remaining twenty patients (64%). They remained disease-free within a median follow up of 25 months. Conclusions. Slightly elevated S100B levels on multiple occasions are uncommon and do not necessarily signify tumour progression. Some healthy people are outside the normal range. In some cases there is another plausible explanation. Close follow up of patients with a slightly elevated S100B is recommended as a recurrence may develop for which treatment with a curative intent is possible.
101
S31 POSTER
Statistical matters relating to sentinel node biopsy in melanoma J.M. Thomas. Royal Marsden Hospital, Surgical Oncology, London, United Kingdom At the time of writing (June 2006) the results of MSLT-1 relating to survival are awaited. However, we know from previous presentations and abstracts that early lymphadenectomy based on sentinel node (SN)- status provides no overall survival advantage. The authors of MSLT-1 have consistently claimed a highly significant disease-free survival (DFS) advantage for patients undergoing early lymphadenectomy. This author has argued that there will inevitably be fewer events in the test arm of MSLT-1 because most patients destined to develop palpable nodal recurrence were SN-positive and treated by lymphadenectomy at the outset. Therefore, any DFS is likely to be a consequence of trial design rather than a genuine palliative benefit for SN-positive patients. Of greater importance, there is a fundamental biological and statistical error at the core of the sentinel node biopsy hypothesis namely that the survival of SN-positive patients who undergo early lymphadenectomy can be compared to the survival of patients who develop palpable nodal recurrence providing that the calculation is made from the date of wide local excision of the primary tumour. In other words, that all positive SNs will inevitably progress to palpable nodal recurrence if not removed. This core assumption is not correct and comparison of survival of these two groups of patients is statistically invalid because they are selected post-randomisation groups. Consequently any conclusion or extrapolation drawn from this analysis is incorrect. Evidence will be presented to show that the incidence of false-positivity within the SN in MSLT-1 is approximately 34% based on figures presented in a preliminary report [1]. These patients are given incorrect prognostic information, undergo unnecessary lymphadenectomy and possibly unnecessary adjuvant therapy. Reference [1] Morton DL, Cochran AJ, Thompson JF et al. Sentinel node biopsy for early-stage melanoma. Ann Surg 2005;242:302-313. 102
POSTER
Interval nodes in sentinel lymph node (SLN) mapping and dissection for melanoma patients M. Matter 1 , M. Alloua 2 , M. Nicod 2 , D. Liénard 3 , O. Gugerli 4 , D. Guggisberg 5 , J. Cerottini 5 , F. Lejeune 3 . 1 Centre Hospitalier Universitaire Vaudois, Visceral Surgery, Lausanne, Switzerland; 2 Centre Hospitalier Universitaire Vaudois, Nuclear Medicine, Lausanne, Switzerland; 3 Centre Hospitalier Universitaire Vaudois, Oncology, Lausanne, Switzerland; 4 Centre Hospitalier Universitaire Vaudois, Pathology, Lausanne, Switzerland; 5 Centre Hospitalier Universitaire Vaudois, Dermatology, Lausanne, Switzerland Introduction: Interval node is defined as an isolated lymph node or a group of lymph nodes located between the primary melanoma and a well anatomically defined lymph node group draining directly the skin and belonging to the following: superficial and deep inguinal, iliac and para-aortic, axillary, subclavian, mammary, cervical. Theoretically and as shown in some reports, these SLN are at same metastatic risk as SLN in usual areas. Material and method: Prospective cohort of consecutive patients between 1997 and 2006 in one centre and one surgical team. Triple technique included lymphoscintigraphy, blue dye and gamma probe. SLN were examined by serial sectioning and immunohistochemistry with Melan A and protein S-100. ENT and mucosal melanoma were excluded. All patients with metastatic SLN were proposed for SLND. When an interval node was discovered, the entire local group was removed. Results: Repartition of 372 patients with primary melanoma showed 171 with trunk localisation, 64 upper limb and 137 lower limb. Overall 91/372 (24,5%) of SLN were metastatic. Interval sentinel lymph node were observed in 17 patients. Nine patients out of 17 (53%) had dorsal, paravertebral or scapular location of the primary, 4 patients had upper limb, 2 lower limb and 2 lumbar melanoma. One patient with a dorsal melanoma had an exclusive drainage in a cervico-scapular area, whose removal showed only a blue lymph channel but no SLN. Apart from the interval SN, 12 patients had 1 other SN area and 4 patients 2 other SN areas. Surprisingly no interval SN disclosed metastatic disease, whereas 3 patients had other metastatic SN: overall 3/17 patients were N+ (17,6%).
S32
Poster session: Melanoma
Discussion and conclusion: Preoperative lymphoscintigraphy must review all known lymphatic areas in order to exclude an interval node. We observed no metastatic interval SLN despite the fact that these SLN are at the same risk as the other SLN. Upper dorsal, paravertebral or scapular location of the primary are at higher risk of presenting interval SLN. 103
POSTER
Isolated limb perfusion with melphalan and tumour necrosis factor alpha demonstrates differential efficacy for advanced melanoma and irresectible extremity soft tissue sarcoma A.J. Hayes, S. Neuhaus, M. Clark, J.M. Thomas. Royal Marsden Hospital, Sarcoma and Melanoma Unit, Department of Surgery, London, United Kingdom Background: Isolated Limb Perfusion (ILP) with melphalan is used in the treatment of advanced in-transit melanoma but has no real efficacy for irresectable soft tissue sarcomas arising in the extremities. The addition of Tumour Necrosis Factor-alpha (TNF-α) may increase response rates for bulky melanoma and for sarcoma. Methods: Between October 2000 and April 2004, 49 ILPs were performed with melphalan and TNF-α. All procedures were performed using continuous leakage monitoring and regional hyperthermia. Outcome was analysed retrospectively. Response rates for patients with sarcoma and melanoma were compared using the Mann-Whitney U test. Curves for progression free survival were constructed using the Kaplan Meier method and the Log Rank test was used to assess statistical significance of differences in progression free survival. Results: Forty nine ILPs were performed for melanoma (n = 30), sarcoma (n = 16) or other tumours (n = 3). The most common indications were widespread in-transit disease for melanoma (n = 29), and irresectable primary disease for sarcoma, (n = 9). None of the patients undergoing resection for sarcoma underwent subsequent marginal resections of residual tumour, as all of the perfused sarcomas were in the distal extremities and were irresectable other than by amputation. Complete and partial responses are shown in Table 1. Progression free survival was significantly less for patients with sarcoma than melanoma (p=0.0476). Four of 16 patients with sarcoma subsequently required amputation for progressive disease but only one has developed distant metastases. Of the 26 patients who underwent perfusions for melanoma without distant metastases at the time of perfusion, nine have subsequently developed distant metastases and five have died of distant disease within the follow up period. None of these patients underwent a palliative amputation after perfusion. Table 1 Complete response Partial response Objective response or stable disease % patients who subsequently progressed locally (after initial response of any form) Median time to progression in months (range)
Melanoma
Sarcoma
40% 37% 15%
20% 33% 40%
40%
71%
Median 6 (1-8)
Median 5 (1-17)
Conclusions: ILP with melphalan and TNF-α is a valuable treatment for advanced in transit melanoma. Significant response rates were also seen in irresectable sarcoma although the duration of response was limited. This data would indicate that when ILP with melphalan and TNF-αis employed purely as regional chemotherapy in the absence of a subsequent surgical resection the benefit for patients with sarcoma may be limited and subsequent amputation may still be required. In contrast, because the patients with melanoma have much more likelihood of developing distant disease in the short term, the relative palliative benefit of ILP is greater than for patients with sarcoma.
104
POSTER
Inguinal lymphatic drainage patterns of patients with a melanoma on the lower trunk or leg M. van Rijk, R. Valdés Olmos, O. Nieweg, C. Hoefnagel, B. Kroon. The Netherlands Cancer Institute, Department of Surgery, Amsterdam, The Netherlands Methods: Between 1993 and 2005, 199 melanoma patients underwent biopsy of a sentinel lymph node in the groin. Their lymphoscintigraphic images were retrospectively evaluated with regard to the location of the sentinel nodes. To this end, the groin was divided in four quadrants and the oval fossa as center zone (Daseler’s zones). Results: Of the 164 patients with a melanoma on the leg, 95% had a sentinel node in the lower quadrants of the groin, 17% had drainage towards the upper quadrants and 7% had a sentinel node in the central region. All 35 patients with a melanoma on the trunk had sentinel nodes in the upper quadrants, one also had a sentinel node in the central zone but none had drainage to the inferior quadrants. Conclusions: Melanomas located on the trunk drain initially towards lymph nodes cranially from the oval fossa and melanomas on the leg drain predominantly to nodes caudally from this anatomical landmark. Knowledge of these pathways may guide the ultrasonographist in the preoperative workup and may aid in the post-operative surveillance if no sentinel lymph node biopsy is performed. 105
POSTER
Sentinel lymph node (SLN) status is an important survival risk factor in melanoma patients M. Matter 1 , D. Roulin 1 , P. Bady 2 , D. Liénard 3 , O. Gugerli 4 , D. Guggisberg 5 , J.P. Cerottini 5 , A. Boubaker 6 , M. Alloua 6 , F. Lejeune 3 . 1 Centre Hospitalier Universitaire Vaudois, Visceral Surgery, Lausanne, Switzerland; 2 Centre Hospitalier Universitaire Vaudois, Clinical Epidemiology, Lausanne, Switzerland; 3 Centre Hospitalier Universitaire Vaudois, Oncology, Lausanne, Switzerland; 4 Centre Hospitalier Universitaire Vaudois, Pathology, Lausanne, Switzerland; 5 Centre Hospitalier Universitaire Vaudois, Dermatology, Lausanne, Switzerland; 6 Centre Hospitalier Universitaire Vaudois, Nuclear Medicine, Lausanne, Switzerland Introduction: While awaiting prospective randomized studies comparing patients’survival with selective (SLND, based on SLN) versus elective radical lymph node dissection (ELND), SLN remains a recognised way to avoid ELND in SLN negative patients. As SLN is already proposed in patients cumulating known risk factors (Breslow /ulceration), the aim of the study is to determine if a metastatic SLN itself is an independent survival risk factor. Material and method: Prospective cohort of consecutive patients between 1997 and 2004 in one centre and one surgical team. Triple technique included lymphoscintigraphy, blue dye and gamma probe. SLN were examined by serial sectioning and immunohistochemistry with Melan A and protein S-100. All patients with metastatic SLN were proposed for SLND. Local, distant and overall survival were correlated with sex, age, type of melanoma, ulceration, T stage (Breslow), localisation, SLN status. Statistical analysis: tables with chi square test, survival analysis with the Cox’ proportional hazards regression test. Results: In 325 patients with primary melanoma 175 (54%) were male, 90 (28%) were older than 65 y-o.; localisation showed 90 (28%) H&N, 167 (51%) upper/lower limbs and 130 (40%) trunk melanomas. Stage repartition: T1-T2: 190 (58%), T3: 94 (29%) and T4: 41 (13%) patients and 88 (27%) had an ulcerated primary. Seventy four patients (23%) had at least one metastatic SLN, which was significantly associated with T stage/Breslow (p<0,001) and ulceration (p=0,003). With a follow-up between 1 and 8 years, disease-free survival was significantly decreased by Breslow index (RR 5,2 p < 0,001), metastatic SLN (RR 4,9 p<0,001) and sex (M: RR 2,9 p 0,001). Disease specific survival was significantly decreased by metastatic SLN (RR 8,4 p<0,001), sex (M: RR 6,1 p <0,001), Breslow index in T3 (RR 3,4 p 0,007, but not for T4) and ulceration (RR 2,5 p0,02). In a second model we show that presence of local and/or distant recurrences remains an important adverse prognostic factor for overall survival (RR 38,2 p<0,001), but dependant of Brelow index and positivity of SLN. Discussion and conclusion: despite a clear association with Breslow index and ulceration, metastatic SLN is an important risk factor for disease-free
POSTER
Does a slightly elevated S100b in melanoma patients without demonstrable disease have clinical significance? T. van der Bom 1 , O.E. Nieweg 1 , C.M. Korse 2 , J.M.G. Bonfrer 2 , B.B.R. Kroon 1 . 1 The Netherlands Cancer Institute / Antoni van Leeuwenhoek Hospital, Department of Surgery, Amsterdam, The Netherlands; 2 The Netherlands Cancer Institute / Antoni van Leeuwenhoek Hospital, Departement of Clinical Chemistry, Amsterdam, The Netherlands Introduction. Some melanoma patients have a slightly elevated tumour marker S100B without demonstrable disease. The aims of this study were to examine the incidence of this unnerving problem, to find possible explanations and to consider clinical implications. Patients and methods. Subjects were identified from a database of all patients who were tested for S100B in our hospital between February 1999 and May 2006. S100B levels were assumed to be slightly elevated between 0.11 μg/l and 0.14 μg/l. When a slightly increased S100B level was found, the test was repeated. Patients were included in the study if both tests showed values within the range described above, if the imaging studies prompted by the elevated S100B did not reveal recurrence, and when there were no signs of tumour progression in the subsequent two months. The charts were reviewed for potential explanations and to document any recurrent disease down the line. Suspected sites of recurrence were biopsied. Results. Thirty-one patients out of 2120 in whom this marker was measured (1.5%) had slightly elevated S100B levels. Five of these 31 patients (16%) developed a recurrence during follow up. The recurrences were detected 10 to 65 months later, median 12 months. Pathological proof was obtained in each patient. Two of them were treated with curative intent, one by inguinal node dissection and the other by local excision. Both of them are alive and free of disease. The other three patients were treated palliatively with chemotherapy. One of them died from brain metastases. Six of the 31 patients (19%) were found to have increased tumour marker levels concurrent with another active disease process that has been linked to S100B elevation in the literature, without signs of tumour progression. The S100B elevation could not be explained in the remaining twenty patients (64%). They remained disease-free within a median follow up of 25 months. Conclusions. Slightly elevated S100B levels on multiple occasions are uncommon and do not necessarily signify tumour progression. Some healthy people are outside the normal range. In some cases there is another plausible explanation. Close follow up of patients with a slightly elevated S100B is recommended as a recurrence may develop for which treatment with a curative intent is possible.
101
S31 POSTER
Statistical matters relating to sentinel node biopsy in melanoma J.M. Thomas. Royal Marsden Hospital, Surgical Oncology, London, United Kingdom At the time of writing (June 2006) the results of MSLT-1 relating to survival are awaited. However, we know from previous presentations and abstracts that early lymphadenectomy based on sentinel node (SN)- status provides no overall survival advantage. The authors of MSLT-1 have consistently claimed a highly significant disease-free survival (DFS) advantage for patients undergoing early lymphadenectomy. This author has argued that there will inevitably be fewer events in the test arm of MSLT-1 because most patients destined to develop palpable nodal recurrence were SN-positive and treated by lymphadenectomy at the outset. Therefore, any DFS is likely to be a consequence of trial design rather than a genuine palliative benefit for SN-positive patients. Of greater importance, there is a fundamental biological and statistical error at the core of the sentinel node biopsy hypothesis namely that the survival of SN-positive patients who undergo early lymphadenectomy can be compared to the survival of patients who develop palpable nodal recurrence providing that the calculation is made from the date of wide local excision of the primary tumour. In other words, that all positive SNs will inevitably progress to palpable nodal recurrence if not removed. This core assumption is not correct and comparison of survival of these two groups of patients is statistically invalid because they are selected post-randomisation groups. Consequently any conclusion or extrapolation drawn from this analysis is incorrect. Evidence will be presented to show that the incidence of false-positivity within the SN in MSLT-1 is approximately 34% based on figures presented in a preliminary report [1]. These patients are given incorrect prognostic information, undergo unnecessary lymphadenectomy and possibly unnecessary adjuvant therapy. Reference [1] Morton DL, Cochran AJ, Thompson JF et al. Sentinel node biopsy for early-stage melanoma. Ann Surg 2005;242:302-313. 102
POSTER
Interval nodes in sentinel lymph node (SLN) mapping and dissection for melanoma patients M. Matter 1 , M. Alloua 2 , M. Nicod 2 , D. Liénard 3 , O. Gugerli 4 , D. Guggisberg 5 , J. Cerottini 5 , F. Lejeune 3 . 1 Centre Hospitalier Universitaire Vaudois, Visceral Surgery, Lausanne, Switzerland; 2 Centre Hospitalier Universitaire Vaudois, Nuclear Medicine, Lausanne, Switzerland; 3 Centre Hospitalier Universitaire Vaudois, Oncology, Lausanne, Switzerland; 4 Centre Hospitalier Universitaire Vaudois, Pathology, Lausanne, Switzerland; 5 Centre Hospitalier Universitaire Vaudois, Dermatology, Lausanne, Switzerland Introduction: Interval node is defined as an isolated lymph node or a group of lymph nodes located between the primary melanoma and a well anatomically defined lymph node group draining directly the skin and belonging to the following: superficial and deep inguinal, iliac and para-aortic, axillary, subclavian, mammary, cervical. Theoretically and as shown in some reports, these SLN are at same metastatic risk as SLN in usual areas. Material and method: Prospective cohort of consecutive patients between 1997 and 2006 in one centre and one surgical team. Triple technique included lymphoscintigraphy, blue dye and gamma probe. SLN were examined by serial sectioning and immunohistochemistry with Melan A and protein S-100. ENT and mucosal melanoma were excluded. All patients with metastatic SLN were proposed for SLND. When an interval node was discovered, the entire local group was removed. Results: Repartition of 372 patients with primary melanoma showed 171 with trunk localisation, 64 upper limb and 137 lower limb. Overall 91/372 (24,5%) of SLN were metastatic. Interval sentinel lymph node were observed in 17 patients. Nine patients out of 17 (53%) had dorsal, paravertebral or scapular location of the primary, 4 patients had upper limb, 2 lower limb and 2 lumbar melanoma. One patient with a dorsal melanoma had an exclusive drainage in a cervico-scapular area, whose removal showed only a blue lymph channel but no SLN. Apart from the interval SN, 12 patients had 1 other SN area and 4 patients 2 other SN areas. Surprisingly no interval SN disclosed metastatic disease, whereas 3 patients had other metastatic SN: overall 3/17 patients were N+ (17,6%).
S32
Poster session: Melanoma
Discussion and conclusion: Preoperative lymphoscintigraphy must review all known lymphatic areas in order to exclude an interval node. We observed no metastatic interval SLN despite the fact that these SLN are at the same risk as the other SLN. Upper dorsal, paravertebral or scapular location of the primary are at higher risk of presenting interval SLN. 103
POSTER
Isolated limb perfusion with melphalan and tumour necrosis factor alpha demonstrates differential efficacy for advanced melanoma and irresectible extremity soft tissue sarcoma A.J. Hayes, S. Neuhaus, M. Clark, J.M. Thomas. Royal Marsden Hospital, Sarcoma and Melanoma Unit, Department of Surgery, London, United Kingdom Background: Isolated Limb Perfusion (ILP) with melphalan is used in the treatment of advanced in-transit melanoma but has no real efficacy for irresectable soft tissue sarcomas arising in the extremities. The addition of Tumour Necrosis Factor-alpha (TNF-α) may increase response rates for bulky melanoma and for sarcoma. Methods: Between October 2000 and April 2004, 49 ILPs were performed with melphalan and TNF-α. All procedures were performed using continuous leakage monitoring and regional hyperthermia. Outcome was analysed retrospectively. Response rates for patients with sarcoma and melanoma were compared using the Mann-Whitney U test. Curves for progression free survival were constructed using the Kaplan Meier method and the Log Rank test was used to assess statistical significance of differences in progression free survival. Results: Forty nine ILPs were performed for melanoma (n = 30), sarcoma (n = 16) or other tumours (n = 3). The most common indications were widespread in-transit disease for melanoma (n = 29), and irresectable primary disease for sarcoma, (n = 9). None of the patients undergoing resection for sarcoma underwent subsequent marginal resections of residual tumour, as all of the perfused sarcomas were in the distal extremities and were irresectable other than by amputation. Complete and partial responses are shown in Table 1. Progression free survival was significantly less for patients with sarcoma than melanoma (p=0.0476). Four of 16 patients with sarcoma subsequently required amputation for progressive disease but only one has developed distant metastases. Of the 26 patients who underwent perfusions for melanoma without distant metastases at the time of perfusion, nine have subsequently developed distant metastases and five have died of distant disease within the follow up period. None of these patients underwent a palliative amputation after perfusion. Table 1 Complete response Partial response Objective response or stable disease % patients who subsequently progressed locally (after initial response of any form) Median time to progression in months (range)
Melanoma
Sarcoma
40% 37% 15%
20% 33% 40%
40%
71%
Median 6 (1-8)
Median 5 (1-17)
Conclusions: ILP with melphalan and TNF-α is a valuable treatment for advanced in transit melanoma. Significant response rates were also seen in irresectable sarcoma although the duration of response was limited. This data would indicate that when ILP with melphalan and TNF-αis employed purely as regional chemotherapy in the absence of a subsequent surgical resection the benefit for patients with sarcoma may be limited and subsequent amputation may still be required. In contrast, because the patients with melanoma have much more likelihood of developing distant disease in the short term, the relative palliative benefit of ILP is greater than for patients with sarcoma.
104
POSTER
Inguinal lymphatic drainage patterns of patients with a melanoma on the lower trunk or leg M. van Rijk, R. Valdés Olmos, O. Nieweg, C. Hoefnagel, B. Kroon. The Netherlands Cancer Institute, Department of Surgery, Amsterdam, The Netherlands Methods: Between 1993 and 2005, 199 melanoma patients underwent biopsy of a sentinel lymph node in the groin. Their lymphoscintigraphic images were retrospectively evaluated with regard to the location of the sentinel nodes. To this end, the groin was divided in four quadrants and the oval fossa as center zone (Daseler’s zones). Results: Of the 164 patients with a melanoma on the leg, 95% had a sentinel node in the lower quadrants of the groin, 17% had drainage towards the upper quadrants and 7% had a sentinel node in the central region. All 35 patients with a melanoma on the trunk had sentinel nodes in the upper quadrants, one also had a sentinel node in the central zone but none had drainage to the inferior quadrants. Conclusions: Melanomas located on the trunk drain initially towards lymph nodes cranially from the oval fossa and melanomas on the leg drain predominantly to nodes caudally from this anatomical landmark. Knowledge of these pathways may guide the ultrasonographist in the preoperative workup and may aid in the post-operative surveillance if no sentinel lymph node biopsy is performed. 105
POSTER
Sentinel lymph node (SLN) status is an important survival risk factor in melanoma patients M. Matter 1 , D. Roulin 1 , P. Bady 2 , D. Liénard 3 , O. Gugerli 4 , D. Guggisberg 5 , J.P. Cerottini 5 , A. Boubaker 6 , M. Alloua 6 , F. Lejeune 3 . 1 Centre Hospitalier Universitaire Vaudois, Visceral Surgery, Lausanne, Switzerland; 2 Centre Hospitalier Universitaire Vaudois, Clinical Epidemiology, Lausanne, Switzerland; 3 Centre Hospitalier Universitaire Vaudois, Oncology, Lausanne, Switzerland; 4 Centre Hospitalier Universitaire Vaudois, Pathology, Lausanne, Switzerland; 5 Centre Hospitalier Universitaire Vaudois, Dermatology, Lausanne, Switzerland; 6 Centre Hospitalier Universitaire Vaudois, Nuclear Medicine, Lausanne, Switzerland Introduction: While awaiting prospective randomized studies comparing patients’survival with selective (SLND, based on SLN) versus elective radical lymph node dissection (ELND), SLN remains a recognised way to avoid ELND in SLN negative patients. As SLN is already proposed in patients cumulating known risk factors (Breslow /ulceration), the aim of the study is to determine if a metastatic SLN itself is an independent survival risk factor. Material and method: Prospective cohort of consecutive patients between 1997 and 2004 in one centre and one surgical team. Triple technique included lymphoscintigraphy, blue dye and gamma probe. SLN were examined by serial sectioning and immunohistochemistry with Melan A and protein S-100. All patients with metastatic SLN were proposed for SLND. Local, distant and overall survival were correlated with sex, age, type of melanoma, ulceration, T stage (Breslow), localisation, SLN status. Statistical analysis: tables with chi square test, survival analysis with the Cox’ proportional hazards regression test. Results: In 325 patients with primary melanoma 175 (54%) were male, 90 (28%) were older than 65 y-o.; localisation showed 90 (28%) H&N, 167 (51%) upper/lower limbs and 130 (40%) trunk melanomas. Stage repartition: T1-T2: 190 (58%), T3: 94 (29%) and T4: 41 (13%) patients and 88 (27%) had an ulcerated primary. Seventy four patients (23%) had at least one metastatic SLN, which was significantly associated with T stage/Breslow (p<0,001) and ulceration (p=0,003). With a follow-up between 1 and 8 years, disease-free survival was significantly decreased by Breslow index (RR 5,2 p < 0,001), metastatic SLN (RR 4,9 p<0,001) and sex (M: RR 2,9 p 0,001). Disease specific survival was significantly decreased by metastatic SLN (RR 8,4 p<0,001), sex (M: RR 6,1 p <0,001), Breslow index in T3 (RR 3,4 p 0,007, but not for T4) and ulceration (RR 2,5 p0,02). In a second model we show that presence of local and/or distant recurrences remains an important adverse prognostic factor for overall survival (RR 38,2 p<0,001), but dependant of Brelow index and positivity of SLN. Discussion and conclusion: despite a clear association with Breslow index and ulceration, metastatic SLN is an important risk factor for disease-free