As previously discussed in this publication, the Wright R3350 radial aircraft piston engines used in the B29 bomber in WW II were subject to catastrophic in-flight failures. These were of the nature of valve breakage, erosion, or sticking, leading in turn to a more of less continuous blast of hot flame direct to the exhaust pipe, without the cooling effect of expansion in the cylinder. This led to burn-through of the pipe, followed by ignition of the magnesium rear accessory case, followed only slightly later by thermal collapse of the wing structure. It is not widely-discussed in the usual hurray-for-our-side publications, but more B29 aircrew were lost to such failures than to enemy action -- quite a scandal at the time.
Now, in reading a paper by (later) Lord Hives of Rolls-Royce, the mechanism of this failure is explained to me. It seems that the heavily-leaded fuels used at the time produced lead oxide, which is a sluggish liquid at high temperatures like those found on exhaust valves. The result is a coating like some fiendish wet paint, covering the exhaust valve, &c. The oxygen in this lead oxide would rather have iron as a partner, or aluminum, so the process proceeds by substitution, generating iron oxide in the case of valves with a lot of iron in them, or perhaps aluminum oxide in the case of the material of the exhaust port. As the tense lads drone on, over the Pacific, chemistry is at work, eating up the valves and/or eroding the material at the apex of the exhaust valve guide boss in the port itself.
Another paper, this one by P & W engineers, provides a measurement of normal cruising temperature at the tip of the exhaust guide boss in a well-designed, fully developed engine of theirs, the R4360; about 800 degrees F. If that temperature moved higher for any reason (such as poor fuel distribution within the complex pipework on these engines, or such as terrible heat overloads resulting from having to use take-off power far into the climb), it's certainly possible that material would begin to disappear at the hottest places. This would lead to valve and guide overheating, sticking, or guttering -- and the failure would compound from there..
Once the first evidence of this process was seen by crew members set to the task of looking for it, it's said there were 30 seconds to get out of the plane.