How Does Time Machine Detect "Runtime Corruption" and How Do I Reset It?

So after months of working quite happily, Time Machine backups to a NAS destination have suddenly decided to stop working due to the annoyingly vague "runtime corruption" in the log.

However I have run every reasonable test I can think of and have found no sign of any corruption; I've run checks of the physical disks, the volumes, and the sparse bundle that Time Machine is using, and none report any errors. My Time Machine log contains no errors; a backup ran happily at 3pm, then at 4pm there is "runtime corruption", no specified cause, no further details. I have also used shasum to compare every file in the latest backup against the originals (where timestamps match) and found no mismatches for the more than 97% of the backup that could be compared.

What is truly ridiculous is that the only "fallback" offered is to delete the entire contents of the backup image (nearly four terabytes of backup plus history!). Why is there no option to just compare the contents and only recreate anything found to be wrong?

At this point I have no reason to believe that any corruption has actually occurred, which means I can only conclude that Time Machine's "runtime corruption" check itself is the cause of the problem and so I would like to somehow reset it so that I can resume backing up normally.

If possible I would like to also force a deep traversal to be safe, but I simply cannot justify the days and days required to recreate a new backup without any history; if that turns out to my only option I will have to abandon Time Machine altogether as I cannot have this happening again if there is no workaround.