Plotting the way forward for {admiralneuro}

[manciniedoardo]

Please use this space to share any thoughts/comments/opinions regarding what we should focus on for our first {admiralneuro} release.

If you have experience in the alzheimer's TA, these leading questions could help guide your thought:

• Are there any specific ADaM datasets that your studies often provide, which are unique to this TA?
  • If so, what are they and why is it necessary to create them?
  • If not (or additionally) do you perhaps need to add any specific variables to your subject-level, BDS or occurrence ADaMs?
  • Similarly
• Is there any TA-specific metadata that drives ADaM creation? Scales, mappings, etc...?
  • If yes, how do you approach using it during your ADaM creation?

The tools we provide in {admiralneuro} will hopefully aim to address any answers to the questions above.

The {admiralneuro} package will also provide test data to play around with said tools and generate ADaMs pertinent to the TA. Is there any TA-specific test data that we should create, over and above what is in pharmaversesdtm? This could include completely new datasets or edited versions of existing ones (e.g. to introduce obese patients).

We can use your comments on the above to inform our way forward 😄

[jwang-lilly]

We frequently develop ADaM datasets ADPET and ADCOG (or separate datasets such as admmse, adcdr etc). These are specialized ADEFF datasets that adhere to the Basic Data Structure (BDS) format.

• ADPET is essential for capturing detailed quantitative or qualitative analyses of PET imaging, focusing specifically on amyloid and tau PET measurements. It tracks Alzheimer's disease pathology over time, including both global and regional values, to assess biological effects of intervention, if available. In observational studies, it allows for the evaluation of pathology accumulation over time.

• ADCOG is designed to store domain-specific or total scores that reflect cognitive or functional efficacy endpoints, either in response to intervention or neurodegeneration with aging.

• Other biomarkers such as CSF or plasma biomarkers, ptau181, ptau217, gfap, nfl etc could be stored in ADLAB

• Corresponding SDTM datasets are NV, FT and LB.

Conversion from SUVR to centiloid for amyloid pathology is often required, with conversion equations varying based on PET tracers (such as PIB, FBP, FBB, etc.) and the imaging processing pipeline. These conversion equations are typically sourced from reference papers.

• A function like compute_centiloid(suvr, tracer, pipeline, custom_equation) can be used for this conversion. We can include common tracers, pipelines and corresponding conversion equations as default.

Depending on the study's purpose, thresholds are applied to determine amyloid positivity, or clearance after treatment. There are well-accepted threshold such as 24.1 CL or study specific ones.

Conversion from cognitively unimpaired to MCI, or from MCI to AD, based on certain criteria, such as two consecutive worsened diagnosis etc. Functions to enable conversion rate calculation will be very helpful.

[manciniedoardo]

@cgpele @khatril how much of this is in common with Roche processes?

[cgpele]

Very similar to Roche, where we also develop separate coa ADaM datasets (adcdr, admmse etc) follwing BDS. Similar also for PET, where we might split even further to ADAPET and ADTPET (amyloid and tau respectively)

Whilst we have typically kept CSF and plasma biomarkers in adlb in the past, my tendency would be towards splitting these into separate datasets now

As well as the already stated ideas for functions (which I totally agree would be very useful), a couple of others to consider:

• Composite calculations such as iADRS or ADCS-ADL (Dependence score)
• Any/All other key endpoint total scores? (Although this is usually a simple sum so we may not need to specifically develop anything)
• Fluid BM ratios and log(10) values (similar to above, we probably don't need to develop something here but could add some documentation?)

[mjiang-lilly]

Agree, and ADARIA or some data structure for MRI findings could be helpful.

[Fanny-Gautier]

From my first research (I am new on Alzheimer TA) I found out that the following data are often collected for the

Cognitive and Functional Assessments:

• Mini-Mental State Examination (MMSE) score (ranging from 0 to 30, lower scores indicate cognitive impairment)
• Functional Assessment score (ranging from 0 to 10, lower scores indicate greater impairment)
• Memory Complaints (flag indicating the presence of memory complaints: 0 indicates No and 1 indicates Yes)
• Behavioral Problems (flag indicating the presence of behavioral problems: 0 indicates No and 1 indicates Yes)
• Activities of Daily Living (ADL) score (ranging from 0 to 10, lower scores indicate greater impairment)

Symptoms:
Flags indicating the presence of the symptoms: 0 indicates No and 1 indicates Yes

• Confusion
• Disorientation
• Personality changes
• Difficulty completing tasks
• Forgetfulness

[jwang-lilly]

ADNI (Alzheimer's disease imaging initiative) will release its first implementation of SDTM datasets in coming weeks.