Abstract Over half of BRAF-mutant melanoma patients with initial response to targeted therapy will recur with resistant disease. It is thought that recurrence arises from the ability of resistant clonal populations to enter and exit a slow-cycling persister state, evading treatment in quiescence before returning to a proliferative state. However, clonal transcriptional profiles and their spatial relationships in the tumor environment are not well characterized. Through longitudinal profiling of tumors from pre-treatment through the growth of resistant tumors, we track clonal lineages and transcriptional state changes to identify recurrent evolutionary patterns. BRAF mutant patient-derived melanoma xenografts were treated with BRAF/MEK inhibitors through maximum tumor size reduction and re-growth. Spatial transcriptomics (ST) and H&E histological staining were performed at five timepoints across 94 and 133 days of PDX models WM4237 and WM4007 treatment, tracking clonal populations with intact tissue structure. Deep learning feature extraction on high-resolution H&E-stained images allowed the detection of histological features that co-localize with expression levels. A novel computational workflow performed integration of samples across time points, pathway enrichment analysis, and pseudotemporal ordering, allowing spatial recreation of copy number variation based clonal phylogenies and cell fate trajectory inference during the treatment course. We characterize the clonal populations by the burden of accumulated copy number variations and establish clonal phylogeny. With the added temporal resolution, ST profiling during melanoma treatment enables observation of a global shift where all lineages begin the transition into the persister transcriptional state. The drug-sensitive cells are eliminated due to treatment, and distinct transcriptional states arise due to metabolic plasticity regardless of the lineage of origin. The clones of the transient persister state constitute the minimal residual disease, where melanoma cells have an increase in invasive capacity and a slowing of the cell cycle. The ability to survive through the persister state and re-emerge into a proliferating clone where the cell cycle is upregulated varies by lineage. We observe that the transcriptional changes during emergence from dormancy show decreased invasiveness and continued reliance on oxidative respiration versus glycolysis. We compare the re-emerged resistant cell populations and pre-treatment samples to pinpoint the changes in the transcriptional state of the MAPK signaling pathway and find that sets of upregulated genes differ by the model, WM4237 and WM4007. We connect deep learning imaging features extracted from longitudinal histology profiling to ST profiles to associate histology data with the cell cycle activation, invasiveness, hypoxia state, and metabolic mode across the timepoints. We observe that the imaging features alone can be potentially used to track the evolution of the tumor cell population phenotypes during drug treatment. Citation Format: Sergii Domanskyi, Todd B. Sheridan, Brian J. Sanderson, SungHee Park, Jessica Kaster, Haiyin Li, Olga Anczukow, Meenhard Herlyn, Jeffrey H. Chuang, Jill C. Rubinstein. Longitudinal histology and spatial transcriptomics profiling in targeted treatment of melanoma patient-derived models interrogates persister and resistant cell populations [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Translating Cancer Evolution and Data Science: The Next Frontier; 2023 Dec 3-6; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_2):Abstract nr A010.