Dose selection and benefit–risk
Dose-comparison and treatment-duration trials read as evidence for choosing a dose: how conclusions change with the weight on benefit and harm, what another comparison would be worth, and how far a prior dose comparison can be borrowed for a new drug.
How far does a dose-comparison trial support dose selection? Arm-level data for sotorasib 960 mg and 240 mg
Published arm-level data from a randomized dose comparison are used to show the decision regions as the weights on benefit and harm change. The difference in severe toxicity is robust, but differences in response rate and OS are uncertain, and the conclusion depends on how harm is defined. No clinical dose is recommended.
- Grade≥3 treatment-related adverse events were 16.0 points more frequent at 960 mg (95% interval +4.2 to +27.8), robust to the prior, data cutoff and correlation.
- The response-rate difference was +7.8 points (−4.3 to +19.8). Weighting one response and one severe toxicity about equally (w ≥ 0.95) favors 240 mg.
- The definition of harm moves the decision most, and the OS HR of 0.75 (0.53 to 1.07) leaves a signal favoring 960 mg. Arm-level public data alone cannot settle the dose.
How much of an existing drug's dose comparison can a new drug borrow?
Borrowing public dose-comparison data from an existing drug through a robust mixture prior helped a new drug's dose selection only slightly when the drugs behaved alike, and cut correct selection sharply when the new drug favored the other dose. With 40 patients per arm the prior-data conflict was hardly detected; a borrowing weight of 0.25 to 0.5 kept the loss within one point.
- 新薬が既存薬と同等なら、既存薬の用量比較を全量借りても期待後悔の改善は約1ポイントで、正しい用量を選ぶ割合は6〜10ポイント上がるにとどまった。
- 新薬が既存薬と逆の用量を有利にする場合、全量の借用で正しい用量を選ぶ割合は28〜42%まで下がり、期待後悔は3.5〜4.3ポイント悪化した。
- 1群40例では頑健な混合事前分布も食い違いをほぼ検出できず、どの条件でも損を1ポイント以内に抑えられた借用の重みは0.25までだった（主解析では0.5まで）。