Molecular staging of the sentinel lymph node in melanoma patients: correlation with clinical outcome

Original article

Annals of Oncology 16: 1832–1840, 2005 doi:10.1093/annonc/mdi372 Published online 17 August 2005

Molecular staging of the sentinel lymph node in melanoma patients: correlation with clinical outcome A. Romanini1*, G. Manca2, D. Pellegrino2, R. Murr1, S. Sarti1, F. Bianchi2, A. AlSharif2, C. Orlandini1, V. Zucchi3, M. Castagna3, D. Gandini4, G. Salimbeni4, F. Ghiara5, P. Barachini5 & G. Mariani2 Division of 1Medical Oncology and 4Plastic Surgery, University Hospital, Pisa; 2Regional Center of Nuclear Medicine, 3Pathology Unit, Department of Surgery and 5 Division of Dermatology, University of Pisa Medical School, Pisa, Italy

Background: This study was designed to determine the debated prognostic significance of reverse transcriptase–polymerase chain reaction (RT–PCR) positivity in melanoma patients’ sentinel lymph node (SLN) negative by conventional histopathology (PATH). Patients and methods: Patients with primary stage I–II cutaneous melanoma underwent radioguided sentinel lymphadenectomy. Their SLNs were assessed for tyrosinase (Tyr) and melanoma antigens recognized by T-cells (MART-1) mRNA expression using RT–PCR, in parallel with hematoxylin and eosin staining and immunohistochemistry. Tyr and MART-1 expression in the SLNs were correlated with PATH assay results, standard prognostic factors, time to progression and overall survival. Results: Twenty-three of the 124 patients (18.5%) had positive SLNs by both PATH and RT–PCR (PATH+/PCR+). Sixteen patients (13%) were negative by PATH and positive by RT–PCR (PATHÿ/ PCR+). Eighty-five patients (68.5%) had SLNs that were negative by both PATH and RT–PCR (PATHÿ/PCRÿ). At a median follow-up of 30 months, recurrence rates among the three cohorts were statistically different (PATH+/PCR+, 60%; PATHÿ/PCR+, 31%; PATHÿ/PCRÿ, 9.4%). Seven of 23 (30%) and two of 16 (12.5%) patients died in the PATH+/PCR+ and PATHÿ/PCR+ SLN groups, respectively, whereas no patient died in the PATHÿ/PCRÿ SLN group. Conclusions: RT–PCR is more sensitive than PATH to detect SLN metastases and it is a reliable predictor of disease relapse in stage I–II melanoma patients. Key words: lymph node metastasis, melanoma, prognosis, RT–PCR, sentinel lymph node

Introduction The number of metastatic nodes and tumor burden of nodal metastasis are the two major predictors of outcome in patients with melanoma of equivalent T classification [1, 2]. Therefore, accurate classification of the nodal status is of compelling prognostic value for patients with early-stage melanoma [1]. The regional tumor-draining lymph node basin is the most frequent site of early melanoma metastases [3]. As originally proposed [4] the sentinel lymph node (SLN) is defined as the first lymph node in the regional basin that receives a cutaneous afferent lymphatic vessel from the primary melanoma. If the SLN is negative for metastases, then the probability for the remainder of the nodes in the basin to harbor melanoma cells is less than 1% [4]. Therefore, full nodal classification can be obtained with a conservative SLN biopsy that is less invasive, less morbid and entails lower costs than elective lymph node

*Correspondence to: Dr A. Romanini, Division of Medical Oncology, ‘Santa Chiara’ University Hospital, Via Roma, 67, 56126 Pisa, Italy. Tel: +39-050-992645; Fax: +39-050-992070; E-mail: [email protected] Ó 2005 European Society for Medical Oncology

dissection. Lymphadenectomy can thus be reserved for only those patients in whom a positive SLN has been detected (selective lymphadenectomy) [5]. Another major advantage of lymphatic mapping is that only one or two SLNs per patient (instead of 10–40 lymph nodes deriving from complete lymphadenectomy) are processed for pathology. Since SLNs are the nodes most likely to contain metastatic disease, it is possible to analyze them more accurately, for instance with serial sectioning (5 lm in thickness) and immunohistochemical (IHC) analysis with antibodies to S-100 [6] or HMB-45 [7]. This procedure has 10% to 30% higher sensitivity for identifying micrometastases compared with conventional hematoxylin and eosin (H&E) staining [8]. Moreover, when SLNs of patients who had developed recurrence within 1.5–3 years were re-evaluated with serial sectioning and/or IHC, occult metastases (that had previously gone undetected with conventional H&E staining) were found in 6% to 11.5% of the cases [9–11]. More recently, application of the reverse transcriptase polymerase chain reaction (RT–PCR) technique enabled the ‘upstaging’ of an additional 13% to 30% of patients whose SLNs were negative when analyzed by H&E and IHC staining (conventional pathology, PATH) [12–15]. Messenger-RNA

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

Received 11 April 2005; revised 10 June 2005; accepted 15 July 2005

1833 (mRNA) codifying for tyrosinase (Tyr, a key enzyme in the synthesis of melanin) and for the melanoma antigen recognized by T-cells (MART-1, expressed in most melanoma cell lines) were chosen as markers for the presence of melanoma cells in SLNs [16]. The purposes of this study were (i) to validate the usefulness of RT–PCR for Tyr and MART-1 mRNA for improving the identification of minimal residual disease in SLNs of melanoma patients versus PATH, and (ii) to determine whether PATHÿ/ PCR+ SLN patients had increased risk of relapse compared with PATHÿ/PCRÿ SLN patients.

counting rate exceeded 20% of the hottest node was considered an additional SLN. Complete removal of the SLNs was confirmed by reduction of the counting rate in the surgical area to background level. In patients who had melanomas in areas that drained into more than one nodal basin, a SLN was removed from each nodal basin. Wide local excision of the primary tumor was usually performed after the removal of all SLNs, except in cases where the nodal basin was close to the primary melanoma than initial excision of the tumor was necessary to reduce interfering radioactivity [18].

Specimen collection

Patients and methods Patients All consecutive clinical stage I–II cutaneous melanoma patients sent to the Nuclear Medicine Division of Pisa University Hospital for detecting SLN were recruited for the study if their primary melanoma was Breslow <1 mm and Clark level III to V (stage I according to the American Joint Committee on Cancer, AJCC) [1] or Breslow ‡1 mm (stage II). Following SLN dissection, patients whose SLNs were metastatic by PATH were scheduled for complete regional basin node dissection. Clinical decisions for further surgical and/or other medical treatment were taken solely on the basis of PATH findings. Stage I patients were observed yearly. Stage II patients (whose risk of recurrence in the first 2 years is about two-fold that of stage I patients) [1] were examined twice yearly for the first 2 years and annually thereafter. Patients in whom SLN metastases were detected by PATH, and therefore were considered stage III, were examined at 3-month intervals for the first year, every 4 months for the second year and every 6 months thereafter up to 5 years. Complete blood counts, liver function tests, chest X-rays and liver ultrasound were performed once a year [17]. Any suspicious clinical finding was further evaluated by computed tomography scan or magnetic resonance imaging. All patients gave written informed consent for the study protocol that was approved by the hospital ethics committee.

Lymphatic mapping Lymphoscintigraphy was performed 4–6 h prior to SLN biopsy to define the draining lymphatic basin at risk for metastatic disease and to identify the corresponding SLN. Aliquots of 0.1–0.2 ml containing 4–8 MBq of 99mTcHSA nanocolloid (NanocollÒ; Amersham Health, Amersham, UK) were injected intradermally around the primary melanoma, or on skin margins of the surgical scar if the primary lesion had previously been excised. A large field-of-view gamma camera (Camstar XRT; GE Medical System, Milwaukee, WI, USA) equipped with a parallel-hole, high-resolution collimator was used for imaging, with a 10% window centered on the 140 keV energy peak of 99mTc. Early dynamic scintigraphic images helped to visualize the lymphatic draining from the primary lesion. Subsequently, delayed static orthogonal and oblique images were acquired to better localize the nodes also depthwise. A hand-held c-probe (Scintiprobe MR 100 Pol.hi. techÒ; Carsoli, Italy) was used to confirm by external counting the approximate location of the SLN, whose skin projection was marked using indelible ink [18].

Radioguided SLN biopsy SLN biopsy was performed under intraoperative hand-held c-probe guidance [18]. The criteria for c-probe identification of a SLN were based on detecting a focal zone of radioactivity with a count ratio (hot spot to background) >10. We define background as the average count rates of the surrounding non-sentinel nodes and lymph node basin. Any additional lymph node whose

Histopathology After fixation in 10% formalin, the dimensions of the SLN specimen and its consistency were evaluated. The hilar region was identified as a landmark for dividing the specimen into two halves along the longitudinal axis, thus obtaining a wider area for sectioning. Several 5-lm thick sections were cut, the first was stained with H&E and the three adjacent ones were immunohistochemically tested for S-100 and HMB-45 proteins, and MART-1 in cases that were negative to the previous evaluations [19]. Histopathologic examination included scanning of the entire slide with magnifications of ·40 and ·100 in order to detect discrete foci of tumor cells. The sinusoids were examined at a magnification of ·250 in order to detect single tumor cells and/or small clumps.

RNA isolation and RT–PCR assay Following surgical resection, SLN specimens for molecular biology analysis were immediately placed in a RNAse-free guanidinium isothiocyanate solution (Trizol; Life Technologies) and stored at ÿ80°C until use. Total RNA was extracted from homogenized tissues according to the manufacturer’s protocol, and integrity and purity of the sample were checked by agaroseformaldehyde gel electrophoresis under denaturing conditions and spectrophotometrical reading. Nucleic acids were then stored in aliquots under ethanol at ÿ80°C until analysis. Total RNA (2 lg) were reverse-transcribed into cDNA in 20 ll reaction mixture containing 100 pmol random hexamers (Boehringer Mannheim), 6 U reverse transcriptase (RT; Promega), 1.5 mm MgCl2, 0.2 mm each dNTP, 8 U RNAse inhibitor (Life Technologies). Samples were incubated at 25°C for 10 min, 42°C for 1 h and RT denatured at 95°C for 5 min in a 2400 Perkin Elmer thermal cycler. cDNA samples were subsequently amplified for the target sequences by using a single-step PCR and the following reaction mixture: 5 ll cDNA, 1.5 mm MgCl2, 0.2 mm each dNTP, 1 U of AmpliTaq Gold DNA polymerase (Perkin Elmer), 20 pmol each primer and PCR buffer 1· in 50 ll final volume. To prevent contamination and false positive results, RNA extraction, preparation of reaction mixtures and amplifications were carried out in separated rooms. Each RT–PCR experiment included a sample with no RNA as negative controls and RNA extracted from SK-mel 23 melanoma cell line as positive control that were added to each run. Each sample was amplified for glyceraldehydes-3-phosphate dehydrogenase (GAPDH) mRNA to assess RNA integrity and absence of DNA polymerase inhibitors, and the melanoma-associated markers, Tyr and MART-1 mRNAs. GAPDH primers were: sense 59 GGT CGG AGT CAA CGG ATT TG 39, antisense 59ATG AGC CCC AGC CTT CTC CAT 39, yielding a 320 bp amplicon [20]. Tyr expression was determined by using HTR3 sense 59 GTC TTT ATG CAA TGG AAC GC 39 and HTR4 antisense 59 GCT ATC CCA GTA AGT GGA CT 39 primers, generating a 207 bp PCR product and used as inner primer set of a previously described nested-PCR reaction [21]. MART-1 was amplified with M1 sense 59 AGA TGC CAA GAG AAG

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

The harvested SLNs were immediately bisected, one half was sent for H&E and IHC, the other half was placed in Trizol (Life Technology) to preserve RNA and sent for RT–PCR molecular analysis.

1834 ATG CTC 39 and M2 antisense 59 GCT CTT AAG GTGAAT AAG GTG G 39, yielding a 364 bp PCR product [22]. The amplification cycling profile was as follows: initial denaturation 95°C for 5 min; cycling: 95°C for 30 s, 55°C for 30 s and 72°C for 30 s (five cycles); 95°C for 20 s, 55°C for 20 s and 72°C for 20 s for 55 cycles (35 cycles for GAPDH); final extension 72°C for 10 min. Amplifications were carried out in a 2400 Perkin Elmer thermal cycler and, after reaction completion, 20 ll of PCR samples were run on 2% agarose gel electrophoresis and samples scored positive when amplicons of the expected size were readily visible under UV light [23].

Adjuvant therapy

Statistical analysis To assess the correlation of clinicopathologic risk factors with PATH and RT–PCR Tyr and MART-1 mRNA expression in the SLN, univariate (using v2-test or Student’s t-test for continuous variables to compare the difference in mean Breslow thickness) and multivariate (using logistic regression model) analyses were used. The Kaplan–Meier method was used to plot disease-free survival (DFS) and overall survival (OS), while the log-rank test was used to compare curves [25, 26]. DFS was calculated from the day of SLN biopsy until the ascertainment of disease progression or the last follow-up visit. OS was calculated from the day of SLN biopsy until death or the last follow-up visit. Cox proportional hazard regression models were performed to examine the association of pathologic and molecular risk factors with DFS. Data were analyzed using SPSS/PC+11.5 statistical software. In all statistical analyses, P < 0.05 was considered significant.

n

%

125

100.0

78

62.4

>60

47

37.6

Median (range)

55.5 (20–81)

Patients Age (years) £60

Sex Male

67

53.6

Female

58

46.4

Lesions

126

100.0

54

42.9

Primary tumor site Extremities Trunk

59

46.8

Head and neck

13

10.3

£1.0

23

18.3

1.01–2.0

56

44.4

2.01–4.0

40

31.7

7

5.6

Breslow thickness (mm)

>4.0 Clark level II

2

1.6

III

41

32.5

IV

79

62.7

V

3

2.4

Unknown

1

0.8

IA

15

11.9

Results

IB

57

45.2

Patients

IIA

39

31.0

IIB

13

10.3

IIC

1

0.8

IIIC

1

0.8

AJCC stage

One hundred and twenty-five consecutive patients entered the study in the period between January 1998 and December 2002. Patient characteristics are summarized in Table 1. In one patient two melanoma lesions were surgically removed and a SLN was detected for both, so that a total of 126 lesions belonging to 125 patients were evaluated. There were 73 stage I (58%) and 53 stage II (42%) melanoma lesions. Median follow-up for the entire patient population was 30 months (range 10–53).

Lymphatic mapping and SLN biopsy SLN was identified in 124 of 125 patients. In the single patient with unsuccessful SLN localization, the draining LN was so massively infiltrated by melanoma cells that it did not retain the radiocolloid. This patient was considered stage IIIC and excluded from the evaluation of the SLN-positive patients. A total of 197 SLNs were excised and sent for pathology, including 167 (84.8%) identified at pre-operative lymphoscintigraphy and 30 (15.2%) additional SLNs identified solely by intra-operative c-probe counting. In the majority of patients (80%) lymphoscin-

Ulceration Ulcerated

17

13.5

109

86.5

Nodular

53

42.0

Superficial spreading

67

53.2

6

4.8

Non-ulcerated Histotype

Other types

AJCC, American Joint Committee on Cancer [1].

tigraphy visualized lymphatic drainage to only one regional lymphatic basin, while in 25 patients (20%) SLNs were detected in two or more basins. A median of two SLNs/lesion was obtained (range one to four). No immediate or long-term complications were observed following radioguided SLN biopsy.

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

Patients with nodal metastases identified by PATH underwent complete node dissection and were offered adjuvant high-dose interferon-a-2b (HD-IFN, Intron-A; Schering-Plough, Kenilworth, NJ, USA) therapy [24]. No lymphadenectomy or adjuvant IFN therapy were offered to patients with SLNs negative by PATH even though they were positive by RT–PCR (PATHÿ/PCR+).

Table 1. Main clinical parameters of the patients enrolled in the study and lesion characteristics

1835

Histology, IHC and RT–PCR analysis Thirty (15%) of the 197 SLNs were positive for metastatic involvement by conventional pathology (H&E, IHC with S-100 and HMB-45). In one SLN, IHC for MART-1 was performed to confirm an uncertain pathological diagnosis. Overall, 144 of 197 SLNs (73%) were negative for metastatic involvement by both PATH and RT–PCR analysis, while 30 of 197 SLNs (15%) were positive by both PATH and RT–PCR analysis (RT–PCR being positive in all SLNs that were positive by PATH). On the other hand, 23 of 197 SLNs (12%) were negative by PATH, yet they turned out positive for metastatic involvement by RT–PCR analysis.

Patient group

Adjuvant therapy All 23 patients with a positive SLN by PATH underwent lymphadenectomy. In 17 patients no further lymph node metastasis was detected, while one lymph node in three patients and two lymph nodes in another three patients resulted positive for melanoma metastases. Fourteen patients subsequently received high-dose (eight patients) or low-dose (six patients) IFN as adjuvant treatment. Of the four PATHÿ patients who recurred (one in the PATHÿ/PCRÿ group and three in the PATHÿ/PCR+ group), three received high-dose IFN upon recurrence. Lymphadenectomy was not performed in patients whose SLNs were negative by PATH and positive by RT–PCR analysis.

Univariate and multivariate analyses After pathologic staging of the primary tumor and SLN, 69 of 125 lesions (55.2%) were stage I, 33 lesions (26.4%) were stage II and 23 (18.4%) were stage III. Age, sex, primary tumor site, Breslow thickness, Clark level, AJCC stage, ulceration and histotype were the variables independently considered as risk factors in univariate analysis (Table 3). Nodular histotype (P = 0.002), Breslow thickness (P = 0.0001), mean Breslow thickness, compared using Student’s t-test for continuous variables

Recurrences Deaths and relapses [n (%)] [n (%)]

Stage IV (M1)a A

B

C

4

–

3

3

1

1

–

10

1

4

15

2

PATHÿ/PCRÿ 85

1

PATHÿ/PCR+ 16 PATH+/PCR+

23 124

Total

Local recurrence or stage at first relapse and number of deaths, on a patient by patient basis, are reported in Table 2. Eighty-five patients (68.5%) had SLNs that were negative by both PATH and RT–PCR (PATHÿ/PCRÿ). Twenty-three of the 124 patients (18.5%) had positive SLNs by PATH, and all of them were also positive by RT–PCR (PATH+/PCR+). Sixteen patients (13%) were negative by PATH and positive by RT–PCR (PATHÿ/ PCR+). Out of the 85 PATHÿ/PCRÿ patients, only one recurred locally (after 9 months) and seven relapsed in distant organs, yet none have died, thus identifying a group with favorable prognosis (9.4% recurrences/relapses, 0% deaths). Out of the 23 PATH+/PCR+ patients, 14 relapsed and seven died, thus identifying a group with poor prognosis (60% recurrences/relapses, 30% deaths). Finally, out of the 16 PATHÿ/PCR+, three recurred (after 9, 16 and 24 months), two relapsed and two died, thus identifying a group with intermediate prognosis (31% recurrences/relapses, 12.5% deaths). The patient with two melanoma lesions with SLNs negative at both PATH and RT–PCR analysis has not recurred after 44 months of follow-up.

Patients Local Relapses (n) (n) recurrences (n)

8 (9.4)

0 (0)

–

5 (31)

2 (12.5)

3

14 (60)

7 (30)

6

27 (21.7)

9 (7.2)

a

According to American Joint Committee on Cancer [1]. RT–PCR, reverse transcriptase polymerase chain reaction; PATHÿ, conventional histopathology negative; PATH+, conventional histopathology positive; PCRÿ, RT–PCR negative; PCR+, RT–PCR positive.

(P = 0.0001), and AJCC stage of the primary lesion (P = 0.0001) were significant predictors of metastatic involvement of SLN as identified by PATH. Multivariate analysis confirmed only AJCC stage II as significant a predictor of SLN metastases [hazard ratio 16; 95% confidence interval (CI) 1.4–186.4; P = 0.03; data not shown].

DFS and OS curves Kaplan–Meier DFS and OS curves (Figures 1 and 2, respectively) show the profound impact of the PATH and PCR status on both parameters. In particular, patients whose SLNs were metastatic at both PATH and PCR analysis had the shortest median DFS (29 months; 95% CI 13–45). Patients whose SLNs were PATHÿ/PCR+ had a slightly, but not significantly, longer median DFS (33 months; 95% CI not determined). Finally, in patients with PATHÿ/PCRÿ SLNs the median DFS was not reached and was anyway significantly longer (P < 0.05) than in patients whose SLNs were PATH+/PCR+ (P < 0.0001) or PATHÿ/PCR+ (P = 0.0057). In a similar fashion, patients whose SLNs were PATH+/PCR+ had significantly shorter (P < 0.0001) median OS (46 months; 95% CI not determined) than patients whose SLNs were PATHÿ/PCRÿ. On the other hand, median OS was not reached in the latter patient group, nor in those patients whose SLNs were PATHÿ/PCR+. Finally, patients whose SLNs were PATHÿ/PCRÿ had significantly longer survival than patients whose SLNs were PATHÿ/PCR+ (P = 0.0001). Cox hazard regression analysis showed that the risk of disease progression was of the same order of magnitude in patients whose micrometastases in SLNs were detected with molecular markers (PCR+) only or with both PATH and RT–PCR assays (P = 0.003) (Table 4). However, most of the relapses and deaths observed in the patients with metastatic SLNs identified only by RT–PCR (PATHÿ/PCR+) occurred relatively later in follow-up versus patients with metastatic involvement detected by both PATH and RT–PCR (PATH+/PCR+).

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

SLN status and clinical outcome

Table 2. Patterns of local and/or distant metastases by number of patients observed in the follow-up as a function of the sentinel lymph node status evaluated by conventional histopathology or by RT–PCR analysis of the molecular markers

1836 100

Table 3. Probability of finding metastatic involvement of SLNs according to the main clinical parameters and lesion characteristics

Age (years)

0.4

£60

65

12

>60

36

11

Sex

0.5

Male

56

10

Female

45

13

Extremities

42

12

Trunk

48

10

Head and neck

11

1

Breslow thickness (mm) 23

0

1.01–2.0

50

5

2.01–4.0

23

17

>4.0

5

1

Mean 6 SD

1.8 6 1.16

2.9 6 0.88

Clark level 2

0

III

37

4

IV

60

18

V

1

1

Unknown

1

0

AJCC stage

0.0001

IA

15

0

IB

53

3

IIA

23

16

IIB

9

4

IIC

1

0

Ulceration

0.7

Ulcerated

12

4

Non-ulcerated

89

19

Nodular

35

17

Superficial spreading

61

5

5

1

Histotype

Other types

0.0001a 0.2

II

PATH-/PCR+ 40

PATH+/PCR+

20

0

10

20

30

40

50

60

Disease Free Survival, months

0.0001

£1.0

P < 0.001

60

0

0.5

PATH-/PCR-

0.002

a

Student’s t-test. SLN, sentinel lymph node; SD, standard deviation; AJCC, American Joint Committee on Cancer [1].

Discussion Among all newly diagnosed melanoma patients, 85% have stage I–II disease [1]. The two major prognostic factors for primary melanoma localized on the skin are thickness of the lesion, assessed pathologically in millimeters (Breslow), and the pres-

Figure 1. Kaplan–Meir plot of disease-free survival (DFS) in patients whose sentinel lymph nodes (SLNs) were negative for metastatic involvement by both conventional pathology (PATH) and RT–PCR (PCR) analysis (upper curve, PATHÿ/PCRÿ), patients whose SLNs were negative by PATH but positive by PCR analysis (middle curve, PATHÿ/ PCR+) and patients whose SLNs were metastatic by both PATH and PCR analysis (PATH+/PCR+). P <0.001 denotes a statistically significant difference in DFS between patients whose SLNs were negative by both PATH and PCR analysis and the other two groups of patients combined. When evaluated separately, statistical significance was P <0.0001 for the difference in DFS between the PATHÿ/PCRÿ SLN patients and the PATH+/PCR+ SLN patients, while it was P = 0.0057 for the difference in DFS between the PATHÿ/PCRÿ and the PATHÿ/PCR+ SLN patients. There was no significant difference in DFS between the PATH+/PCR+ and the PATHÿ/PCR+ SLN patients.

ence or absence of ulceration (evaluated pathologically) [27]. Once spread has occurred to the regional lymphatic basin, the number of metastatic lymph nodes, and whether they are microscopically or macroscopically involved, become the most important prognostic factors [1, 28]. In fact, regardless of thickness, the 5-year survival of patients with non-ulcerated melanomas with a single, clinically occult nodal metastasis is much greater than that of patients with four or more clinically evident nodal metastases (70% versus 27%) [2]. It is now recommended that a search be made for micrometastases in the SLN of patients with lesions ‡1.0 mm in Breslow thickness, or <1.0 mm but Clark level IV or V, and/or ulcerated [2]. Also, among patients with melanoma lesions >4 mm thick, the detection of a positive SLN discriminates a subset of patients with worse prognosis (30% 5-year survival versus 61% in patients with negative SLN) [29]. To date, the only recommendation after the detection of a pathologically positive SLN is to complete lymphadenectomy of the lymphatic basin where the metastatic SLN has been detected, since this procedure seems to improve the 5-year OS from 51% to 73% [15]. To increase sensitivity in the detection of nodal metastases, the molecular biology technique RT–PCR has been introduced to ascertain the presence of mRNA linked to genes highly expressed in malignant melanoma [12–14, 30–32]. Tyr, the key enzyme in melanin biosynthesis, is expressed both in normal melanocytes and in melanoma cells [33]. Nerve cells have also been reported as a source of Tyr expression [34]. On the other hand, MART-1, a melanoma-associated antigen

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

Primary tumor site

80

Probability, %

SLN negative SLN positive P value patients (n = 101) patients (n = 23) (v2-test)

Probability, %

1837 100

PATH-/PCR-

80

P < 0.001

60

PATH+/PCR+ PATH-/PCR+

40

20

Factor

Hazard ratio 95% CI

v2

df P value

Histopathology and RT–PCR status of SLN

0.003

PATHÿ/PCRÿ

1 (Ref)

PATHÿ/PCR+

4.6

1.5–14.4

PATH+/PCR+

4.4

1.7–11.4

11.4

2

Breslow thickness (mm)

0

10

20

30

40

50

60

Overall Survival, months Figure 2. Kaplan–Meir plot of overall survival (OS) in patients whose sentinel lymph nodes (SLNs) were negative for metastatic involvement by both conventional pathology (PATH) and RT–PCR (PCR) analysis (upper curve, PATHÿ/PCRÿ), patients whose SLNs were negative by PATH but positive by PCR analysis (PATHÿ/PCR+) and patients whose SLNs were metastatic by both PATH and PCR analysis (PATH+/PCR+). P < 0.001 denotes a statistically significant difference in OS between patients whose SLNs were negative by both PATH and PCR analysis and the other two groups of patients combined. When evaluated separately, statistical significance was P < 0.0001 for the difference in OS between the PATHÿ/ PCRÿ SLN patients and the PATH+/PCR+ SLN patients, while it was P = 0.0001 for the difference in OS between the PATHÿ/PCRÿ and the PATHÿ/PCR+ SLN patients. There was no significant difference in OS between the PATH+/PCR+ and the PATHÿ/PCR+ SLN patients.

recognized by T-cells, is a 118 amino acid transmembrane protein (13 kDa) with tissue specificity for melanoma, melanocytes and retina. In order to increase sensitivity, we considered both markers, Tyr and MART-1, to assess the prognostic value of the SLN status [35]. By applying this approach, we identified an additional 12% metastatic SLNs over those already detected by conventional pathology. In our patient population, local recurrence occurred in three of 16 and distant metastases in two of 16 patients whose SLN was negative by PATH but positive by PCR analysis. On the other hand, only one of 85 patients with SLN negative by PCR (and obviously also by PATH, since no false-negative results were observed for PCR) had local recurrences and seven other patients had distant metastases, indicating that PCR is both sensitive and accurate in identifying patients more likely to recur because of their melanoma. Moreover, the patients with SLNs negative by PATH but positive by PCR analysis who would have been classified in a low-risk category if based solely on PATH analysis, had an approximate four-fold risk of developing metastases compared with patients whose SLNs were negative both by PATH and PCR analysis (Table 4). The fact that 10 out of these 16 patients had AJCC stage I disease (data not shown) indicates that the molecular assay is more accurate in detecting micrometastases even in early-stage melanomas. Out of the nine patients who have died in our entire group, seven had SLN metastases detected by both techniques (PATH and PCR), while in two patients the SLN metastasis was detected only by PCR (Table 2). These considerations further support the statement

0.7

<2

1 (Ref)

>2

0.76

0.16–3.6

0.12 1

Clark level

0.1

II+III

1 (Ref)

IV+V

2.6

0.72–9.0

2.1

1

AJCC stage

0.09

IA + IB

1 (Ref)

IIA + IIB + IIC

3.9

0.77–20.1

2.7

1

RT–PCR, reverse transcriptase polymerase chain reaction; PATHÿ, conventional histopathology negative; PATH+, conventional histopathology positive; PCRÿ, RT–PCR negative; PCR+, RT–PCR positive; AJCC, American Joint Committee on Cancer; CI, confidence interval; df, degrees of freedom; Ref, reference value.

that, among those patients whose SLNs are negative by PATH, molecular analysis identifies a subset of patients who could benefit from closer follow-up or from an adjuvant treatment. The fact that most of the relapses and deaths observed in patients with PATHÿ/PCR+ SLNs occurred relatively later in follow-up compared with PATH+/PCR+ SLN patients, is possibly correlated to the lower tumor burden at diagnosis in the former group. In fact, the majority of such patients had melanomas with <2 mm Breslow thickness and AJCC stage IA or IB, while the group with both PATH and PCR positive SLNs had a significantly higher proportion of patients with >2 mm Breslow thickness (P = 0.009) and AJCC stage II (P = 0.004) melanomas (data not shown). On the other hand, patients with melanoma Breslow thickness >2 mm and AJCC stage IIA–IIB had worse DFS and OS compared with patients with lower Breslow thickness lesions and lower AJCC stage disease (P < 0.05; data not shown). We identified Breslow thickness, AJCC stage and nodular histotype as significant predictors of disease progression (Table 3). Of the three factors only nodular histotype was never reported as significant predictor of disease progression in other studies evaluating SLN status. Owing to the low number of events observed, correlation of these variables with survival was not performed. The present results are consistent with, and extend data from, recent studies stratifying patients according to combined PATH and PCR analysis (Table 5) [12–14, 30–32]. In this regard, it should be emphasized that, at variance with some, but not all reports, our study includes two separate RT–PCR markers in addition to conventional PATH, no false-negative case was found at PCR analysis, and the correlation between RT–PCR

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

0

Table 4. Results of multivariate analysis of factors possibly affecting disease-free survival

1838 Table 5. Review of literature and results of the present study concerning correlation of sentinel lymph node status evaluated by conventional histopathology and by molecular markers with recurrences in patients with melanoma Author (year)

Patients (n)

Molecular marker(s)

Median follow-up (months)

Recurrences PATH+/PCR+ (%)

Shivers [12] (1998)

114

Bostick [13] (1999)

72

PATHÿ/PCR+ (%)

PATHÿ/PCRÿ (%)

Tyr

28

14/23 (61)

6/47 (13)

1/44 (2)

MMa

12

5/16 (31)

3/20 (15)

0/35 (0)

Li [30] (2000)

233

Tyr

20

18/49 (36.7)

Blaheta [14] (2000)

116

Tyr

19

10/15 (67)

b

12/115 (10) 9/36 (25)

1/64 (1.6) 4/65 (6)

57

MM

36

6/16 (38)

3/19 (16)

2/22 (9)

Goydos [32] (2003)

175

Tyr

34

17/34 (50)

14/68 (21)

0/73 (0)

Kammula [40] (2004)

112

Tyr

42a

8/15 (53)

8/58 (14)

0/39 (0)

67b

10/15 (67)

14/58 (24)

6/39 (15)

30

14/23 (60)

5/16 (31)

8/85 (9.4)

Present study

124

Tyr MART-1

a

Early follow-up. Late follow-up. RT–PCR, reverse transcriptase polymerase chain reaction; PATHÿ, conventional histopathology negative; PATH+, conventional histopathology positive; PCRÿ, RT–PCR negative; PCR+, RT–PCR positive; Tyr, tyrosinase; MMa, multiple markers (Tyr, MART-1, MAGE-3); MMb, multiple markers (Tyr, Melan-A/MART-1). b

results and DFS is based on 30 month follow-up. In fact, consistent with data in the literature [36], in this period 85% of our SLN-positive patients recurred, with median time to recurrence of 17 months (range 6–45). Out of the seven studies including over 50 melanoma patients and evaluating fresh SLNs, only three included a larger patient population than ours (but PCR analysis was based on one marker only), only two utilized multiple molecular markers (based, however, on much smaller groups of patients, or on a much shorter follow-up), and finally three followed patients for >30 months (but with one molecular marker only, or with a smaller group of patients). All eight studies, including our own, reported statistically significant differences in the proportion of patients with recurring disease between groups with PATH and PCR both negative and groups with both parameters positive. Only in three studies (including our own) was there a statistically significant difference in OS between the groups with PATH and PCR both negative and, respectively, negative PATH but positive PCR (Table 5). Other recently published papers on this topic are not included in Table 5 because of different patient selection [37] or PCR techniques [38, 39] and/or study design [15]. The recent paper by Kammula et al. [40] reports a statistically significant difference in DFS between PATHÿ/PCR+ and PATHÿ/PCRÿ patients at the 42month follow-up, while the difference in DFS was no longer statistically significant at extended follow-up (67 months). Based on these findings, the authors conclude that ‘future studies evaluating molecular staging will require approximately 5 years of median follow-up to accurately define outcome for patients with occult melanoma metastases’. However, such conclusions are hampered by a number of considerations. In fact, in a consistent proportion of their PATHÿ/PCRÿ patients (28%) the ulceration status was unknown. Given the small sample size, the ensuing misclassification based on SLN status alone might

heavily affect survival analysis, as ulceration is an independent adverse prognostic factor in patients with 2–4 mm Breslow thickness lesions [41]. Furthermore, Kammula et al. did not report their false-positive rate at RT–PCR of SLNs, a possible occurrence when employing Tyr alone, which is expressed also in melanocytes and nerve cells, normal cells that can be found in non-metastatic lymph nodes. Thus, the study by Kammula et al. should not jeopardize the importance of molecular staging in distinguishing a subgroup of patients with intermediate prognosis between the PATHÿ/PCRÿ and PATH+/PCR+ groups. In fact, even if this advantage is short-lived, this information could guide clinicians in performing a more intensive follow-up after the third year of diagnosis or offering patients adjuvant highdose IFN or a more aggressive chemoimmunotherapy treatments in the attempt to achieve a longer survival. The study by Morton et al. [15] pioneered the design of the MSLT-1 trial, the third interim analysis of which was recently reported at the ASCO meeting: SLN status definitely has a statistically prognostic significance, and lymphadenectomy after detection of clinically occult SLN metastases significantly prolongs survival compared with therapeutic lymphadenectomy when the draining lymph nodes grow to a palpable size [42]. Uncertainty still remains as to how to treat PATHÿ/PCR+ SLN patients, a still open issue that is being addressed by the ‘Sunbelt melanoma’ trial that compares the effect of regional lymphadenectomy (with or without HD-IFN) on DFS and OS in patients with early nodal metastases detected by PATH or by PCR analysis [43].

Acknowledgements The authors wish to thank Dr M. Chiarugi and Dr G. Evangelista, Dr S. Lorenzi, Dr G. Meucci, M. Rondini, S. Doria,

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

Rimoldi [31] (2003)

1839 G. Giovacchini and Dr M. Roncella for their precious collaboration. Special thanks are due to Dr J. R. Bertino for his encouragement, precious comments and advice while preparing this manuscript.

References

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

1. Balch CM, Buzaid AC, Soong SJ et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 2001; 19: 3635–3648. 2. Balch CM, Soong SJ, Gershenwald JE et al. Prognostic factors analysis of 17,600 melanoma patients: validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 2001; 19: 3622–3634. 3. Reintgen D, Cruse CW, Wells K et al. The orderly progression of melanoma nodal metastases. Ann Surg 1994; 220: 759–767. 4. Morton DL, Wen DR, Wong JH et al. Technical details of intraoperative lymphatic mapping for early stage melanoma. Arch Surg 1992; 127: 392–399. 5. Essner R, Conforti A, Kelley MC et al. Efficacy of lymphatic mapping, sentinel lymphadenectomy, and selective complete lymph node dissection as a therapeutic procedure for early-stage melanoma. Ann Surg Oncol 1999; 6: 442–449. 6. Cho KH, Hashimoto K, Taniguchi Y et al. Immunohistochemical study of melanocytic nevus and malignant melanoma with monoclonal antibodies against S-100 subunits. Cancer 1990; 66: 765–771. 7. Walts AE, Said JW, Shintaku IP. Cytodiagnosis of malignant melanoma. Immunoperoxidase staining with HMB-45 antibody as an aid to diagnosis. Am J Clin Pathol 1988; 90: 77–80. 8. Gibbs JF, Huang PP, Zhang PJ et al. Accuracy of pathologic techniques for the diagnosis of metastatic melanoma in sentinel lymph nodes. Ann Surg Oncol 1999; 6: 699–704. 9. Gadd MA, Cosimi AB, Yu J et al. Outcome of patients with melanoma and histologically negative sentinel lymph nodes. Arch Surg 1999; 134: 381–387. 10. Gershenwald JE, Colome MI, Lee JE et al. Patterns of recurrence following a negative sentinel lymph node biopsy in 243 patients with stage I or II melanoma. J Clin Oncol 1998; 16: 2253–2260. 11. Chao C, Wong SL, Ross MI et al. Patterns of early recurrence after sentinel lymph node biopsy for melanoma. Am J Surg 2002; 184: 520–524. 12. Shivers SC, Wang X, Li W et al. Molecular staging of malignant melanoma: correlation with clinical outcome. JAMA 1998; 280: 1410–1415. 13. Bostick PJ, Morton DL, Turner RR et al. Prognostic significance of occult metastases detected by sentinel lymphadenectomy and reverse transcriptase-polymerase chain reaction in early-stage melanoma patients. J Clin Oncol 1999; 17: 3238–3244. 14. Blaheta HJ, Ellwanger U, Schittek B et al. Examination of regional lymph nodes by sentinel node biopsy and molecular analysis provides new staging facilities in primary cutaneous melanoma. J Invest Dermatol 2000; 114: 637–642. 15. Morton DL, Hoon DS, Cochran AJ et al. Lymphatic mapping and sentinel lymphadenectomy for early-stage melanoma: therapeutic utility and implications of nodal microanatomy and molecular staging for improving the accuracy of detection of nodal micrometastases. Ann Surg 2003; 238: 538–549. 16. Hoon DS, Wang Y, Dale PS et al. Detection of occult melanoma cells in blood with a multiple-marker polymerase chain reaction assay. J Clin Oncol 1995; 13: 2109–2116.

17. Poo-Hwu WJ, Ariyan S, Lamb L et al. Follow-up recommendations for patients with American Joint Committee on Cancer Stages I–III malignant melanoma. Cancer 1999; 86: 2252–2258. 18. Mariani G, Gipponi M, Moresco L et al. Radioguided sentinel lymph node biopsy in malignant cutaneous melanoma. J Nucl Med 2002; 43: 811–827. 19. Prieto VG, Clark SH. Processing of sentinel lymph nodes for detection of metastatic melanoma. Ann Diagn Pathol 2002; 6: 257–264. 20. Reinhold U, Ludtke-Handjery HC, Schnautz S et al. The analysis of tyrosinase-specific mRNA in blood samples of melanoma patients by RT–PCR is not a useful test for metastatic tumor progression. J Invest Dermatol 1997; 108: 166–169. 21. Smith B, Selby P, Southgate J et al. Detection of melanoma cells in peripheral blood by means of reverse transcriptase and polymerase chain reaction. Lancet 1991; 338: 1227–1229. 22. Noppen C, Spagnoli GC, Schaefer C. Isolation of multiple mRNAs from a few eukaryotic cells: a fast method to obtain templates for RT–PCR. Biotechniques 1996; 21: 394–396. 23. Pellegrino D, Bellina CR, Manca G et al. Detection of melanoma cells in peripheral blood and sentinel lymph nodes by RT–PCR analysis: a comparative study with immunohistochemistry. Tumori 2000; 86: 336–338. 24. Kirkwood JM, Strawderman MH, Ernstoff MS et al. Interferon alfa-2b adjuvant therapy of high-risk resected cutaneous melanoma: the Eastern Cooperative Oncology Group Trial EST 1684. J Clin Oncol 1996; 14: 7–17. 25. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481. 26. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966; 50: 163–170. 27. Balch CM, Wilkerson JA, Murad TM et al. The prognostic significance of ulceration of cutaneous melanoma. Cancer 1980; 45: 3012–3017. 28. Carlson GW, Murray DR, Lyles RH et al. The amount of metastatic melanoma in a sentinel lymph node: does it have prognostic significance? Ann Surg Oncol 2003; 10: 575–581. 29. Zettersten E, Sagebiel RW, Miller JR 3rd et al. Prognostic factors in patients with thick cutaneous melanoma (>4 mm). Cancer 2002; 94: 1049–1056. 30. Li W, Stall A, Shivers SC et al. Clinical relevance of molecular staging for melanoma: comparison of RT–PCR and immunohistochemistry staining in sentinel lymph nodes of patients with melanoma. Ann Surg 2000; 231: 795–803. 31. Rimoldi D, Lemoine R, Kurt AM et al. Detection of micrometastases in sentinel lymph nodes from melanoma patients: direct comparison of multimarker molecular and immunopathological methods. Melanoma Res 2003; 13: 511–520. 32. Goydos JS, Patel KN, Shih WJ et al. Patterns of recurrence in patients with melanoma and histologically negative but RT–PCR-positive sentinel lymph nodes. J Am Coll Surg 2003; 196: 196–204. 33. Ghossein RA, Carusone L, Bhattacharya S. Review: polymerase chain reaction detection of micrometastases and circulating tumor cells: application to melanoma, prostate, and thyroid carcinomas. Diagn Mol Pathol 1999; 8: 165–175. 34. Cronin CA, Ryan AB, Talley EM, Scrable H. Tyrosinase expression during neuroblast divisions affects later pathfinding by retinal ganglion cells. J Neurosci 2003; 23: 11692–11697. 35. Schittek B, Bodingbauer Y, Ellwanger U et al. Amplification of MelanA messenger RNA in addition to tyrosinase increases sensitivity of melanoma cell detection in peripheral blood and is associated with the

1840

36. 37.

38.

39.

40. Kammula US, Ghossein R, Bhattacharya S, Coit DG. Serial follow-up and the prognostic significance of reverse transcriptase-polymerase chain reaction-staged sentinel lymph nodes from melanoma patients. J Clin Oncol 2004; 22: 3989–3996. 41. Eigentler TK, Buettner PG, Leiter U, Garbe C. Impact of ulceration in stages I to III cutaneous melanoma as staged by the American Joint Committee on Cancer Staging System: an analysis of the German Central Malignant Melanoma Registry. J Clin Oncol 2004; 22: 4376–4383. 42. Morton DL, Thompson JF, Cochran AJ et al. Interim results of the Multicenter Selective Lymphadenectomy Trial (MSLT-I) in clinical stage I melanoma. J Clin Oncol 2005; 23 (Suppl): 710s (Abstr). 43. McMasters KM. The Sunbelt Melanoma Trial. Ann Surg Oncol 2001; 8: 41S–43S.

Downloaded from http://annonc.oxfordjournals.org/ at National Institute of Education Library, Serials Unit on June 4, 2015

clinical stage and prognosis of malignant melanoma. Br J Dermatol 1999; 141: 30–36. Reintgen DS, Vollmer R, Tso CY, Seigler HF. Prognosis for recurrent stage I malignant melanoma. Arch Surg 1987; 122: 1338–1342. Ulrich J, Bonnekoh B, Bockelmann R et al. Prognostic significance of detecting micrometastases by tyrosinase RT/PCR in sentinel lymph node biopsies: lessons from 322 consecutive melanoma patients. Eur J Cancer 2004; 40: 2812–2819. Giese T, Engstner M, Mansmann U et al. Quantification of melanoma micrometastases in sentinel lymph nodes using real-time RT–PCR. J Invest Dermatol 2005; 124: 633–637. Kuo CT, Hoon DS, Takeuchi H et al. Prediction of disease outcome in melanoma patients by molecular analysis of paraffin-embedded sentinel lymph nodes. J Clin Oncol 2003; 21: 3566–3572.