Table 1.
Meta‐analysis results using Royston‐Parmar models
| Comparison | FTE* | Cochran's Q | Global Non‐PH Test | Schoenfeld Residuals |
|---|---|---|---|---|
| RT vs CTRT | −0.215 (−0.336, −0.086) | 12.71, 15 df, P=.625 | χ 2=0.161, 1 df, P=.688 | χ 2=25.64, 16 df, P=.059 |
| RT vs CT+RT | −0.191 (−0.375, −0.007) | 20.69, 6 df, P=.002 | χ 2=2.522, 1 df, P=.112 | χ 2=10.34, 7 df, P=.170 |
| ⩽14 days | ||||
| RT vs CT+RT | 0.227 (0.073, 0.385) | 12.34, 10 df, P=.263 | χ 2=0.006, 1 df, P=.944 | χ 2=7.65, 11 df, P=.744 |
| >14 days | ||||
| RT vs CT+S | −0.447 (−0.654, ‐0.243) | 8.85, 4 df, P=.065 | χ 2=0.118, 1 df, P=.731 | χ 2=8.65, 5 df, P=.124 |
| CT+RT vs CT+S | −0.444 (−0.830, −0.061) | 0.01, 1 df, P=.939 | χ 2=0.164, 1 df, P=.686 | χ 2=0.49, 2 df, P=.783 |
Abbreviations: CTRT, chemoradiation; CT+RT, neoadjuvant chemotherapy plus radiotherapy; CT+S, neoadjuvant chemotherapy plus surgery; FTE, fixed treatment effect; RT, radiotherapy; PH, proportional hazard.* Values are log hazard ratios and 95% credible intervals.