Time-dependent changes in human tumors: Implications for diagnosis and clinical behavior

Stephen B. Baylin, Geoffrey Mendelsohn

Research output: Contribution to journalArticlepeer-review


In our above discussions, we have tried to illustrate, in three important human cancers, selected features of apparent tumor progression. In each of the situations examined, the changes appear to imvolve shifts in populations of cells that alter the morphological, biochemical, and ultrastructural characteristics of the tumors with time. We have tried to illustrate that all of these changes would contribute to difficulties in precisely classifying the tumors as to cell and/or tissue of origin. Furthermore, we have speculated how changes in small cell lung carcinoma with time could contribute to increasing resistance to currently employed therapeutic modalities; and how, in medullary thyroid carcinoma, deviation of the tumor away from the parent cell pheno-type appears to correlate with increasing aggressiveness of tumor behavior. We have suggested that an improved understanding of the events in tumor progression for human cancers, and particularly an increased ability to biochemically characterize certain stages of tumor change, could have important consequences for enhanced diagnostic capabilities. Also, new means for providing therapeutically important classifications of human cancers might be forthcoming. Finally, we have tried to put our selected observations in human cancers into perspective with the classic concept of tumor progression which has been articulated for non-human tumor models. While our results agree with many of the points set forth in Table 1, we suggest that certain alterations might be considered. First, we would propose that the direction of change in human cancers can be quite vectorial in nature. In other words, a series of unit characters appear to change in a given direction with time. The simultaneous loss of neuroendocrine features and the change in small cell lung cancer histology to that of large cell undifferentiated cancer is used as an illustration of this point. The possibility that resistance to therapeutic modalities could also evolve in tandem with such changes has been discussed. This more orderly picture of change in tumor parameters than would be predicted for the classic concept of tumor progression given in Table 1, would give promise that predicted events in progression of human tumors might be delineated and utilized in diagnostic classifications. Our suggestions concerning the concept of tumor progression as it relates to important human cancers must serve solely, at this point in time, as a working construct. It is obvious that much detailed work must be carried out, not only in the tumors we have discussed, but in other important human cancers. Particularly, we must come to understand the biochemical and molecular parameters of the changes observed and how they relate to the differentiation processes of normal parent cells from which the tumors arise. As such knowledge increases, it seems likely that the progressive changes that occur in human cancers will be recognized as important events for the treatment of cancer; and even as critical target points for manipulating the differentiation of human neoplasms for therapeutic benefit.

Original languageEnglish (US)
Pages (from-to)504-512
Number of pages9
JournalSeminars in oncology
Issue number4
StatePublished - Dec 1982

ASJC Scopus subject areas

  • Hematology
  • Oncology


Dive into the research topics of 'Time-dependent changes in human tumors: Implications for diagnosis and clinical behavior'. Together they form a unique fingerprint.

Cite this