Photoacoustic Physio-Chemical Analysis For Prostate Cancer Diagnosis

Guan Xu,Qian Cheng,Shengsong Huang,Ming Qin,Thomas Hopkins,Chang H. Lee,Raoul Kopelman,Wan-Yu Chao,Evan T. Keller,Denglong Wu,Xueding Wang
DOI: https://doi.org/10.1117/12.2252262
2017-01-01
Abstract:Photoacoustic physio-chemical analysis (PAPCA) is a recently developed technology capable of simultaneously quantifying the content of molecular components and the corresponding microarchitectures in biological tissue. We have successfully quantified the diagnostic information in livers with PAPCA. In this study, we implemented PAPCA to the diagnosis of prostate cancers. 4 human prostates were scanned ex vivo. The PA signals from normal and cancerous regions in the prostates were acquired by an interstitial needle PA probe. A total of 14 interstitial measurements, including 6 within the normal regions and 8 in the cancerous regions, were acquired. The observed changes in molecular components, including lipid, collagen and hemoglobin were consistent with the findings by other research groups. The changes were quantified by PA spectral analysis (PASA) at wavelengths where strong optical absorption of the relevant molecular components was found. Statistically significant differences among the PASA parameters were observed (p=0.025 at significance of 0.05). A support vector machine model for differentiating the normal and cancerous tissue was established. With the limited number of samples, an 85% diagnostic accuracy was found. The diagnostic information in the PCPCA can be further enriched by targeted optical contrast agents visualizing the microarchitecture in PCa tissues. F3 PAA-PEG nanoparticles was employed to stain the PCa cells in a transgenic mouse model, in which the microarchitectures of normal and cancerous prostate tissues are comparable to that in human. Statistically significant differences were observed between the contrast-enhanced normal and cancerous regions (p=0.038 at a significance of 0.05).
What problem does this paper attempt to address?