Neuro, you indicated that analyzing completers is not statistically appropriate. It would be reasonable for a proof-of-concept (POC) study and treatment-validating results would be worth publishing.
One could think of an intent-to-treat (ITT) analysis as estimating how well a marketed treatment might work (on average, due to compliance, etc), which is preferable for a phase-3 confirmatory study. Similarly, a per-protocol (PP) analysis more accurately estimates the actual treatment effect and justifies putting resources into additional studies.
The PP analysis does not necessarily have to be so strict as to analyze completers only, but could be limited to subjects for whom the study protocol was reasonably followed. Procedural problems that substantially lowered the reliability of a given subject’s results could be justification for excluding such a subject from a PP analysis. Respectively, sotb.