Larry Norton and Joan Massagué recently proposed a provocative model
of cancer growth (1, 2) that is based on a unified process of growth and
metastasis including self-seeding of the primary tumor by circulating
tumor cells. As pointed out in the first paper (1), new insights in cancer
growth can focus our therapies to mechanisms involved in metastasis,
including metastasis to self.
Clinicians usually require some testable prediction before accepting
new interpretation of the previously defined phenomenon and cancer growth
is no exception. An obvious candidate for testing is the model’s
prediction that continuous exchange of malignant cells between metastatic
sites and the primary tumor would make all tumor sites similar in
malignant cells subpopulations. Based on this prediction, we have tried to
statistically compare immunohistochemical (IHC) features of breast cancers
and their metastasis on data of 60 ductal invasive breast cancer patients,
as presented elsewhere (3). IHC profiles of six parameters (ER, PR, Ki-67,
bcl-2, p53, and cathepsin D) for primary tumors and their node metastasis
were done. Observed values were compared by the Wilcoxon matched pairs
test and a p-value was calculated for the each patient. Wilcoxon p-values
ranged in our patients from 0.0431 to 0.999 with median of 0.593 (eight
cases), mean value of 0.571 and standard deviation of 0.297. Six cases had
p-values below 0.10 and 10 cases above 0.90.
If we can use the Wilcoxon p-value as a measure of similarity in IHC
profiles of primary and metastatic breast cancer sites, higher values
might suggest higher rates of self seeding that prevent any site-specific
evolution of malignant subclones. On the contrary, the low values can be
expected in patients with low rates of self seeding, thus allowing
separate tumor sites to take separate routes of development. According to
the described distribution of p-values, it seems that our group of
patients involved both expected types.
To check out whether p-values are dependent on tumor size, or on the
share of the axillary node involvement (number of positive nodes divided
by the number of extirpated nodes), we have performed an EM clustering of
our cases based on tumor size and their node involvement (StatSoft, Inc.
(2004). STATISTICA (data analysis software system), version 7;
www.statsoft.com). Two clusters were identified: the first one with 39
cases and the second with 21 cases, as shown in Figure 1. The first cluster is characterized by
smaller tumors and lower axillary involvement, while the contrary can be
said for the second one. It turned out that the Wilcoxon p-values were
significantly higher in the first cluster, tested with the Mann-Whitney U
test (Z=2.378, p=0.0173).
This result suggests that higher similarity in IHC profiles of breast
cancer and its axillary metastasis can be expected in small aggressive
tumors that, despite their small size, have already produced a regional
metastasis. Larger tumors might be less continuous in self-feeding, so
local evolution of subclones can take place in primary and secondary sites
and this can explain the lower Wilcoxon p-values.
References
1. Norton L. Conceptual and practical implications of breast tissue
geometry: toward a more effective, less toxic therapy. The Oncologist 2005;10:370-381.
2. Norton L, Massagué J. Is cancer a disease of self-seeding? Nat Med.
2006;12:875-878.
3. Kristek J, Dmitrovic B, Kurbel S et al. Tumor growth fraction, expression of estrogen and
progesterone receptors, p53, bcl-2 and cathepsin D activity in primary
ductal invasive breast carcinoma and their axillary lymph node metastases.
Coll Antropol 2007;31:1043-1047.
Disclosures: No actual or potential conflicts of interest were disclosed by the authors of this eLetter.
Read 