Limits...
Identification of predictive pathways for non-hodgkin lymphoma prognosis.

Han X, Li Y, Huang J, Zhang Y, Holford T, Lan Q, Rothman N, Zheng T, Kosorok MR, Ma S - Cancer Inform (2010)

Bottom Line: Unlike in existing studies, we targeted at identifying pathways with significant additional predictive power beyond clinical factors.In addition, we accounted for the joint effects of multiple SNPs within pathways, whereas some existing studies drew pathway-level conclusions based on separate analysis of individual SNPs.They may provide further insights into the biological mechanisms underlying the prognosis of NHL.

View Article: PubMed Central - PubMed

Affiliation: Department of Nutrition, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA.

ABSTRACT
Despite decades of intensive research, NHL (non-Hodgkin lymphoma) still remains poorly understood and is largely incurable. Recent molecular studies suggest that genomic variants measured with SNPs (single nucleotide polymorphisms) in genes may have additional predictive power for NHL prognosis beyond clinical risk factors. We analyzed a genetic association study. The prognostic cohort consisted of 346 patients, among whom 138 had DLBCL (diffuse large B-cell lymphoma) and 101 had FL ( follicular lymphoma). For DLBCL, we analyzed 1229 SNPs which represented 122 KEGG pathways. For FL, we analyzed 1228 SNPs which represented 122 KEGG pathways. Unlike in existing studies, we targeted at identifying pathways with significant additional predictive power beyond clinical factors. In addition, we accounted for the joint effects of multiple SNPs within pathways, whereas some existing studies drew pathway-level conclusions based on separate analysis of individual SNPs. For DLBCL, we identified four pathways, which, combined with the clinical factors, had medians of the prediction logrank statistics as 2.535, 2.220, 2.094, 2.453, and 2.512, respectively. As a comparison, the clinical factors had a median of the prediction logrank statistics around 0.552. For FL, we identified two pathways, which, combined with the clinical factors, had medians of the prediction logrank statistics as 4.320 and 3.532, respectively. As a comparison, the clinical factors had a median of the prediction logrank statistics around 1.212. For NHL overall, we identified three pathways, which, combined with the clinical factors, had medians of the prediction logrank statistics as 5.722, 5.314, and 5.441, respective. As a comparison, the clinical factors had a median of the prediction logrank statistics around 4.411. The identified pathways have sound biological bases. In addition, they are different from those identified using existing approaches. They may provide further insights into the biological mechanisms underlying the prognosis of NHL.

No MeSH data available.


Related in: MedlinePlus

Densities of PIC (blue dashed line) and PIC+G for a predictive pathway (black dash-dotted line) and a non-predictive pathway (green solid line).
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3021201&req=5

f2-cin-2010-281: Densities of PIC (blue dashed line) and PIC+G for a predictive pathway (black dash-dotted line) and a non-predictive pathway (green solid line).

Mentions: Representative plots of PIC+G and PIC are shown in Figure 2. For FL, two pathways are used as examples: the Endometrial cancer pathway which has significant additional predictive power, and the Glycerolipid metabolism pathway which does not. For a better view, only the estimated densities of the logrank statistics are plotted. It is easy to see that, for a predictive pathway, the estimated densities of PIC + G and PIC are well separated. However, for a pathway without predictive power, the estimated densities are almost completely overlapped.


Identification of predictive pathways for non-hodgkin lymphoma prognosis.

Han X, Li Y, Huang J, Zhang Y, Holford T, Lan Q, Rothman N, Zheng T, Kosorok MR, Ma S - Cancer Inform (2010)

Densities of PIC (blue dashed line) and PIC+G for a predictive pathway (black dash-dotted line) and a non-predictive pathway (green solid line).
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3021201&req=5

f2-cin-2010-281: Densities of PIC (blue dashed line) and PIC+G for a predictive pathway (black dash-dotted line) and a non-predictive pathway (green solid line).
Mentions: Representative plots of PIC+G and PIC are shown in Figure 2. For FL, two pathways are used as examples: the Endometrial cancer pathway which has significant additional predictive power, and the Glycerolipid metabolism pathway which does not. For a better view, only the estimated densities of the logrank statistics are plotted. It is easy to see that, for a predictive pathway, the estimated densities of PIC + G and PIC are well separated. However, for a pathway without predictive power, the estimated densities are almost completely overlapped.

Bottom Line: Unlike in existing studies, we targeted at identifying pathways with significant additional predictive power beyond clinical factors.In addition, we accounted for the joint effects of multiple SNPs within pathways, whereas some existing studies drew pathway-level conclusions based on separate analysis of individual SNPs.They may provide further insights into the biological mechanisms underlying the prognosis of NHL.

View Article: PubMed Central - PubMed

Affiliation: Department of Nutrition, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA.

ABSTRACT
Despite decades of intensive research, NHL (non-Hodgkin lymphoma) still remains poorly understood and is largely incurable. Recent molecular studies suggest that genomic variants measured with SNPs (single nucleotide polymorphisms) in genes may have additional predictive power for NHL prognosis beyond clinical risk factors. We analyzed a genetic association study. The prognostic cohort consisted of 346 patients, among whom 138 had DLBCL (diffuse large B-cell lymphoma) and 101 had FL ( follicular lymphoma). For DLBCL, we analyzed 1229 SNPs which represented 122 KEGG pathways. For FL, we analyzed 1228 SNPs which represented 122 KEGG pathways. Unlike in existing studies, we targeted at identifying pathways with significant additional predictive power beyond clinical factors. In addition, we accounted for the joint effects of multiple SNPs within pathways, whereas some existing studies drew pathway-level conclusions based on separate analysis of individual SNPs. For DLBCL, we identified four pathways, which, combined with the clinical factors, had medians of the prediction logrank statistics as 2.535, 2.220, 2.094, 2.453, and 2.512, respectively. As a comparison, the clinical factors had a median of the prediction logrank statistics around 0.552. For FL, we identified two pathways, which, combined with the clinical factors, had medians of the prediction logrank statistics as 4.320 and 3.532, respectively. As a comparison, the clinical factors had a median of the prediction logrank statistics around 1.212. For NHL overall, we identified three pathways, which, combined with the clinical factors, had medians of the prediction logrank statistics as 5.722, 5.314, and 5.441, respective. As a comparison, the clinical factors had a median of the prediction logrank statistics around 4.411. The identified pathways have sound biological bases. In addition, they are different from those identified using existing approaches. They may provide further insights into the biological mechanisms underlying the prognosis of NHL.

No MeSH data available.


Related in: MedlinePlus