Clinical and Pathological Features of Cutaneous Malignant Melanoma: A Retrospective Analysis of 124 Japanese Patients

Transcription

1 Clinical and Pathological Features of Cutaneous Malignant Melanoma: A Retrospective Analysis of 24 Japanese Patients Yoshie Kuno, Kazuyuki Ishihara, Naoya Yamazaki and Kiyoshi Mukai 2 'Dermatology Division and 2 Clinical Laboratory Division, National Cancer Center Hospital, Tokyo A review of mainly histopathologic factors associated with the survival of patients with malignant melanoma was carried out in a retrospective study of 24 Japanese patients treated at the National Cancer Center Hospital between July 962 and December 992. There were 6 females and 64 males, and the median follow-up period was 52.7 months (range,. to 25. months). The histologic features included tumor thickness, level of invasion, histologic subtype, ulceration, pigmentation, and cell type. Melanomas thicker than.5 mm (P<.) and with ulceration (P<.) had a significantly worse prognosis. With regard to histologic type, ten-year survival was 65.% for acral lentiginous melanoma, 5.7% for nodular melanoma, and 47.% for superficial spreading melanoma (SSM), there being no significant differences among them. We suggest that the prognosis was affected not by histologic type but by the large radial or vertical growth component. With regard to clinical features, the clinicopathological stage, patient age, and year when the disease was diagnosed were reflected in the prognosis (P<.). Multifactorial analysis showed that the most significant prognostic variables were histopathologic type (SSM or other), stage (I and II or III and IV), and patient age (<7 or 2 7yr). (Jpn J Clin Oncol 26: 44-5, 996) Key words: Cutaneous malignant melanoma Histopathology Prognosis Introduction Since 787, when John Hunter wrote the first description of a patient with malignant melanoma, a 5-year-old man with a recurrent mass behind the angle of the lower jaw, knowledge of the epidemiology, pathology, and treatment of malignant melanoma has advanced considerately. A major step forward was Clark's recognition in 969 that the anatomic level of invasion of malignant melanoma could separate groups of patients with low, intermediate, and high risks of dying due to the disease. Subsequently, Breslow 2 ' recognized that the thickness of the primary tumor could be used to obtain similar information. Because the incidence of cutaneous melanoma in Japan is much lower than that in Western countries with predominantly white population, it has not been possible to amass large series like those which have been used to assess prognostic factors of Received: August, 995 Accepted: November 6, 995 For reprints and all correspondence: Naoya Yamazaki, Dermatology Division, National Cancer Center Hospital, -, Tsukiji 5-chome, Chuo-ku, Tokyo 4 melanoma in white populations. We present here a review of the prognostic factors of primary cutaneous melanoma based on a Japanese patient population. Patients and Methods Between July 962 and December 992, 5 patients with cutaneous melanoma were referred to the National Cancer Center Hospital, Tokyo. Among them, we reanalyzed 2 patients with histopathologically proven primary cutaneous malignant melanoma. We concentrated mainly on histologic features, such as maximum tumor thickness, level of invasion, histologic subtype, ulceration, pigmentation, and cell type, and correlated them with clinical features including patient sex and age, primary site distribution, staging, and changes in survival rate. The age and sex distribution of the patients is shown in Table I, and the primary site and histologic subtype distribution is shown in Table II. Melanomas were divided into the following groups based on total vertical height: Group I melanoma in situ Group II melanomas 5=.75 mm thick 44 Jpn J Clin Oncol 26() 996 on 29 November 27

2 CLINICOPATHOLOGICAL ANALYSIS OF MELANOMA Table I. Patient Sex and Age Table II. Primary Location and Histologic Subtype Age 8 < Total Number of patients Male Female Head and neck Upper limbs Lower limbs Trunk Hand Foot Number of patients SSM LMM ALM NM Undiff Total SSM, superficial spreading melanoma; LMM, lentigo maligna melanoma; ALM, acral lentiginous melanoma; NM, nodular melanoma; Undiff., undifferentiated. Total 4(.%) 8 (6.5%) (.5%) 8 (4.5%) 2 (8.5%) 48 (8.7%) Fig.. Survival of 24 melanoma patients. Five-year overall survival rate was 6.6% and tenyear survival 56.9%. Group III melanomas >.75 mm and ^.5 mm thick Group IV melanomas >.5 mm and ^. mm thick Group V melanomas >. mm and ^4. mm thick Group VI melanomas >4. mm thick Tumor thickness was measured according to Breslow's method using on ocular micrometer. Level of invasion was determined according to Clark." Histologic subtypes were classified into four: (I) superficial spreading melanoma (SSM), (II) lentigo maligna melanoma (LMM), (III) acral lentiginous melanoma (ALM), and (IV) nodular melanoma (NM). All patients were retrospectively staged clinicopathologically as follows: Stage I localized primary melanoma, sl.5 mm thick Stage II localized primary melanoma, >.5 mm and <4. mm thick Stage HI metastasis to regional lymph nodes or melanomas thicker than 4. mm Stage IV disseminated melanoma Survival curves were calculated based on the method of Kaplan and Meier. ' The generalized Wilcoxon method 4 and long-rank method 5 ' were used to determine whether differences between curves were significant. Results The median follow-up period was 52.7 months (range,.-25. mo) for 24 patients and 87.9 months (range, mo) for 77 patients still alive at the time of the study. The overall survival rate for the 24 patients was 6.6% at 5 years and 56.9% at years (Fig. ). 45 on 29 November 27

3 KUNO ET AL Fig. 2. Overall survival curve according to age., <4 years (2);, 4-49 years (25);, 5-59 years ();, 6-69 years (26);, 7 years and over (24). The survival rate at 5 years was.6 for patients aged <4 years,.66 for those aged 4-49 years,.67 for those aged 5-59 years,.77 for those aged 6-69 years, and.6 for those aged 7 years and over. There was a significant difference between patients aged 7 years and over and those younger than 7 years (P<.). Clinical Factors (A) Patient Age and Sex: The median age of the 24 patients was 54 years. Those aged 7 years and over had a worse survival rate than those younger than 7 (P<.) (Fig. 2). Among stage III cases, the probability of survival was similar, the survival rates at 5 years being 6.% for patients < 4 years old, 62.5<Vo for those aged 4-49 years, 66.4% for those aged 5-59 years, 58.7% for those aged 6-69 years, and 2.% for those aged 7 years or over. All but one patient older than 7 died of melanoma. Overall, melanomas affected females and males almost equally (5.6% versus 48.4%). In our analysis, there was no significant difference in survival between the sex (Fig. ). (B) Anatomic Location of the Primary Lesion: Melanomas were divided evenly between two major anatomic locations: 49.2% were on the lower extremities, and 5.2% occurred at other sites (Table II). The five-year survival rate of patients with melanomas on the lower extremities (including the foot) was 64.8%, whereas that of patients with tumor at other sites was 49.7%. Patients with melanomas of the trunk had the worst 5-year survival rate of 7.9%. However, there was no significant difference in survival among any of the primary sites of melanoma. In our analysis, patients with melanomas on the sole showed better survival than those with lesions at other sites (Table III). (C) Staging: The overall survival of the 24 patients according to melanoma stage is shown in Fig. 4. Nine lesions of melanoma in situ or Clark's level I Fig.. Overall survival curve according to sex. male (64);, female (6). The survival rate at 5 years and years was.69 and.69 for females and.62 and.48 for males, respectively. There was no significant difference. Table III. Anatomic Location of Primary Melanoma Lower extremity (except foot and sole) Foot (except sole) Sole Upper extremity (except hand) Hand Trunk Head and neck No., number of patients. No Overall survival (%) 5-year -year L.L IC..:J..L JLL I Fig. 4. Overall survival curve according to staging. -, stage I (27);, stage II (25);, stage III (54);, stage IV (9). There was a significant difference between stages I+ and stages HI + IV (P<.). 46 Jpn J Clin Oncol 26() 996 on 29 November 27

4 CLINICOPATHOLOGICAL ANALYSIS OF MELANOMA Table IV. Staging Analysis Sex male female Pigmentation present absent Margin Ulcer positive negative present absent Age at diagnosis <7 7 and over Year of treatment < Stages I and II No. P value P<.5 No., number of patients;, not significant Stages III and IV No. P value P<. Fig. 5. Overall survival curve of 54 patients with stage III. -, T4aN(2);, T4aNl ();, any T (without T4a)Nl (2). The survival rate at 5 years was.79 with T4aN and.6 with any TNI and T4aNl. There was a significant difference between patients with localized disease and those with regional disease (P<.). were excluded. The five-year survivals for patients with stage I, II, III and IV lesions were 86.7%, 74.%, 5.% and 22.5%, respectively. There was a significant difference in survival between stages II and HI (P<.5) and between stages HI and IV (P<.5), and there was a more significant difference between stages (I+ ) and (III + IV) (P<.). Univariate analysis showed a very significant difference between stages, indicating that patients with favorable disease features belonged to stages I and II. Therefore, we compared six factors between stages I and II combined and stages III and IV combined. There was no significant difference in the distribution of patients in terms of stage (Table IV). Because stage III was defined as melanoma with metastasis to regional lymph nodes or melanoma thicker than 4. mm, tumors at this stage were divided into the following three groups; (I) melanomas 4. mm or thicker without lymph node metastasis (2 tumors), (II) melanomas 4. mm or thicker with lymph node metastasis (2), and (III) melanomas thinner than 4. mm with lymph node metastasis (). There was a significant difference between groups I and II (P<.) and groups I and III (P<.), but not between groups II and III (Fig. 5). (D) Metastasis: Eighty-four patients with localized melanomas had a -year survival rate of 77.8%. Among patients with regional lymph node Fig. 6. Overall survival curve according to metastasis. -, NM (84);, NM ();, Ml (9). The survival rate at 5 years was.8 for patients with localized melanoma,.6 for those with regional metastasis, and.2 for those with disseminated metastasis. There was a significant difference between patients with localized melanomas and those with metastasis at first diagnosis (P<.). metastasis but no disseminated metastasis, 26 who underwent lymph node dissection (4 patients received radiotherapy and one immunotherapy postoperatively) had a -year survival rate of 2.8%. Nine patients with disseminated melanomas had a 5-year survival rate of 22.2% (Fig. 6). The first sites of distant metastasis were distant lymph nodes (4 cases), skin (), lung (), and brain (). There was a significant difference in survival between patients with localized melanoma and those with melanoma with metastasis (P<.). However, there was no significant difference in survival between melanoma patients with only regional lymph node metastasis and those with disseminated metastasis. 47 on 29 November 27

5 KUNO ET AL Years Years 8 Fig. 7. Overall survival curve according to Breslow level. -, <.75 mm ();, mm (8);,.5-. mm (26);,.-4. mm ();, >4. mm (48). The survival rate at 5 years was.8 for patients with <.75 mm melanoma,.92 for those with mm,.55 for those with.5-. mm tumors,.55 for those with.-4. mm tumors, and.5 for those with >4. mm tumors. Tumor thickness of.5 mm was an important cut off value for prognosis (P<.). Three melanomas thicker than the size of the specimen were excluded. Fig. 8. Overall survival curve according to Clark's level., level I ();, level II ();, level III (6);, level IV ();, level V (4). The survival rate at 5 years was.9 for patients with level,.8 for those with level II,.65 for those with level III,.5 for those with level IV, and.42 for those with level V. There was a significant difference between levels I + II + III and levels IV + V (P<.). Twenty melanomas occurring in the nail or genital regions were not classified. Table V. Five-year Survival Rate (%) in Relation to Histologic Subtype and Tumor Thickness Thickness <.75 mm mm.5-. mm.-4. mm 4. mm< SSM LMM... _ 5. ALM NM _ SSM, superficial spreading melanoma; LMM, lentigo maligna melanoma; ALM, acral lentiginous melanoma; NM, nodular melanoma. Fig. 9. Overall survival curve according to histologic subtype., superficial spreading melanoma (SSM) (4);, lentigo maligna melanoma (LMM) (6);, acral lentiginous melanoma (ALM) (44);, nodular melanoma (NM) (2). The survival rate at 5 years was.5 for SSM,.79 for LMM,.7 for ALM, and.52 for NM. On univariate analysis there was no significant difference. One melanoma had unclassified histology. Histopathological Factors (A) Tumor Thickness and Level of Invasion: Nine patients were classsified as group I, in group II, 8 in group III, 26 in group IV, 4 in group V, and 48 in group VI. Ten patients were classified as having level I melanoma, level II, 6 level III, level IV, and 4 level V. Twenty cases could not be classified; they included melanomas of the nail, a genital lesion and cases in which incomplete excision was performed. Figure 7 shows the actuarial survival curve according to tumor thickness. As demonstrated by this curve, a significant difference in survival was found between patients with melanomas thinner than.5 mm and those with melanomas.5 mm or thicker {P<.). Figure 8 shows the actuarial survival curve according to Clark's level. Survival time tended to be shortened for more advanced levels of invasion, but there was no significant difference between the levels. When level I, II and HI patients were collected into one group and level IV and V patients in another, there was a significant difference between the groups (P<.). (B) Histologic Subtype: Figure 9 shows the actuarial survival curve according to histologic subtype. LMM and ALM carried a better prognosis than NM or SSM. In each thickness group, analysis of the histologic subtype also revealed that SSM had a worse prognosis than the other subtypes (Table V), but there were no significant differences between the types. 48 Jpn J Clin Oncol 26() 996 on 29 November 27

6 CLINICOPATHOLOGICAL ANALYSIS OF MELANOMA too 9 Fig.. Overall survival curve according to ulceration. -, present (27);, absent (97). The survival rate at 5 and years was.4 and.27 for melanomas with ulceration and.7 and.66 for those with no ulceration, respectively. On univariable analysis there was a significant difference between presence and absence of ulceration (P<.). Fig.. Overall survival curve according to surgical margin., positive ();, negative (). There was no significant difference. (C) Ulceration (Histopathologic): Twenty-seven (2.8%) melanomas, including about one quarter of all melanomas of the foot, hand, trunk and upper extremities, were ulcerated. More than half of the patients with ulcerated lesions had melanomas thicker than 4. mm. The median thickness of the ulcerated lesions was.8 mm, compared with.8 mm for non-ulcerated lesions. The presence of melanoma ulceration was a significant risk factor for survival (P<.) (Fig. ). (D) Pigmentation: Six melanomas (4.8%) lacked cytoplasmic melanin [stage I (2 melanomas), stage II (), stage III (2), stage IV ()]. The presence of melanin in two melanomas could not be ascertained because of decalcification. The ten-year survival rate of patients with amelanotic lesions was 6.%, compared with 5.9% for patients with pigmented lesions. (E) Cell Type: The cell types constituting the bulk of the primary melanomas were divided into four groups: epitheloid, spindle, polymorphic and mixed. Ninety-six melanomas (77.4%) were of the epitheloid cell type. There was no significant difference in patient survival between the cell types. (F) Surgical Margin (Histopathologic): Histologic examination demonstrated the presence of malignant cells at the margin in cases which did not include incisional biopsy cases. Excluding four cases of disseminated melanoma, wide excision was performed for seven tumors. None of these showed documented local recurrence. The ten-year survival rate of patients with positive surgical margins was 2.%, whereas that of patients with negative margins was 6.%. There was no significant difference in survival in terms of surgical margin (Fig. ). 9 SO S 7 S 6 < I 5 "a I 2 6 Years Fig. 2. Overall survival curve according to period of initial presentation., < 977 (29);, (29)., (6);, (). There was a significant difference in survival between the period before 977 and each period after 977 (P<.). Other Factors We divided the patients into four chronological groups so that certain comparisons could be made concerning the behavior of melanoma over a -year period. In these groups, the proportion of stage I and II patients increased gradually with time from 24.% to 7.9%, 44.4% and 6.%, respectively, suggesting that the number of patients treated while the melanoma was still at an early stage had increased in recent years. For the last 5 years, we partitioned off three groups, one for every 5-year period; between 978 and 982 for the first period, between 98 and 987 for the second, and between 988 and 992 for the most recent. There was a significant difference in survival between the first group and the other groups (P<.) (Fig. 2). However, for 84 melanoma patients.49 on 29 November 27

7 KUNO ET AL. 9 8 S* 7 I 6 2 Fig.. Overall survival curve according to period of initial presentation and metastasis., <977 with no metastasis ();, < 977 with metastasis (8);, > 978 with no metastasis (7);, > 978 with metastasis (22). There was no significant difference among the groups. Table VI. Multivariate Analysis P value Hazard Lower Upper rate ratio Type (SSM or other) Stage (I and II or III and IV) Age(<7or and over) Breslow (<.5 or and over) Ulcer (+ or-) SSM, superficial spreading melanoma. without metastasis, and for 4 with metastasis, there was no significant difference in survival before and after 977 (Fig. ). Multivariate analysis showed that only histologic type (SSM or other), stage (I and II or III and IV), and patient age (<7 or 7 and over) were significant factors (Table VI). Discussion Tumor thickness has consistently been reported to be the most important independent prognostic variable in melanoma, 6 " 8 ' and in fact has been used as a quantitative parameter for defining subsets of patients with different survival rates. Patients with melanoma thicker than.5 mm had high mortality rates in the present study, similar to previous reports from Japan and Western countries. Although pathologic ulceration was shown to be highly significant by univariate analysis (P<.), it was not significant by multivariate analysis. Ulceration has been shown to be dependent on tumor thickness. Controversy exists as to whether the levels proposed by Clark" have any prognostic significance. In our examination of the -year survival rate, an increase in Clark's level was correlated with a shortened survival time. Although Clark's level was shown to be highly significant by univariate analysis (P<.), no significance was demonstrable by multivariate analysis. Its prognostic importance was replaced by the Breslow thickness. Among the four major histologic types, several studies have shown that SSM and LMM have a better patient survival rate than NM and ALM. 8 " " In the present study, there was no significant difference in survival between these types, although tumor thickness and stage for four types did not differ significantly. The prognosis of patients with ALM, which represents about 46% of all melanomas in Japanese, I2) has been reported to be very poor. ' However in our analysis, the survival rate of patients with ALM,-which accounts for 5.5% of all melanomas, was not poor. It has been suggested that the prognosis of ALM is worse because many cases are at an advanced stage at detection. In our series, however, the prognosis of ALM was rather better than that of SSM at the advanced stage. The year survival of patients with acral SSM and NM was 4.% and 6.4%, respectively, reflecting the lack of a significant difference between ALM and SSM or NM,(the -year survival of ALM patients was 65.%). It was suggested that acral melanoma with a large radical or vertical growth component has a worse prognosis. In surgical therapy for acral lesions, decisions about removal are made on the basis of tumor thickness. Fundamentally, patients with subungual melanoma underwent amputation and those with tumors at other subsites received wide excision with a margin of to 5 cm. If melanomas were thicker than level IV and 4. mm, elective lymph node dissection was performed. Patient age was correlated with melanoma thickness, and older patients presented with thicker lesions. 5 ' In our study, patients aged 7 years and over had a less favorable prognosis, and all except one died of malignant melanoma. The median tumor thickness for melanoma patients in their 7th, 4th and 6th decades was 2. mm, compared with.5 mm for those in their 5th decade or in their rd decade or youngers. Thus, the distribution of melanoma depth was similar in each age group, as were the histologic subtype and stage. It was sug- 5 Jpn J Clin Oncol 26() 996 on 29 November 27

8 CLINICOPATHOLOGICAL ANALYSIS OF MELANOMA gested that age at initial diagnosis is an independent prognostic factor. Patients with stage III melanoma had a 5-year survival rate of 5.%. If regional metastasis was excluded, the rate was 78.5%. Patients with regional lymph node metastases at stage III had a 5-year survival rate of 6.%. Our analysis indicated that patients with stage III disease are better divided into two groups according to regional metastasis, as recommended by Worth et al. 5) or the American Joint Committee on Cancer. 6 * Over a -year period, the 5-year survival rate gradually improved, possibly because of the increased proportion of cases detected at an early stage. Various single- and multiple-agent chemotherapy regimens, including Dacarbazine (DTIC) [5-(,-dimethyl- -trazeno)-imidazole-4-carboxamide], have been used since 977. In fact, 7.7% of melanomas detected after 977 were treated by DTIC chemotherapy. From our data, it is unclear whether chemotherapy is a significant factor in survival. The role of combination chemotherapy in the treatment of melanoma is still controversial, 7 ' 8) and it is necessary to analyze further the correlation between treatment and survival. A review of the literature shows that certain histologic features are significant prognostic factors for survival. From our histopathologic analysis, melanomas thicker than.5 mm, with ulceration, and those showing a large radial or vertical growth component, had a worse prognosis, and clinically, melanoma patients aged 7 years and over and those with lesion on the trunk had a poor prognosis. Multifactorial analysis showed that the dominant factors affecting survival were (I) histologic type, (II) stage, and (III) age. References ) Clark WH Jr: The histogenesis and biologic behavior of primary human malignant melanomas of the skin. Cancer Res 29: , 969 2) Breslow A: Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 72: 92-98, 97 ) Kaplan EL, Meier P: Nonparametric estimation from incomplete observation. I. Am Stat Assoc 5: , 958 4) Gehan E: A generalized Wilcoxon test for comparing arbitrarily single-censored samples! Biometrika 52: 2-224, 965 5) Peto R, Pike MC: Conservation of the approximation a (O-E)2/E in the log-rank test for survival data or tumor incidence data. Biometrics 29: , 97 6) Balch CM, Murad TM, Soong SJ, Ingalls AL, Halpern NB, Maddox WA: A multifactorial analysis of melanoma: prognostic histopathological features comparing Clark's and Breslow's staging methods. Ann Surg 88: , 978 7) Prade M, Bognel C, Charpentier P, Gadenne C, Duvillard P, Sancho-Garnier H, Petit JY: Malignant melanoma of the skin: prognostic factors derived from a multifactorial analysis of 29 cases. Am J Dermatopathol 4: 4-42, 982 8) Balch CM, Soong S-J, Shaw HM, Urist MM, McCarthy WH: An analysis of prognostic factors in 85 patients with cutaneous melanoma. In Cutaneous Melanoma, 2nd ed, Balch CM, JB Lippencott, Philadelphia. pl65-87, 992 9) Harrist TJ, Rigel DS, Day CL Jr, Sober AJ, Lew RA, Phodes AR, Harris MN, Kopf AW, Friedman RJ, Golomb FM Cosimi B, Gorstein F, Malt RA, Wood WC, Postel Hennesy P, Gumport SL, Roses DF, Minitzis MM, Raker JW, Fitzpatrick TB, Mihm MC: "Microscopic satellites" are more highly associated with regional lymph node metastases than is primary melanoma thickness. Cancer 5: , 984 ) Van der Esch EP, Cascinelli N, Preda F, Morabito A, Bufalino R: Stage I melanoma of the skin: evaluation of prognosis according to histologic characteristics. Cancer 48: , 98 ) Hacene K, Le Doussal V, Brunet M, Lemoine F, Guerin P, Hebert H: Prognostic index for clinical stage I cutaneous malignant melanoma. Cancer Res 4: , 98 2) Seiji M, Takematsu H, Hosokawa M, Obata M, Tomita Y, Kato T, Takahashi M, Mihm MC Jr: Acral melanoma in Japan. / Invest Dermatol 8: 56-6, 98 ) KcGovern VJ, Murad TM: Pathology of melanoma: clinical management and treatment results worldwide. In Cutaneous Melanoma, Balch CM, JB Lippencott, Philadelpha. p29, 985 4) Patterson RH, Helwig EB: Subungual malignant melanoma: a clinical-pathologic study. Cancer 46: , 98 5) Worth AJ, Gallagher RP, Elwood JM, Yang PC, Lamb C, Spinelli JJ, Wood WS, Threlfall WJ, Hill GB: Pathologic prognostic factors for cutaneous malignant melanoma: the Western Canada Melanoma Study. Int J Cancer 4: 7-75, 989 6) Beahrs OH, Myers MH: Manual for Staging of Cancer: American Joint Committee on Cancer. JB Lippencott, Philadelphia, p.7, 98 7) Houghton AN, Legha S, Bajorin DF: Chemotherapy for Metastatic Melanoma. In Cutaneous Melanoma. 2nd ed, Balch CM, JB Lippencott, Philadelphia. P498-58, 992 8) David R. Umberto V. John M: Malignant melanoma. Proceedings of ASCO 2: 8-94, 99 5 on 29 November 27