Limits...
An online gene expression assay for determining adjuvant therapy eligibility in patients with stage 2 or 3 colon cancer.

Van Laar RK - Br. J. Cancer (2010)

Bottom Line: the decision whether to treat patients with non-metastatic colon cancer with adjuvant chemotherapy is determined by clinical staging, frequently resulting in over or undertreatment. gene expression data and clinical information from 232 stage 1-4 colon cancer patients were analysed to identify expression patterns predictive of recurrence.The signature was evaluated on an independent series of 60 stage 2 and 3 patients.Multivariate analysis showed the classifier to be associated with approximately three- to fourfold increased risk of recurrence. the prognostic gene expression signature is able to stratify stage 2 and 3 colon cancer patients into groups with significant differences in 5-year DFS, information that may ultimately reduce deaths from colon cancer.

View Article: PubMed Central - PubMed

Affiliation: ChipDX LLC, PO Box 286874, New York, NY 10128, USA. ryan.vanlaar@chipdx.com

Show MeSH

Related in: MedlinePlus

Kaplan–Meier analysis of stage 2 and 3 colon cancer patients stratified by clinical staging and the 163-gene prognostic signature. Risk-group predictions for those patients who were part of the training series (n=144) were determined by cross-validation analysis. The final 163-gene classifier was then applied to the independent analysis series (n=60) of patients, who were not involved in the gene selection or algorithm development process. P-values generated by the log rank test.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3008603&req=5

fig3: Kaplan–Meier analysis of stage 2 and 3 colon cancer patients stratified by clinical staging and the 163-gene prognostic signature. Risk-group predictions for those patients who were part of the training series (n=144) were determined by cross-validation analysis. The final 163-gene classifier was then applied to the independent analysis series (n=60) of patients, who were not involved in the gene selection or algorithm development process. P-values generated by the log rank test.

Mentions: In order to reflect the intended use of the assay, risk-group predictions for the subset of the training series with stage 2 or 3 colon cancer (n=144) were compared using Kaplan–Meier analysis for DFS and DSS. These predictions were generated using LOOCV, as part of the gene selection and algorithm training process, in order to minimise over fitting of the data (Simon, 2005). Log rank testing revealed a significant difference between the high- and low-risk groups for both DFS (P=0.0008, HR: 4.08 95% CI: 1.99–8.34) and DSS (P<0.0001, HR 19.59 95% CI: 8.33–46.07), and also for stratification by risk group and clinical staging (Figure 3). For comparison purposes, Kaplan–Meier analysis of these patients stratified by staging was performed, however the result was not statistically significant for either DFS (P=0.75) or DSS (P=0.30).


An online gene expression assay for determining adjuvant therapy eligibility in patients with stage 2 or 3 colon cancer.

Van Laar RK - Br. J. Cancer (2010)

Kaplan–Meier analysis of stage 2 and 3 colon cancer patients stratified by clinical staging and the 163-gene prognostic signature. Risk-group predictions for those patients who were part of the training series (n=144) were determined by cross-validation analysis. The final 163-gene classifier was then applied to the independent analysis series (n=60) of patients, who were not involved in the gene selection or algorithm development process. P-values generated by the log rank test.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3008603&req=5

fig3: Kaplan–Meier analysis of stage 2 and 3 colon cancer patients stratified by clinical staging and the 163-gene prognostic signature. Risk-group predictions for those patients who were part of the training series (n=144) were determined by cross-validation analysis. The final 163-gene classifier was then applied to the independent analysis series (n=60) of patients, who were not involved in the gene selection or algorithm development process. P-values generated by the log rank test.
Mentions: In order to reflect the intended use of the assay, risk-group predictions for the subset of the training series with stage 2 or 3 colon cancer (n=144) were compared using Kaplan–Meier analysis for DFS and DSS. These predictions were generated using LOOCV, as part of the gene selection and algorithm training process, in order to minimise over fitting of the data (Simon, 2005). Log rank testing revealed a significant difference between the high- and low-risk groups for both DFS (P=0.0008, HR: 4.08 95% CI: 1.99–8.34) and DSS (P<0.0001, HR 19.59 95% CI: 8.33–46.07), and also for stratification by risk group and clinical staging (Figure 3). For comparison purposes, Kaplan–Meier analysis of these patients stratified by staging was performed, however the result was not statistically significant for either DFS (P=0.75) or DSS (P=0.30).

Bottom Line: the decision whether to treat patients with non-metastatic colon cancer with adjuvant chemotherapy is determined by clinical staging, frequently resulting in over or undertreatment. gene expression data and clinical information from 232 stage 1-4 colon cancer patients were analysed to identify expression patterns predictive of recurrence.The signature was evaluated on an independent series of 60 stage 2 and 3 patients.Multivariate analysis showed the classifier to be associated with approximately three- to fourfold increased risk of recurrence. the prognostic gene expression signature is able to stratify stage 2 and 3 colon cancer patients into groups with significant differences in 5-year DFS, information that may ultimately reduce deaths from colon cancer.

View Article: PubMed Central - PubMed

Affiliation: ChipDX LLC, PO Box 286874, New York, NY 10128, USA. ryan.vanlaar@chipdx.com

Show MeSH
Related in: MedlinePlus