Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Poster Display session

355P - The role of deep learning image reconstruction algorithm with small field of view in spectral computed tomography in the diagnosis of lymph nodes in prostate cancer

Date

07 Dec 2024

Session

Poster Display session

Presenters

Meiting Chen

Citation

Annals of Oncology (2024) 35 (suppl_4): S1531-S1543. 10.1016/annonc/annonc1690

Authors

M. Chen1, S. Li2

Author affiliations

  • 1 Medical Oncology, Sun Yat-sen University Cancer Center, 510060 - Guangzhou/CN
  • 2 Radiology, Sun Yat-Sen University Cancer Center, 510060 - Guangzhou/CN

Resources

This content is available to ESMO members and event participants.

Abstract 355P

Background

Our study aimed to explore the improvement of image quality with spectral computed tomography (CT) of prostate cancer by combining the deep learning image reconstruction (DLIR) algorithm with small field of view (FOV) and the diagnostic value of lymph node metastasis. Our study aimed to explore the improvement of image quality with spectral computed tomography (CT) of prostate cancer by combining the deep learning image reconstruction (DLIR) algorithm with small field of view (FOV) and the diagnostic value of lymph node metastasis.

Methods

22 patients who underwent pelvic spectral CT scans for prostate cancer at Sun Yat-sen University Cancer Centre between March and December 2022 were enrolled. Axial images were reconstructed using DLIR-H, DLIR-M, DLIR-L, and the adaptive statistical iterative reconstruction algorithm with 50% weight (ASIR-V 50%). To evaluate the image quality, the signal-to-noise ratios (SNR) of spectral parameter images were measured. Lymph nodes were classified into benign and malignant groups based on surgical pathology results.

Results

With increasing in reconstruction strengths, both image SNR and subjective scores increased progressively, with DLIR-H having the highest subjective scores and SNR. Regarding the diagnostic performance, significant differences were showed between benign and malignant lymph node in the measured values of mono-energy images at 40-80 keV, iodine (water) and iodine (fat) based material decomposition images, and effective atomic number images, regardless of the reconstruction algorithm. Notably, compared to ASIR-V 50% and DLIR-L, the measurements of water (iodine) concentration and fat (iodine) concentration in DLIR-M and DLIR-H presented significant differences between benign and malignant groups. The maximum AUCs for distinguishing benign from malignant prostate cancer lymph nodes for the four reconstruction algorithms were 0.916, 0.907, 0.909, and 0.913, respectively (ASIR 50%, DLIR-L/M/H).

Conclusions

Compared to ASIR-V 50%, DLIR-H significantly enhanced image quality and provided more spectral parameters that can be considered as risk factors between benign and malignant lymph nodes of prostate cancer.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.