MIL normalization -- prerequisites for accurate MRI radiomics analysis

Anti-HER2 therapy response assessment for guiding treatment (de-)escalation in early HER2-positive breast cancer using a novel deep learning radiomics model

OBJECTIVES

Anti-HER2 targeted therapy significantly reduces risk of relapse in HER2 + breast cancer. New measures are needed for a precise risk stratification to guide (de-)escalation of anti-HER2 strategy.

METHODS

A total of 726 HER2 + cases who received no/single/dual anti-HER2 targeted therapies were split into three respective cohorts. A deep learning model (DeepTEPP) based on preoperative breast magnetic resonance (MR) was developed. Patients were scored and categorized into low-, moderate-, and high-risk groups. Recurrence-free survival (RFS) was compared in patients with different risk groups according to the anti-HER2 treatment they received, to validate the value of DeepTEPP in predicting treatment efficacy and guiding anti-HER2 strategy.

RESULTS

DeepTEPP was capable of risk stratification and guiding anti-HER2 treatment strategy: DeepTEPP-Low patients (60.5%) did not derive significant RFS benefit from trastuzumab (p = 0.144), proposing an anti-HER2 de-escalation. DeepTEPP-Moderate patients (19.8%) significantly benefited from trastuzumab (p = 0.048), but did not obtain additional improvements from pertuzumab (p = 0.125). DeepTEPP-High patients (19.7%) significantly benefited from dual HER2 blockade (p = 0.045), suggesting an anti-HER2 escalation.

CONCLUSIONS

DeepTEPP represents a pioneering MR-based deep learning model that enables the non-invasive prediction of adjuvant anti-HER2 effectiveness, thereby providing valuable guidance for anti-HER2 (de-)escalation strategies. DeepTEPP provides an important reference for choosing the appropriate individualized treatment in HER2 + breast cancer patients, warranting prospective validation.

CLINICAL RELEVANCE STATEMENT

We built an MR-based deep learning model DeepTEPP, which enables the non-invasive prediction of adjuvant anti-HER2 effectiveness, thus guiding anti-HER2 (de-)escalation strategies in early HER2-positive breast cancer patients.

KEY POINTS

• DeepTEPP is able to predict anti-HER2 effectiveness and to guide treatment (de-)escalation. • DeepTEPP demonstrated an impressive prognostic efficacy for recurrence-free survival and overall survival. • To our knowledge, this is one of the very few, also the largest study to test the efficacy of a deep learning model extracted from breast MR images on HER2-positive breast cancer survival and anti-HER2 therapy effectiveness prediction.

Rapid intraoperative multi-molecular diagnosis of glioma with ultrasound radio frequency signals and deep learning

BACKGROUND

Molecular diagnosis is crucial for biomarker-assisted glioma resection and management. However, some limitations of current molecular diagnostic techniques prevent their widespread use intraoperatively. With the unique advantages of ultrasound, this study developed a rapid intraoperative molecular diagnostic method based on ultrasound radio-frequency signals.

METHODS

We built a brain tumor ultrasound bank with 169 cases enrolled since July 2020, of which 43483 RF signal patches from 67 cases with a pathological diagnosis of glioma were a retrospective cohort for model training and validation. IDH1 and TERT promoter (TERTp) mutations and 1p/19q co-deletion were detected by next-generation sequencing. We designed a spatial-temporal integration model (STIM) to diagnose the three molecular biomarkers, thus establishing a rapid intraoperative molecular diagnostic system for glioma, and further analysed its consistency with the fifth edition of the WHO Classification of Tumors of the Central Nervous System (WHO CNS5). We tested STIM in 16-case prospective cohorts, which contained a total of 10384 RF signal patches. Two other RF-based classical models were used for comparison. Further, we included 20 cases additional prospective data for robustness test (ClinicalTrials.govNCT05656053).

FINDINGS

In the retrospective cohort, STIM achieved a mean accuracy and AUC of 0.9190 and 0.9650 (95% CI, 0.94-0.99) respectively for the three molecular biomarkers, with a total time of 3 s and a 96% match to WHO CNS5. In the prospective cohort, the diagnostic accuracy of STIM is 0.85 ± 0.04 (mean ± SD) for IDH1, 0.84 ± 0.05 for TERTp, and 0.88 ± 0.04 for 1p/19q. The AUC is 0.89 ± 0.02 (95% CI, 0.84-0.94) for IDH1, 0.80 ± 0.04 (95% CI, 0.71-0.89) for TERTp, and 0.85 ± 0.06 (95% CI, 0.73-0.98) for 1p/19q. Compared to the second best available method based on RF signal, the diagnostic accuracy of STIM is improved by 16.70% and the AUC is improved by 19.23% on average.

INTERPRETATION

STIM is a rapid, cost-effective, and easy-to-manipulate AI method to perform real-time intraoperative molecular diagnosis. In the future, it may help neurosurgeons designate personalized surgical plans and predict survival outcomes.

FUNDING

A full list of funding bodies that contributed to this study can be found in the Acknowledgements section.

Research on application of radiomics in glioma: a bibliometric and visual analysis

Background

With the continuous development of medical imaging informatics technology, radiomics has become a new and evolving field in medical applications. Radiomics aims to be an aid to support clinical decision making by extracting quantitative features from medical images and has a very wide range of applications. The purpose of this study was to perform a bibliometric and visual analysis of scientific results and research trends in the research application of radiomics in glioma.

Methods

We searched the Web of Science Core Collection (WOScc) for publications related to glioma radiomics. A bibliometric and visual analysis of online publications in this field related to countries/regions, authors, journals, references and keywords was performed using CiteSpace and R software.

Results

A total of 587 relevant literature published from 2012 to September 2022 were retrieved in WOScc, and finally a total of 484 publications were obtained according to the filtering criteria, including 393 (81.20%) articles and 91 (18.80%) reviews. The number of relevant publications increases year by year. The highest number of publications was from the USA (171 articles, 35.33%) and China (170 articles, 35.12%). The research institution with the highest number of publications was Chinese Acad Sci (24), followed by Univ Penn (22) and Fudan Univ (21). WANG Y (27) had the most publications, followed by LI Y (22), and WANG J (20). Among the 555 co-cited authors, LOUIS DN (207) and KICKINGEREDER P (207) were the most cited authors. FRONTIERS IN ONCOLOGY (42) was the most published journal and NEURO-ONCOLOGY (412) was the most co-cited journal. The most frequent keywords in all publications included glioblastoma (187), survival (136), classification (131), magnetic resonance imaging (113), machine learning (100), tumor (82), and feature (79), central nervous system (66), IDH (57), and radiomics (55). Cluster analysis was performed on the basis of keyword co-occurrence, and a total of 16 clusters were formed, indicating that these directions are the current hotspots of radiomics research applications in glioma and may be the future directions of continuous development.

Conclusion

In the past decade, radiomics has received much attention in the medical field and has been widely used in clinical research applications. Cooperation and communication between countries/regions need to be enhanced in future research to promote the development of radiomics in the field of medicine. In addition, the application of radiomics has improved the accuracy of pre-treatment diagnosis, efficacy prediction and prognosis assessment of glioma and helped to promote the development into precision medicine, the future still faces many challenges.

Performance of deep learning algorithms to distinguish high-grade glioma from low-grade glioma: A systematic review and meta-analysis

This study aims to evaluate deep learning (DL) performance in differentiating low- and high-grade glioma. Search online database for studies continuously published from 1st January 2015 until 16th August 2022. The random-effects model was used for synthesis, based on pooled sensitivity (SE), specificity (SP), and area under the curve (AUC). Heterogeneity was estimated using the Higgins inconsistency index (I²). 33 were ultimately included in the meta-analysis. The overall pooled SE and SP were 94% and 93%, with an AUC of 0.98. There was great heterogeneity in this field. Our evidence-based study shows DL achieves high accuracy in glioma grading. Subgroup analysis reveals several limitations in this field: 1) Diagnostic trials require standard method for data merging for AI; 2) small sample size; 3) poor-quality image preprocessing; 4) not standard algorithm development; 5) not standard data report; 6) different definition of HGG and LGG; and 7) poor extrapolation.

Toward deep learning replacement of gadolinium in neuro-oncology: A review of contrast-enhanced synthetic MRI

Gadolinium-based contrast agents (GBCAs) have become a crucial part of MRI acquisitions in neuro-oncology for the detection, characterization and monitoring of brain tumors. However, contrast-enhanced (CE) acquisitions not only raise safety concerns, but also lead to patient discomfort, the need of more skilled manpower and cost increase. Recently, several proposed deep learning works intend to reduce, or even eliminate, the need of GBCAs. This study reviews the published works related to the synthesis of CE images from low-dose and/or their native -non CE- counterparts. The data, type of neural network, and number of input modalities for each method are summarized as well as the evaluation methods. Based on this analysis, we discuss the main issues that these methods need to overcome in order to become suitable for their clinical usage. We also hypothesize some future trends that research on this topic may follow.

Correlation of T1- to T2-weighted signal intensity ratio with T1- and T2-relaxation time and IDH mutation status in glioma

The current study aimed to test whether the ratio of T1-weighted to T2-weighted signal intensity (T1W/T2W ratio: rT1/T2) derived from conventional MRI could act as a surrogate relaxation time predictive of IDH mutation status in histologically lower-grade gliomas. Strong exponential correlations were found between rT1/T2 and each of T1- and T2-relaxation times in eight subjects (rT1/T2 = 1.63exp^{-0.0005T1-relax} + 0.30 and rT1/T2 = 1.27exp^{-0.0081T2-relax} + 0.48; R² = 0.64 and 0.59, respectively). In a test cohort of 25 patients, mean rT1/T2 (mrT1/T2) was significantly higher in IDHwt tumors than in IDHmt tumors (p < 0.05) and the optimal cut-off of mrT1/T2 for discriminating IDHmt was 0.666-0.677, (AUC = 0.75, p < 0.05), which was validated in an external domestic cohort of 29 patients (AUC = 0.75, p = 0.02). However, this result was not validated in an external international cohort derived from TCIA/TCGA (AUC = 0.63, p = 0.08). The t-Distributed Stochastic Neighbor Embedding analysis revealed a greater diversity in image characteristics within the TCIA/TCGA cohort than in the two domestic cohorts. The failure of external validation in the TCIA/TCGA cohort could be attributed to its wider variety of original imaging characteristics.

Intensity standardization of MRI prior to radiomic feature extraction for artificial intelligence research in glioma-a systematic review

OBJECTIVES

Radiomics is a promising avenue in non-invasive characterisation of diffuse glioma. Clinical translation is hampered by lack of reproducibility across centres and difficulty in standardising image intensity in MRI datasets. The study aim was to perform a systematic review of different methods of MRI intensity standardisation prior to radiomic feature extraction.

METHODS

MEDLINE, EMBASE, and SCOPUS were searched for articles meeting the following eligibility criteria: MRI radiomic studies where one method of intensity normalisation was compared with another or no normalisation, and original research concerning patients diagnosed with diffuse gliomas. Using PRISMA criteria, data were extracted from short-listed studies including number of patients, MRI sequences, validation status, radiomics software, method of segmentation, and intensity standardisation. QUADAS-2 was used for quality appraisal.

RESULTS

After duplicate removal, 741 results were returned from database and reference searches and, from these, 12 papers were eligible. Due to a lack of common pre-processing and different analyses, a narrative synthesis was sought. Three different intensity standardisation techniques have been studied: histogram matching (5/12), limiting or rescaling signal intensity (8/12), and deep learning (1/12)-only two papers compared different methods. From these studies, histogram matching produced the more reliable features compared to other methods of altering MRI signal intensity.

CONCLUSION

Multiple methods of intensity standardisation have been described in the literature without clear consensus. Further research that directly compares different methods of intensity standardisation on glioma MRI datasets is required.

KEY POINTS

• Intensity standardisation is a key pre-processing step in the development of robust radiomic signatures to evaluate diffuse glioma. • A minority of studies compared the impact of two or more methods. • Further research is required to directly compare multiple methods of MRI intensity standardisation on glioma datasets.

Multi-task learning-based feature selection and classification models for glioblastoma and solitary brain metastases

Purpose

To investigate the diagnostic performance of feature selection via a multi-task learning model in distinguishing primary glioblastoma from solitary brain metastases.

Method

The study involved 187 patients diagnosed at Xiangya Hospital, Yunnan Provincial Cancer Hospital, and Southern Cancer Hospital between January 2010 and December 2018. Radiomic features were extracted from conventional magnetic resonance imaging including T1-weighted, T2-weighted, and contrast-enhanced T1-weighted sequences. We proposed a new multi-task learning model using these three sequences as three tasks. Multi-series fusion was performed to complement the information from different dimensions in order to enhance model robustness. Logical loss was used in the model as the data-fitting item, and the feature weights were expressed in the logical loss space as the sum of shared weights and private weights to select the common features of each task and the characteristics having an essential impact on a single task. A diagnostic model was constructed as a feature selection method as well as a classification method. We calculated accuracy, recall, precision, and area under the curve (AUC) and compared the performance of our new multi-task model with traditional diagnostic model performance.

Results

A diagnostic model combining the support vector machine algorithm as a classification algorithm and our model as a feature selection method had an average AUC of 0.993 in the training set, with AUC, accuracy, precision, and recall rates respectively of 0.992, 0.920, 0.969, and 0.871 in the test set. The diagnostic model built on our multi-task model alone, in the training set, had an average AUC of 0.987, and in the test set, the AUC, accuracy, precision, and recall rates were 0.984, 0.895, 0.954, and 0.838.

Conclusion

It is feasible to implement the multi-task learning model developed in our study using logistic regression to differentiate between glioblastoma and solitary brain metastases.

Phantom Study on the Robustness of MR Radiomics Features: Comparing the Applicability of 3D Printed and Biological Phantoms

The objectives of our study were to (a) evaluate the feasibility of using 3D printed phantoms in magnetic resonance imaging (MR) in assessing the robustness and repeatability of radiomic parameters and (b) to compare the results obtained from the 3D printed phantoms to metrics obtained in biological phantoms. To this end, three different 3D phantoms were printed: a Hilbert cube (5 × 5 × 5 cm³) and two cubic quick response (QR) code phantoms (a large phantom (large QR) (5 × 5 × 4 cm³) and a small phantom (small QR) (4 × 4 × 3 cm³)). All 3D printed and biological phantoms (kiwis, tomatoes, and onions) were scanned thrice on clinical 1.5 T and 3 T MR with 1 mm and 2 mm isotropic resolution. Subsequent analyses included analyses of several radiomics indices (RI), their repeatability and reliability were calculated using the coefficient of variation (CV), the relative percentage difference (RPD), and the interclass coefficient (ICC) parameters. Additionally, the readability of QR codes obtained from the MR images was examined with several mobile phones and algorithms. The best repeatability (CV ≤ 10%) is reported for the acquisition protocols with the highest spatial resolution. In general, the repeatability and reliability of RI were better in data obtained at 1.5 T (CV = 1.9) than at 3 T (CV = 2.11). Furthermore, we report good agreements between results obtained for the 3D phantoms and biological phantoms. Finally, analyses of the read-out rate of the QR code revealed better texture analyses for images with a spatial resolution of 1 mm than 2 mm. In conclusion, 3D printing techniques offer a unique solution to create textures for analyzing the reliability of radiomic data from MR scans.

Radiomics software comparison using digital phantom and patient data: IBSI-compliance does not guarantee concordance of feature values

Introduction Radiomics is a promising imaging-based tool which could enhance clinical observation and identify representative features. To avoid different interpretations, the Image Biomarker Standardisation Initiative (IBSI) imposed conditions for harmonisation. This study evaluates IBSI-compliant radiomics applications against a known benchmark and clinical datasets for agreements. Materials and methods The applications compared were RadiomiX Research Toolbox, LIFEx v7.0.0, and syngo.via Frontier Radiomics v1.2.5 (based on PyRadiomics v2.1). Basic assessment included comparing feature names and their formulas. The IBSI digital phantom was used for evaluation against reference values. For agreement evaluation (including same software but different versions), two clinical datasets were used: 27 contrast-enhanced computed tomography (CECT) of colorectal liver metastases and 39 magnetic resonance imaging (MRI) of breast cancer, including intravoxel incoherent motion and dynamic contrast-enhanced (DCE) MRI. The lower 95% confidence interval of intraclass correlation coefficient was used to define excellent agreement (>0.9). Results The three radiomics applications share 41 (3 shape, 8 intensity, 30 texture) out of 172, 84 and 110 features for RadiomiX, LIFEx and syngo.via, respectively, as well as wavelet filtering. The naming convention is, however, different between them. Syngo.via had excellent agreement with the IBSI benchmark, while LIFEx and RadiomiX showed slightly worse agreement. Excellent reproducibility was achieved for shape features only, while intensity and texture features varied considerably with the imaging type. For intensity, excellent agreement ranged from 46% for the DCE maps to 100% for CECT, while this lowered to 44% and 73% for texture features, respectively. Wavelet features produced the greatest variation between applications, with an excellent agreement for only 3% to 11% features. Conclusion Even with IBSI-compliance, the reproducibility of features between radiomics applications is not guaranteed. To evaluate variation, quality assurance of radiomics applications should be performed and repeated when updating to a new version or adding a new modality.

Synthetic MRI improves radiomics-based glioblastoma survival prediction

Glioblastoma is an aggressive and fast-growing brain tumor with poor prognosis. Predicting the expected survival of patients with glioblastoma is a key task for efficient treatment and surgery planning. Survival predictions could be enhanced by means of a radiomic system. However, these systems demand high numbers of multicontrast images, the acquisitions of which are time consuming, giving rise to patient discomfort and low healthcare system efficiency. Synthetic MRI could favor deployment of radiomic systems in the clinic by allowing practitioners not only to reduce acquisition time, but also to retrospectively complete databases or to replace artifacted images. In this work we analyze the replacement of an actually acquired MR weighted image by a synthesized version to predict survival of glioblastoma patients with a radiomic system. Each synthesized version was realistically generated from two acquired images with a deep learning synthetic MRI approach based on a convolutional neural network. Specifically, two weighted images were considered for the replacement one at a time, a T2w and a FLAIR, which were synthesized from the pairs T1w and FLAIR, and T1w and T2w, respectively. Furthermore, a radiomic system for survival prediction, which can classify patients into two groups (survival >480 days and ≤ 480 days), was built. Results show that the radiomic system fed with the synthesized image achieves similar performance compared with using the acquired one, and better performance than a model that does not include this image. Hence, our results confirm that synthetic MRI does add to glioblastoma survival prediction within a radiomics-based approach.

A Radiomics Approach to Assess High Risk Carotid Plaques: A Non-invasive Imaging Biomarker, Retrospective Study

Objective

This study aimed to construct a radiomics-based MRI sequence from high-resolution magnetic resonance imaging (HRMRI), combined with clinical high-risk factors for non-invasive differentiation of the plaque of symptomatic patients from asyptomatic patients.

Methods

A total of 115 patients were retrospectively recruited. HRMRI was performed, and patients were diagnosed with symptomatic plaques (SPs) and asymptomatic plaques (ASPs). Patients were randomly divided into training and test groups in the ratio of 7:3. T2WI was used for segmentation and extraction of the texture features. Max-Relevance and Min-Redundancy (mRMR) and least absolute shrinkage and selection operator (LASSO) were employed for the optimized model. Radscore was applied to construct a diagnostic model considering the T2WI texture features and patient demography to assess the power in differentiating SPs and ASPs.

Results

SPs and ASPs were seen in 75 and 40 patients, respectively. Thirty texture features were selected by mRMR, and LASSO identified a radscore of 16 radiomics features as being related to plaque vulnerability. The radscore, consisting of eight texture features, showed a better diagnostic performance than clinical information, both in the training (area under the curve [AUC], 0.923 vs. 0.713) and test groups (AUC, 0.989 vs. 0.735). The combination model of texture and clinical information had the best performance in assessing lesion vulnerability in both the training (AUC, 0.926) and test groups (AUC, 0.898).

Conclusion

This study demonstrated that HRMRI texture features provide incremental value for carotid atherosclerotic risk assessment.