Synthetic correlated diffusion imaging hyperintensity delineates clinically significant prostate cancer

Alexander Wong, Hayden Gunraj, Vignesh Sivan, Masoom A. Haider

Abstract

Prostate cancer (PCa) is the second most common cancer in men worldwide and the most frequently diagnosed cancer among men in more developed countries. The prognosis of PCa is excellent if detected at an early stage, making early screening crucial for detection and treatment. In recent years, a new form of diffusion magnetic resonance imaging called correlated diffusion imaging (CDI) was introduced, and preliminary results show promise as a screening tool for PCa. In the largest study of its kind, we investigate the relationship between PCa presence and a new variant of CDI we term synthetic correlated diffusion imaging ($CDI^s$), as well as its performance for PCa delineation compared to current standard MRI techniques [T2-weighted (T2w) imaging, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) imaging] across a cohort of 200 patient cases. Statistical analyses reveal that hyperintensity in $CDI^s$ is a strong indicator of PCa presence and achieves strong delineation of clinically significant cancerous tissue compared to T2w, DWI, and DCE. These results suggest that $CDI^s$ hyperintensity may be a powerful biomarker for the presence of PCa, and may have a clinical impact as a diagnostic aid for improving PCa screening.

Introduction

Prostate cancer (PCa) screening currently relies on a combination of Prostate-Specific Antigen (PSA) blood tests and multi-parametric MRI (mpMRI). However, both modalities present significant challenges. PSA tests frequently lead to over-diagnosis and unnecessary invasive biopsies. Standard MRI techniques—such as T2-weighted (T2w) imaging and Diffusion-Weighted Imaging (DWI)—often yield ambiguous results; for instance, inflammation or benign prostatic hyperplasia (BPH) can mimic the appearance of cancer on T2w scans.

While Correlated Diffusion Imaging (CDI) previously showed promise by analyzing water molecule diffusion across varying magnetic strengths, it was hindered by low signal-to-noise ratios (SNR) and prohibitive scan times. This study introduces Synthetic Correlated Diffusion Imaging ($CDI^s$), a novel MRI variant designed to enhance the "diffusion fingerprint" of cancerous tissue, making it more robust, consistent across different scanners, and accurate in distinguishing significant tumors from healthy tissue.

Methodology

The innovation of $CDI^s$ lies in its synthetic approach to signal acquisition and processing. Rather than attempting to capture every magnetic gradient—which is time-consuming and noisy—the system utilizes a few standard native DWI scans. A mathematical model (least-squares ADC estimate) then "synthesizes" tissue appearance at extremely high magnetic strengths (b-values up to 7000 $s/mm^2$), which are typically impossible to image clearly with standard hardware.

These native and synthetic signals are blended using a calibrated function that includes a factor ($Z$) to ensure cross-scanner consistency (e.g., between Siemens Skyra and Trio machines). The researchers further optimized a "tuned" version ($CDI^s_t$) to maximize the contrast between cancerous and healthy tissue. The study was validated using a cohort of 200 patient cases from the PROSTATEx repository, utilizing histopathology as the ground truth for clinically significant PCa (Gleason Score $\ge$ 7).

Results and Discussion

A primary advantage of $CDI^s$ is its intuitive visual representation. In traditional Apparent Diffusion Coefficient (ADC) maps, cancer typically appears as a dark "hypointense" spot, which can be difficult to differentiate from shadows or dense healthy tissue. Conversely, $CDI^s$ represents cancer as a bright "hyperintense" region.

Comparison of normalized frequency distributions for $CDI^s$ SI, ADC, T2w SI, and $K^{trans}$ across healthy, insPCa, and csPCa tissues.

Histogram analysis reveals that while signal distributions for healthy and cancerous tissue overlap significantly in T2w and ADC scans, they are markedly separated in $CDI^s$. This increased separation suggests a substantially lower risk of false positives. Quantitatively, $CDI^s$ achieved an Area Under the Curve (AUC) of approximately 0.916 for identifying significant cancer, outperforming standard ADC (AUC ~0.886), T2w (AUC ~0.713), and DCE (AUC ~0.746).

Conclusion

$CDI^s$ hyperintensity appears to be a powerful biomarker for the presence of clinically significant prostate cancer. By bypassing hardware limitations through signal synthesis, the technique provides superior contrast and delineation accuracy compared to current clinical standards. Future work will focus on multi-center clinical trials and exploring the application of the $CDI^s$ framework to other malignancies, such as breast and brain tumors.

Summary updated

Abstract

Prostate cancer (PCa) is the second most common cancer in men worldwide and the most frequently diagnosed cancer among men in more developed countries. The prognosis of PCa is excellent if detected at an early stage, making early screening crucial for detection and treatment. In recent years, a new form of diffusion magnetic resonance imaging called correlated diffusion imaging (CDI) was introduced, and preliminary results show promise as a screening tool for PCa. In the largest study of its kind, we investigate the relationship between PCa presence and a new variant of CDI we term synthetic correlated diffusion imaging (CDI[Formula: see text]), as well as its performance for PCa delineation compared to current standard MRI techniques [T2-weighted (T2w) imaging, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) imaging] across a cohort of 200 patient cases. Statistical analyses reveal that hyperintensity in CDI[Formula: see text] is a strong indicator of PCa presence and achieves strong delineation of clinically significant cancerous tissue compared to T2w, DWI, and DCE. These results suggest that CDI[Formula: see text] hyperintensity may be a powerful biomarker for the presence of PCa, and may have a clinical impact as a diagnostic aid for improving PCa screening.
BibTeX
@inproceedings{wong2022synthetic,
  author = {Alexander Wong and Hayden Gunraj and Vignesh Sivan and Masoom A. Haider},
  title = {{Synthetic correlated diffusion imaging hyperintensity delineates clinically significant prostate cancer}},
  booktitle = {Scientific},
  year = {2022},
  doi = {10.1038/s41598-022-06872-7},
}
Figure