Sse. Fernando et al., IMMUNOHISTOCHEMICAL ANALYSIS OF CUTANEOUS MALIGNANT-MELANOMA - COMPARISON OF S-100 PROTEIN, HMB-45 MONOCLONAL-ANTIBODY AND NKI C3 MONOCLONAL-ANTIBODY/, Pathology, 26(1), 1994, pp. 16-19
Immunohistochemical analysis was carried out on 144 formalin fixed par
affin embedded cutaneous melanomas to ascertain the value of 3 differe
nt immune markers. The sensitivity and staining patterns regarding int
ensity and distribution, as well as correlation to pigment content, ce
ll type, surface ulceration and host response was noted. The stains us
ed were monoclonal HMB-45 (Dako product No M634), NKI/C3 antibodies (B
iogenex product No MU077) and polyclonal rabbit anti S-100A protein (D
ako product No L1845). Of the lesions tested, 63 were malignant melano
ma with an adjacent component of superficial spreading type, 61 were m
alignant melanoma with no adjacent component, 2 were malignant melanom
a with adjacent lentigo maligna, 8 were in situ with 7 superficial spr
eading melanoma and one lentigo maligna (HMF) and 10 were metastatic m
elanoma. All 144 lesions stained for S-100 (100% sensitivity). One hun
dred and thirty-seven stained for NKI/C3 (95%); 132 stained for HMB-45
(92% sensitivity). S-100 was the most sensitive marker and stained tu
mor cells diffusely. With HMB-45 and NKI/C3, though marginally less se
nsitive, staining was stronger and patchy. In addition, NKI/C3 showed
a tendency for peripheral (membrane) staining. HMB-45 staining was dir
ectly proportional to the pigment content, with stronger staining of r
adial growth phase melanoma and negative staining of those lesions whe
re pigment content was minimal or absent. Also, with HMB-45 a decrease
in staining intensity with depth or vertical growth phase was observe
d. There was no relationship to cell type with HMB-45 but with NKI/C3,
5 out of the 7 that failed to stain showed spindle cell differentiati
on. Surface ulceration or the degree of host response played no part i
n the staining properties. The authors advocate that all 3 stains be u
sed in the assessment of nevomelanocytic lesions. No specific single m
arker is available for the diagnosis of malignant melanoma where pigme
nt content is minimal or absent.