Affordable Access

Publisher Website

Evaluating the Number of Stages in Development of Squamous Cell and Adenocarcinomas across Cancer Sites Using Human Population-Based Cancer Modeling

Public Library of Science
Publication Date
DOI: 10.1371/journal.pone.0037430
  • Research Article
  • Biology
  • Developmental Biology
  • Molecular Development
  • Population Biology
  • Epidemiology
  • Population Modeling
  • Medicine
  • Dermatology
  • Skin Neoplasms
  • Malignant Skin Neoplasms
  • Squamous Cell Carcinomas
  • Cancer Epidemiology
  • Oncology
  • Cancers And Neoplasms
  • Lung And Intrathoracic Tumors
  • Adenocarcinoma Of The Lung
  • Basic Cancer Research
  • Biology


Background Adenocarcinomas (ACs) and squamous cell carcinomas (SCCs) differ by clinical and molecular characteristics. We evaluated the characteristics of carcinogenesis by modeling the age patterns of incidence rates of ACs and SCCs of various organs to test whether these characteristics differed between cancer subtypes. Methodology/Principal Findings Histotype-specific incidence rates of 14 ACs and 12 SCCs from the SEER Registry (1973–2003) were analyzed by fitting several biologically motivated models to observed age patterns. A frailty model with the Weibull baseline was applied to each age pattern to provide the best fit for the majority of cancers. For each cancer, model parameters describing the underlying mechanisms of carcinogenesis including the number of stages occurring during an individual’s life and leading to cancer (m-stages) were estimated. For sensitivity analysis, the age-period-cohort model was incorporated into the carcinogenesis model to test the stability of the estimates. For the majority of studied cancers, the numbers of m-stages were similar within each group (i.e., AC and SCC). When cancers of the same organs were compared (i.e., lung, esophagus, and cervix uteri), the number of m-stages were more strongly associated with the AC/SCC subtype than with the organ: 9.79±0.09, 9.93±0.19 and 8.80±0.10 for lung, esophagus, and cervical ACs, compared to 11.41±0.10, 12.86±0.34 and 12.01±0.51 for SCCs of the respective organs (p<0.05 between subtypes). Most SCCs had more than ten m-stages while ACs had fewer than ten m-stages. The sensitivity analyses of the model parameters demonstrated the stability of the obtained estimates. Conclusions/Significance A model containing parameters capable of representing the number of stages of cancer development occurring during individual’s life was applied to the large population data on incidence of ACs and SCCs. The model revealed that the number of m-stages differed by cancer subtype being more strongly associated with ACs/SCCs histotype than with organ/site.

There are no comments yet on this publication. Be the first to share your thoughts.