CT Modalities Research Articles

Deep Learning-Based Multi-Modality Segmentation of Primary Gross Tumor Volume in CT and MRI for Nasopharyngeal Carcinoma

The delineation of primary gross tumor volume (GTV) of nasopharyngeal carcinoma (NPC) is an essential step for radiotherapy planning. In clinical practice, radiation oncologists manually delineate the GTV in planning CT with the help of diagnostic MRI. This is because NPC tumors are closely adjacent to many important anatomic structures, and CT and MRI provide complementary strength to accurately determine the tumor extension boundary. Manual delineation is time-consuming with the potential registration errors between MRI and CT decreasing the delineation accuracy. In this study, we propose a fully automated GTV segmentation method based on CT and MRI by first aligning MRI to CT, and then, segmenting the GTV using a multi-modality deep learning model. We collected 104 nasopharyngeal carcinoma patients with both planning CT and diagnostic MRI scans (T1 & T2 phases). An experienced radiation oncologists manually delineated the GTV, which was further examined by another senior radiation oncologist. Then, a coarse to fine cross-modality registration from MRI to CT was conducted as follows: (1) A rigid transformation was performed on MRI to roughly align MRI to CT with similar anatomic position. (2) Then, the region of interest (RoI) on both CT and rigid-transformed MRI were cropped. (3) A leading cross-modality deformable registration algorithm, named DEEDS, was applied on the cropped MRI and CT RoIs to find an accurate local alignment. Next, using CT and registered MRI as the combined input, a multi-modality deep segmentation network based on nnUNet was trained to generate the GTV prediction. 20% patients were randomly selected as the unseen testing set to quantitatively evaluate the performance. The quantitative NPC GTV segmentation performance is summarized in Table 1. The deep segmentation model using CT alone achieved reasonable high performance with 76.6% Dice score and 1.34mm average surface distance (ASD). When both CT and registered MRI were used, the segmentation model further improved the performance by 0.9% Dice score increase and 11% relative ASD error reduction, demonstrating the complementary strength of CT and MRI in determining NPC GTV. Notably, the achieved 77.5% Dice score and 1.19mm ASD by the multimodality model is among the top performing results reported in recent automatic NPC GTV segmentation using either CT or MRI modality. We developed a fully automated multi-modal deep-learning model for NPC GTV segmentation. The developed model can segment the NPC GTV in high accuracy. With further optimization and validation, this automated model has potential to standardize the NPC GTV segmentation and significantly decrease the workload of radiation oncologists in clinical practice.

A Comparison of Imaging Outcomes From 2 Weightbearing CT Modalities.

The use of weightbearing images to diagnose foot and ankle injuries continues to offer hope for improved insight into pathologies, but weightbearing CT imaging has been limited by availability. The ability to apply force to the lower limb in a horizontal bore CT system may offer an adaptation to currently available imaging systems that provides access to weightbearing images without the acquisition of additional expensive imaging space or equipment. In order to determine whether a horizontal CT system could produce the same results as a standing CT, 3 images of one foot from 10 subjects was obtained and standard measures were calculated. Each subject underwent a standing CT scan, a scan in a horizontal bore CT machine while the subject pressed against a pedal with spring resistance and a finally a scan with the foot placed on the pedal but without any pressure. No statistically significant difference between the standing and pedal-based CTs resulted. Navicular height and Meary angle (axial) were statistically different from nonweightbearing for both standing and horizontal systems. The horizontal results were statistically different from nonweightbearing in IM angle, talocalcaneal angle, and talonavicular coverage. No differences from nonweightbearing were found for either system in talar tilt, talocrural angle, or the lateral Meary angle. The results in this initial study of normal control subjects suggest that a pedal-based loading mechanism may adapt a horizontal-bore CT system for the acquisition of weightbearing images. The ability to acquire a weightbearing CT from a horizontal bore CT machine can make these images more available.

Development and external validation of the multichannel deep learning model based on unenhanced CT for differentiating fat-poor angiomyolipoma from renal cell carcinoma: a two-center retrospective study.

There are undetectable levels of fat in fat-poor angiomyolipoma. Thus, it is often misdiagnosed as renal cell carcinoma. We aimed to develop and evaluate a multichannel deep learning model for differentiating fat-poor angiomyolipoma (fp-AML) from renal cell carcinoma (RCC). This two-center retrospective study included 320 patients from the First Affiliated Hospital of Sun Yat-Sen University (FAHSYSU) and 132 patients from the Sun Yat-Sen University Cancer Center (SYSUCC). Data from patients at FAHSYSU were divided into a development dataset (n = 267) and a hold-out dataset (n = 53). The development dataset was used to obtain the optimal combination of CT modality and input channel. The hold-out dataset and SYSUCC dataset were used for independent internal and external validation, respectively. In the development phase, models trained on unenhanced CT images performed significantly better than those trained on enhanced CT images based on the fivefold cross-validation. The best patient-level performance, with an average area under the receiver operating characteristic curve (AUC) of 0.951 ± 0.026 (mean ± SD), was achieved using the "unenhanced CT and 7-channel" model, which was finally selected as the optimal model. In the independent internal and external validation, AUCs of 0.966 (95% CI 0.919-1.000) and 0.898 (95% CI 0.824-0.972), respectively, were obtained using the optimal model. In addition, the performance of this model was better on large tumors (≥ 40mm) in both internal and external validation. The promising results suggest that our multichannel deep learning classifier based on unenhanced whole-tumor CT images is a highly useful tool for differentiating fp-AML from RCC.

Towards MR contrast independent synthetic CT generation

The use of synthetic CT (sCT) in the radiotherapy workflow would reduce costs and scan time while removing the uncertainties around working with both MR and CT modalities. The performance of deep learning (DL) solutions for sCT generation is steadily increasing, however most proposed methods were trained and validated on private datasets of a single contrast from a single scanner. Such solutions might not perform equally well on other datasets, limiting their general usability and therefore value. Additionally, functional evaluations of sCTs such as dosimetric comparisons with CT-based dose calculations better show the impact of the methods, but the evaluations are more labor intensive than pixel-wise metrics.To improve the generalization of an sCT model, we propose to incorporate a pre-trained DL model to pre-process the input MR images by generating artificial proton density, T1 and T2 maps (i.e. contrast-independent quantitative maps), which are then used for sCT generation. Using a dataset of only T2w MR images, the robustness towards input MR contrasts of this approach is compared to a model that was trained using the MR images directly. We evaluate the generated sCTs using pixel-wise metrics and calculating mean radiological depths, as an approximation of the mean delivered dose.On T2w images acquired with the same settings as the training dataset, there was no significant difference between the performance of the models. However, when evaluated on T1w images, and a wide range of other contrasts and scanners from both public and private datasets, our approach outperforms the baseline model.Using a dataset of T2w MR images, our proposed model implements synthetic quantitative maps to generate sCT images, improving the generalization towards other contrasts. Our code and trained models are publicly available.

Effective hybrid attention network based on pseudo-color enhancement in ultrasound image segmentation

Ultrasound image segmentation plays a vital role in the early diagnosis of human diseases. It helps diagnose many diseases, such as breast cancer, hemangioma, and other gynecological disorders. However, the intrinsic imaging characteristics of ultrasound images result in substantially lower resolution and clarity than CT, MRI, and other imaging modalities, and they are sensitive to interference from external influences. With its inherent artifacts, blurred lesion boundaries, and uneven intensity distribution, ultrasound images present a challenging task when it comes to segmenting lesion areas accurately. In recent years, convolutional neural networks (CNNs) have achieved remarkable results in medical image segmentation tasks. However, CNNs are limited in capturing the remote dependencies of the input image, leading to degraded accuracy in segmenting ultrasound lesions. In this paper, we developed a deep convolutional neural network that incorporates the pseudo-color enhancement algorithm and hybrid attention modules that enhance the network’s ability to extract fine features and remote modeling capabilities. We propose a novel multi-scale channel attention-based decoder that efficiently uses the feature maps from the encoder as a complement and fuses them with the upsampled feature maps. The hybrid attention combination captures cross-channel interactions efficiently and enhances the context modeling capability, further improving the extraction of coarse and delicate features, and resulting in significant performance improvements. We found that the dice performance improved by 2.54%, 2.47%, 1.39%, 0.99%, and 1.23% on the BUL, BUSI, Hemangioma, BP, and VUI. Results from four public datasets and one self-collected dataset indicate that the proposed method outperforms other medical image segmentation methods for ultrasound image lesion segmentation.

SWM-DE: Statistical wavelet model for joint denoising and enhancement for multimodal medical images

Medical images are usually degraded by numerous noises during acquisition or transmission, which often causes low contrast leading to deterioration of image quality. As such, medical image denoising and enhancement has become a paramount routine task. To overcome this problem, we propose a cutting-edge joint statistical and morphological model for the denoising and enhancement operation. Firstly, we propose a statistical model in formulating the marginal distribution of the wavelet coefficients. This model is integrated into a Bayesian inference framework to develop a maximum a posterior (MAP) estimator of the noise-free coefficient. Based on the statistical model, we eliminate the need for noise level estimation, and allows the model to automatically adapts to the observed image data. Secondly, we propose an adjustable morphological reconstruction model to eliminate known and unknown noises associated with medical images, while preserving the image details. After these operations, the image is decomposed into several wavelet subbands to extract the illumination and detail components. The image is then reconstructed based on the inverse wavelet to generate the enhanced noise-free image.Experimental results show that the proposed framework obtained high EME values of 41.04, 48.81, 47.81, and 45.75 for OCTA, FFA, CT, and X-ray imaging modalities, and performs better than the state-of-the-art methods. The proposed algorithm can effectively and efficiently enhance medical images, which will assist the clinicians in disease diagnosis, monitoring, and treatment.

Head and neck tumor segmentation convolutional neural network robust to missing PET/CT modalities using channel dropout

Objective. Radiation therapy for head and neck (H&N) cancer relies on accurate segmentation of the primary tumor. A robust, accurate, and automated gross tumor volume segmentation method is warranted for H&N cancer therapeutic management. The purpose of this study is to develop a novel deep learning segmentation model for H&N cancer based on independent and combined CT and FDG-PET modalities. Approach. In this study, we developed a robust deep learning-based model leveraging information from both CT and PET. We implemented a 3D U-Net architecture with 5 levels of encoding and decoding, computing model loss through deep supervision. We used a channel dropout technique to emulate different combinations of input modalities. This technique prevents potential performance issues when only one modality is available, increasing model robustness. We implemented ensemble modeling by combining two types of convolutions with differing receptive fields, conventional and dilated, to improve capture of both fine details and global information. Main Results. Our proposed methods yielded promising results, with a Dice similarity coefficient (DSC) of 0.802 when deployed on combined CT and PET, DSC of 0.610 when deployed on CT, and DSC of 0.750 when deployed on PET. Significance. Application of a channel dropout method allowed for a single model to achieve high performance when deployed on either single modality images (CT or PET) or combined modality images (CT and PET). The presented segmentation techniques are clinically relevant to applications where images from a certain modality might not always be available.

Derivation and validation of three intracerebral hemorrhage expansion scores using different CT modalities.

To derivate and validate three scores for the prediction of intracerebral hemorrhage (ICH) expansion depending on the use of non-contrast CT (NCCT), single-phase CTA, or multiphase CTA markers of hematoma expansion, and to evaluate the added value of single-phase and multiphase CTA over NCCT. After prospectively deriving NCCT, single-phase CTA, and multiphase CTA hematoma expansion scores in 156 patients with ICH < 6h, we validated them in 120 different patients. Discrimination and calibration of the three scores was assessed. Primary outcome was substantial hematoma expansion > 6mL or > 33% at 24h. The evaluation of single-phase and multiphase CTA markers gave a steadily increase in discrimination for substantial hematoma expansion over NCCT markers. The C-index (95% confidence interval) in derivation and validation cohorts was 0.69 (0.58-0.80) and 0.59 (0.46-0.72) for NCCT score, significantly lower than 0.75 ([0.64-0.87], p = 0.038) and 0.72 ([0.59-0.84], p = 0.016) for single-phase CTA score, and than 0.79 ([0.68-0.89], p = 0.033) and 0.73 ([0.62-0.85], p = 0.031) for multiphase CTA score, respectively. The three scores showed good calibration in both derivation and validation cohorts: NCCT (χ2 statistic 0.389, p = 0.533; and χ2 statistic 0.352, p = 0.553), single-phase CTA (χ2 statistic 2.052, p = 0.359; and χ2 statistic 2.230, p = 0.328), and multiphase CTA (χ2 statistic 0.559, p = 0.455; and χ2 statistic 0.020, p = 0.887) scores, respectively. This study shows the added prognostic value of more advanced CT modalities in acute ICH evaluation. NCCT, single-phase CTA, and multiphase CTA scores may help to refine the selection of patients at risk of expansion in different decision-making scenarios. • This study shows the added prognostic value of more advanced CT modalities in acute intracerebral hemorrhage evaluation. • The evaluation of single-phase and multiphase CTA markers provides a steadily increase in discrimination for intracerebral hemorrhage expansion over non-contrast CT markers. • Non-contrast CT, single-phase CTA, and multiphase CTA scores may help clinicians and researchers to refine the selection of patients at risk of intracerebral hemorrhage expansion in different decision-making scenarios.

Can We Detect Peritoneal Adhesions With MRI and CT Prior to Abdominal Surgery?

This article is part 2 of our review assessing diagnostic modalities. In the first part, we analyzed the ultrasonography capabilities to detect abdominal adhesions. The second part assesses the effectiveness of MRI and CT.MRI and CT modalities are much less commonly used in detecting adhesions than ultrasonography. MRI has approximately the same prognostic capabilities as ultrasonography in detecting adhesions between the viscera and the abdominal wall, and it additionally detects adhesions between the abdominal organs. Yet, MRI is significantly more time-consuming and highly dependent on the radiologist experience. Like ultrasonography, MRI determines the extent of adhesions but not the type.Native CT does not visualize abdominal adhesions, though its findings in artificial pneumoperitoneum are superior in diagnostic value to those of ultrasonography and MRI, so this modality should be considered.

Prevalence, diagnosis, and impact on clinical outcomes of dural ossification in the thoracic ossification of the ligamentum flavum: a systematic review.

Systematic review. Thoracic ossification of the ligamentum flavum (TOLF) has become the principal cause of thoracic spinal stenosis. Dural ossification (DO) was a common clinical feature accompanying with TOLF. However, on account of the rarity, we know little about the DO in TOLF so far. This study was conducted to elucidate the prevalence, diagnostic measures, and impact on the clinical outcomes of DO in TOLF by integrating the existing evidence. PubMed, Embase, and Cochrane Database were comprehensively searched for studies relevant to the prevalence, diagnostic measures, or impact on the clinical outcomes of DO in TOLF. All retrieved studies meeting the inclusion and criterion were included into this systematic review. The prevalence of DO in TOLF treated surgically was 27% (281/1046), ranging from 11 to 67%. Eight diagnostic measures have been put forward to predict the DO in TOLF using the CT or MRI modalities, including "tram track sign", "comma sign", "bridge sign", "banner cloud sign", "T2 ring sign", TOLF-DO grading system, CSAOR grading system, and CCAR grading system. DO did not affect the neurological recovery of TOLF patients treated with the laminectomy. The rate of dural tear or CSF leakage in TOLF patients with DO was approximately 83% (149/180). The prevalence of DO in TOLF treated surgically was 27%. Eight diagnostic measures have been put forward to predict the DO in TOLF. DO did not affect the neurological recovery of TOLF treated with laminectomy but was associated with high risk of complications.

Imaging of Central Nervous System Autoimmune, Paraneoplastic, and Neuro-rheumatologic Disorders.

This article provides an overview of the imaging modalities used in the evaluation of central nervous system (CNS) autoimmune, paraneoplastic, and neuro-rheumatologic disorders. An approach is outlined for interpreting imaging findings in this context, synthesizing a differential diagnosis based on certain imaging patterns, and choosing further imaging for specific diseases. The rapid discovery of new neuronal and glial autoantibodies has revolutionized the autoimmune neurology field and has elucidated imaging patterns characteristic of certain antibody-associated diseases. Many CNS inflammatory diseases, however, lack a definitive biomarker. Clinicians should recognize neuroimaging patterns suggestive of inflammatory disorders, as well as the limitations of imaging. CT, MRI, and positron emission tomography (PET) modalities all play a role in diagnosing autoimmune, paraneoplastic, and neuro-rheumatologic disorders. Additional imaging modalities such as conventional angiography and ultrasonography can be helpful for further evaluation in select situations. Knowledge of imaging modalities, both structural and functional, is critical in identifying CNS inflammatory diseases quickly and can help avoid invasive testing such as brain biopsy in certain clinical scenarios. Recognizing imaging patterns suggestive of CNS inflammatory diseases can also facilitate the early initiation of appropriate treatments to diminish morbidity and future disability.

Abstract TP120: Noncontrast Ct, Single-phase Cta, And Multiphase Cta Intracerebral Hemorrhage Expansion Scores

Background: Several noncontrast computed tomography (NCCT), single-phase computed tomography angiography (CTA), and multiphase CTA markers of intracerebral hemorrhage (ICH) expansion have been previously proposed. We derived and validated three scores for the prediction of hematoma expansion depending on the use of NCCT, single-phase CTA, or multiphase CTA markers of hematoma expansion. Methods: We prospective studied 276 consecutive patients with ICH within 6 hours from symptom onset. After deriving NCCT, single-phase CTA, and multiphase CTA scores in a 5-year period population (n=156), we validated them in a different 3-year period population (n=120). Outcome parameters included substantial hematoma expansion &gt;6 mL or &gt;33% at 24 hours (primary outcome) and poor outcome (mRS score &gt;2) at 90 days. Results: The most accurate marker of hematoma expansion was spot sign in phase 1 of multiphase CTA (80.3%). The four independent predictors of substantial hematoma expansion included in the different scores were ultraearly hematoma growth (uHG) &gt;5 mL/h, heterogeneous density, spot sign in phase 1 of multiphase CTA, and spot sign in any phase of multiphase CTA (Table). On each of the three scores, the proportion of patients that experienced substantial hematoma expansion increased with each point increase. C-index for both substantial hematoma expansion and poor outcome in the derivation and validation cohort was lower in NCCT expansion score than in single-phase CTA expansion score which, in turn, was lower than in multiphase CTA expansion score (Table). Conclusions: This study demonstrates the added prognostic value of more advanced CT modalities in acute ICH evaluation. Single-phase CTA score and, especially, multiphase CTA score, are more robust than NCCT score in the prediction of hematoma expansion and poor outcome. These scores may help to refine the selection of patients at risk of expansion and poorest outcomes in different decision-making scenarios.

A Novel Lung Nodule Accurate Segmentation of PET-CT Images Based on Convolutional Neural Network and Graph Model

Positron Emission Tomography and Computed Tomography(PET/CT) imaging could obtain functional metabolic feature information and anatomical localization information of the patient body. However, tumor segmentation in PET/CT images is significantly challenging for fusing of dual-modality characteristic information. In this work, we have proposed a novel deep learning-based graph model network which can automatically fuse dual-modality information for tumor area segmentation. Our method rationally utilizes the advantage of each imaging modality(PET: the superior contrast, CT: the superior spatial resolution). We formulate this task as a Conditional Random Field(CRF) based on multi-scale fusion and dual-modality co-segmentation of object image with a normalization term which balances the segmentation divergence between PET and CT. This mechanism considers that the spatial varying characteristics acquire different scales, which encode various feature information over different modalities. The ability of our method was evaluated to detect and segment tumor regions with different fusion approaches using a dataset of PET/CT clinical tumor images. The results illustrated that our method effectively integrates both PET and CT modalities information, deriving segmentation accuracy result of 0.86 in DSC and the sensitivity of 0.83, which is 3.61% improvement compared to the W-Net.

Fully Automatic End-to-End Convolutional Neural Networks-Based Pancreatic Tumor Segmentation on CT Modality

The pancreas is one of the vital organs in the human body. Early diagnosis of a disease in the pancreas is critical. In this way, the effects of pancreas diseases, especially pancreatic cancer on the person are decreased. With this purpose, artificial intelligence-assisted pancreatic cancer segmentation was performed for early diagnosis in this paper. For this aim, several state-of-the-art segmentation networks, UNet, LinkNet, SegNet, SQ-Net, DABNet, EDANet, and ESNet were used in this study. In the comparative analysis, the best segmentation performance has been achieved by SQ-Net. SQ-Net has achieved a 0.917 dice score, 0.847 IoU score, 0.920 sensitivity, 1.000 specificity, 0.914 precision, and 0.999 accuracy. Considering these results, an artificial intelligence-based decision support system was created in the study.

Comparison of computed tomography in relation to ventilation perfusion scan in the diagnosis of pulmonary embolism

Abstract&#x0D; Introduction: Pulmonary embolism is still a challenge in diagnosis due to its variable and nonspecific symptoms. Computed tomography and ventilation/perfusion scanning are the modalities most commonly used in the diagnosis of pulmonary embolism, and both modalities have their advantages and disadvantages. One of the most important factors in the assessment and localization of pulmonary embolism is the diagnostic accuracy of these modalities, which serves to model different diagnostic strategies in the diagnosis of pulmonary embolism.&#x0D; Material and methods: The research was conducted as a review of professional literature available in scientific research databases. A selection of 20 professional papers was made, based on which an analysis was conducted and a database was formed. Criteria for inclusion in the research were scientific research papers that report on the sensitivity and specificity of diagnostic modalities of CT and V/P scanning as well as the results of diagnostic tests based on which the comparison of data from two modalities determined a diagnostically more accurate modality.&#x0D; Results: Sensitivity of 91.89% and specificity of 98.86% and diagnostic accuracy of 94.83% were determined in the case of computed tomography. Sensitivity of 90.58% and specificity of 98.33% and diagnostic accuracy of 96.43% were determined for the ventilation perfusion scanning method. We can conclude that the application of ventilation/perfusion scanning in the diagnosis of pulmonary embolism is a little more accurate compared to computed tomography.&#x0D; Conclusion: Ventilation/perfusion lung scanning will more accurately identify healthy individuals while on the other hand we can conclude that computed tomography is more accurate in diagnosing embolism in sick individuals. . Given that this difference between the two modalities is very small, the question is whether it is statistically significant at all. We can conclude that both diagnostic procedures have a high level of accuracy.

Noise correlation and its impact on the performance of multi-material decomposition-based spectral imaging in photon-counting CT.

It has been known that noise correlation plays an important role in the determination of the performance of spectral imaging based on two-material decomposition (2-MD). To further understand the basics of spectral imaging in photon-counting CT toward optimal design and implementation, we study the noise correlation in multi-MD (m-MD) and its impact on the performance of spectral imaging. We derive the equations that characterize the noise and noise correlation in the material-specific (basis) images in m-MD, followed by a simulation study to verify the derived equations and study the noise correlation's impact on the performance of spectral imaging. Using a specially designed digital phantom, the study of noise correlation runs over the cases of two-, three-, and four-MD (2-MD, 3-MD, and 4-MD). Then, the noise correlation's impact on the performance of spectral imaging in photon-counting CT is investigated, using a modified Shepp-Logan phantom. The results in 2-MD show that, in-line with what has been reported in the literature, the noise correlation coefficient between the material-specific images corresponding to the basis materials approaches -1. The results in m-MD (m≥3) are more complicated and interesting, as the noise correlation coefficients between a pair of the material-specific images alternate between ±1, and so do in the case of 4-MD. The m-MD data show that the noise in virtual monochromatic imaging (a form of spectral imaging) is moderate even though the noises in material-specific (basis) images vary drastically. The observation of noise correlation in 3-MD, 4-MD, and beyond (i.e., m-MD) is informative to the community. The relationship between noise correlation and the performance of spectral imaging revealed in this work may help clinical medical physicists understand the fundamentals of spectral imaging based on MD and optimize the performance of spectral imaging in photon-counting CT and other X-ray imaging modalities.

Towards smart diagnostic methods for COVID-19: Review of deep learning for medical imaging.

The infectious disease known as COVID-19 has spread dramatically all over the world since December 2019. The fast diagnosis and isolation of infected patients are key factors in slowing down the spread of this virus and better management of the pandemic. Although the CT and X-ray modalities are commonly used for the diagnosis of COVID-19, identifying COVID-19 patients from medical images is a time-consuming and error-prone task. Artificial intelligence has shown to have great potential to speed up and optimize the prognosis and diagnosis process of COVID-19. Herein, we review publications on the application of deep learning (DL) techniques for diagnostics of patients with COVID-19 using CT and X-ray chest images for a period from January 2020 to October 2021. Our review focuses solely on peer-reviewed, well-documented articles. It provides a comprehensive summary of the technical details of models developed in these articles and discusses the challenges in the smart diagnosis of COVID-19 using DL techniques. Based on these challenges, it seems that the effectiveness of the developed models in clinical use needs to be further investigated. This review provides some recommendations to help researchers develop more accurate prediction models.

Using Radiomics to Improve the Diagnostic Accuracy of Indeterminate Residual Primary Disease on Restaging PET/CT Imaging Following Radiation Therapy for Head and Neck Cancers

<h3>Purpose/Objective(s)</h3> Identifying residual disease (RD) in the primary site of head and neck cancer (HNC) patients treated with radiotherapy (RT) on the 3-month post RT PET/CT scans can be challenging. Radiologists have to account for RT-related inflammation versus RD. Accurate identification is key to determine the need for morbid salvage versus close observation. In this study, we hypothesized a clinical/radiomics model could complement a radiologist's interpretation in these uncertain cases. <h3>Materials/Methods</h3> We identified 182 patients with HNC treated at our institution from 2016-2020. We reviewed the 3-month post-RT PET/CT scan interpretations to obtain the radiologists' RD assessment. We classified assessments as "certain" or "uncertain." "Certain" assessments included words like "complete metabolic response," "no RD," "RD present" while "uncertain" assessments provided a differential. For uncertain cases, predictions were inferred as the most likely diagnosis ("likely radiation changes but can't rule out RD" was no RD present). If no clarification was provided, then the prediction was marked as RD present. We extracted radiomics features from the CT and PET modalities using open-source software. We gathered clinical data such as HPV status, T stage, etc. Feature reduction was performed by assessing single feature performance on a logistic regression model. Features with high correlation (>0.9) measured using Pearson coefficient were discarded. Recursive feature elimination was performed on the remaining features (only 5-10 features were allowed in the final model). We trained an explainable boosting machine model using the "certain" data and tested with "uncertain" data. Ground truth was obtained from follow-up imaging or pathology. <h3>Results</h3> Ninety-six assessments were "certain" and 86 were "uncertain." RD was present in 10 and 9 cases, respectively. The performances for the radiologists and our models are summarized in Table 1. The Clinical/CT radiomics model reduced the number of false positives from 22 to 8 at the cost of only one additional false negative. The ensemble of our CT and PET models did not improve performance. Table 1 <h3>Conclusion</h3> Even expert radiologists may struggle with identifying RD given post-RT inflammatory changes. Our Clinical/CT radiomics model greatly improved specificity with only a minor drop in sensitivity in this uncertain cohort. Machine learning models may have utility in complementing clinical experience to improve diagnostic accuracy for this challenging scenario.

Cost-effectiveness of Contrast-Enhanced Ultrasound for Diagnosis and Active Surveillance of Complex Cystic Renal Lesions.

To determine the cost-effectiveness of Contrast-Enhanced Ultrasound (ceUS) for the active surveillance of complex renal masses compared to the more established imaging modalities of CT and MRI. A decision-analytic Markov state microsimulation model was constructed in TreeAge Pro. We simulated independent cohorts of 100,000 60-year-old individuals with either a Bosniak IIF or Bosniak III complex renal mass who were followed for 10 years or until death. The model compared three imaging strategies: (1) ceUS, (2) contrast-enhanced magnetic-resonance imaging (ceMRI), and (3) contrast-enhanced computed tomography (ceCT) for active surveillance of a complex renal mass. For 60-year-old patients with either Bosniak IIF or III renal masses, ceUS was the most cost-effective strategy even after varying rates of active surveillance from 10-100%. ceUS is a viable and cost-effective option in the active surveillance of Bosniak class IIF and III renal cysts. Even after varying the rates of active surveillance usage, ceUS was robust and remained the most dominant strategy. For patients who have impaired kidney functions, ceUS is can be a safer alternative than non-contrast enhanced CT or MRI in the management of patients with Bosniak III renal cysts.

Pancreas Segmentation Using U-Net Based Segmentation Networks in CT Modality: A Comparative Analysis

The pancreas is one of the small size organs in the abdomen. Moreover, anatomical differences make it difficult to detect the pancreas. This project aims to automatically segmentation of pancreas. For this purpose, NIH-CT82 data set, which includes CT images from 82 patients was used. U-Net which is state-of-the-art model and its different versions, namely Attention U-Net, Residual U-Net, Attention Residual U-Net, and Residual U-Net++ were tested. Best predict performance was achieved by Residual U-Net with the dice of 0.903, IoU of 0.823, sensitivity of 0.898, specificity of 1.000, precision of 0.908, and accuracy of 0.999. Consequently, an artificial intelligence (AI) supported decision support system was created for pancreas segmentation.