With the aid of good biocompatibility and stability with hydroxyapatite (HAp) in protein separation and adsorption fields, we developed a novel extraction-isolation albumin analysis method by relying on the specific adsorption capacity of HAp, combining with surface-enhanced Raman spectroscopy (SERS) for prostate cancer screening. Two different nanostructures of HAp particles, including the HAp flower and HAp sphere, were synthesized with a hydrothermal method, and the targeted binding and extraction abilities of serum albumin of these two HAp particles were compared. By changing the morphology of the nanostructure, the albumin-adsorption capacity of HAp varied significantly. Compared with spherical HAp particles, HAp flower particles have more albumin binding sites per unit area. Thus, the HAp flower displayed the superior capacity for adsorption-release of albumin, which was further employed for clinical prostate cancer screening. Based on the superior adsorption-extraction ability of albumin of HAp flower, serum albumin was adsorbed and extracted by HAp flower from serum samples of prostate cancer patients (n = 30) and healthy volunteers (n = 30), and mixed with silver colloids to perform SERS spectral analysis. The partial least square-support vector machines (PLS-SVM) model is used to analyze the obtained serum albumin SERS spectra and establish the diagnostic model, the diagnostic accuracy was up to 95.00% for differentiating the normal volunteer from prostate patient groups. The results demonstrate that the PLS-SVM model provides superior performance in the classification of a prostate cancer diagnosis. Due to the advantages of simplicity and rapidness, the HAp flower-adsorbed-released albumin combined with SERS was expected to become a promising tool for prostate cancer detection.

借助与羟基磷灰石（HAp）在蛋白质分离和吸附领域良好的生物相容性和稳定性，我们利用HAp的比吸附能力，结合表面增强拉曼光谱（SERS），开发了一种新型的提取-分离白蛋白分析方法。用于前列腺癌筛查。采用水热法合成了两种不同纳米结构的HAp颗粒，包括HAp花和HAp球，并比较了这两种HAp颗粒对血清白蛋白的靶向结合和提取能力。通过改变纳米结构的形态，HAp的白蛋白吸附能力发生了显着变化。与球形HAp颗粒相比，HAp花颗粒每单位面积具有更多的白蛋白结合位点。因此，HAp 花表现出优异的白蛋白吸附释放能力，进一步用于临床前列腺癌筛查。基于HAp花白蛋白优越的吸附-提取能力，HAp花从前列腺癌患者（n = 30）和健康志愿者（n = 30）的血清样本中吸附提取血清白蛋白，并与银胶体混合进行 SERS 光谱分析。采用偏最小二乘支持向量机(PLS-SVM)模型对得到的血清白蛋白SERS谱进行分析，建立诊断模型，区分正常志愿者和前列腺患者组的诊断准确率高达95.00%。结果表明，PLS-SVM 模型在前列腺癌诊断的分类中提供了卓越的性能。