"Diet doesn't affect the underlying pathology" in Crohn's Disease is definitely a simplification, but I believe it's a healthy and important one, as the harm caused by an elimination diet (or the cost of any given special diet) can easily outweigh its benefits. It's not in the human nature to accept that certain things are outside of our control, and IBD is definitely a case where people tend to delude themselves into thinking that some dietary intervention worked when it did not.
In principle, we understand that an important part of the aetiology of CD has to do with interactions of the gut microbiome with a dysregulated immune system and a perturbation of the intestinal barrier. We know that the gut microbiome is somewhat plastic to dietary changes, though research findings regularly frustrate attempts at intentional modification.
It would be more correct to say that there is no dietary change which is known to causally and intentionally alter the course of the underlying disease.
That is not the same as saying that no relationships or correlations have ever been observed, or that dietary changes do not provide some symptomatic relief: on the contrary, we know that people who have a certain diet have a lower prevalence, or better subjective or even objective outcomes. Unfortunately, most attempts at taking patients from the worse group and changing their diet to match that of the better-off group fall flat on their face. This is a few years old and is not specific to CD, but it is a great example of the type of results we see over and over again in dietary intervention studies of this nature: https://www.cell.com/fulltext/S0092-8674(15)01111-3
Obviously this is an area of research of great interest, and if we could figure out a way to improve outcomes by making specific and robust dietary change recommendations, we would. It really is not that simple, unfortunately.
I don't dispute the results of the studies you link, and I find the Cochrane meta analysis particularly notable, but I would caution that interpreting the results doesn't seem to me to be as straightforward as it might seem.
I wonder whether controls are adequate (there is no placebo group here, for ethical reasons), I wonder whether the results can speak of longer-term outcomes, whether there is a selection bias (would you do this study with patients who have severe or advanced disease, or mostly with people who have milder clinical presentations?), whether the results might indicate to some degree the natural progression of largely self-limiting flareups, whether it's appropriate to compare induction of remission using corticosteroids with diet interventions (which I'd expect would have more of an effect on maintenance of remission, if anything)... I just don't think it's that simple. Remember that people with untreated CD will have flare-ups that occur and resolve spontaneously for some time before the disease starts progressing and surgery is required.
That being said, non-elimination, generally-healthy diets, on the other hand, are always recommended, as is regular physical exercise.