Question
Question Posted 10/28/26:
Dear ADNI MRI Core,
We are auditing the UCSF Cross-Sectional FreeSurfer 7.x table (UCSFFSX7) for a multimodal study.
The table-specific dictionary describes STATUS="partial" as not yet QC-reviewed or finalized. We understand that this differs from OVERALLQC="Partial", which is a quality rating.
Many records relevant to our predefined pre-diagnosis windows have STATUS="partial" and missing segmentation QC ratings. Could you please clarify:
1. Is there an updated UCSFFSX7 release or a separate table containing finalized segmentation QC for these records? Please provide its exact name and download location.
2. How should QC be linked to the exact derived measurements: IMAGEUID alone, or IMAGEUID together with FSVER, RUNDATE, or another processing identifier?
3. Which methods document and regional QC-to-feature mapping apply to FreeSurfer 7.x? For finalized records with OVERALLQC="Partial", how should usable and excluded measurements be identified?
4. If finalized QC is unavailable for STATUS="partial" records, is there a recommended procedure for reviewing their segmentation outputs, and are those outputs available through IDA?
We have read the January 2014 UCSF FreeSurfer Methods document, but would appreciate confirmation of which guidance applies to 7.x. We do not want to treat acquisition-level MRI QC or nonmissing numerical values as evidence that derived segmentation measurements have passed QC.
Thank you for your guidance.
Best regards,
Yunsong Long
Dear ADNI MRI Core,
We are auditing the UCSF Cross-Sectional FreeSurfer 7.x table (UCSFFSX7) for a multimodal study.
The table-specific dictionary describes STATUS="partial" as not yet QC-reviewed or finalized. We understand that this differs from OVERALLQC="Partial", which is a quality rating.
Many records relevant to our predefined pre-diagnosis windows have STATUS="partial" and missing segmentation QC ratings. Could you please clarify:
1. Is there an updated UCSFFSX7 release or a separate table containing finalized segmentation QC for these records? Please provide its exact name and download location.
2. How should QC be linked to the exact derived measurements: IMAGEUID alone, or IMAGEUID together with FSVER, RUNDATE, or another processing identifier?
3. Which methods document and regional QC-to-feature mapping apply to FreeSurfer 7.x? For finalized records with OVERALLQC="Partial", how should usable and excluded measurements be identified?
4. If finalized QC is unavailable for STATUS="partial" records, is there a recommended procedure for reviewing their segmentation outputs, and are those outputs available through IDA?
We have read the January 2014 UCSF FreeSurfer Methods document, but would appreciate confirmation of which guidance applies to 7.x. We do not want to treat acquisition-level MRI QC or nonmissing numerical values as evidence that derived segmentation measurements have passed QC.
Thank you for your guidance.
Best regards,
Yunsong Long
Response posted 10/28/26 by ADNI MRI Core:
Thank you for your question to the ADNI MRI Core. We have forwarded it to Dr. Tosun (Duygu.Tosun@ucsf.edu) at UCSF so that her team can assist.
ADNI MRI
ADNI MRI
