Deep learning-based imaging reconstruction for MRI after neoadjuvant chemoradiotherapy for rectal cancer: effects on image quality and assessment of treatment response

Kim, Bona; Lee, Chul Min; Jang, Jong Keon; Kim, Jihun; Lim, Seok‐Byung; Kim, Ah Young

doi:10.1007/s00261-022-03701-3

Cited by 13 publications

(6 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(4) Deep learning: Deep learning algorithms, such as convolutional neural networks, have shown promise in automatically extracting and analyzing features from medical images. They can capture complex patterns and relationships, allowing for more accurate prediction and classification of CRC [59][60][61][62][63]. These are just a few examples of the new analytical techniques being explored in the field of radiomics for CRC.…”

Section: Discussionmentioning

confidence: 99%

Radiomics in colorectal cancer

Wu,

et al. 2023

iRADIOLOGY

View full text Add to dashboard Cite

Colorectal cancer (CRC) is a global health challenge with high morbidity and mortality. Radiomics, an emerging field, utilizes quantitative imaging features extracted from medical images for CRC diagnosis, staging, treatment response assessment, and prognostication. This review highlights the potential of radiomics for personalized CRC management. Radiomics enables noninvasive tumor characterization, aiding in early detection and accurate diagnosis, and it can be used to predict tumor stage, lymph node involvement, and prognosis. Furthermore, radiomics guides personalized therapies by assessing the treatment response and identifying patients who could benefit. Challenges include standardizing imaging protocols and analysis techniques. Robust validation frameworks and user‐friendly software are needed for the integration of radiomics into clinical practice. Despite challenges, radiomics offers valuable insights into tumor biology, treatment response, and prognosis in CRC. Overcoming technical and clinical hurdles will unlock its full potential in CRC management.

show abstract

Section: Discussionmentioning

confidence: 99%

Radiomics in colorectal cancer

Wu,

et al. 2023

iRADIOLOGY

View full text Add to dashboard Cite

show abstract

“…Therefore, under the existing hardware conditions, image reconstruction technology can break through the inherent limitations of hardware, improve the quality of medical images, and reduce operating costs but also provide medical personnel with clear images and further improve the accurate diagnosis of diseases. Technology based on deep learning improves the speed, accuracy, and robustness of medical image reconstruction [472,473].…”

Section: Image Reconstructionmentioning

confidence: 99%

Deep Learning for Medical Image-Based Cancer Diagnosis

Jiang

Wang

et al. 2023

Cancers

View full text Add to dashboard Cite

(1) Background: The application of deep learning technology to realize cancer diagnosis based on medical images is one of the research hotspots in the field of artificial intelligence and computer vision. Due to the rapid development of deep learning methods, cancer diagnosis requires very high accuracy and timeliness as well as the inherent particularity and complexity of medical imaging. A comprehensive review of relevant studies is necessary to help readers better understand the current research status and ideas. (2) Methods: Five radiological images, including X-ray, ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), positron emission computed tomography (PET), and histopathological images, are reviewed in this paper. The basic architecture of deep learning and classical pretrained models are comprehensively reviewed. In particular, advanced neural networks emerging in recent years, including transfer learning, ensemble learning (EL), graph neural network, and vision transformer (ViT), are introduced. Five overfitting prevention methods are summarized: batch normalization, dropout, weight initialization, and data augmentation. The application of deep learning technology in medical image-based cancer analysis is sorted out. (3) Results: Deep learning has achieved great success in medical image-based cancer diagnosis, showing good results in image classification, image reconstruction, image detection, image segmentation, image registration, and image synthesis. However, the lack of high-quality labeled datasets limits the role of deep learning and faces challenges in rare cancer diagnosis, multi-modal image fusion, model explainability, and generalization. (4) Conclusions: There is a need for more public standard databases for cancer. The pre-training model based on deep neural networks has the potential to be improved, and special attention should be paid to the research of multimodal data fusion and supervised paradigm. Technologies such as ViT, ensemble learning, and few-shot learning will bring surprises to cancer diagnosis based on medical images.

show abstract

“…10 Deep learning reconstruction (DLR), which uses a deep convolutional neural network trained to reconstruct images with higher quality, can improve the signal-to-noise ratio (SNR), sharpness, and spatial resolution of MRI and has been used effectively in various diseases. [15][16][17][18][19][20][21][22] In rectal MRI, the usefulness of DLR in predicting pathological complete response after neoadjuvant treatment has been reported. 16 However, the value of DLR in assessing local tumor extent has not been reported, and the impact of DLR on radiologists' judgments remains to be elucidated despite the direct impact of MRI-positive diagnoses on patient management.…”

mentioning

confidence: 99%

“…[15][16][17][18][19][20][21][22] In rectal MRI, the usefulness of DLR in predicting pathological complete response after neoadjuvant treatment has been reported. 16 However, the value of DLR in assessing local tumor extent has not been reported, and the impact of DLR on radiologists' judgments remains to be elucidated despite the direct impact of MRI-positive diagnoses on patient management. To address the 2 main problems of rectal T2WI, we hypothesized that ultra-high-resolution T2WI obtained with DLR could reduce the partial volume effects, and the combination with motion robust sequences such as PROPELLER would potentially add significant value, as DLR cannot reduce motion artifacts.…”

mentioning

confidence: 99%

“…Deep learning reconstruction (DLR), which uses a deep convolutional neural network trained to reconstruct images with higher quality, can improve the signal-to-noise ratio (SNR), sharpness, and spatial resolution of MRI and has been used effectively in various diseases 15–22 . In rectal MRI, the usefulness of DLR in predicting pathological complete response after neoadjuvant treatment has been reported 16 .…”

mentioning

confidence: 99%

See 1 more Smart Citation

Ultra-High-Resolution T2-Weighted PROPELLER MRI of the Rectum With Deep Learning Reconstruction

Matsumoto,

Tsuboyama,

Onishi

et al. 2023

Invest Radiol

View full text Add to dashboard Cite

Objective The aim of this study was to evaluate the impact of ultra-high-resolution acquisition and deep learning reconstruction (DLR) on the image quality and diagnostic performance of T2-weighted periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) imaging of the rectum. Materials and Methods This prospective study included 34 patients who underwent magnetic resonance imaging (MRI) for initial staging or restaging of rectal tumors. The following 4 types of oblique axial PROPELLER images perpendicular to the tumor were obtained: a standard 3-mm slice thickness with conventional reconstruction (3-CR) and DLR (3-DLR), and 1.2-mm slice thickness with CR (1.2-CR) and DLR (1.2-DLR). Three radiologists independently evaluated the image quality and tumor extent by using a 5-point scoring system. Diagnostic accuracy was evaluated in 22 patients with rectal cancer who underwent surgery after MRI without additional neoadjuvant therapy (median interval between MRI and surgery, 22 days). The signal-to-noise ratio and tissue contrast were measured on the 4 types of PROPELLER imaging. Results 1.2-DLR imaging showed the best sharpness, overall image quality, and rectal and lesion conspicuity for all readers (P < 0.01). Of the assigned scores for tumor extent, extramural venous invasion (EMVI) scores showed moderate agreement across the 4 types of PROPELLER sequences in all readers (intraclass correlation coefficient, 0.60–0.71). Compared with 3-CR imaging, the number of cases with MRI-detected extramural tumor spread was significantly higher with 1.2-DLR imaging (19.0 ± 2.9 vs 23.3 ± 0.9, P = 0.03), and the number of cases with MRI-detected EMVI was significantly increased with 1.2-CR, 3-DLR, and 1.2-DLR imaging (8.0 ± 0.0 vs 9.7 ± 0.5, 11.0 ± 2.2, and 12.3 ± 1.7, respectively; P = 0.02). For the diagnosis of histopathologic extramural tumor spread, 3-CR and 1.2-CR had significantly higher specificity than 3-DLR and 1.2-DLR imaging (0.75 and 0.78 vs 0.64 and 0.58, respectively; P = 0.02), and only 1.2-CR had significantly higher accuracy than 3-CR imaging (0.83 vs 0.79, P = 0.01). The accuracy of MRI-detected EMVI with reference to pathological EMVI was significantly lower for 3-CR and 3-DLR compared with 1.2-CR (0.77 and 0.74 vs 0.85, respectively; P < 0.01), and was not significantly different between 1.2-CR and 1.2-DLR (0.85 vs 0.80). Using any pathological venous invasion as the reference standard, the accuracy of MRI-detected EMVI was significantly the highest with 1.2-DLR, followed by 1.2-CR, 3-CR, and 3-DLR (0.71 vs 0.67 vs 0.59 vs 0.56, respectively; P < 0.01). The signal-to-noise ratio was significantly highest with 3-DLR imaging (P < 0.05). There were no significant differences in tumor-to-muscle contrast between the 4 types of PROPELLER imaging. Conclusions Ultra-high-resolution PROPELLER T2-weighted imaging of the rectum combined with DLR improved image quality, increased the number of cases with MRI-detected extramural tumor spread and EMVI, but did not improve diagnostic accuracy with respect to pathology in rectal cancer, possibly because of false-positive MRI findings or false-negative pathologic findings.

show abstract

Deep learning-based imaging reconstruction for MRI after neoadjuvant chemoradiotherapy for rectal cancer: effects on image quality and assessment of treatment response

Cited by 13 publications

References 35 publications

Radiomics in colorectal cancer

Radiomics in colorectal cancer

Deep Learning for Medical Image-Based Cancer Diagnosis

Ultra-High-Resolution T2-Weighted PROPELLER MRI of the Rectum With Deep Learning Reconstruction

Contact Info

Product

Resources

About