Skip to main content
. 2018 Mar 7;2:7. doi: 10.1038/s41698-018-0051-4

Fig. 4.

Fig. 4

MPAS is a negative prognostic for survival in colorectal cancer. a Multivariate Cox proportional hazard regression models for both metastatic CRC samples (mCRC) and for samples from AVANT adjuvant phIII trial parameterized using MPAS, KRAS, and BRAF mutation status (mCRC) in addition to MSI status and treatment condition; FOLFOX vs. XELOX (Tx FOL vs. XEL) and Avastin for the AVANT samples. Hazard ratios, 95% confidence intervals, and associated P-values associated with each feature are shown. Kaplan–Meier curves of overall survival (OS) were plotted using the median value of MPAS to classify b mCRC patients or c AVANT patients as either MAPK-high vs. -low for all patients or for patients with BRAF mutations. Cox proportional hazard regression models were used to fit each treatment arm separately, using MPAS as a univariate predictor of overall survival. Hazard ratios, 95% confidence intervals, and associated P-values for MPAS as a univariate predictor are shown within each graph. Kaplan–Meier scans showing hazard ratios of overall survival (OS) over a range of thresholds for classifying MAPK-high vs. -low patients from b mCRC patients or c AVANT patients are also shown. Blue line represents the hazard computed at each threshold, and black lines show 95% confidence intervals