Lipponen P, Eskelinen M, Collan Y
Patologian laitos ja kirurgian klinikka, Suomi.
Ital J Gastroenterol. 1990 Jun;22(3):124-8.
A retrospective study was performed on 63 pancreatic cancer patients diagnosed during the years 1970-1988. The mean follow-up time of all individual patients was 6.9 months (range 0-37 months). Histological grade, clinical stage, and 12 morphometric nuclear variables were correlated to the survival of patients. Clinical stage (p = 0.066), histological grade (p = 0.095), and morphometric variables (p = 0.155) predicted survival in survival analysis. Clinical stage (p = 0.029), morphometric variables (p = 0.157), and histological grade (p = 0.306) predicted survival at one year after diagnosis. Morphometric variables divided grade II tumours into two prognostically different groups (p = 0.008) when the mean survival time was used as a classifier. Also the metastasizing potential of pancreatic cancer and lymph node involvement were associated with morphometric variables. Histological grade and morphometric variables were positively correlated (p less than 0.001). On the basis of our results, it seems that nuclear morphometric variables are more efficient or equal to subjective histological grade in predicting survival of pancreatic cancer patients. This result suggests the potential use of nuclear morphometric variables in grading pancreatic adenocarcinomas, and selecting patients for different modes of therapy.