Bioanalytical Community

Would ISR acceptance criteria (i.e. >67% of samples within 20% for small molecule 30% for LBA) still be considered as an acceptable approach to evaluate cross validation data although it is not mentioned in ICH M10?

Thanks!

I have seen labs use the ISR criteria for incurred samples in cross-validation. If it's good enough to show that the results are comparable within a lab, applying the same criteria to results between labs seems appropriate as well.

That said, I prefer to also use the Bland-Altman approach described in Rocci et al (DOI: 10.1208/aapsj0903040). As an AAPS member, you can download AAPS Journal articles at no additional cost. I like this approach because it provides a confidence interval for the mean and tolerance intervals for the difference between individual samples. It also incorporates the outliers in a more statistically robust way than the 2/3 rule which does not take into account how far outside of acceptance criteria they are.

------------------------------
Joleen White Ph.D.
AAPS 2023 Global Health Community Past Chair
Bioanalytical 101 Course Development
Head of Bioassay Development
Gates Medical Research Institute
Cambridge MA
[email protected]

Disclaimer: Opinions expressed are solely my own and do not express the views or opinions of my employer.

Original Message:
Sent: 03-06-2023 08:27
From: Anonymous Member
Subject: Acceptance criteria for cross validation

This message was posted by a user wishing to remain anonymous

Would ISR acceptance criteria (i.e. >67% of samples within 20% for small molecule 30% for LBA) still be considered as an acceptable approach to evaluate cross validation data although it is not mentioned in ICH M10?

Thanks!

@Joleen White

Do we have consensus on the acceptance criteria for B & A plots? According Giavarina (2015), The B&A plot method only defines the intervals of agreements, it does not say whether those limits are acceptable or not. Acceptable limits must be defined a priori, based on clinical necessity, biological considerations or other goals. Do we propose that 2/3rd samples should be within the limit of agreement?

------------------------------
Ritankar Majumdar
Senior Lead Scientist
Covance Inc. - Chantilly, Va
Chantilly VA
[email protected]

Disclaimer: Opinions expressed are solely my own and do not express the views or opinions of my employer.

Original Message:
Sent: 03-06-2023 11:41
From: Joleen White
Subject: Acceptance criteria for cross validation

I have seen labs use the ISR criteria for incurred samples in cross-validation. If it's good enough to show that the results are comparable within a lab, applying the same criteria to results between labs seems appropriate as well.

That said, I prefer to also use the Bland-Altman approach described in Rocci et al (DOI: 10.1208/aapsj0903040). As an AAPS member, you can download AAPS Journal articles at no additional cost. I like this approach because it provides a confidence interval for the mean and tolerance intervals for the difference between individual samples. It also incorporates the outliers in a more statistically robust way than the 2/3 rule which does not take into account how far outside of acceptance criteria they are.

------------------------------
Joleen White Ph.D.
AAPS 2023 Global Health Community Past Chair
Bioanalytical 101 Course Development
Head of Bioassay Development
Gates Medical Research Institute
Cambridge MA
[email protected]

Disclaimer: Opinions expressed are solely my own and do not express the views or opinions of my employer.

Original Message:
Sent: 03-06-2023 08:27
From: Anonymous Member
Subject: Acceptance criteria for cross validation

This message was posted by a user wishing to remain anonymous

Would ISR acceptance criteria (i.e. >67% of samples within 20% for small molecule 30% for LBA) still be considered as an acceptable approach to evaluate cross validation data although it is not mentioned in ICH M10?

Thanks!

Would ISR acceptance criteria (i.e. >67% of samples within 20% for small molecule 30% for LBA) still be considered as an acceptable approach to evaluate cross validation data although it is not mentioned in ICH M10?

Thanks!