LEK is sort of the case that brought me to Websleuths and without anything new in awhile I thought I'd post a few thoughts. I've tried to keep up with all the threads, but I may have missed something, so apologies for any unintentional rehashing. There will be some intentional rehashing.
I spent the 90's in Dallas and Austin and can't get over the sensation that LEK looks familiar to me, unsure why that is, I get that feeling sometimes without a good reason, maybe she just has a familiar face, but the picture of her in the sweater looks very familiar to me for some reason. Useless, I know. Maybe that's another reason I'm drawn to this case.
It seems entirely possible that anyone who knew her before she swapped identities could be dead, or not want that association to be known. I know someone mentioned her love of Cuban food might tie her to Miami/FL... I don't know how much Florida sleuthing has been done, but looking at a map of Cuban populations in the US, seems a high % in Southern California as well.
Mostly, I wonder about the DNA... I wonder what sort of baseline sequencing they do and maybe some experienced people around here can help. I'm sure they're doing STRs (short tandem repeats) and the like to do direct comparisons in matching people with unknowns. I wonder in LEK because they mention this cardiac disease related gene. I wonder if it is possible to contact a company like 23andme, and if they've ever helped in these "unidentified" cases. 23andme does a whole genome single nucleotide polymorphism (SNP) array of over 500k distinct SNPs. They then have a "relative finder" service that will compare your profile to other customers. I've found verifiable 3rd cousins by doing this, it could possibly provide some surnames or other hints as to her true identity. It will also give you maternal haplogroup (probably known from STRs) and some data on ethnic heritage which might be useful. As more and more people have their genomes sequenced, there is some evidence that "surname inference" may be very easily achieved from whole genome sequencing data.
Also certain communities that are big into geneology have been extensively sequenced, particularly in Utah. Given some of the theories related to different groups she might have come from, in this case it might be particularly useful. I'd be interested in any Marfan's related screening as well.
Also, related to implant serial numbers, I know someone asked if they would've been checked at autopsy. I doubt it. I don't have any forensic autopsy experience, but I have attended some medical autopsies and in general, the extremities are given a surface examination but not dissected without a very specific reason. Now when implants are removed, any surgical hardware, it is often sent to histology for a gross identification and logging of serial numbers, but not at autopsy. Only hardware I remember identifying in an autopsy were pacemakers.
Just a few random thoughts. A very haunting mystery, this case is... so much information yet we still don't know the story.