Several investigators, the SEER data, and the ECOG/Intergroup study ha
ve shown that patients with small tumors (<0.5 cm) have a recurrence r
ate of less than 2%, compared to 20-25% for large tumors (greater than
or equal to 5 cm). Nuclear grade and tumor differentiation are establ
ished indicators; however, the interobserver lack of concordance has t
hwarted their use in clinical trials. The presence of peritumoral lymp
hatic and blood vessel invasion (PLBI) is associated with a relative r
isk of recurrence of 4.7. The predictive value of the presence of horm
one receptors in tumors is associated with a favorable disease free an
d overall survival difference of 8-10%; however, this advantage is bei
ng eroded by the early appearance of other factors, such as the epider
mal growth factor receptor (EGFR), proliferative capacity (S-phase), n
uclear grade, and HER-2/neu oncogene. Concordance among the different
methods of hormone-receptor assay (immunocytochemical, sucrose gradien
t, and dextran-coated charcoal) is essential to refine the true value
of these factors. DNA flow cytometry measurements of ploidy (DNA conte
nt) and S-phase fraction are the most characterized of the prognostic
factors. There are conflicting reports regarding the clinical signific
ance of ploidy status, while measurements of S-phase fraction clearly
indicate a robust association with disease free and overall survival.
Our data continue to show that S-phase, but not ploidy, can predict ti
me to recurrence significantly in untreated patients, even when data a
re stratified for tumor size. HER-2/neu oncogene is expressed in about
50% of ductal carcinoma in situ and 14% of invasive ductal carcinoma.
The presence of this oncogene at high copy number may be a useful ind
ependent marker of poor prognosis and may be associated with drug resi
stance and correlated with tumor recurrence and shorter survival. EGFR
could be measured in most breast tumors, and the level of its express
ion has inversely correlated with estrogen receptor protein expression
. The value of EGFR as a predictor of prognosis remains controversial
and is still being investigated. Cathepsin-D provides a provocative bi
ologic rationale but is hindered by different and incongruent methods
of analysis. The majority of large studies with more than 3-years' fol
low-up suggests that high cathepsin-D levels may be predictive of grea
ter recurrence and lower survival. Angiogenesis has been implicated as
a critical component of the metastatic process. Early studies show th
at tumor angiogenesis is an independent and highly significant prognos
tic indicator, and its presence may suggest the selection of ''anti-an
giogenic therapy.'' Mutations in the tumor suppressor p53 gene have be
en detected in 13-15% of primary breast cancers, resulting in overexpr
ession of mutant p53, which is associated with poor prognosis. There a
lso may be a strong relationship between mutant p53 overexpression and
high S-phase activity. Although there is a growing understanding of t
he prognostic significance of individual prognostic factors, the topic
of how practically to use multiple prognostic factors in concert is s
till largely unexplored. The development of practical systems for clin
ical application will require improvements in the areas of (1) standar
dization of methodologies and interlaboratory quality control for prog
nostic factor determinations, (2) definition of a limited set of progn
ostic markers that are independently predictive, and (3) staging syste
ms or tools that integrate this information. Large studies based on co
operative group material and large data bases will play a vital role i
n this process.