Stochastic Modeling of Temporal Enhanced Ultrasound: Impact of Temporal Properties on Prostate Cancer Characterization

Abstract

Objectives: Temporal Enhanced Ultrasound (TeUS) is a new ultrasound-based imaging technique that provides tissue-specific information. Recent studies have shown the potential of TeUS for improving tissue characterization in prostate cancer diagnosis. We study the temporal properties of TeUS – temporal order and length – and present a new framework to assess their impact on tissue information. Methods: We utilize a probabilistic modeling approach using Hidden Markov Models (HMMs) to capture the temporal signatures of malignant and benign tissues from TeUS signals of 9 patients. We model signals of benign and malignant tissues (284 and 286 signals, respectively) in their original temporal order as well as under order permutations. We then compare the resulting models using the Kullback-Liebler divergence and assess their performance differences in characterization. Moreover, we train HMMs using TeUS signals of different durations and compare their model performance when differentiating tissue types. Results: Our findings demonstrate that models of order-preserved signals perform statistically significantly better (85% accuracy) in tissue characterization compared to models of order-altered signals (62% accuracy). The performance degrades as more changes in signal-order are introduced. Additionally, models trained on shorter sequences perform as accurately as models of longer sequences. Conclusion: The work presented here strongly indicates that temporal order has substantial impact on TeUS performance, thus it plays a significant role in conveying tissue-specific information. Further-more, shorter TeUS signals can relay sufficient information to accurately distinguish between tissue types. Significance: Under-standing the impact of TeUS properties facilitates the process of its adopting in diagnostic procedures and provides insights on improving its acquisition.

Stochastic Sequential Modeling: Toward Improved Prostate Cancer Diagnosis Through Temporal-Ultrasound

“…Different imaging modalities have been employed for tissue characterization including ultrasound-based techniques, 2,12,14,[17][18][19][20] magnetic resonance sequences. 22,28 Despite the low resolution of ultrasound images, ultrasound-based techniques have the advantage of using a low-cost technology that is already integrated into standard diagnostic procedures.…”

“…Previous research on TeUS utilized frequency-domain analysis and classifiers such as support vector machines, deep belief networks and others. 2,12,17 In our recent studies, we have proposed to directly represent the temporality of TeUS and employ it to reach more accurate tissue-typing, [18][19][20] and showed preliminary results when applied to a small dataset.…”

Artificial Intelligence in Quantitative Ultrasound Imaging: A Survey

“…Temporal-Enhanced US Temporal-enhanced US extracts information from the temporal sequence of backscattered US RF data of the ROI. 62,63 In 2016, Azizi et al used a DBN to learn the high-level latent features of RF data and a SVM classifier to differentiate cancerous versus benign tissue. 64 Given that RF data are not available on all commercial US systems, this group then used transfer learning from RF to B-mode features for detecting prostate cancer.…”

“…Temporal‐enhanced US extracts information from the temporal sequence of backscattered US RF data of the ROI 62,63 . In 2016, Azizi et al used a DBN to learn the high‐level latent features of RF data and a SVM classifier to differentiate cancerous versus benign tissue 64 .…”

Artificial Intelligence in Quantitative Ultrasound Imaging: A Review

“…Temporal-enhanced ultrasound extracts information from the temporal sequence of backscattered US RF data of the region of interest (ROI) [64,65]. In 2016, Azizi et al used a deep belief network (DBN) to learn the high-level latent features and a SVM classifier to differentiate cancerous versus benign tissue [66].…”