PURPOSE: The Bethesda guidelines were developed for selection of patients w
hose tumors should be tested for high microsatellite instability. This stud
y examined the validity of the different Bethesda criteria in relation to m
icrosatellite instability status to simplify their use in clinical practice
. METHODS: A total of 164 patients with colorectal or hereditary nonpolypos
is colorectal cancer-associated cancers were registered on the basis of the
Amsterdam criteria without age limitations (11 cases), multiple tumors (2
cases), the accumulation of colorectal cancer in the family (no first-degre
e relatives affected or the index patient's age up to 50 years; 45 cases),
an early age at onset up to 50 years (13 cases), morphologic and histopatho
logic manifestations (right-sided colorectal cancer, mucinous undifferentia
ted histology; 1 case), and the Bethesda criteria (92 cases). The microsate
llite instability status of tumors was determined using the International C
ollaborative Group on Hereditary Non-Polyposis Colorectal Cancer marker ref
erence panel. RESULTS: When applying all Bethesda criteria, high microsatel
lite instability tumors were identified in our hereditary nonpolyposis colo
rectal cancer registry with a sensitivity of 87 percent. Twenty-nine percen
t (27/92) of the Bethesda-positive patients displayed high microsatellite i
nstability compared with 6 percent of patients (4/72) not meeting these cri
teria (P < 0.001). Only Bethesda Criteria 1, 3, and 4 showed a significantl
y different distribution of the microsatellite instability status when comp
ared with those of the remaining patients registered (P less than or equal
to 0.001). These three criteria detected high microsatellite instability tu
mors in 48 percent (10/21), 50 percent (18/36), and 31 percent (21/67) of p
atients, respectively. When applying these criteria only, a cumulative dete
ction rate of 77 percent of all (24/31) high microsatellite instability cas
es was found, thereby identifying 89 percent of high microsatellite instabi
lity tumors among the Bethesda-positive patients. Patients matching Criteri
a 1, 3, and 4 frequently showed hMSH2 or hMLH1 germline mutations and tumor
-specific loss of protein expression. CONCLUSION: In our hereditary nonpoly
posis colorectal cancer registry the complete Bethesda criteria showed the
highest sensitivity to identify patients with high microsatellite instabili
ty tumors. However, for general medical practice outside academic centers,
three criteria are reasonably accurate for adequate high microsatellite ins
tability tumor selection.