Feed

Purpose To develop a photoacoustic microscopy (PAM) method for classifying colorectal liver metastasis (CRLM) microvascular networks as angiogenesis-type or vessel co-option-type at different metastatic stages and to evaluate their responses to bevacizumab. Materials and Methods Seventy-two BALB/c nude mice received intrasplenic injections of HCT116-luc or HT29-luc cells to establish CRLM models. PAM was used to visualize CRLM vascular architecture, and its diagnostic accuracy was validated through histopathologic analysis. Intergroup differences in PAM parameters were assessed with one-way analysis of variance or independent-samples t tests. Results PAM differentiated angiogenesis-type from vessel-co-option-type CRLMs according to distinct vascular architectures: Angiogenesis-type CRLMs exhibited lower vessel density (15.66% ± 4.38 vs 31.44% ± 6.95; P < .001) and vessel junction density (1.19 × 10^-3 n/px² ± 0.52 vs 3.12 × 10^-3 n/px² ± 1.11; P = .001) and higher vessel diameter (31.45 µm ± 4.59 vs 13.68 µm ± 1.61; P < .001) and lacunarity (0.294 arbitrary units [au] ± 0.095 vs 0.148 au ± 0.059; P = .003) than vessel co-option-type CRLMs. These differences persisted throughout the 8-week progression period. A 33.3% reduction in vessel density at 3 weeks following bevacizumab treatment was observed in angiogenesis-type CRLMs (untreated, 15.66% ± 4.38 vs treated at 3 weeks, 10.44% ± 4.91; P = .047), with no evidence of a difference in vessel co-option-type CRLMs (untreated, 31.44% ± 6.95 vs treated at 3 weeks, 28.82% ± 3.56; P = .56). Conclusion PAM quantified microvascular signatures differentiating CRLM blood supply patterns and identified vessel co-option-type metastases as resistant to bevacizumab. Keywords: Angiogenesis, Bevacizumab, Colorectal Cancer, Histopathologic Growth Pattern, Liver Metastases, Photoacoustic Imaging, Photoacoustic Microscopy, Vessel Co-Option Supplemental material is available for this article. © RSNA, 2026 See also editorial by Wang and Chen in this issue.