growing colorectal cancer model [1, 2], whereas KPC tumors arise from genetically engineered mutations in Kras and Trp53, resulting in a dense, desmoplastic stroma which is characteristic of the human pancreatic cancer [3, 4]. A prerequisite for successful cancer treatment is effective delivery of therapeutic agents to tumor cells. However, a major limitation of chemotherapy is the low accumulation of injected drug [5]. To address this, multiple strategies have been proposed to enhance drug delivery. Among these, focused ultrasound combined with intravenously administered microbubbles has demonstrated improved delivery and effect of both small molecular drugs and nanoparticles in clinical [6-8] and preclinical [9-12] trials. In studies of CT26 and KPC tumors, ultrasound and microbubbles enhanced the delivery of polymeric nanoparticles in the soft and well-vascularized CT26 tumors, but not in the stiff, poorly vascularized KPC tumors [13]. Development of novel treatment strategies relies on the understanding of the tumor model at multiple scales, including macroscopic features such as extracellular matrix (ECM) composition, tumor stiffness, and perfusion, as well as molecular and protein-level characteristics. Proteomic profiling provides a powerful approach for characterizing such differences at the molecular level. In particular, liquid chromatography-mass spectrometry (LC-MS)-based label-free quantification (LFQ) proteomics enables high-throughput, unbiased measurement of thousands of proteins within complex biological samples [14, 15]. Through gene ontology (GO) enrichment analysis, this comprehensive measurement of proteins can offer valuable insight into biological processes, molecular functions and cellular components [16, 17]. Despite the widespread use of CT26 and KPC models, the characterization of their proteomic landscapes remains insufficient. In this study, we applied LC-MS-based LFQ proteomics to identify molecular, cellular, and biological features that distinguish CT26 and KPC tumors. In addition, we used the reference database SpLICe [18] and CIBERSORTx [19, 20] to create a signature matrix for murine immune cells, and immune cell deconvolution identified important immune cell populations. Reproducibility is critical in preclinical cancer research. It is known that several environmental factors, such as temperature, housing conditions, and handling influence the outcomes of animal experiments by affecting phenotypes, behavior and physiology [21-23]. In addition, measurement techniques such as mass spectrometry–based proteomics introduce further variability [24]. To our knowledge, variation in the proteomic landscape of tumors across independent experimental replicates has not been systematically reported. Therefore, we employed MS-based LFQ proteomics to characterize proteomic profiles across three independent experimental replicates using KPC and CT26 tumor models. This approach enabled us to assess differences in biological, molecular, and cellular features between tumors across independent experiments.