Submit to Applied Sciences Review for Applied Sciences Propose a Special Issue

Journal Menu

Journal Browser

Applications of Radiomics and Deep Learning in Medical Image Analysis

Special Issue Editors
Special Issue Information
Keywords
Published Papers

A special issue of Applied Sciences (ISSN 2076-3417). This special issue belongs to the section "Applied Biosciences and Bioengineering".

Deadline for manuscript submissions: closed (31 December 2022) | Viewed by 10166

Share This Special Issue

Special Issue Editor

Dr. Abdalla Ibrahim

E-Mail Website
Guest Editor

Department of Radiology, Columbia University Irving Medical Center, New York, NY 10032, USA
Interests: radiology
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear Colleagues,

This Special Issue is devoted to the applications and harmonization of handcrafted radiomics and deep learning methods in medical imaging analysis. Many studies have shown that radiomics has the potential to decode clinical information. However, many studies have also reported on the limitations of radiomics techniques due to the heterogenous nature of medical imaging data. Variations in factors like imaging vendors, imaging parameters, and patient characteristics have been reported to significantly affect the reproducibility and generalizability of image-derived quantitative features. Therefore, in this Special Issue we would like to focus on different methods, especially novel ones, to improve the reproducibility of image-derived quantitative features across data acquired heterogeneously, and the generalizability of developed signatures. Particularly, we welcome clinical and phantom studies that assess the impacts on reproducibility and generalizability, as well as those promoting new techniques. The overarching goal is to provide additional evidence of the potential of these methods for robust clinical applications.

Dr. Abdalla Ibrahim
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Applied Sciences is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

radiomics harmonization
image harmonization
deep learning
clinical radiomics

Published Papers (5 papers)

Download All Papers

Order results

Result details

Show export options Show export options

Select all

Export citation of selected articles as:

Research

14 pages, 1693 KiB

Open AccessArticle

Artificial Intelligence-Based Patient Selection for Deep Inspiration Breath-Hold Breast Radiotherapy from Respiratory Signals

by Alessandra Vendrame, Cristina Cappelletto, Paola Chiovati, Lorenzo Vinante, Masud Parvej, Angela Caroli, Giovanni Pirrone, Loredana Barresi, Annalisa Drigo and Michele Avanzo

Appl. Sci. 2023, 13(8), 4962; https://doi.org/10.3390/app13084962 - 14 Apr 2023

Cited by 1 | Viewed by 1816

Abstract

Purpose: to predict eligibility for deep inspiration breath-hold (DIBH) radiotherapy (RT) treatment of patients with left breast cancer from analysis of respiratory signal, using Deep Bidirectional Long Short-Term Memory (BLSTM) recurrent neural networks. Methods: The respiratory traces from 36 patients who underwent DIBH RT were collected. The patients’ RT treatment plans were generated for both DIBH and free-breathing (FB) modalities. The patients were divided into two classes (patient eligible or not), based on the decrease of maximum dose to the left anterior descending (LAD) artery achieved with DIBH, compared to that achieved with FB and ΔD_L. Patients with ΔD_L > median value of ΔD_L within the patient cohort were assumed to be those selected for DIBH. A BLSTM-RNN was trained for classification of patients eligible for DIBH by analysis of their respiratory signals, as acquired during acquisition of the pre-treatment computed tomography (CT), for selecting the window for DIBH. The dataset was split into training (60%) and test groups (40%), and the hyper-parameters, including the number of hidden layers, the optimizer, the learning rate, and the number of epochs, were selected for optimising model performance. The BLSTM included 2 layers of 100 neural units, each followed by a dropout layer with 20% dropout, and was trained in 35 epochs using the Adam optimizer, with an initial learning rate of 0.0003. Results: The system achieved accuracy, specificity, and sensitivity of, F1 score and area under the receiving operating characteristic curve (AUC) of 71.4%, 66.7%, 80.1%, 72.4%, and 69.4% in the test dataset, respectively. Conclusions: The proposed BLSTM-RNN classified patients in the test set eligible for DIBH with good accuracy. These results look promising for building an accurate and robust decision system to provide automated assistance to the radiotherapy team in assigning patients to DIBH. Full article

(This article belongs to the Special Issue Applications of Radiomics and Deep Learning in Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 1138 KiB

Open AccessArticle

Machine Learning Algorithm Accuracy Using Single- versus Multi-Institutional Image Data in the Classification of Prostate MRI Lesions

by Destie Provenzano, Oleksiy Melnyk, Danish Imtiaz, Benjamin McSweeney, Daniel Nemirovsky, Michael Wynne, Michael Whalen, Yuan James Rao, Murray Loew and Shawn Haji-Momenian

Appl. Sci. 2023, 13(2), 1088; https://doi.org/10.3390/app13021088 - 13 Jan 2023

Cited by 4 | Viewed by 1856

Abstract

(1) Background: Recent studies report high accuracies when using machine learning (ML) algorithms to classify prostate cancer lesions on publicly available datasets. However, it is unknown if these trained models generalize well to data from different institutions. (2) Methods: This was a retrospective study using multi-parametric Magnetic Resonance Imaging (mpMRI) data from our institution (63 mpMRI lesions) and the ProstateX-2 challenge, a publicly available annotated image set (112 mpMRI lesions). Residual Neural Network (ResNet) algorithms were trained to classify lesions as high-risk (hrPCA) or low-risk/benign. Models were trained on (a) ProstateX-2 data, (b) local institutional data, and (c) combined ProstateX-2 and local data. The models were then tested on (a) ProstateX-2, (b) local and (c) combined ProstateX-2 and local data. (3) Results: Models trained on either local or ProstateX-2 image data had high Area Under the ROC Curve (AUC)s (0.82–0.98) in the classification of hrPCA when tested on their own respective populations. AUCs decreased significantly (0.23–0.50, p < 0.01) when models were tested on image data from the other institution. Models trained on image data from both institutions re-achieved high AUCs (0.83–0.99). (4) Conclusions: Accurate prostate cancer classification models trained on single-institutional image data performed poorly when tested on outside-institutional image data. Heterogeneous multi-institutional training image data will likely be required to achieve broadly applicable mpMRI models. Full article

(This article belongs to the Special Issue Applications of Radiomics and Deep Learning in Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 783 KiB

Open AccessFeature PaperArticle

Effects of Contrast Enhancement Phase on the Reproducibility and Predictivity of CT-Based Renal Lesions Radiomic Features

by Hayder Alkhafaji and Abdalla Ibrahim

Appl. Sci. 2022, 12(24), 12599; https://doi.org/10.3390/app122412599 - 8 Dec 2022

Cited by 1 | Viewed by 897

Abstract

The extraction of quantitative medical imaging features, or radiomics, has been an exponentially growing research field in recent decades. Nonetheless, more studies are investigating the limitations of the quantitative imaging features, especially the reproducibility of RFs across different scanning settings. In this experiment, we investigate the reproducibility of renal cell carcinoma (RCC) RFs between the non-contrast, arterial, and late phases contrast-enhanced computed tomography (CE-CT) scans; and the ability of ComBat technique to harmonize these RFs. In addition, we assessed the predictive performance of the RFs extracted from the different phases. A total of 69 CECT scans with the three different phases were analyzed. Original RFs were extracted from the segmented lesions on each phase using Pyradiomics toolbox. The agreement in RF values before and after harmonization was evaluated with the concordance correlation coefficient (CCC). Our results show that most RFs are not reproducible across different imaging phases. In addition, ComBat harmonization did not significantly increase the number of reproducible RFs in any of the three scenarios. Furthermore, RFs extracted from the arterial phase were, on average, the most predictive of overall survival in RCC patients. The findings can guide the analysis of retrospective RCC heterogeneous data acquired in different phases and add to the call for radiomics-specific harmonization techniques. Full article

(This article belongs to the Special Issue Applications of Radiomics and Deep Learning in Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1177 KiB

Open AccessArticle

The Impact of Image Acquisition Parameters and ComBat Harmonization on the Predictive Performance of Radiomics: A Renal Cell Carcinoma Model

by Abdalla Ibrahim, Lin Lu, Hao Yang, Oguz Akin, Lawrence H. Schwartz and Binsheng Zhao

Appl. Sci. 2022, 12(19), 9824; https://doi.org/10.3390/app12199824 - 29 Sep 2022

Cited by 2 | Viewed by 2007

Abstract

Radiomics, one of the potential methods for developing clinical biomarker, is one of the exponentially growing research fields. In addition to its potential, several limitations have been identified in this field, and most importantly the effects of variations in imaging parameters on radiomic features (RFs). In this study, we investigate the potential of RFs to predict overall survival in patients with clear cell renal cell carcinoma, as well as the impact of ComBat harmonization on the performance of RF models. We assessed the robustness of the results by performing the analyses a thousand times. Publicly available CT scans of 179 patients were retrospectively collected and analyzed. The scans were acquired using different imaging vendors and parameters in different medical centers. The performance was calculated by averaging the metrics over all runs. On average, the clinical model significantly outperformed the radiomic models. The use of ComBat harmonization, on average, did not significantly improve the performance of radiomic models. Hence, the variability in image acquisition and reconstruction parameters significantly affect the performance of radiomic models. The development of radiomic specific harmonization techniques remain a necessity for the advancement of the field. Full article

(This article belongs to the Special Issue Applications of Radiomics and Deep Learning in Medical Image Analysis)

► Show Figures

Figure 1

23 pages, 6702 KiB

Open AccessArticle

Deep and Hybrid Learning Technique for Early Detection of Tuberculosis Based on X-ray Images Using Feature Fusion

by Suliman Mohamed Fati, Ebrahim Mohammed Senan and Narmine ElHakim

Appl. Sci. 2022, 12(14), 7092; https://doi.org/10.3390/app12147092 - 14 Jul 2022

Cited by 26 | Viewed by 3066

Abstract

Tuberculosis (TB) is a fatal disease in developing countries, with the infection spreading through direct contact or the air. Despite its seriousness, the early detection of tuberculosis by means of reliable techniques can save the patients’ lives. A chest X-ray is a recommended screening technique for locating pulmonary abnormalities. However, analyzing the X-ray images to detect abnormalities requires highly experienced radiologists. Therefore, artificial intelligence techniques come into play to help radiologists to perform an accurate diagnosis at the early stages of TB disease. Hence, this study focuses on applying two AI techniques, CNN and ANN. Furthermore, this study proposes two different approaches with two systems each to diagnose tuberculosis from two datasets. The first approach hybridizes two CNN models, which are Res-Net-50 and GoogLeNet techniques. Prior to the classification stage, the approach applies the principal component analysis (PCA) algorithm to reduce the features’ dimensionality, aiming to extract the deep features. Then, the SVM algorithm is used for classifying features with high accuracy. This hybrid approach achieved superior results in diagnosing tuberculosis based on X-ray images from both datasets. In contrast, the second approach applies artificial neural networks (ANN) based on the fused features extracted by ResNet-50 and GoogleNet models and combines them with the features extracted by the gray level co-occurrence matrix (GLCM), discrete wavelet transform (DWT) and local binary pattern (LBP) algorithms. ANN achieved superior results for the two tuberculosis datasets. When using the first dataset, the ANN, with ResNet-50, GLCM, DWT and LBP features, achieved an accuracy of 99.2%, a sensitivity of 99.23%, a specificity of 99.41%, and an AUC of 99.78%. Meanwhile, with the second dataset, ANN, with the features of ResNet-50, GLCM, DWT and LBP, reached an accuracy of 99.8%, a sensitivity of 99.54%, a specificity of 99.68%, and an AUC of 99.82%. Thus, the proposed methods help doctors and radiologists to diagnose tuberculosis early and increase chances of survival. Full article

(This article belongs to the Special Issue Applications of Radiomics and Deep Learning in Medical Image Analysis)

► Show Figures

Journal Menu

Journal Browser

Applications of Radiomics and Deep Learning in Medical Image Analysis

Share This Special Issue

Special Issue Editor

Special Issue Information

Keywords

Published Papers (5 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI