Background
Methods
Data acquisition from the CGP screen
The HBCCL screen and culture conditions
CHO dose response experiments
RNA microarray analysis
Clinical cohorts
Statistical analysis
CHO dose response analysis
Microarray pre-processing
Establishment of in vitro based REGS
Regularised logistic regression for establishment of REGS classifiers
Regularised linear regression for establishment of REGS predictors based on HBCCL
Evaluation in clinical data
Comparison of REGS developed using CGP and HBCCL
Retrospective validation on clinical samples
Differential expression between sensitive and resistant patient samples
GO enrichment
Results
Developing the HBCCL resistance index
Cross validating the elastic net logistic regression models
Cross validating the elastic net linear regression models
Comparison of REGSs developed using CGP and HBCCL
HBCCL | CGP | |||||
---|---|---|---|---|---|---|
N | HR (95% CI) | P-value | N | HR (95% CI) | P-value | |
REGS classifier (univariate)
| ||||||
IDRC (PFS) | 470 | 2.58 (1.72,3.86) | 4.37E-06 | 470 | 1.54 (1.07,2.22) | 0.0211 |
LLMPP (PFS) | 220 | 2.28 (1.10,4.73) | 0.0269 | 220 | 1.70 (0.89,3.22) | 0.105 |
MDFCI (OS) | 67 | 4.56 (1.29,16.19) | 0.0188 | 67 | 0.93 (0.28,3.04) | 0.899 |
REGS classifier (adjusted for IPI)
| ||||||
IDRC (PFS) | 424 | 2.52 (1.64,3.87) | 2.65E-05 | 424 | 1.46 (0.99,2.16) | 0.0562 |
LLMPP (PFS) | 180 | 2.52 (1.13,5.64) | 0.0237 | 180 | 1.37 (0.68,2.74) | 0.375 |
MDFCI (OS) | 63 | 4.05 (1.13,14.51) | 0.0318 | 63 | 0.84 (0.25,2.76) | 0.769 |
REGS predictor (univariate)
| ||||||
IDRC (PFS) | 470 | 1.11 (1.04,1.17) | 0.000572 | 470 | 1.07 (1.01,1.13) | 0.0131 |
LLMPP (PFS) | 220 | 1.10 (1.00,1.20) | 0.0438 | 220 | 1.09 (0.99,1.19) | 0.0794 |
MDFCI (OS) | 67 | 1.32 (1.14,1.54) | 0.000318 | 67 | 0.96 (0.83,1.12) | 0.63 |
REGS predictor (adjusted for IPI)
| ||||||
IDRC (PFS) | 424 | 1.09 (1.03,1.16) | 0.00595 | 424 | 1.05 (1.00,1.11) | 0.0733 |
LLMPP (PFS) | 180 | 1.11 (1.00,1.23) | 0.0433 | 180 | 1.04 (0.94,1.15) | 0.467 |
MDFCI (OS) | 63 | 1.34 (1.15,1.57) | 0.000237 | 63 | 0.95 (0.83,1.07) | 0.387 |
Retrospective validation on clinical samples
Univariate (N = 690) | Multivariate (N = 604) | |||
---|---|---|---|---|
HR (95% CI) | P-value | HR (95% CI) | P-value | |
REGS classifier
| ||||
CHO | 2.33 (1.64,3.30) | 2.05e-06 | 2.30 (1.57,3.37) | 1.78e-05 |
Cyclophosphamide (C) | 0.95 (0.70,1.30) | 0.764 | 0.99 (0.71,1.37) | 0.949 |
Doxorubicin (H) | 2.52 (1.77,3.59) | 2.81e-07 | 2.55 (1.74,3.72) | 1.29e-06 |
Vincristine (O) | 2.07 (1.48,2.88) | 2.04e-05 | 1.69 (1.19,2.40) | 0.003 |
REGS predictor
| ||||
CHO | 1.22 (1.09,1.36) | 0.000353 | 1.22 (1.09,1.37) | 0.0009 |
Cyclophosphamide (C) | 0.97 (0.92,1.02) | 0.184 | 0.98 (0.93,1.03) | 0.438 |
Doxorubicin (H) | 1.10 (1.05,1.16) | 5.53e-05 | 1.10 (1.04,1.16) | 0.0005 |
Vincristine (O) | 1.67 (1.38,2.01) | 9.3e-08 | 1.59 (1.30,1.93) | 4.79e-06 |