Osteosarcoma recurrence, proximal humerus — second case, second algorithmic story.

The findings, in plain language

Chemotherapy pressure changed the neoantigen landscape.

Compared with the original tumor (sample bank archive), the recurrent deposit showed a distinct neoantigen profile. Two of the top epitopes from the original ranking were no longer expressed at the RNA level in the recurrent sample — consistent with subclonal selection under carboplatin pressure.

The top new epitope, GLLLPTDEC, arose from a hotspot mutation in TP53 (R273H) and showed strong predicted binding to the dog's DLA-88 haplotype (IC50 38 nM by DarkScan’s canine-DLA scoring layer). Two adjacent neoantigens in the top five came from the same mutated allele — a clustered set that survives many chemotherapy regimens because TP53 loss is foundational to OSA biology.

The clinical implication we’re NOT making: that a vaccine targeting these epitopes would have prevented recurrence. We don’t have that evidence and won’t imply it. The clinical implication we ARE making: post-recurrence rebiopsy + re-ranking is technically inexpensive and surfaces information that frontline sequencing alone cannot.