Method Profile
Compare patient-derived colorectal tumoroids under perfusion with unconditioned versus IGFBP5-conditioned HUVEC networks.
Investigate how vascular phenotype affects antiangiogenic combination-drug response measurements.
Patient-derived colorectal tumoroids and HUVEC vascular networks in a PDMS chip; conditioned and unconditioned endothelial versions.
50 ng/mL IGFBP5 conditioning; vascular self-assembly, 1:1 mixed medium and perfusion; four-day drug exposure after establishment.
Model components
Patient-derived tumoroids, HUVECs, IGFBP5, fibrin matrix and central tumoroid compartment.
Measurements
Dextran permeability, microbead perfusion, AngioTool vessel metrics and ImageJ tumoroid diameter.
Conditioned and original chips are compared within each patient. Clinical resistance is the positive class for sensitivity, and thresholds were fitted within the 22-case cohort.
Compare FOLFOX/bevacizumab chip responses with RECIST-derived clinical response groups.
22 patient-derived models with clinical information; experimental assays performed in triplicate per patient.
Original versus conditioned-HUVEC chips; complete/partial response versus stable/progressive disease.
Optimized model sensitivity 81.82%, specificity 90.91%, accuracy 86.36%; original 54.55%, 90.91%, 72.73%, with resistance as the positive class.
A colorectal tumoroid chip pairs patient-derived cultures with perfusable endothelial networks. Conditioning endothelial cells with IGFBP5 produced more tumor-like vascular features and improved agreement with treatment outcomes in a 22-patient comparison.
