It has long been appreciated that both tumor size and the presence of cancer in the regional lymph nodes are indicators of invasive breast cancer outcome, although it has not been obvious as to how to integrate these two qualities into an overall assessment of prognosis. The unifying concept, which makes these calculations possible, was our finding that the lethal distant spread of cancer cells, which render patients incurable by local treatment, occurs with a definable probability per cell. This lead to a simple expression relating tumor size to survival:
(1)
where F is the fraction of patients surviving, e is the exponential constant, and D is the tumor diameter, and Z and Q are coefficients whose values can be determined from survival data. We have found that this expression accurately predicts the relationship between tumor size and population-wide survival for both breast cancer and melanoma.

Eq. (1) offers the possibility of making improved population-wide estimates of cancer survival from data on tumor size, but among individuals, additional factors may need to be taken into account, especially the presence of cancer in the local lymph nodes. To address this question, we examined Kaplan-Meier survival estimates (15-years), by tumor size and nodal status, for 1352 women with invasive breast cancer seen at the Van Nuys Breast Center before 1991. To isolate the individual contributions to survival of tumor size and nodal status, the data were sorted by both tumor size and survival. These survival values revealed that for women with equivalent nodal status, tumor size was associated with increased lethality, and that for women with tumors of equivalent size, lethality increased with the number of positive nodes. Notably, it was not node positivity, per se, but the number of positive nodes that appeared to indicate extra risk of death. Women that had only a single positive node, or had either one or two positive nodes, had very similar survivals to node negative women with tumors of the same size. However, as the number of positive nodes increased, there was a graded increase in lethality with each unitary increase in the number of positive nodes, such that there was approximately an extra 3.5% chance of death for each positive lymph node. The lethal contribution from the primary site was found to be well fit to eq. (1), with the values of Q and Z estimatable from survival data of women with node negative cancer. The presence of each positive node was found to contribute an extra 3.5% lethality. The overall lethality was found to be the sum of the two components, and this provided the potential to make improved estimates of survival for individual patients.

Future studies will concern the determination of how accurately survival estimates can be made among various populations of patients, estimation of the role of local metastasis in the prediction of melanoma survival, and the extension of these methods to the estimation of lung and colon cancer survival.

If you would like additional information on this research project, please contact:
James S. Michaelson, PhD at 617-724-3868