[525] Interobserver Agreement of Assessment of Desmoplasia and Neurotropism in Melanoma

Rajmohan Murali, Elyn R Riedel, Klaus J Busam, Wendy A Cooper, Maria Garrido-Ruiz, Rooshdiya Z Karim, Tamazin Leecy, Melissa P Pulitzer, Richard A Scolyer. Memorial Sloan-Kettering Cancer Center, New York; Royal Prince Alfred Hospital, Sydney, Australia; Melanoma Institute Australia, Sydney, Australia

Background: A proposed scheme for classification of desmoplastic melanoma (DM) divides melanomas that exhibit desmoplasia into pure DM [pDM, in which ≥90% of invasive tumor is desmoplastic] and combined DM [cDM, in which <90% of invasive tumor is desmoplastic]. This classification of DM type correlated with survival in some studies but not in others. One possible reason for these conflicting findings is that the assessment and classification of desmoplasia is not reproducible. We sought to determine the level of interobserver agreement in the classification of desmoplasia and neurotropism in melanoma.
Design: Hematoxylin-eosin-stained slides from 50 cases of melanoma exhibiting varying degrees of desmoplasia and neurotropism were scored by 6 pathologists. Each observer scored the following: DM type (pDM or cDM); nerve involvement (absent, focal or extensive); location of nerve involvement (within or outside main tumor mass, or both); and whether the tumor should be termed “neurotropic melanoma”. The level of interobserver agreement for each parameter was assessed using kappa (k) statistic.
Results: Agreement for assessment of DM type, nerve involvement, location of nerve involvement, and whether the observer would classify the tumor as “neurotropic melanoma” were moderate (k=0.47), fair (k=0.31), slight (k=0.13) and fair (0.39), respectively (Table).

Observer->1*2*3*4*5*6*Kappa
DM type      0.47
Pure DM22(44)37(74)30(60)33(66)32(64)23(46) 
Combined DM18(36)13(26)12(24)13(26)15(30)19(38) 
Not DM10(20)08(16)4(8)3(6)8(16) 
Nerve involvement      0.31
Absent17(34)26(52)30(60)24(48)10(20)29(58) 
Focal9(18)23(46)3(6)5(10)6(12)2(4) 
Extensive24(48)1(2)17(34)21(42)34(68)19(38) 
Location of nerve involvement      0.13
Within main tumor mass11(33)3(13)12(60)6(23)15(37)5(24) 
External to main tumor mass3(9)18(75)1(5)8(31)7(18)2(10) 
Both within and external19(58)3(13)7(35)12(46)18(45)14(67) 
Neurotropic melanoma      0.39
No30(60)27(54)32(64)24(48)10(20)29(58) 
Yes20(40)23(46)18(36)26(52)40(80)21(42) 
*no. of cases (%)


Conclusions: There was only moderate agreement in the assignment of DM type, which may explain the disparate findings of studies assessing the prognostic significance of this classification system. Agreement in the assessment of neurotropism was poorer. Clearer definitions and criteria are required to improve the reproducibility of assessment of desmoplasia and neurotropism in melanoma, and to evaluate their true prognostic value.
Category: Dermatopathology

Wednesday, March 21, 2012 1:00 PM

Poster Session VI # 121, Wednesday Afternoon

 

Close Window