Machine learning methodology for high throughput personalized neutron dose reconstruction in mixed neutron + photon exposures.
Machine learning methodology for high throughput personalized neutron dose reconstruction in mixed neutron + photon exposures.
复制标题
DOI：
10.1038/s41598-021-83575-5
复制
发表时间：
2021-02-17
影响因子：
4.6
通讯作者：
Brenner DJ
中科院分区：
文献类型：
--
作者：
Shuryak I;Turner HC;Pujol-Canadell M;Perrier JR;Garty G;Brenner DJ
我们在辐射生物剂量学领域实施了机器学习，以定量重建混合中子+光子暴露中的中子剂量，这在简易核装置爆炸中是预期的。这种个性化的重建对于分类和治疗至关重要，因为中子比光子更具生物破坏性。我们使用了一种高通量微核试验和自动扫描/成像技术，对来自人血的淋巴细胞进行离体照射，采用44种不同的0-4戈伊中子和0-15戈伊光子组合（542份血样），其中包括对过去实验的再分析。我们开发了几个描述双核细胞中微核/细胞概率分布的指标，并将其用作随机森林（RF）和XGboost机器学习分析中的预测因子，以重建每个样本中的中子剂量。通过对随机选择的数据子集进行重复交叉验证来训练两种算法，并对其余数据进行性能测量，从而最大限度地减少“过拟合”的可能性。RF实现了最佳性能。在300次随机训练/测试分割中，实际中子剂量与重建中子剂量的平均R2为0.869（范围为0.761至0.919），均方根误差为0.239（0.195至0.351）戈伊。这些结果表明，机器学习在临床相关的复杂辐射暴露场景中重建中子剂量分量的潜力很大。
We implemented machine learning in the radiation biodosimetry field to quantitatively reconstruct neutron doses in mixed neutron + photon exposures, which are expected in improvised nuclear device detonations. Such individualized reconstructions are crucial for triage and treatment because neutrons are more biologically damaging than photons. We used a high-throughput micronucleus assay with automated scanning/imaging on lymphocytes from human blood ex-vivo irradiated with 44 different combinations of 0–4 Gy neutrons and 0–15 Gy photons (542 blood samples), which include reanalysis of past experiments. We developed several metrics that describe micronuclei/cell probability distributions in binucleated cells, and used them as predictors in random forest (RF) and XGboost machine learning analyses to reconstruct the neutron dose in each sample. The probability of “overfitting” was minimized by training both algorithms with repeated cross-validation on a randomly-selected subset of the data, and measuring performance on the rest. RF achieved the best performance. Mean R2 for actual vs. reconstructed neutron doses over 300 random training/testing splits was 0.869 (range 0.761 to 0.919) and root mean squared error was 0.239 (0.195 to 0.351) Gy. These results demonstrate the promising potential of machine learning to reconstruct the neutron dose component in clinically-relevant complex radiation exposure scenarios.
登录
查看更多内容
影响因子：
3.4
作者：
Lue SW;Repin M;Mahnke R;Brenner DJ
通讯作者：
Brenner DJ
影响因子：
--
作者：
Liu J;Li Y;Wilkins R;Flegal F;Knoll JHM;Rogan PK
通讯作者：
Rogan PK
影响因子：
2.6
作者：
Jacobs, Aviva R.;Guyon, Timothy;Terbrueggen, Robert
通讯作者：
Terbrueggen, Robert
影响因子：
1
作者：
Coleman, C. Norman;Koerner, John F.
通讯作者：
Koerner, John F.
影响因子：
3.4
作者：
Garty, Guy;Xu, Yanping;Brenner, David J.
通讯作者：
Brenner, David J.