Sarcomatoid renal cell carcinoma is not a distinct histologic entity and re
presents high-grade transformation in different subtypes of renal cell carc
inoma. It is not known whether any particular histologic type has a predile
ction for sarcomatoid change or whether the primary histologic type of rena
l carcinoma undergoing sarcomatoid change affects prognosis. Of 952 consecu
tively histologically subtyped renal cell carcinomas, the incidence of sarc
omatoid differentiation was 8% in conventional (clear cell) renal carcinoma
, 3% in papillary renal carcinoma, 9% in chromophobe renal carcinoma, 29% i
n collecting duct carcinoma, and 11% in unclassified renal cell carcinoma.
One hundred one renal cell carcinomas with sarcomatoid change were studied,
and clinicopathologic parameters were correlated with outcome. The mean ag
e of patients was 60 years (range, 33-80 years), and the male-to-female rat
io was 1.6:1. The median tumor size was 9.2 cm (range, 3-25 cm). The primar
y histologic subtype of the carcinoma component was conventional (clear cel
l) renal carcinoma in 80 cases, papillary renal carcinoma in eight, chromop
hobe renal carcinoma in seven, collecting duct carcinoma in two, and unclas
sified renal cell carcinoma in four. The sarcomatoid component resembled fi
brosarcoma in 54 cases, malignant fibrous histiocytoma in 44, undifferentia
ted sarcoma (not otherwise specified) in three with focal rhabdomyosarcomat
ous component in two of them. The spindled elements accounted for 1% to 99%
of the sampled tumor (median, 40%; mean 45%). The histologic grade of the
spindled elements was intermediate to high in 92 cases and low in nine case
s. Most cases were TNM stages III and IV (seven stage I, six stage II, 63 s
tage III, and 25 stage IV). Follow-up was available in 88 patients; 61 (69%
) patients died of disease and had a median survival time of 19 months. Dis
tant metastases, most frequently to the lungs, were documented in 51 (66%)
of 77 patients who had available clinical information regarding distant met
astasis. The disease-specific survival rate was 22% and 13% after 5 and 10
years, respectively, compared with a cohort of renal cell carcinomas withou
t sarcomatoid change with a 5- and 10-year disease-specific survival of 79%
and 76%, respectively. Kaplan-Meier survival analysis showed that tumors w
ith high TNM stage (p = 0.0027), at least 50% sarcomatoid component (p = 0.
0453), and angiolymphatic invasion (p = 0.0282) were associated with decrea
sed survival rates. The primary histologic subtype of the carcinoma compone
nt and the type and grade of the sarcomatoid component did not affect survi
val. In a Cox proportional hazard regression model, TNM stage appeared to b
e the only significant variable in predicting outcome among renal cell carc
inomas with sarcomatoid change (p = 0.018; risk ratio, 6.984 and 8.439). Co
mpared with a cohort of renal cell carcinomas without sarcomatoid change, s
arcomatoid tumors tended to present at a more advanced stage (p = 0.0001).
Also, when adjusted for stage, necrosis, and tumor size, patients with tumo
rs with sarcomatoid differentiation had a worse prognosis than did patients
with tumors without sarcomatoid change(p = 0.0001). In conclusion, sarcoma
toid change in renal cell carcinoma portends a worse prognosis. Because tum
ors with even a small component of sarcomatoid change may have an adverse o
utcome, this finding, when present, should be noted in the surgical patholo
gy report.