Handcrafted MRI radiomics and machine learning: Classification of indeterminate solid adrenal lesions

Adrenal Mass Evaluation: Suspicious Radiological Signs of Malignancy

An adrenal mass discovered incidentally during imaging for unrelated clinical reasons is termed an "adrenal incidentaloma" (AI). AIs can be categorized as primary or metastatic, functioning or non-functioning, and benign or malignant. The primary goal of radiological evaluation is to exclude malignancy by differentiating between benign and malignant lesions. Most AIs are benign, with adenomas and macronodular bilateral adrenal hyperplasia being the most common types. Less common benign lesions include myelolipomas, pheochromocytomas, cysts, and hematomas. Malignant adrenal masses account for less than 10% of cases and often include metastases from other cancers or primary adrenal diseases, such as adrenocortical carcinoma and pheochromocytoma. Computed Tomography (CT) remains the gold standard for diagnosing adrenal incidentalomas, while Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) are utilized for indeterminate cases. Additionally, innovative imaging techniques such as texture analysis are gaining importance, as they can assess quantitative parameters that are not visible to the human eye. This review aims to provide an updated overview of malignant adrenal lesions on CT and MRI, emphasizing key imaging features suspicious for malignancy to aid in distinguishing between benign and malignant lesions. Furthermore, it highlights the growing role of radiomics as a supportive tool for radiologists.

Delta-Radiomics Using Machine Learning Classifiers With Auxiliary Data Sets to Predict Disease Progression During Magnetic Resonance-Guided Radiotherapy in Adrenal Metastases

PURPOSE

Adaptive radiotherapy accounts for interfractional anatomic changes. We hypothesize that changes in the gross tumor volumes identified during daily scans could be analyzed using delta-radiomics to predict disease progression events. We evaluated whether an auxiliary data set could improve prediction performance.

MATERIALS AND METHODS

We analyzed 108 patients (n = 90 internal; n = 18 external) who received ablative radiotherapy. The internal data set included 42 patients with adrenal cancer, 23 patients with lung cancer, and 25 patients with pancreatic cancer, with the clinical end point of progression-free survival events. The median dose was 50 Gy, which was delivered over five fractions. The delta features are the ratio of the features of the last to first treatment fraction, F5/F1, and the concatenation of the first and last fraction features, F1||F5. Decision tree classifier with and without auxiliary data sets, and the external data set was used exclusively for independent testing of the final models.

RESULTS

During internal training, for the F1||F5 model, the inclusion of the lung data set increased our AUC receiver operator characteristic curve (ROC) from 0.53 ± 0.12 to 0.61 ± 0.11, whereas the pancreatic data set increased our AUC-ROC to 0.60 ± 0.14. For the F5/F1 model, the inclusion of the lung auxiliary data increased our AUC-ROC from 0.52 ± 0.13 to 0.65 ± 0.11, whereas it modestly changed by 0.62 ± 0.13 with the pancreas. During external testing, for the F5/F1 model, we reported an AUC-ROC of 0.60 with the lung auxiliary data and 0.43 with the pancreatic data. Also, for the F5||F1 model, we reported an AUC-ROC of 0.70 with the lung auxiliary and 0.60 with the pancreatic data.

CONCLUSION

Decision trees provided an explainable model on the external data set. The validation of our model on an external data set may be the first step to biologically adapted radiotherapy recognizing radiomics signals for potential recurrence.

Comparing Texture Analysis of Apparent Diffusion Coefficient MRI in Hepatocellular Adenoma and Hepatocellular Carcinoma

AIM

This study aimed to assess the effectiveness of using MRI-apparent diffusion coefficient (ADC) map-driven radiomics to differentiate between hepatocellular adenoma (HCA) and hepatocellular carcinoma (HCC) features.

MATERIALS AND METHODS

The study involved 55 patients with liver tumors (20 with HCA and 35 with HCC), featuring 106 lesions equally distributed between hepatic carcinoma and hepatic adenoma who underwent texture analysis on ADC map MR images. The analysis identified several imaging features that significantly differed between the HCA and HCC groups. Four classification models were compared for distinguishing HCA from HCC including linear support vector machine (linear-SVM), radial basis function SVM (RBF-SVM), random forest (RF), and k-nearest neighbor (KNN).

RESULTS

The k-nearest neighbor (KNN) classifier displayed the top accuracy (0.89) and specificity (0.90). Linear-SVM and KNN classifiers showcased the leading sensitivity (0.88) for both, with the KNN classifier achieving the highest precision (0.9). In comparison, the conventional interpretation had lower sensitivity (70.1%) and specificity (77.9%).

CONCLUSION

The study found that utilizing ADC maps for texture analysis in MR images is a viable method to differentiate HCA from HCC, yielding promising results in identified texture features.

Role of artificial intelligence in the characterization of indeterminate pancreatic head mass and its usefulness in preoperative diagnosis

Artificial intelligence (AI) has been used in various fields of day-to-day life and its role in medicine is immense. Understanding of oncology has been improved with the introduction of AI which helps in diagnosis, treatment planning, management, prognosis, and follow-up. It also helps to identify high-risk groups who can be subjected to timely screening for early detection of malignant conditions. It is more important in pancreatic cancer as it is one of the major causes of cancer-related deaths worldwide and there are no specific early features (clinical and radiological) for diagnosis. With improvement in imaging modalities (computed tomography, magnetic resonance imaging, endoscopic ultrasound), most often clinicians were being challenged with lesions that were difficult to diagnose with human competence. AI has been used in various other branches of medicine to differentiate such indeterminate lesions including the thyroid gland, breast, lungs, liver, adrenal gland, kidney, etc. In the case of pancreatic cancer, the role of AI has been explored and is still ongoing. This review article will focus on how AI can be used to diagnose pancreatic cancer early or differentiate it from benign pancreatic lesions, therefore, management can be planned at an earlier stage.

An Integrated CT and MRI Imaging Model to Differentiate between Adrenal Adenomas and Pheochromocytomas

PURPOSE

to perform an external validation of our predictive model to rule out pheochromocytoma (PHEO) based on unenhanced CT in a cohort of patients with PHEOs and adenomas who underwent adrenalectomy.

METHODS

The predictive model was previously developed in a retrospective cohort of 1131 patients presenting with adrenal lesions. In the present study, we performed an external validation of the model in another cohort of 214 patients with available histopathological results.

RESULTS

For the external validation, 115 patients with PHEOs and 99 with adenomas were included. Our previously described predictive model combining the variables of high lipid content and tumor size in unenhanced CT (AUC-ROC: 0.961) had a lower diagnostic accuracy in our current study population for the prediction of PHEO (AUC: 0.750). However, when we excluded atypical adenomas (with Hounsfield units (HU) > 10, n = 39), the diagnostic accuracy increased to 87.4%. In addition, in the whole cohort (including atypical adenomas), when MRI information was included in the model, the diagnostic accuracy increased to up to 85% when the variables tumor size, high lipid content in an unenhanced CT scan, and hyperintensity in the T2 sequence in MRI were included. The probability of PHEO was <0.3% for adrenal lesions <20 mm with >10 HU and without hyperintensity in T2.

CONCLUSION

Our study confirms that our predictive model combining tumor size and lipid content has high reliability for the prediction of PHEO when atypical adrenal lesions are excluded. However, for atypical adrenal lesions with >10 HU in an unenhanced CT scan, MRI information is necessary for a proper exclusion of the PHEO diagnosis.

Prediction of 2-[18F]FDG PET-CT SUVmax for Adrenal Mass Characterization: A CT Radiomics Feasibility Study

BACKGROUND

Indeterminate adrenal masses (AM) pose a diagnostic challenge, and 2-[¹⁸F]FDG PET-CT serves as a problem-solving tool. Aim of this study was to investigate whether CT radiomics features could be used to predict the 2-[¹⁸F]FDG SUVmax of AM.

METHODS

Patients with AM on 2-[¹⁸F]FDG PET-CT scan were grouped based on iodine contrast injection as CT contrast-enhanced (CE) or CT unenhanced (NCE). Two-dimensional segmentations of AM were manually obtained by multiple operators on CT images. Image resampling and discretization (bin number = 16) were performed. 919 features were calculated using PyRadiomics. After scaling, unstable, redundant, and low variance features were discarded. Using linear regression and the Uniform Manifold Approximation and Projection technique, a CT radiomics synthetic value (RadSV) was obtained. The correlation between CT RadSV and 2-[¹⁸F]FDG SUVmax was assessed with Pearson test.

RESULTS

A total of 725 patients underwent PET-CT from April 2020 to April 2021. In 150 (21%) patients, a total of 179 AM (29 bilateral) were detected. Group CE consisted of 84 patients with 108 AM (size = 18.1 ± 4.9 mm) and Group NCE of 66 patients with 71 AM (size = 18.5 ± 3.8 mm). In both groups, 39 features were selected. No statisticallyf significant correlation between CT RadSV and 2-[¹⁸F]FDG SUVmax was found (Group CE, r = 0.18 and p = 0.058; Group NCE, r = 0.13 and p = 0.27).

CONCLUSIONS

It might not be feasible to predict 2-[¹⁸F]FDG SUVmax of AM using CT RadSV. Its role as a problem-solving tool for indeterminate AM remains fundamental.

A machine learning approach to distinguishing between non-functioning and autonomous cortisol secreting adrenal incidentaloma on magnetic resonance imaging using texture analysis

PURPOSE

To investigate the possibility of distinguishing between nonfunctioning adrenal incidentalomas (NFAI) and autonomous cortisol secreting adrenal incidentalomas (ACSAI) with a model created with magnetic resonance imaging (MRI)-based radiomics and clinical features.

METHODS

In this study, 100 adrenal lesions were evaluated. The lesions were segmented on unenhanced T1-weighted in-phase (IP) and opposed-phase (OP) as well as on T2-weighted (T2-W) 3Tesla MRIs. The LASSO regression model was used to select potential predictors from 108 texture features for each sequence. Subsequently, a combined radiomics score and clinical features were created and compared.

RESULTS

A significant difference was found between median rad-scores for ACSAI and NFAI in training and test sets (p < 0.05 for all sequences). Multivariate logistic regression analysis revealed that the length of the tumor (OR = 1.09, p = 0.007) was an independent risk factor related to ACSAI. Multivariate logistic regression analysis was used for building clinical-radiomics (combined) models. The Op, IP, and IP plus T2-W model had a higher performance with area under curve (AUC) 0.758, 0.746, and 0.721 on the test dataset, respectively.

CONCLUSION

ACSAI can be distinguished from NFAI with high accuracy on unenhanced MRI. Radiomics analysis and the model constructed by machine learning algorithms seem superior to another radiologic assessment method. The inclusion of chemical shift MRI and the length of the tumor in the radiomics model could increase the power of the test.

Automated Adrenal Gland Disease Classes Using Patch-Based Center Symmetric Local Binary Pattern Technique with CT Images

Incidental adrenal masses are seen in 5% of abdominal computed tomography (CT) examinations. Accurate discrimination of the possible differential diagnoses has important therapeutic and prognostic significance. A new handcrafted machine learning method has been developed for the automated and accurate classification of adrenal gland CT images. A new dataset comprising 759 adrenal gland CT image slices from 96 subjects were analyzed. Experts had labeled the collected images into four classes: normal, pheochromocytoma, lipid-poor adenoma, and metastasis. The images were preprocessed, resized, and the image features were extracted using the center symmetric local binary pattern (CS-LBP) method. CT images were next divided into 16 × 16 fixed-size patches, and further feature extraction using CS-LBP was performed on these patches. Next, extracted features were selected using neighborhood component analysis (NCA) to obtain the most meaningful ones for downstream classification. Finally, the selected features were classified using k-nearest neighbor (kNN), support vector machine (SVM), and neural network (NN) classifiers to obtain the optimum performing model. Our proposed method obtained an accuracy of 99.87%, 99.21%, and 98.81% with kNN, SVM, and NN classifiers, respectively. Hence, the kNN classifier yielded the highest classification results with no pathological image misclassified as normal. Our developed fixed patch CS-LBP-based automatic classification of adrenal gland pathologies on CT images is highly accurate and has low time complexity [Formula: see text]. It has the potential to be used for screening of adrenal gland disease classes with CT images.

Artificial intelligence in adrenal imaging: A critical review of current applications

In the elective field of adrenal imaging, artificial intelligence (AI) can be used for adrenal lesion detection, characterization, hypersecreting syndrome management and patient follow-up. Although a perfect AI tool that includes all required steps from detection to analysis does not exist yet, multiple AI algorithms have been developed and tested with encouraging results. However, AI in this setting is still at an early stage. In this regard, most published studies about AI in adrenal gland imaging report preliminary results that do not have yet daily applications in clinical practice. In this review, recent developments and current results of AI in the field of adrenal imaging are presented. Limitations and future perspectives of AI are discussed.

Clinical, pathophysiologic, genetic and therapeutic progress in Primary Bilateral Macronodular Adrenal Hyperplasia

Patients with primary bilateral macronodular adrenal hyperplasia (PBMAH) usually present bilateral benign adrenocortical macronodules at imaging and variable levels of cortisol excess. PBMAH is a rare cause of primary overt Cushing's syndrome, but may represent up to one third of bilateral adrenal incidentalomas with evidence of cortisol excess. The increased steroidogenesis in PBMAH is often regulated by various G-protein coupled receptors aberrantly expressed in PBMAH tissues; some receptor ligands are ectopically produced in PBMAH tissues creating aberrant autocrine/paracrine regulation of steroidogenesis. The bilateral nature of PBMAH and familial aggregation, led to the identification of germline heterozygous inactivating mutations of the ARMC5 gene, in 20-25% of the apparent sporadic cases and more frequently in familial cases; ARMC5 mutations/pathogenic variants can be associated with meningiomas. More recently, combined germline mutations/pathogenic variants and somatic events inactivating the KDM1A gene were specifically identified in patients affected by GIP-dependent PBMAH. Functional studies demonstrated that inactivation of KDM1A leads to GIP-receptor (GIPR) overexpression and over or down-regulation of other GPCRs. Genetic analysis is now available for early detection of family members of index cases with PBMAH carrying identified germline pathogenic variants. Detailed biochemical, imaging, and co-morbidities assessment of the nature and severity of PBMAH is essential for its management. Treatment is reserved for patients with overt or mild cortisol/aldosterone or other steroid excesses taking in account co-morbidities. It previously relied on bilateral adrenalectomy; however recent studies tend to favor unilateral adrenalectomy, or less frequently, medical treatment with cortisol synthesis inhibitors or specific blockers of aberrant GPCR.

Diagnostic performance of radiomics in adrenal masses: A systematic review and meta-analysis

Objectives

(1) To assess the methodological quality and risk of bias of radiomics studies investigating the diagnostic performance in adrenal masses and (2) to determine the potential diagnostic value of radiomics in adrenal tumors by quantitative analysis.

Methods

PubMed, Embase, Web of Science, and Cochrane Library databases were searched for eligible literature. Methodological quality and risk of bias in the included studies were assessed by the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) and Radiomics Quality Score (RQS). The diagnostic performance was evaluated by pooled sensitivity, specificity, diagnostic odds ratio (DOR), and area under the curve (AUC). Spearman’s correlation coefficient and subgroup analysis were used to investigate the cause of heterogeneity. Publication bias was examined using the Deeks’ funnel plot.

Results

Twenty-eight studies investigating the diagnostic performance of radiomics in adrenal tumors were identified, with a total of 3579 samples. The average RQS was 5.11 (14.2% of total) with an acceptable inter-rater agreement (ICC 0.94, 95% CI 0.93–0.95). The risk of bias was moderate according to the result of QUADAS-2. Nine studies investigating the use of CT-based radiomics in differentiating malignant from benign adrenal tumors were included in the quantitative analysis. The pooled sensitivity, specificity, DOR and AUC with 95% confidence intervals were 0.80 (0.68-0.88), 0.83 (0.73-0.90), 19.06 (7.87-46.19) and 0.88 (0.85–0.91), respectively. There was significant heterogeneity among the included studies but no threshold effect in the meta-analysis. The result of subgroup analysis demonstrated that radiomics based on unenhanced and contrast-enhanced CT possessed higher diagnostic performance, and second-order or higher-order features could enhance the diagnostic sensitivity but also increase the false positive rate. No significant difference in diagnostic ability was observed between studies with machine learning and those without.

Conclusions

The methodological quality and risk of bias of studies investigating the diagnostic performance of radiomics in adrenal tumors should be further improved in the future. CT-based radiomics has the potential benefits in differentiating malignant from benign adrenal tumors. The heterogeneity between the included studies was a major limitation to obtaining more accurate conclusions.

Systematic Review Registration

https://www.crd.york.ac.uk/PROSPERO/ CRD 42022331999 .

The stability of oncologic MRI radiomic features and the potential role of deep learning: a review

The use of MRI radiomic models for the diagnosis, prognosis and treatment response prediction of tumors has been increasingly reported in literature. However, its widespread adoption in clinics is hampered by issues related to features stability. In the MRI radiomic workflow, the main factors that affect radiomic features computation can be found in the image acquisition and reconstruction phase, in the image pre-processing steps, and in the segmentation of the region of interest on which radiomic indices are extracted. Deep Neural Networks (DNNs), having shown their potentiality in the medical image processing and analysis field, can be seen as an attractive strategy to partially overcome the issues related to radiomic stability and mitigate their impact. In fact, DNN approaches can be prospectively integrated in the MRI radiomic workflow to improve image quality, obtain accurate and reproducible segmentations and generate standardized images. In this review, DNN methods that can be included in the image processing steps of the radiomic workflow are described and discussed, in the light of a detailed analysis of the literature in the context of MRI radiomic reliability.

Radiomics in Cross-Sectional Adrenal Imaging: A Systematic Review and Quality Assessment Study

In this study, we aimed to systematically review the current literature on radiomics applied to cross-sectional adrenal imaging and assess its methodological quality. Scopus, PubMed and Web of Science were searched to identify original research articles investigating radiomics applications on cross-sectional adrenal imaging (search end date February 2021). For qualitative synthesis, details regarding study design, aim, sample size and imaging modality were recorded as well as those regarding the radiomics pipeline (e.g., segmentation and feature extraction strategy). The methodological quality of each study was evaluated using the radiomics quality score (RQS). After duplicate removal and selection criteria application, 25 full-text articles were included and evaluated. All were retrospective studies, mostly based on CT images (17/25, 68%), with manual (19/25, 76%) and two-dimensional segmentation (13/25, 52%) being preferred. Machine learning was paired to radiomics in about half of the studies (12/25, 48%). The median total and percentage RQS scores were 2 (interquartile range, IQR = -5-8) and 6% (IQR = 0-22%), respectively. The highest and lowest scores registered were 12/36 (33%) and -5/36 (0%). The most critical issues were the absence of proper feature selection, the lack of appropriate model validation and poor data openness. The methodological quality of radiomics studies on adrenal cross-sectional imaging is heterogeneous and lower than desirable. Efforts toward building higher quality evidence are essential to facilitate the future translation into clinical practice.

Adrenal Mass Characterization in the Era of Quantitative Imaging: State of the Art

Simple Summary

Non-invasive characterization of adrenal lesions requires a rigorous approach. Although CT is the cornerstone of adrenal lesion characterization, a multimodality multiparametric imaging approach helps improve confidence in adrenal lesion characterization.

Abstract

Detection and characterization of adrenal lesions have evolved during the past two decades. Although the role of imaging in adrenal lesions associated with hormonal secretion is usually straightforward, characterization of non-functioning adrenal lesions may be challenging to confidently identify those that need to be resected. Although many adrenal lesions can be readily diagnosed when they display typical imaging features, the diagnosis may be challenging for atypical lesions. Computed tomography (CT) remains the cornerstone of adrenal imaging, but other morphological or functional modalities can be used in combination to reach a diagnosis and avoid useless biopsy or surgery. Early- and delayed-phase contrast-enhanced CT images are essential for diagnosing lipid-poor adenoma. Ongoing studies are evaluating the capabilities of dual-energy CT to provide valid virtual non-contrast attenuation and iodine density measurements from contrast-enhanced examinations. Adrenal lesions with attenuation values between 10 and 30 Hounsfield units (HU) on unenhanced CT can be characterized by MRI when iodinated contrast material injection cannot be performed. ¹⁸F-FDG PET/CT helps differentiate between atypical benign and malignant adrenal lesions, with the adrenal-to-liver maximum standardized uptake value ratio being the most discriminative variable. Recent studies evaluating the capabilities of radiomics and artificial intelligence have shown encouraging results.

Qualitative Heterogeneous Signal Drop on Chemical Shift (CS) MR Imaging: Correlative Quantitative Analysis between CS Signal Intensity Index and Contrast Washout Parameters Using T1-Weighted Sequences

The aim of this study was to calculate MRI quantitative parameters extracted from chemical-shift (CS) and dynamic contrast-enhanced (DCE) T1-weighted (T1-WS) images of adrenal lesions (AL) with qualitative heterogeneous signal drop on CS T1-WS and compare them to those of AL with homogeneous or no signal drop on CS T1-WS. On 3 T MRI, 65 patients with a total of 72 AL were studied. CS images were qualitatively assessed for grouping AL as showing homogeneous (Group 1, n = 19), heterogeneous (Group 2, n = 23), and no (Group 3, n = 30) signal drop. Histopathology or follow-up data served as reference standard to classify AL. ROIs were drawn both on CS and DCE images to obtain adrenal CS signal intensity index (ASII), absolute (AWO), and relative washout (RWO) values. Quantitative parameters (QP) were compared with ANOVA analysis and post hoc Dunn's test. The performance of QP to classify AL was assessed with receiver operating characteristic analysis. CS ASII values were significantly different among the three groups (p < 0.001) with median values of 71%, 53%, and 3%, respectively. AWO/RWO values were similar in Groups 1 (adenomas) and 2 (benign AL) but significantly (p < 0.001) lower in Group 3 (20 benign AL and 10 malignant AL). With cut-offs, respectively, of 60% (Group 1 vs. 2), 20% (Group 2 vs. 3), and 37% (Group 1 vs. 3), CS ASII showed areas under the curve of 0.85, 0.96, and 0.93 for the classification of AL, overall higher than AWO/RWO. In conclusion, AL with qualitative heterogeneous signal drop at CS represent benign AL with QP by DCE sequence similar to those of AL with homogeneous signal drop at CS, but different to those of AL with no signal drop at CS; ASII seems to be the only quantitative parameter able to differentiate AL among the three different groups.