I wonder what logic makes Lelit come to that conclusion.
That it needs better water, and perhaps the occasional removal and soak to descale, sure. But what do they expect a chemical backflush to do to a needle valve located upstream of the brew valve? Since the brew valve is closed tight long before the exhaust valve is open, the needle valve never gets a drop of chemical cleaning during any type of backflush activities. I'm confident a backflush would be as effective at cleaning the needle valve as it is cleaning the pump...
Backflushing with Cafiza or similar detergents will help remove coffee oils from the all the water flow areas upstream of the filter screen, to the cam and the exhaust valves. As Jim points out, the brew valve is open for the 10 second flush time, so some turbulent flow and mixing may (or may not) occur up inside the mushroom and back to the needle valve. Cycling the brew valve stem up and down 5 times (backflush cycles) in this pressurised fluid will probably result in some detergent solution being injected into the mushroom valve chamber and hence the needle valve.
None of this will descale any part of the E61 group or needle valve though, these detergent solutions don't affect mineral scaling.
I've posted here some photos of my Bianca valve assembly, after 1014 shot cycles. Mine is a medium hardness water area and I've used the Lelit water filters and changed them as the counter on the water tank calls up. They are chunky little water cartridges and I'm not so dismissive of them. I've dried and weighed them before and after their life cycle and they do accumulate a significant load of non solubles. As you can see the valve and surrounding parts are building a fine film of scale, but its not enough to affect valve flow rates yet (I measure them for set valve positions) which are still as original.
Answering (eventually) a question from Jake_G, the valve has 1mm of travel in 180 degrees of paddle movement, so quite a coarse pitch, but the valve is actually a taper in a jet, with a ground taper on the stem, with 0.1mm of reduction in diameter, over this travel (not super accurate measurement, but sufficient for user info).