Correlative Discrimination into Sorting Heroin Based on Inductively Coupled Plasma Mass Spectrometry
CUI Xuezi1,2, CHEN Yongsheng2, CHEN Guiliang3, XU Jinlun2, WANG Rong2, ZHANG Yurong2,*
1. Shanghai Institute of Pharmaceutical Industry & China State Institute of Pharmaceutical Industry & State Key Laboratory of Innovative Pharmaceuticals and Pharmaceutical Industry, Shanghai 201203, China;
2. Institute of Forensic Evidence, Shanghai Public Security Bureau & Shanghai Key Laboratory of Scene Evidence, Shanghai 200083, China;
3. Shanghai Center for Pharmaceuticals Evaluation and Inspection, Shanghai 201203, China
Objective To analyze the elements in seized heroin samples with inductively coupled plasma-mass spectrometry (ICP-MS) to determine the contents of tested elements and therewith categorize the samples so as to have the relevant cases clustered.Methods Ultrasonic-assisted dissolution was used to prepare the heroin samples from which ICP-MS was optimized to simultaneously quantify 28 elements and their isotopes. Sixty-five statistical combinations were compared and consecutively analyzed into all the obtained data with hierarchical cluster analysis (HCA).Results Standard curves were established for determination of 28 elements and their isotopes. For all the heroin samples analyzed through ICP-MS, their harbored elements were obtained of each one’s respective concentration range. Data pre-processing of normalization followed by Euclidean distance was found to be the best statistical compromise for optimal discrimination on whether the samples are correlated. Resultant from HCA to analyze the data of 33 heroin samples, the corresponding cases were classified.Conclusions Multiple elements harbored in heroin can be determined with ICP-MS simultaneously and quickly, valuable of providing intelligence for cases connection and/or relevance with their disclosed correlation of the involving samples.
Key words:
drug analysis; ICP-MS; heroin; elements; HCA
表1 决定系数, 标准曲线方程, 检出限, 海洛因检材浓度范围, 方法的日内精密度和日间精密度Table 1 The determination coefficients, standard curves, LODs, concentration ranges of elements in the tested heroin samples, and the intra- and inter-day precisions of the method adopted
质量数
名称
决定系数(R٢)
标准曲线方程
检出限/(µ g/L)
海洛因检材中各元素的浓度范围/(mg/kg)
日内精密度(n=5)
日间精密度(n=3)
9
Be
0.9999
y=0.0011x+5.4740× 10-6
0.0089
n.d.~0.0001978
0.66%
3.31%
11
B
0.9998
y=0.0010x+3.3288× 10-4
0.0423
n.d.~1.155
0.35%
3.98%
23
Na
0.9993
y=0.0065x+0.0350
2.5029
18.834~1000
0.33%
2.35%
24
Mg
1.0000
y=0.0040x+0.0033
0.1413
24.737~214.395
3.22%
2.60%
26
Mg
0.9999
y=6.4400× 10-4x+5.1738× 10-4
0.1933
24.629~211.843
0.43%
2.48%
27
Al
1.0000
y=0.0048x+0.0053
0.1605
0.682~15.312
3.14%
1.98%
39
K
1.0000
y=0.0060x+0.2178
5.3790
1.333~141.064
0.39%
1.57%
43
Ca
0.9997
y=1.3630× 10-5x+4.5868× 10-4
5.0430
1065.96~4106.65
0.28%
1.61%
44
Ca
0.9990
y=1.8786× 10-4x+0.0061
4.3620
1038.35~4836.94
0.34%
1.56%
47
Ti
1.0000
y=5.8073× 10-4x+3.7409× 10-5
0.0422
0.0531~0.9156
2.65%
2.45%
51
V
1.0000
y=0.0083x+2.8377× 10-4
0.0040
0.0054~0.156
0.63%
1.21%
52
Cr
1.0000
y=0.0074x+0.0017
0.0350
0.1~11.14
0.71%
0.51%
53
Cr
0.9999
y=9.0394× 10-4x+1.169× 10-4
0.0760
0.117~9.864
0.58%
2.44%
55
Mn
1.0000
y=0.0097x+7.2391× 10-4
0.0190
0.8739~89.63
0.59%
1.05%
56
Fe
0.9997
y=0.0087x+0.4171
3.2490
0.4517~52.588
2.03%
1.94%
57
Fe
0.9998
y=2.2003× 10-4x+1.2001× 10-4
1.3287
4.45~59.705
0.75%
3.63%
59
Co
1.0000
y=0.0086x+6.0143× 10-5
0.0081
0.0072~0.0823
0.36%
1.31%
60
Ni
0.9998
y=0.0019x+1.1117× 10-4
0.0101
0.0151~0.9998
0.22%
2.08%
63
Cu
0.9999
y=0.0045x+2.1167× 10-4
0.0133
0.0499~3.6619
0.29%
1.05%
65
Cu
0.9997
y=0.0022x+1.2001× 10-4
0.0339
0.0537~3.698
0.12%
1.91%
66
Zn
0.9993
y=0.0012x+0.0017
0.1274
7.612~151.576
0.06%
1.51%
75
As
0.9999
y=0.0010x+3.5907× 10-5
0.0064
0.0062~0.9282
0.28%
1.48%
77
As
1.0000
y=6.9561× 10-5x+1.2491× 10-5
0.0551
n.d.~0.6618
0.41%
1.92%
78
Se
1.0000
y=2.2217× 10-4x+0.0012
0.7284
n.d.~0.0149
0.41%
1.90%
82
Se
1.0000
y=9.1192× 10-5x+1.0911× 10-4
0.0650
n.d.~0.037
0.31%
0.98%
95
Mo
0.9999
y=0.0022x+4.3738× 10-5
0.0162
n.d.~0.133
0.09%
1.50%
107
Ag
0.9999
y=0.0060x+5.3088× 10-5
0.0161
n.d.~0.1363
2.35%
0.93%
108
Cd
1.0000
y=8.8154× 10-5x+1.0911× 10-4
0.5961
n.d.~0.1882
7.47%
2.79%
109
Ag
0.9999
y=0.0057x+4.4242× 10-5
0.0144
n.d.~0.1361
2.37%
0.64%
111
Cd
1.0000
y=0.0013x+6.7778× 10-6
0.0046
n.d.~0.1072
0.32%
0.72%
114
Cd
0.9999
y=0.0030x+2.2092× 10-5
0.0085
n.d.~0.1143
0.33%
0.80%
121
Sb
1.0000
y=0.0041x+1.6072× 10-4
0.0231
n.d.~0.0956
0.63%
1.60%
123
Sb
1.0000
y=0.0031x+1.2502× 10-4
0.0224
n.d.~0.0948
0.59%
1.55%
127
I
0.9998
y=0.0029x+0.0938
0.5978
n.d.~5.3848
12.13%
23.29%
137
Ba
0.9998
y=0.0017x+0.0023
0.1459
n.d.~1.0109
0.22%
0.45%
200
Hg
0.9997
y=9.69× 10-4x+6.545× 10-6
0.0052
0.0002~0.0576
4.41%
14.45%
201
Hg
0.9997
y=5.486× 10-4x+3.804× 10-6
0.0097
n.d.~0.0575
4.31%
14.56%
202
Hg
0.9996
y=1.23× 10-4x+7.808× 10-6
0.0044
n.d.~0.0574
4.13%
14.90%
205
Tl
0.9999
y=0.0101x+5.2807× 10-5
0.0093
n.d.~0.0007
0.37%
1.35%
206
Pb
0.9998
y=0.0034x+8.4652× 10-5
0.0076
0.0098~2.397
0.13%
0.66%
207
Pb
0.9998
y=0.0030x+7.3518× 10-5
0.0150
n.d.~1.8736
0.29%
0.54%
208
Pb
0.9999
y=0.0136x+8.1864× 10-4
0.0123
n.d.~2.259
0.38%
0.34%
238
U
0.9997
y=0.0128x+2.2448× 10-5
0.0010
n.d.~0.0135
0.58%
0.34%
注:n.d. 表示未检测到。
表1 决定系数, 标准曲线方程, 检出限, 海洛因检材浓度范围, 方法的日内精密度和日间精密度Table 1 The determination coefficients, standard curves, LODs, concentration ranges of elements in the tested heroin samples, and the intra- and inter-day precisions of the method adopted
通过测定7个相关和8个不相关的已知海洛因检材, 确定不同组合方式的最佳分析结果。数据预处理(表2)和距离/相关性(表3)的每个组合应用于所有检材。最后, 观察到总共21个相关的和28个不相关的距离/相关性数值。由此获得每种组合的平均值(µ )和标准偏差(σ )。通过下面的公式计算判别值Eq(可以对统计组合进行判别力评估, 值越大, 说明组合方式越好)[12]。Eq值从大到小的五种组合方式依次是S / Squ Dij, N / Euc Dij, N / Cheb Dij, S / Euc Dij, S / Cheb Dij, 这五种组合方式比其他组合方法更好地区分相关和不相关的样本。
Eq: Estimation of the discrimination =$\frac{μ(不相关)-σ(不相关)}{μ(相关)+σ(相关)}$
2.3.2 ROC曲线
对每个目标元素的浓度值进行统计预处理之后, 测定相关和不相关海洛因检材的距离/相关性值。计算每个统计组合的ROC曲线和AUC(ROC曲线下面积), AUC值越大, 越能证明数据统计组合方式能很好地分开相关样本和不相关样本。表4列出了5种组合N / Cheb Dij, N / Euc Dij, N + 2R / Euc Dij, S / Cheb Dij, S / Euc Dij(统计组合分析中5个最好的组合方式)的AUC值。根据AUC值, N/ Euc Dij是最优的组合方式。
2.3.3 判定阈值
表4展现了N / Cheb Dij, N / Euc Dij, N + 2R / Euc Dij, S / Cheb Dij, S / Euc Dij的五种统计组合相关检材和不相关检材的距离/相关性数值的范围。距离/相关性数值低于不相关样本的最小值被认为是相关, 高于相关样本的最大值被认为不相关。为保证测定结果准确性, 选择同时满足低于不相关样本的最小值和相关样本的最大值的距离值作为判定标准。
表4
Table 4
表4(Table 4)
表4 相关检材和不相关检材的不同结合方式的ROC曲线下面积及阈值Table 4 The areas and threshold values below the ROC curves about different statistical combinations from the relevant/irrelevant samples
统计组合方式
ROC曲线下面积
距离/相关性数值
相关检材
不相关检材
N+2R/ Euc Dij
1.1049
0.1181~0.8013
0.8089~4.1903
N/ Euc Dij
2.2806
0.0791~0.3997
0.4827~5.0594
N/ Cheb Dij
2.2056
0.0381~0.2611
0.2797~0.9980
S/ Euc Dij
1.9576
0.7618~5.0995
5.9646~24.1210
S/ Cheb Dij
1.7269
0.3770~2.4275
2.9362~5.4798
表4 相关检材和不相关检材的不同结合方式的ROC曲线下面积及阈值Table 4 The areas and threshold values below the ROC curves about different statistical combinations from the relevant/irrelevant samples
LIUC, HUAZ, BAIY, et al. Profiling and classification of illicit heroin by ICP-MS analysis of inorganic elements[J]. , 2014, 239: 37-43. [本文引用:3]
[2]
CASALE JF. CASALE ES, TOSKE SG, et al. Characterization and origin of the ‘B’ and ‘C’ compounds in the acid/neutral forensic signatures of heroin - products from the acylation of porphyroxine and subsequent hydrolysis[J]. , 2017, 9(3): 462-469. [本文引用:1]
[3]
CHANK, TANG, WONG RC. ICP-MS method validation for the analysis of trace elements in illicit heroin[J]. , 2012, 45(9): 1122-1132. [本文引用:1]
[4]
宋小娇, 郭洪玲, 梅宏成, 等. 电感耦合等离子体质谱检验弹头研究进展[J]. , 2015, 40(3): 173-178. (SONGXiaojiao, GUOHongling, MEIHongcheng, et al, Advances in the application of ICP-MS in bullet identification[J]. , 2015, 40(3): 173-178. )[本文引用:1]
[5]
张更宇, 吴超, 邓宇杰. 电感耦合等离子体质谱(ICP-MS)联用技术的应用及展望[J]. , 2016, 13: 19-26. (ZHANGGengyu, WUChao, DENGYujie. Application and outlook of the hyphenated ICP-MS technique[J]. , 2016, 13: 19-26. )[本文引用:1]
[6]
郭洪玲, 乔婷, 宋小娇, 等. 电感耦合等离子体质谱在微量物证检验中的应用(英文)[J]. , 2016, 41(6): 470-475. (GUOHongling, QIAOTing, SONGXiaojiao, et al. The application of inductively coupled plasma mass spectrometry in trace evidence examination[J]. , 2016, 41(6): 470-475. )[本文引用:1]
[7]
KOPERC, VAN D BC, WIARDAW, et al. Elemental analysis of 3, 4-methylenedioxymethamphetamine (MDMA): a tool to determine the synthesis method and trace links[J]. , 2007, 171(2-3): 171-179. [本文引用:1]
[8]
LIUC, HUAZ, XINM. Profiling of illicit cocaine seized in China by ICP-MS analysis of inorganic elements[J]. , 2017, 276: 77-84. [本文引用:1]
[9]
XUZ, DUP, LIK, et al. Tracing methamphetamine and amphetamine sources inwastewater and receiving waters via concentration and enantiomeric profiling[J]. , 2017, 601-602: 159-166. [本文引用:1]
[10]
CHANK, TANG, WONG R CS. Investigation of trace inorganic elements in street doses of heroin[J]. , 2013, 53(1): 73-80. [本文引用:2]
[11]
NICDAÉIDN, JAYARAMS, KERR WJ. Elemental profiling using ICPMS of methylamphetamine hydrochloride prepared from proprietary medication using the Moscow and hypophosphorous synthesis[J]. , 2013, 53(3): 278-285. [本文引用:1]
[12]
LOCICIROS, ESSEIVAP, HAYOZP, et al. Cocaine profiling for strategic intelligence, a cross-border project between France and Switzerland [J]. , 2008, 177(2-3): 199-206. [本文引用:1]