June 1996

Ploidy as a Prognostic Feature in Colonic Adenocarcinoma

Author Affiliations

From the Departments of Surgery (Drs Takanishi, Covarelli, and Michelassi), and Pathology (Dr Hart), The University of Chicago, Chicago, Ill; and the Departments of Statistics and Biostatistics, University of Wisconsin, Madison (Dr Chappell).

Arch Surg. 1996;131(6):587-592. doi:10.1001/archsurg.1996.01430180013002

Objective:  To determine whether DNA content and cellcycle kinetic characteristics in Dukes stage B colonic adenocarcinomas provide additional prognostic information in the context of clinicopathologic variables of known significance.

Design:  Archival, paraffin-embedded tissue blocks from 210 Dukes B colonic adenocarcinomas were retrieved. After confirming stage, tumor cell nuclei were extracted, suspended, and stained. Cell nuclei from adjacent normal colon mucosa were used as controls.

Setting:  University-based, tertiary cancer referral center.

Interventions:  Samples obtained from tumors resected at our institution between 1965 and 1984 were analyzed by flow cytometry for DNA index (DI) and percentages of cells in synthesis (S) phase (%S) and in G2 and mitosis (M) phases (%G2M). The data were correlated with 5-year survival. Follow-up was complete in all patients to at least 5 years.

Results:  Univariate analysis showed that the highest survival rates were associated with DI values near 1 and 2 (diploid and tetraploid tumors, P=.02) and the lowest %G2M values (tumors with fewer mitoses; P=.01). Five-year survival rates also differed significantly between patients with diploid (DI<1.1) and those with aneuploid (1.1<DI<2) tumors (80% vs 64%, respectively; P=.02). Multivariate analysis revealed that race (P<.01), lymphatic or capillary microinvasion (P<.03), and ploidy (P<.05) were significantly associated with outcome. The influence of ploidy, race, and microinvasion on 5-year survival was estimated with logistic regression, and 8 subgroups of patients emerged with 5-year survival probabilities ranging from 39% for black patients with aneuploid tumors and microinvasion to 88% for white patients with diploid tumors and no microinvasion.

Conclusions:  Tumor DNA content provides additional independent information that allows further refinement of our prognostic ability in patients with Dukes B colonic adenocarcinoma. This may aid in the identification of a cohort of patients who may potentially benefit from aggressive adjuvant therapy.(Arch Surg. 1996;131:587-592)