[3D Slicer : Universal 3D visualization software]

An unsupervised learning model based on CT radiomics features accurately predicts axillary lymph node metastasis in breast cancer patients: diagnostic study

BACKGROUND

The accuracy of traditional clinical methods for assessing the metastatic status of axillary lymph nodes (ALNs) is unsatisfactory. In this study, the authors propose the use of radiomic technology and three-dimensional (3D) visualization technology to develop an unsupervised learning model for predicting axillary lymph node metastasis in patients with breast cancer (BC), aiming to provide a new method for clinical axillary lymph node assessment in patients with this disease.

METHODS

In this study, we retrospectively analyzed the data of 350 patients with invasive BC who underwent lung-enhanced computed tomography (CT) and axillary lymph node dissection surgery at the Department of Breast Surgery of the Second Xiangya Hospital of Central South University. The authors used 3D visualization technology to create a 3D atlas of ALNs and identified the region of interest for the lymph nodes. Radiomic features were subsequently extracted and selected, and a prediction model for ALNs was constructed using the K-means unsupervised algorithm. To validate the model, the authors prospectively collected data from 128 BC patients who were clinically evaluated as negative at our center.

RESULTS

Using 3D visualization technology, we extracted and selected a total of 36 CT radiomics features. The unsupervised learning model categorized 1737 unlabeled lymph nodes into two groups, and the analysis of the radiomic features between these groups indicated potential differences in lymph node status. Further validation with 1397 labeled lymph nodes demonstrated that the model had good predictive ability for axillary lymph node status, with an area under the curve of 0.847 (0.825-0.869). Additionally, the model's excellent predictive performance was confirmed in the 128 axillary clinical assessment negative cohort (cN0) and the 350 clinical assessment positive (cN+) cohort, for which the correct classification rates were 86.72 and 87.43%, respectively, which were significantly greater than those of clinical assessment methods.

CONCLUSIONS

The authors created an unsupervised learning model that accurately predicts the status of ALNs. This approach offers a novel solution for the precise assessment of ALNs in patients with BC.

Utilizing radiomics for differential diagnosis of inverted papilloma and chronic rhinosinusitis with polyps based on unenhanced CT scans

To evaluate whether radiomics models based on unenhanced paranasal sinuses CT images could be a useful tool for differentiating inverted papilloma (IP) from chronic rhinosinusitis with polyps (CRSwNP). This retrospective study recruited 240 patients with CRSwNP and 106 patients with IP from three centers. 253 patients from Qilu Hospital were randomly divided into the training set (n = 151) and the internal validation set (n = 102) with a ratio of 6:4. 93 patients from the other two centers were used as the external validation set. The patients with the unilateral disease (n = 115) from Qilu Hospital were selected to further develop a subgroup analysis. Lesion segmentation was manually delineated in CT images. Least absolute shrinkage and selection operator algorithm was performed for feature reduction and selection. Decision tree, support vector machine, random forest, and adaptive boosting regressor were employed to establish the differential diagnosis models. 43 radiomic features were selected for modeling. Among the models, RF achieved the best results, with an AUC of 0.998, 0.943, and 0.934 in the training set, the internal validation set, and the external validation set, respectively. In the subgroup analysis, RF achieved an AUC of 0.999 in the training set and 0.963 in the internal validation set. The proposed radiomics models offered a non-invasion and accurate differential approach between IP and CRSwNP and has some significance in guiding clinicians determining the best treatment plans, as well as predicting the prognosis.

Influence of build orientation and support structure on additive manufacturing of human knee replacements: a computational study

Developing patient-specific implants has an increasing interest in the application of emerging additive manufacturing (AM) technologies. On the other hand, despite advances in total knee replacement (TKR), studies suggest that up to 20% of patients with elective TKR are dissatisfied with the outcome. By creating 3D objects from digital models, AM enables the production of patient-specific implants with complex geometries, such as those required for knee replacements. Previous studies have highlighted concerns regarding the risk of residual stresses and shape distortions in AM parts, which could lead to structural failure or other complications. This article presents a computational framework that uses CT images to create patient-specific finite element models for optimizing AM knee replacements. The workflow includes image processing in the open-source software 3DSlicer and MeshLab and AM process simulations in the commercial platform 3DEXPERIENCE. The approach is demonstrated on a distal femur replacement for a 50-year-old male patient from the open-access Natural Knee Data. The results show that build orientations have a significant impact on both shape distortions and residual stresses. Support structures have a marginal effect on residual stresses but strongly influence shape distortions, whereas conical support exhibits a maximum distortion of 18.5 mm. Future research can explore how these factors affect the functionality of AM knee replacements under in-service loading.

Deep learning radiomics-based preoperative prediction of recurrence in chronic rhinosinusitis

Chronic rhinosinusitis (CRS) is characterized by poor prognosis and propensity for recurrence even after surgery. Identification of those CRS patients with high risk of relapse preoperatively will contribute to personalized treatment recommendations. In this paper, we proposed a multi-task deep learning network for sinus segmentation and CRS recurrence prediction simultaneously to develop and validate a deep learning radiomics-based nomogram for preoperatively predicting recurrence in CRS patients who needed surgical treatment. 265 paranasal sinuses computed tomography (CT) images of CRS from two independent medical centers were analyzed to build and test models. The sinus segmentation model achieved good segmentation results. Furthermore, the nomogram combining a deep learning signature and clinical factors also showed excellent recurrence prediction ability for CRS. Our study not only facilitates a technique for sinus segmentation but also provides a noninvasive method for preoperatively predicting recurrence in patients with CRS.

Whole-volume ADC histogram of the brain as an image biomarker in evaluating disease severity of neonatal hypoxic-ischemic encephalopathy

Purpose

To examine the diagnostic significance of the apparent diffusion coefficient (ADC) histogram in quantifying neonatal hypoxic ischemic encephalopathy (HIE).

Methods

An analysis was conducted on the MRI data of 90 HIE patients, 49 in the moderate-to-severe group, and the other in the mild group. The 3D Slicer software was adopted to delineate the whole brain region as the region of interest, and 22 ADC histogram parameters were obtained. The interobserver consistency of the two radiologists was assessed by the interclass correlation coefficient (ICC). The difference in parameters (ICC > 0.80) between the two groups was compared by performing the independent sample t-test or the Mann–Whitney U test. In addition, an investigation was conducted on the correlation between parameters and the neonatal behavioral neurological assessment (NBNA) score. The ROC curve was adopted to assess the efficacy of the respective significant parameters. Furthermore, the binary logistic regression was employed to screen out the independent risk factors for determining the severity of HIE.

Results

The ADCmean, ADCmin, ADCmax,10th−70th, 90th percentile of ADC values of the moderate-to-severe group were smaller than those of the mild group, while the group's variance, skewness, kurtosis, heterogeneity, and mode-value were higher than those of the mild group (P < 0.05). All the mentioned parameters, the ADCmean, ADCmin, and 10th−70th and 90th percentile of ADC displayed positive correlations with the NBNA score, mode-value and ADCmax displayed no correlations with the NBNA score, the rest showed negative correlations with the NBNA score (P < 0.05). The area under the curve (AUC) of variance was the largest (AUC = 0.977; cut-off 972.5, sensitivity 95.1%; specificity 87.8%). According to the logistic regression analysis, skewness, kurtosis, variance, and heterogeneity were independent risk factors for determining the severity of HIE (OR > 1, P < 0.05).

Conclusions

The ADC histogram contributes to the HIE diagnosis and is capable of indicating the diffusion information of the brain objectively and quantitatively. It refers to a vital method for assessing the severity of HIE.

Integration of Deep Learning and Active Shape Models for More Accurate Prostate Segmentation in 3D MR Images

Magnetic resonance imaging (MRI) has a growing role in the clinical workup of prostate cancer. However, manual three-dimensional (3D) segmentation of the prostate is a laborious and time-consuming task. In this scenario, the use of automated algorithms for prostate segmentation allows us to bypass the huge workload of physicians. In this work, we propose a fully automated hybrid approach for prostate gland segmentation in MR images using an initial segmentation of prostate volumes using a custom-made 3D deep network (VNet-T2), followed by refinement using an Active Shape Model (ASM). While the deep network focuses on three-dimensional spatial coherence of the shape, the ASM relies on local image information and this joint effort allows for improved segmentation of the organ contours. Our method is developed and tested on a dataset composed of T2-weighted (T2w) MRI prostatic volumes of 60 male patients. In the test set, the proposed method shows excellent segmentation performance, achieving a mean dice score and Hausdorff distance of 0.851 and 7.55 mm, respectively. In the future, this algorithm could serve as an enabling technology for the development of computer-aided systems for prostate cancer characterization in MR imaging.

Combination of Preoperative Multimodal Image Fusion and Intraoperative Dyna CT in Percutaneous Balloon Compression of Trigeminal Ganglion for Primary Trigeminal Neuralgia: Experience in 24 Patients

Objective

We retrospectively assessed the surgical results of PBC with preoperative multimodal image fusion and intraoperative Dyna Computed Tomography (CT) in 24 patients with primary trigeminal neuralgia (PTN) to explore a valuable aid for Percutaneous balloon compression (PBC).

Methods

We studied the data of 24 patients with PTN. All patients underwent PBC and were assessed with preoperative multimodal image fusion [computed tomography (CT) and magnetic resonance imaging (MRI)] and intraoperative Dyna CT in the Department of Neurosurgery of Zhuhai People’s Hospital between October 2020 and September 2021. Multimodal image fusion—three-dimensional (3D) reconstruction of CT and MRI data—was performed using 3D-Slicer software, and preoperative evaluation was performed according to the results of image fusion. Dyna CT was used to dynamically observe the position and shape of the metallic hollow introducer and Fogarty catheter and balloon during the operation to guide the operation in real time. We performed follow-up assessments each month and summarized the clinical characteristics, surgical effects, and complications in all patients.

Results

Surgery was successful for all patients; the patients reported immediate pain relief. Surgical complications included facial numbness in 24 patients (100%), mild masseter weakness in three (12.5%), herpes zoster in three (12.5%), and balloon rupture in one (4.2%). None of the patients had serious surgical complications. The mean follow-up time was 9.6 ± 2.7 months. During the follow-up period, 22 patients (91.7%) experienced no recurrence of pain, and two patients (8.3%) experienced recurrence of pain, of which one underwent secondary PBC surgery.

Conclusions

Preoperative multimodal image reconstruction can help fully evaluate PBC surgery, clarify the etiology, and predict the volume of contrast medium required during the operation. It provided important assistance for PBC treatment of trigeminal neuralgia patients when preoperative multimodal image fusion is combined with intraoperative Dyna CT.

A new method of preoperative assessment of correct electrode array alignment based on post-operative measurements in a cochlear implanted cohort

PURPOSE

During cochlear implantation surgery, a range of complications may occur such as tip fold-over. We recently developed a method to estimate the insertion orientation of the electrode array. The aim of the study was to determine the optimal angle of orientation in a cohort of cochlear implanted patients.

METHODS

On eighty-five CT scans (80 uncomplicated insertions and 5 cases with tip fold-over), location of the electrode array's Insertion Guide (IG), Orientation marker (OM) and two easily identifiable landmarks (the round window (RW) and the incus short process (ISP)) were manually marked. The angle enclosed by ISP-RW line and the Cochlear™ Slim Modiolar electrode array's OM line determined the electrode array insertion angle.

RESULTS

The average insertion angle was 45.0-47.2° ± 10.4-12° SD and was validated with 98% confidence interval. Based on the measurements obtained, patients' sex and age had no impact on the size of this angle. Although the angles of the tip fold-over cases (44.9°, 46.9°, 34.2°, 54.3°, 55.9°) fell within this average range, the further it diverted from the average it increased the likelihood for tip fold-over.

CONCLUSION

Electrode array insertion in the individually calculated angle relative to the visible incus short process provides a useful guide for the surgeon when aiming for the optimal angle, and potentially enhances good surgical outcomes. Our results show that factors other than the orientation angle may additionally contribute to failures in implantation when the Slim Modiolar electrode is used.

Rib fracture detection system based on deep learning

Rib fracture detection is time-consuming and demanding work for radiologists. This study aimed to introduce a novel rib fracture detection system based on deep learning which can help radiologists to diagnose rib fractures in chest computer tomography (CT) images conveniently and accurately. A total of 1707 patients were included in this study from a single center. We developed a novel rib fracture detection system on chest CT using a three-step algorithm. According to the examination time, 1507, 100 and 100 patients were allocated to the training set, the validation set and the testing set, respectively. Free Response ROC analysis was performed to evaluate the sensitivity and false positivity of the deep learning algorithm. Precision, recall, F1-score, negative predictive value (NPV) and detection and diagnosis were selected as evaluation metrics to compare the diagnostic efficiency of this system with radiologists. The radiologist-only study was used as a benchmark and the radiologist-model collaboration study was evaluated to assess the model's clinical applicability. A total of 50,170,399 blocks (fracture blocks, 91,574; normal blocks, 50,078,825) were labelled for training. The F1-score of the Rib Fracture Detection System was 0.890 and the precision, recall and NPV values were 0.869, 0.913 and 0.969, respectively. By interacting with this detection system, the F1-score of the junior and the experienced radiologists had improved from 0.796 to 0.925 and 0.889 to 0.970, respectively; the recall scores had increased from 0.693 to 0.920 and 0.853 to 0.972, respectively. On average, the diagnosis time of radiologist assisted with this detection system was reduced by 65.3 s. The constructed Rib Fracture Detection System has a comparable performance with the experienced radiologist and is readily available to automatically detect rib fracture in the clinical setting with high efficacy, which could reduce diagnosis time and radiologists' workload in the clinical practice.

Individualized Prediction of Colorectal Cancer Metastasis Using a Radiogenomics Approach

Objectives: To evaluate whether incorporating the radiomics, genomics, and clinical features allows prediction of metastasis in colorectal cancer (CRC) and to develop a preoperative nomogram for predicting metastasis.

Methods: We retrospectively analyzed radiomics features of computed tomography (CT) images in 134 patients (62 in the primary cohort, 28 in the validation cohort, and 44 in the independent-test cohort) clinicopathologically diagnosed with CRC at Dazhou Central Hospital from February 2018 to October 2019. Tumor tissues were collected from all patients for RNA sequencing, and clinical data were obtained from medical records. A total of 854 radiomics features were extracted from enhanced venous-phase CT of CRC. Least absolute shrinkage and selection operator regression analysis was utilized for data dimension reduction, feature screen, and radiomics signature development. Multivariable logistic regression analysis was performed to build a multiscale predicting model incorporating the radiomics, genomics, and clinical features. The receiver operating characteristic curve, calibration curve, and decision curve were conducted to evaluate the performance of the nomogram.

Results: The radiomics signature based on 16 selected radiomics features showed good performance in metastasis assessment in both primary [area under the curve (AUC) = 0.945, 95% confidence interval (CI) 0.892–0.998] and validation cohorts (AUC = 0.754, 95% CI 0.570–0.938). The multiscale nomogram model contained radiomics features signatures, four-gene expression related to cell cycle pathway, and CA 19-9 level. The multiscale model showed good discrimination performance in the primary cohort (AUC = 0.981, 95% CI 0.953–1.000), the validation cohort (AUC = 0.822, 95% CI 0.635–1.000), and the independent-test cohort (AUC = 0.752, 95% CI 0.608–0.896) and good calibration. Decision curve analysis confirmed the clinical application value of the multiscale model.

Conclusion: This study presented a multiscale model that incorporated the radiological eigenvalues, genomics features, and CA 19-9, which could be conveniently utilized to facilitate the individualized preoperatively assessing metastasis in CRC patients.

A Novel Model Based on CXCL8-Derived Radiomics for Prognosis Prediction in Colorectal Cancer

Introduction: Prognosis prediction is essential to improve therapeutic strategies and to achieve better clinical outcomes in colorectal cancer (CRC) patients. Radiomics based on high-throughput mining of quantitative medical imaging is an emerging field in recent years. However, the relationship among prognosis, radiomics features, and gene expression remains unknown. Methods: We retrospectively analyzed 141 patients (from study 1) diagnosed with CRC from February 2018 to October 2019 and randomly divided them into training (N = 99) and testing (N = 42) cohorts. Radiomics features in venous phase image were extracted from preoperative computed tomography (CT) images. Gene expression was detected by RNA-sequencing on tumor tissues. The least absolute shrinkage and selection operator (LASSO) regression model was used for selecting imaging features and building the radiomics model. A total of 45 CRC patients (study 2) with immunohistochemical (IHC) staining of CXCL8 diagnosed with CRC from January 2014 to October 2018 were included in the independent testing cohort. A clinical model was validated for prognosis prediction in prognostic testing cohort (163 CRC patients from 2014 to 2018, study 3). We performed a combined radiomics model that was composed of radiomics score, tumor stage, and CXCL8-derived radiomics model to make comparison with the clinical model. Results: In our study, we identified the CXCL8 as a hub gene in affecting prognosis, which is mainly through regulating cytokine-cytokine receptor interaction and neutrophil migration pathway. The radiomics model incorporated 12 radiomics features screened by LASSO according to CXCL8 expression in the training cohort and showed good performance in testing and IHC testing cohorts. Finally, the CXCL8-derived radiomics model combined with tumor stage performed high ability in predicting the prognosis of CRC patients in the prognostic testing cohort, with an area under the curve (AUC) of 0.774 [95% confidence interval (CI): 0.674-0.874]. Kaplan-Meier analysis of the overall survival probability in CRC patients stratified by combined model revealed that high-risk patients have a poor prognosis compared with low-risk patients (Log-rank P < 0.0001). Conclusion: We demonstrated that the radiomics model reflected by CXCL8 combined with tumor stage information is a reliable approach to predict the prognosis in CRC patients and has a potential ability in assisting clinical decision-making.