Three-dimensional oxygen maps of tumors in real time – analysis in the context of active tumor vasculature - replication data

The dataset was generated in a Mel270 uveal melanoma xenograft model in CB-17/Icr-Prkdc scid/R female mice (N = 12). For each animal, 1.5 × 10⁶ Mel270 cells in 50 μl of growth factor–reduced Geltrex™ were injected into the intrascapular fat pad. Tumors were allowed to grow under standardized housing conditions and imaged once they reached ~30 mm³. All procedures were performed under approved institutional animal protocols.
Each mouse contributed a multimodal set of quantitative tumor descriptors derived from:
3D B-mode ultrasound (Vevo F2, 57–25 MHz): sagittal and axial tumor scans with 0.10 mm step size were used to manually delineate the tumor border and reconstruct the volume. This yielded average tumor volume (mm³) based on combined axial and sagittal segmentations.
Power Doppler ultrasound: 3D Doppler acquisitions in both planes (velocity 1.5 kHz, low wall filter, priority 75%, medium persistence, 0.10 mm step size) were analyzed to compute average tumor vascularity, expressed as the percentage of tumor volume exhibiting Doppler signal.
Pulse EPR oxygen imaging (EPROI) with OX071: 4D pO₂ maps were reconstructed from IRESE measurements (JIVA-25 imager), after registration with 2D ultrasound slices in ArbuzGUI and transformation of the ultrasound-derived tumor mask into pO₂ space. From the tumor voxels, the following parameters were extracted: mean OX071 amplitude, median pO₂, HF20 (fraction of tumor volume with pO₂ < 20 mmHg), and VC40/VC60 (fractions of tumor volume with pO₂ > 40 or > 60 mmHg, respectively).
L-band CW EPR redox measurements: time-resolved spectra (~1 h acquisition, spectra every ~2 min) after intraperitoneal administration of 3-carboxyproxyl were fit with the Empirical Mathematical Model, yielding kinetic parameters that describe probe uptake and washout (e.g. α, inflow rate; β, overall decay rate), used as functional readouts of tumor vascular and redox status.
Histology-based metastasis assessment: H&E-stained lung sections were scanned and metastasis burden was scored semi-quantitatively as categorical bins representing the percentage of lung parenchyma occupied by tumor tissue, providing a dimensionless metastasis category per mouse.
For downstream analyses (correlation mapping, PCA, and MANOVA), the dataset was organized at the level of individual tumors/mice. Each row in the analytic table contained the EPROI-derived oxygenation metrics (median pO₂, HF20, VC40, mean amplitude), ultrasound-derived metrics (average tumor volume, average Doppler vascularity), selected redox kinetic parameters, and the histology-based metastasis score, enabling integrated assessment of 3D tumor oxygenation in the context of active vasculature and metastatic outcome.