PERFORMANCE EVALUATION OF Q-FLAG OF HEMATOLOGY ANALYZER, SYSMEX XN-20
(Abstract release date: 05/19/16)
EHA Library. Kim H. 06/09/16; 134753; PB1853
Prof. Dr. Hye Ryoun Kim
Contributions
Contributions
Abstract
Abstract: PB1853
Type: Publication Only
Background
The Sysmex XN modular analyzer (Sysmex, Kobe, Japan), which were introduced in 2011, applied several new analysis channels compared to previous analyzer, XE series. White cell precursor channel (WPC), one of the new channels, adopted fluorescence flow cytometry technology and developed especially for detecting myeloblasts or lymphoblasts more accurately than any other channels applied in previous modules.
Aims
Our study aim to investigate the flagging performance of Sysmex XN-20 modular analyzer focusing on the Q values.
Methods
We investigated the performance of Q values flagging according to interinstrumental agreement, intrainstrumental precision, and diagnostic accuracy. Tested suspect flags are “blasts/abnormal lymphocytes”, “atypical lymphocytes” from white cell differential channel, and “blasts” from white cell precursor channel.
Results
In interinstrumental agreement evaluation, two XN modules showed significant Q value correlation for each suspect flags although Pearson’s correlation coefficient for “blasts/abnormal lymphocytes” and “blasts” flags were too low; they were 0.28 and 0.29. However, from Pearson’s chi-square tests, there were significant difference in flagging performance between two analyzers for all of each flags (P<0.01 for all of tested flags). Absolute agreement rates were ranged from 65.0% to 83.5%, and kappa values were ranged from 0.305 to 0.501. In intrainstrumental precision evaluation, for the specimens with Q value of 50-150, standard deviations were ranged from 4.8 to 23.9 for “blasts/abnormal lymphocytes” flag, from 18.7 to 59.1 for “blasts” flag, and from 11.0 to 23.0 for “atypical lymphocytes” flag during 10 replicates. For diagnostic accuracy, in the default Q value cutoff setting (Q value of 100), sensitivities and specificities for “blasts/abnormal lymphocytes” flag were 90.1% and 73.8%, the values for “blasts” flag were 92.1% and 86.3%, and the values for “atypical lymphocytes” flag were 70.0% and 89.9%, respectively.
Conclusion
We recommend to adjust the Q value threshold setting for “blast/abnormal lymphocyte”, “blast” and “atypical lymphocytes” because of their inadequate diagnostic accuracy of default threshold setting. However, their poor intrainstrumental precision and interinstrumental agreement weaken the reliability of Q value flags.
Session topic: E-poster
Keyword(s): Automation, Diagnosis, Leukemia cells, Quality control
Type: Publication Only
Background
The Sysmex XN modular analyzer (Sysmex, Kobe, Japan), which were introduced in 2011, applied several new analysis channels compared to previous analyzer, XE series. White cell precursor channel (WPC), one of the new channels, adopted fluorescence flow cytometry technology and developed especially for detecting myeloblasts or lymphoblasts more accurately than any other channels applied in previous modules.
Aims
Our study aim to investigate the flagging performance of Sysmex XN-20 modular analyzer focusing on the Q values.
Methods
We investigated the performance of Q values flagging according to interinstrumental agreement, intrainstrumental precision, and diagnostic accuracy. Tested suspect flags are “blasts/abnormal lymphocytes”, “atypical lymphocytes” from white cell differential channel, and “blasts” from white cell precursor channel.
Results
In interinstrumental agreement evaluation, two XN modules showed significant Q value correlation for each suspect flags although Pearson’s correlation coefficient for “blasts/abnormal lymphocytes” and “blasts” flags were too low; they were 0.28 and 0.29. However, from Pearson’s chi-square tests, there were significant difference in flagging performance between two analyzers for all of each flags (P<0.01 for all of tested flags). Absolute agreement rates were ranged from 65.0% to 83.5%, and kappa values were ranged from 0.305 to 0.501. In intrainstrumental precision evaluation, for the specimens with Q value of 50-150, standard deviations were ranged from 4.8 to 23.9 for “blasts/abnormal lymphocytes” flag, from 18.7 to 59.1 for “blasts” flag, and from 11.0 to 23.0 for “atypical lymphocytes” flag during 10 replicates. For diagnostic accuracy, in the default Q value cutoff setting (Q value of 100), sensitivities and specificities for “blasts/abnormal lymphocytes” flag were 90.1% and 73.8%, the values for “blasts” flag were 92.1% and 86.3%, and the values for “atypical lymphocytes” flag were 70.0% and 89.9%, respectively.
Conclusion
We recommend to adjust the Q value threshold setting for “blast/abnormal lymphocyte”, “blast” and “atypical lymphocytes” because of their inadequate diagnostic accuracy of default threshold setting. However, their poor intrainstrumental precision and interinstrumental agreement weaken the reliability of Q value flags.
Session topic: E-poster
Keyword(s): Automation, Diagnosis, Leukemia cells, Quality control
Abstract: PB1853
Type: Publication Only
Background
The Sysmex XN modular analyzer (Sysmex, Kobe, Japan), which were introduced in 2011, applied several new analysis channels compared to previous analyzer, XE series. White cell precursor channel (WPC), one of the new channels, adopted fluorescence flow cytometry technology and developed especially for detecting myeloblasts or lymphoblasts more accurately than any other channels applied in previous modules.
Aims
Our study aim to investigate the flagging performance of Sysmex XN-20 modular analyzer focusing on the Q values.
Methods
We investigated the performance of Q values flagging according to interinstrumental agreement, intrainstrumental precision, and diagnostic accuracy. Tested suspect flags are “blasts/abnormal lymphocytes”, “atypical lymphocytes” from white cell differential channel, and “blasts” from white cell precursor channel.
Results
In interinstrumental agreement evaluation, two XN modules showed significant Q value correlation for each suspect flags although Pearson’s correlation coefficient for “blasts/abnormal lymphocytes” and “blasts” flags were too low; they were 0.28 and 0.29. However, from Pearson’s chi-square tests, there were significant difference in flagging performance between two analyzers for all of each flags (P<0.01 for all of tested flags). Absolute agreement rates were ranged from 65.0% to 83.5%, and kappa values were ranged from 0.305 to 0.501. In intrainstrumental precision evaluation, for the specimens with Q value of 50-150, standard deviations were ranged from 4.8 to 23.9 for “blasts/abnormal lymphocytes” flag, from 18.7 to 59.1 for “blasts” flag, and from 11.0 to 23.0 for “atypical lymphocytes” flag during 10 replicates. For diagnostic accuracy, in the default Q value cutoff setting (Q value of 100), sensitivities and specificities for “blasts/abnormal lymphocytes” flag were 90.1% and 73.8%, the values for “blasts” flag were 92.1% and 86.3%, and the values for “atypical lymphocytes” flag were 70.0% and 89.9%, respectively.
Conclusion
We recommend to adjust the Q value threshold setting for “blast/abnormal lymphocyte”, “blast” and “atypical lymphocytes” because of their inadequate diagnostic accuracy of default threshold setting. However, their poor intrainstrumental precision and interinstrumental agreement weaken the reliability of Q value flags.
Session topic: E-poster
Keyword(s): Automation, Diagnosis, Leukemia cells, Quality control
Type: Publication Only
Background
The Sysmex XN modular analyzer (Sysmex, Kobe, Japan), which were introduced in 2011, applied several new analysis channels compared to previous analyzer, XE series. White cell precursor channel (WPC), one of the new channels, adopted fluorescence flow cytometry technology and developed especially for detecting myeloblasts or lymphoblasts more accurately than any other channels applied in previous modules.
Aims
Our study aim to investigate the flagging performance of Sysmex XN-20 modular analyzer focusing on the Q values.
Methods
We investigated the performance of Q values flagging according to interinstrumental agreement, intrainstrumental precision, and diagnostic accuracy. Tested suspect flags are “blasts/abnormal lymphocytes”, “atypical lymphocytes” from white cell differential channel, and “blasts” from white cell precursor channel.
Results
In interinstrumental agreement evaluation, two XN modules showed significant Q value correlation for each suspect flags although Pearson’s correlation coefficient for “blasts/abnormal lymphocytes” and “blasts” flags were too low; they were 0.28 and 0.29. However, from Pearson’s chi-square tests, there were significant difference in flagging performance between two analyzers for all of each flags (P<0.01 for all of tested flags). Absolute agreement rates were ranged from 65.0% to 83.5%, and kappa values were ranged from 0.305 to 0.501. In intrainstrumental precision evaluation, for the specimens with Q value of 50-150, standard deviations were ranged from 4.8 to 23.9 for “blasts/abnormal lymphocytes” flag, from 18.7 to 59.1 for “blasts” flag, and from 11.0 to 23.0 for “atypical lymphocytes” flag during 10 replicates. For diagnostic accuracy, in the default Q value cutoff setting (Q value of 100), sensitivities and specificities for “blasts/abnormal lymphocytes” flag were 90.1% and 73.8%, the values for “blasts” flag were 92.1% and 86.3%, and the values for “atypical lymphocytes” flag were 70.0% and 89.9%, respectively.
Conclusion
We recommend to adjust the Q value threshold setting for “blast/abnormal lymphocyte”, “blast” and “atypical lymphocytes” because of their inadequate diagnostic accuracy of default threshold setting. However, their poor intrainstrumental precision and interinstrumental agreement weaken the reliability of Q value flags.
Session topic: E-poster
Keyword(s): Automation, Diagnosis, Leukemia cells, Quality control
{{ help_message }}
{{filter}}