Clinical significance of the genetic landscape of pancreatic cancer and implications for identification of potential long-term survivors

Abstract

Purpose: Genetic alterations of KRAS, CDKN2A, TP53, and SMAD4 are the most frequent events in pancreatic cancer. We determined the extent to which these 4 alterations are coexistent in the same carcinoma, and their impact on patient outcome.

Experimental design: Pancreatic cancer patients who underwent an autopsy were studied (n = 79). Matched primary and metastasis tissues were evaluated for intragenic mutations in KRAS, CDKN2A, and TP53 and immunolabeled for CDKN2A, TP53, and SMAD4 protein products. The number of altered driver genes in each carcinoma was correlated to clinicopathologic features. Kaplan-Meier estimates were used to determine median disease free and overall survival, and a Cox proportional hazards model used to compare risk factors.

Results: The number of genetically altered driver genes in a carcinoma was variable, with only 29 patients (37%) having an alteration in all 4 genes analyzed. The number of altered driver genes was significantly correlated with disease free survival (P = 0.008), overall survival (P = 0.041), and metastatic burden at autopsy (P = 0.002). On multivariate analysis, the number of driver gene alterations in a pancreatic carcinoma remained independently associated with overall survival (P = 0.046). Carcinomas with only 1 to 2 driver alterations were enriched for those patients with the longest survival (median 23 months, range 1 to 53).

Conclusions: Determinations of the status of the 4 major driver genes in pancreatic cancer, and specifically the extent to which they are coexistent in an individual patients cancer, provides distinct information regarding disease progression and survival that is independent of clinical stage and treatment status.

Figure 1

Deviant SMAD4 and CDKN2A immunohistochemical labeling patterns in pancreatic cancer tissues. A. Intact Smad4 immunolabeling in a primary carcinoma. Both nuclear and cytoplasmic labeling for SMAD4 is present within the neoplastic glands (Ca) in an area of perineural invasion. B. Loss of SMAD4 immunolabeling in a liver metastasis derived from the carcinoma shown in A. In this example, no labeling of SMAD4 is seen within the neoplastic glands (Ca). By contrast, positive labeling of surrounding stromal cells (Str) is present. C. Hematoxylin and eosin stained section of infiltrating pancreatic carcinoma. D. CK19 labeling of the carcinoma shown in C indicating strong positive labeling throughout the neoplastic epithelium. E. Example of focal loss of CDKN2A immunolabeling in the carcinoma shown in C. No labeling of CDKN2A is seen in the neoplastic epithelium within the upper half of the shown section, whereas strong positive labeling is seen within scattered neoplastic glands in the lower half.

Figure 2

Kaplan-Meier survival curves demonstrating the relationship of number of driver gene alterations (1–2 versus 3–4) to disease free survival in 20 Stage I/II patients specifically (A) and overall survival among all 79 patients (B). Survival curves were compared by a log rank test. The percent of patients alive at interval time points are also indicated for each arm.