)
)
)
)
)
)
)
)
)
MICROSCOPIC DESCRIPTION:
The tumor cells were arranged in organoid clusters or sheets and composed of round to polygonal cells with central round nucleus, abundant eosinophilic or clear cytoplasm and distinct cell membrane (IMAGES 1, 2). Myxoid matrix was also present between the tumor cells (IMAGE 3). The tumor had increased cellularity with nuclear pleomorphism and significant mitotic activity (IMAGES 4, 5). Immunohistochemical stains showed an unusual pattern of staining for CD117 (c-kit) (IMAGE 6). Instead of cytoplasmic staining as typically seen in GIST, there was predominantly nuclear staining. CD34 (IMAGE 7), smooth muscle specific actin (IMAGE 8) and S-100 (IMAGE 9) were immunostain negative. Atypical staining for CD117 and absence of positive immunohistochemical staining for CD34 in the present tumor tissue prevent a definitive diagnosis of GIST at this time. Loss of positive staining for these characteristic GIST markers could be consistent with tumor dedifferentiation over time associated with recurrence. Without the availability of original tissue sections from 1997 to confirm immunohistochemical staining for GIST, the latter diagnosis was suspected but could not be definitively established. To attempt to address this issue of increasing anaplasia over time we undertook a microdissection-based mutational investigational analysis using formalin-fixed paraffin-embedded tissue sections.
MOLECULAR GENETIC ANALYSIS:
) |
Image 10: The left and middle slides show the different tumor areas with blank spaces which have been microdissected directly from the histologic sections. The right side is the H&E slide for the comparison. |
) |
Image 11: Molecular genetic analysis using two microsatellite markers, MCC.E10 (Lanes 1-4) and D9S254 (Lanes 6-8), for loss of heterozygosity (LOH) determination. Lanes 1-4 show heterozygosity (2 bands) in both normal and 3 different tumor tissue areas. Lane 5 shows heterozygosity (2 bands) in normal tissue area. Lanes 6-8 show LOH (1 band) (arrow) in three different tumor areas, which indicated allelic loss. |
Multiple sites within the tumor were microdissected directly off 4 micron thick fixed histologic sections for mutational analysis (IMAGE 10). Normal surrounding mesenchymal tissue was also microdissected to serve as an internal negative control. Loss of heterozygosity (LOH) analysis was carried out using a panel of 14 microsatellite markers situated adjacent to common tumor suppressor genes. These LOH markers target genomic regions including 1p34, 3p26, 5q21, 9p21, 10q25, 17p13 and 18q21. Among the 7 informative markers available for analysis in this case, 4 showed LOH (IMAGE 11) consistent with tumor suppressor gene loss at 9p21 (D9S254). The panel LOH rate (fractional mutation rate) was thus 57%. We have found that stromal tumors with a low fractional mutation rate below 20% behave in a benign fashion while those with a high rate above 80% behave aggressively with early recurrence and short survival (see discussion). The fractional mutation rate for this patient is intermediate consistent with significant tumor anaplasia and associated recurrent disease.
  |
