Uploaded August 2025 | Updated September 2026, 19 minutes ago
NOTE: this Q&A session followed Dave's presentation on preliminary data for Keto-CTA following all four analyses now in hand -- you can watch it here: youtube.com/watch?v=aJobOSRdIOM
In this Q&A session, Dave Feldman addresses audience questions about the Keto-CTA study, including participant adherence, cardiovascular risk in Lean Mass Hyper-Responders (LMHRs), imaging variability, plaque regression, and quality control protocols. He discusses insights from multiple imaging analyses, emphasizes the importance of personalized risk assessment, and explains why rigorous blinding and longer-term follow-ups are essential for interpreting results.
0:01 — Did the Citizen Science Foundation pay Cleerly for their analysis?
An audience member asked whether CSF funded Cleerly’s imaging work. Dave clarified that the foundation funded the study itself through Lundquist and offered to fully pay for a blinded quality control (QC) review. Payment was not a limiting factor.
0:40 — How was dietary adherence ensured?
Dave explained that Keto-Mojo provided glucose and ketone meters for all 100 participants, enabling continuous adherence tracking throughout the year.
1:30 — Does the study prove LMHRs are high- or low-risk?
An audience member asked whether the findings show LMHRs are inherently high-risk. Dave emphasized that neither extreme applies: as a group, LMHRs are not defined as high-risk, though some participants showed progression. Imaging highlights significant individual variability.
3:36 — The role of imaging in risk assessment
Dave stressed that imaging, particularly CAC scoring, remains a powerful risk-prediction tool. However, plaque analysis technologies are still developing and can produce variability across repeated scans.
5:20 — Why do some develop atherosclerosis while others don’t?
From the waterfall plots, the data showed both regression and progression but no clear association with LDL-C or ApoB. Dave agreed this underscores that, at a population level, we still cannot reliably predict who develops disease — supporting the value of personalized, imaging-driven care.
6:07 — Where does functional testing fit in?
One question raised concerns about overtreatment based solely on imaging. Dave acknowledged the risk but stressed that well-conducted randomized controlled trials remain critical for guiding care.
7:03 — Lowering LDL/ApoB despite no observed association
An audience member asked whether lowering LDL or ApoB could still benefit those with higher plaque burden despite no association in this study. Dave said it’s plausible but unproven, requiring a separate, targeted study.
10:01 — Variability in plaque regression signals
Dave cautioned that small regression changes may fall within imaging noise and highlighted the importance of a third scan at 4–5 years to better confirm true trajectories.
12:07 — How reliable is a CAC score of zero?
While a CAC score of zero is highly prognostic, Dave noted it’s not absolute — rare cases involve non-calcified plaque or occlusions.
13:02 — Was exercise frequency tracked?
Dave confirmed exercise patterns weren’t tracked but expressed concern that over-exercisers within the LMHR group might impact vascular health if recovery isn’t prioritized.
14:18 — Vitamin K2 and calcification changes
Asked about calcification shifts, Dave confirmed changes occurred in both directions but highlighted discrepancies in the Cleerly dataset, favoring corroborated results from other analyses.
15:38 — Understanding what drives plaque regression
Dave confirmed that sub-analyses are planned but emphasized that three-scan trajectories will be needed to distinguish true biological change from measurement variability.
17:11 — Imaging variability and AI vs. human reads
Dave explained that contrast timing, scanner settings, and algorithms all introduce variability, making rigorous blinding and standardized protocols essential.
19:44 — Quality control and double-blinding protocols
Dave revealed that metadata errors compromised full blinding in initial Cleerly reads, which is why a fully blinded QC pass remains a priority.
22:03 — Are Cleerly’s imaging methods reliable?
Asked if Cleerly’s technology is flawed, Dave clarified he doesn't assume there's inherent issues with the technology. The discrepancies may simply relate to this specific dataset, underscoring the need for independent validation.
NOTE: this Q&A session followed Dave's presentation on preliminary data for Keto-CTA following all four analyses now in hand -- you can watch it here: youtube.com/watch?v=aJobOSRdIOM
In this Q&A session, Dave Feldman addresses audience questions about the Keto-CTA study, including participant adherence, cardiovascular risk in Lean Mass Hyper-Responders (LMHRs), imaging variability, plaque regression, and quality control protocols. He discusses insights from multiple imaging analyses, emphasizes the importance of personalized risk assessment, and explains why rigorous blinding and longer-term follow-ups are essential for interpreting results.
0:01 — Did the Citizen Science Foundation pay Cleerly for their analysis?
An audience member asked whether CSF funded Cleerly’s imaging work. Dave clarified that the foundation funded the study itself through Lundquist and offered to fully pay for a blinded quality control (QC) review. Payment was not a limiting factor.
0:40 — How was dietary adherence ensured?
Dave explained that Keto-Mojo provided glucose and ketone meters for all 100 participants, enabling continuous adherence tracking throughout the year.
1:30 — Does the study prove LMHRs are high- or low-risk?
An audience member asked whether the findings show LMHRs are inherently high-risk. Dave emphasized that neither extreme applies: as a group, LMHRs are not defined as high-risk, though some participants showed progression. Imaging highlights significant individual variability.
3:36 — The role of imaging in risk assessment
Dave stressed that imaging, particularly CAC scoring, remains a powerful risk-prediction tool. However, plaque analysis technologies are still developing and can produce variability across repeated scans.
5:20 — Why do some develop atherosclerosis while others don’t?
From the waterfall plots, the data showed both regression and progression but no clear association with LDL-C or ApoB. Dave agreed this underscores that, at a population level, we still cannot reliably predict who develops disease — supporting the value of personalized, imaging-driven care.
6:07 — Where does functional testing fit in?
One question raised concerns about overtreatment based solely on imaging. Dave acknowledged the risk but stressed that well-conducted randomized controlled trials remain critical for guiding care.
7:03 — Lowering LDL/ApoB despite no observed association
An audience member asked whether lowering LDL or ApoB could still benefit those with higher plaque burden despite no association in this study. Dave said it’s plausible but unproven, requiring a separate, targeted study.
10:01 — Variability in plaque regression signals
Dave cautioned that small regression changes may fall within imaging noise and highlighted the importance of a third scan at 4–5 years to better confirm true trajectories.
12:07 — How reliable is a CAC score of zero?
While a CAC score of zero is highly prognostic, Dave noted it’s not absolute — rare cases involve non-calcified plaque or occlusions.
13:02 — Was exercise frequency tracked?
Dave confirmed exercise patterns weren’t tracked but expressed concern that over-exercisers within the LMHR group might impact vascular health if recovery isn’t prioritized.
14:18 — Vitamin K2 and calcification changes
Asked about calcification shifts, Dave confirmed changes occurred in both directions but highlighted discrepancies in the Cleerly dataset, favoring corroborated results from other analyses.
15:38 — Understanding what drives plaque regression
Dave confirmed that sub-analyses are planned but emphasized that three-scan trajectories will be needed to distinguish true biological change from measurement variability.
17:11 — Imaging variability and AI vs. human reads
Dave explained that contrast timing, scanner settings, and algorithms all introduce variability, making rigorous blinding and standardized protocols essential.
19:44 — Quality control and double-blinding protocols
Dave revealed that metadata errors compromised full blinding in initial Cleerly reads, which is why a fully blinded QC pass remains a priority.
22:03 — Are Cleerly’s imaging methods reliable?
Asked if Cleerly’s technology is flawed, Dave clarified he doesn't assume there's inherent issues with the technology. The discrepancies may simply relate to this specific dataset, underscoring the need for independent validation.
