Optical imaging techniques for biofilm observation, like laser scanning microscopy, are not applicable when investigating biofilm formation in opaque porous media. X‐ray micro‐tomography (X‐ray CMT) might be an alternative but it finds limitations in similarity of X‐ray absorption coefficients for the biofilm and aqueous phases. To overcome this difficulty, barium sulphate was used in Davit et al. (2011) to enable high‐resolution 3D imaging of biofilm via X‐ray CMT. However, this approach lacks comparison with well‐established imaging methods, which are known to capture the fine structures of biofilms, as well as uncertainty quantification. Here, we compare two‐photon laser scanning microscopy (TPLSM) images of Pseudomonas Aeruginosa biofilm grown in glass capillaries against X‐ray CMT using an improved protocol where barium sulphate is combined with low‐gelling temperature agarose to avoid sedimentation. Calibrated phantoms consisting of mono‐dispersed fluorescent and X‐ray absorbent beads were used to evaluate the uncertainty associated with our protocol along with three different segmentation techniques, namely hysteresis, watershed and region growing, to determine the bias relative to image binarization. Metrics such as volume, 3D surface area and thickness were measured and comparison of both imaging modalities shows that X‐ray CMT of biofilm using our protocol yields an accuracy that is comparable and even better in certain respects than TPLSM, even in a nonporous system that is largely favourable to TPLSM.

中文翻译：

---

通过 X 射线显微断层扫描和双光子激光扫描显微镜成像的生物膜的定量 3D 比较

用于生物膜观察的光学成像技术，如激光扫描显微镜，在研究不透明多孔介质中的生物膜形成时不适用。X 射线显微断层扫描（X 射线 CMT）可能是一种替代方法，但它发现生物膜和水相的 X 射线吸收系数的相似性存在局限性。为了克服这个困难，Davit 等人使用了硫酸钡。(2011) 通过 X 射线 CMT 实现生物膜的高分辨率 3D 成像。然而，这种方法缺乏与众所周知的捕获生物膜精细结构以及不确定性量化的成熟成像方法的比较。这里，我们使用改进的方案比较了玻璃毛细管中生长的铜绿假单胞菌生物膜的双光子激光扫描显微镜 (TPLSM) 图像与 X 射线 CMT，其中硫酸钡与低胶凝温度琼脂糖结合以避免沉淀。由单分散荧光和 X 射线吸收珠组成的校准模型用于评估与我们的协议相关的不确定性以及三种不同的分割技术，即滞后、分水岭和区域增长，以确定相对于图像二值化的偏差。测量了体积、3D 表面积和厚度等指标，两种成像方式的比较表明，使用我们的协议的生物膜 X 射线 CMT 产生的准确性与 TPLSM 相当，甚至在某些方面甚至更好，