Quick guide on radiology image pre-processing for deep learning applications in prostate cancer research

Masoudi, Samira; Harmon, Stephanie; Mehralivand, Sherif; Walker, Stephanie M.; RaviPrakash, Harish; Bağcı, Ulaş; Choyke, Peter L.; Türkbey, Barış

doi:10.1117/1.jmi.8.1.010901

Cited by 49 publications

(25 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The three-dimensional equivalent of a pixel is a voxel. 25 Radiodensity in CT Scans is quantified by means of the Hounsfield Unit (HU), which depends on the absorption/attenuation coefficient of the tissue under consideration. Distilled water (at a standardized temperature and pressure) is arbitrarily assigned a HU value of 0, whereas air is designated as À1000.…”

Section: A Windowingmentioning

confidence: 99%

Methods to address metal artifacts in post-processed CT images – A do-it-yourself guide for orthopedic surgeons

Sharma¹,

Kaushal²,

Patel

et al. 2021

Journal of Clinical Orthopaedics and Trauma

View full text Add to dashboard Cite

Computed tomography (CT) scans are often used for postoperative imaging in orthopedics. In the presence of metallic hardware, artifacts are generated, which can hamper visualization of the CT images, and also render the study ineffective for 3-D printing. Various solutions are available to minimize metal artifacts, and radiologists can employ these before or after processing the CT study. However, the orthopedic surgeon may be faced with situations where the metal artifacts were not addressed. To counter such problems, we present three do-it-yourself (DIY) techniques that can be used to manage metal artifacts.

show abstract

Section: A Windowingmentioning

confidence: 99%

Methods to address metal artifacts in post-processed CT images – A do-it-yourself guide for orthopedic surgeons

Sharma¹,

Kaushal²,

Patel

et al. 2021

Journal of Clinical Orthopaedics and Trauma

View full text Add to dashboard Cite

show abstract

Section: Methodsmentioning

confidence: 99%

“…Now, we discuss the network in detail: Feature Generation: During manual analysis, a radiologist uses HU windowing to adjust CT intensity values to focus on organs/tissues of interest. Inspired by Masoudi et al[15], we mimic the radiologist's behaviour in our ULD and highlight multiple organs of interest with heuristically determined 5 HU windows[16]:U 1 = [400, 2000], U 2,3 = [−600, 1500], [50, 350], U 4 = [30, 150], U 5 = [50, 400] for bones, chest region including lungs & mediastinum, abdomen including liver & kidney, and soft-tissues, respectively. After windowing, 3-channel multi-intensity image (I Ui ) is passed as input to the ResNeXt-101 shared backbone with feature…”

mentioning

confidence: 99%

An Efficient Anchor-free Universal Lesion Detection in CT-scans

Sheoran¹,

Dani²,

Sharma³

et al. 2022

Preprint

View full text Add to dashboard Cite

Existing universal lesion detection (ULD) methods utilize compute-intensive anchor-based architectures which rely on predefined anchor boxes, resulting in unsatisfactory detection performance, especially in small and mid-sized lesions. Further, these default fixed anchor-sizes and ratios do not generalize well to different datasets. Therefore, we propose a robust one-stage anchor-free lesion detection network that can perform well across varying lesions sizes by exploiting the fact that the box predictions can be sorted for relevance based on their center rather than their overlap with the object. Furthermore, we demonstrate that the ULD can be improved by explicitly providing it the domain-specific information in the form of multi-intensity images generated using multiple HU windows, followed by self-attention based feature-fusion and backbone initialization using weights learned via selfsupervision over CT-scans. We obtain comparable results to the state-of-the-art methods, achieving an overall sensitivity of 86.05% on the DeepLesion dataset, which comprises of approximately 32K CT-scans with lesions annotated across various body organs.

show abstract

“…The different Python packages used during this study can be found in Supplementary Table S1. Pre-processing of MRI is essential for ML purposes, for reducing scanner dependence, and for ensuring reproducibility (39)(40)(41). As there is, to date, no consensus regarding the best way to pre-process MRI for our purposes, three different pre-processing workflows were applied and compared: " minimalist" , standardization, and "harmonization".…”

Section: Pre-processing Of Brain Mri Datamentioning

confidence: 99%

Predicting Adverse Radiation Effects in Brain Tumors After Stereotactic Radiotherapy With Deep Learning and Handcrafted Radiomics

et al. 2022

View full text Add to dashboard Cite

IntroductionThere is a cumulative risk of 20–40% of developing brain metastases (BM) in solid cancers. Stereotactic radiotherapy (SRT) enables the application of high focal doses of radiation to a volume and is often used for BM treatment. However, SRT can cause adverse radiation effects (ARE), such as radiation necrosis, which sometimes cause irreversible damage to the brain. It is therefore of clinical interest to identify patients at a high risk of developing ARE. We hypothesized that models trained with radiomics features, deep learning (DL) features, and patient characteristics or their combination can predict ARE risk in patients with BM before SRT.MethodsGadolinium-enhanced T1-weighted MRIs and characteristics from patients treated with SRT for BM were collected for a training and testing cohort (N = 1,404) and a validation cohort (N = 237) from a separate institute. From each lesion in the training set, radiomics features were extracted and used to train an extreme gradient boosting (XGBoost) model. A DL model was trained on the same cohort to make a separate prediction and to extract the last layer of features. Different models using XGBoost were built using only radiomics features, DL features, and patient characteristics or a combination of them. Evaluation was performed using the area under the curve (AUC) of the receiver operating characteristic curve on the external dataset. Predictions for individual lesions and per patient developing ARE were investigated.ResultsThe best-performing XGBoost model on a lesion level was trained on a combination of radiomics features and DL features (AUC of 0.71 and recall of 0.80). On a patient level, a combination of radiomics features, DL features, and patient characteristics obtained the best performance (AUC of 0.72 and recall of 0.84). The DL model achieved an AUC of 0.64 and recall of 0.85 per lesion and an AUC of 0.70 and recall of 0.60 per patient.ConclusionMachine learning models built on radiomics features and DL features extracted from BM combined with patient characteristics show potential to predict ARE at the patient and lesion levels. These models could be used in clinical decision making, informing patients on their risk of ARE and allowing physicians to opt for different therapies.

show abstract

Quick guide on radiology image pre-processing for deep learning applications in prostate cancer research

Cited by 49 publications

References 38 publications

Methods to address metal artifacts in post-processed CT images – A do-it-yourself guide for orthopedic surgeons

Methods to address metal artifacts in post-processed CT images – A do-it-yourself guide for orthopedic surgeons

An Efficient Anchor-free Universal Lesion Detection in CT-scans

Predicting Adverse Radiation Effects in Brain Tumors After Stereotactic Radiotherapy With Deep Learning and Handcrafted Radiomics

Contact Info

Product

Resources

About