Context: a customer in the united kingdom notified biomérieux of obtaining misidentification results in association with the vitek® ms instrument (ref.(b)(4), serial# (b)(6)) using knowledge base v3.2 when testing three (3) patient isolates.Vitek ms mode : ivd.Kb version : 3.2 (cli).Issue type : misidentification to brucella spp (3 samples).Vitek ms results : note : vitek pickme is not used on this site.Sample (b)(6).30 may : one no id result.31 may : one no id result and one single choice to brucella spp.Other methods : gram staining = gram positive cocci which is not in favor to brucella spp expected id: unknown because identification was not confirmed with a reference method (sequencing).There was no confirmatory method performed.Culture conditions: origin : tissue -culture media : coh plate.Incubation: 24 hours.Sample (b)(6).31 may: one no id result, one single choice to brucella spp and one single choice to staphylococcus aureus.Other methods : chromogenic media = positive on mrsa plate, it is potentially a staphylococcus aureus.Expected id: unknown because identification was not confirmed with a reference method (sequencing).There was no confirmatory method performed.Culture conditions: origin : genital vaginal.Culture media : cap plate.Incubation: 5 days.Sample (b)(6).30 may: one no id result and one single choice to brucella spp.31 may one no id result.Other methods : unknown.Expected id: unknown because identification was not confirmed with a reference method (sequencing).There was no confirmatory method performed.Culture conditions: origin : unknown -culture media : s.Aureus plate.Incubation: 24 hours issue date: 30 and 31 may 2022.Fine tuning date before the issue: 20 may 2022.Investigation: batch history record and complaint trend analysis** there is no capa, no non-conformity on vitek ms linked with customer 's complaint.A trend analysis has been done on the period from june 2021 to august 2021 regarding the complaints recorded for a misidentification ; no trend has been identified.Investigation results fine tuning according to the vilink alert tool criteria, no fine tuning was needed but is was at the limit during the tests made on 30 and 31 may 2022.Spot preparation quality: the sample ¿all peaks¿ values are quite heterogeneous.Based on this finding, the spot preparation quality needs to be verified with the customer.Kb review : the sample ¿all peaks¿ values are quite heterogeneous.Based on this finding, the spot preparation quality needs to be verified with the customer.Sample data analysis: sample (b)(6) the sample « all peaks » values are quite heterogeneous, it varies between 27 and 61.Sample spot preparation needs to be verified with the customer.Reprocessing the customer data with vitek ms kb v3.2 show heterogeneous results : one spectra led to single choice to brucella spp.Others gave no identification.The potential misidentification to brucella spp was obtained with a very low number of peaks 38.When the number of peaks is just over 30, the risk to get ¿doubtful¿ results is high due to a lack of information (missing peaks, not enough peaks to eliminate candidate identification).One no identification result was due to ¿not enough peaks¿, it could be due to a bad organism quantity deposit on the spots (too thick or too fine organism deposit).Consequently, the amount of ¿good peaks¿ detected during acquisition will be not sufficient to allow a good identification.Sample (b)(6) the sample « all peaks » values are quite heterogeneous, it varies between 31 and 290.Sample spot preparation needs to be verified with the customer.Reprocessing the customer data with vitek ms kb v3.2 show heterogeneous results : one spectra led to single choice to brucella spp.Others gave no identification.The potential misidentification to brucella spp was obtained with a very low number of peaks 31.When the number of peaks is just over 30, the risk to get ¿doubtful¿ results is high due to a lack of information (missing peaks, not enough peaks to eliminate candidate identification).One no identification results was due to ¿too many peaks¿, it could be due to a bad organism quantity deposit on the spots (too thick or too fine organism deposit).Consequently, the amount of ¿good peaks¿ detected during acquisition will be not sufficient to allow a good identification.Sample (b)(6).The sample « all peaks » values are quite heterogeneous, it varies between 71 and 99.Sample spot preparation needs to be verified with the customer.Reprocessing the customer data with vitek ms kb v3.2 show heterogeneous results : one spectra led to single choice to brucella spp one spectra led to single choice to staphylococcus aureus one spectra led to no identification the potential misidentification to brucella spp was obtained with a very low score.The score limit is at -0.4 and in this case the score was at -0.39, so just above the limit.In this condition, the risk to get ¿doubtful¿ results is high due to a lack of information (missing peaks, not enough peaks to eliminate candidate identification).Reprocessing the customer data with next vitek ms kb under development allows to eliminate all misidentification to brucella spp and to staphylococcus aureus.There is no misidentification anymore.For information, the cause of the misidentification issue as brucella spp is a kb weakness with a bad quality of spectra (linked to a non-optimal spot preparation or/and non-optimal fine tuning).Significant improvements (change request #2264) have already been made by r&d for next vitek ms kb version to limit misidentification to brucella spp.In addition, vitek® pickme could be used to optimize the quality of deposit.In addition, the expected species seems to be species out of the vitek ms knowledge base.The identifications should be confirmed by the customer with a reference method.This is a system limitation as explain below.For information, the following limitation is mentioned in the user manual supplements - 161150-924 - a - en - vitek ms clinical use - v3.2 knowledge base : interpretation of results and use of the vitek ms system require a competent laboratorian who should judiciously make use of experience, specimen information, and other pertinent procedures before reporting the identification of test organisms.Additional information known to the user, such as gram stain reaction, colonial and cellular morphology, and growth aerobically or in co2 should be considered when accepting vitek ms results.¿ testing of species not found in the database may result in an unidentified result or a misidentification.Brucella spp is an ¿highly pathogenic organism¿.Handle isolate with extreme caution and send it to a reference laboratory for further identification according to your laboratory¿s protocol and/or country regulations.Conclusion: the most probable root causes of this issue are: knowledge base weakness linked with the bad quality of spectra non optimal spot preparation -system limitation (species out of kb v3.2).According to data above, there is no reconsideration of vitek ms instrument (ref.410895, serial# (b)(6) performance.
|