Next Article in Journal
Beyond the Prognostic Value of 2-[18F]FDG PET/CT in Prostate Cancer: A Case Series and Literature Review Focusing on the Diagnostic Value and Impact on Patient Management
Next Article in Special Issue
An Explainable Machine Learning Pipeline for Stroke Prediction on Imbalanced Data
Previous Article in Journal
The Composition and Physical Properties of Clots in COVID-19 Pathology
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Systematic Review

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

Department of Advanced Biomedical Sciences, University of Naples “Federico II”, 80131 Naples, Italy
Department of Clinical Medicine and Surgery, University of Naples “Federico II”, 80131 Naples, Italy
Interdepartmental Research Center on Management and Innovation in Healthcare-CIRMIS, University of Naples “Federico II”, 80100 Naples, Italy
Laboratory of Augmented Reality for Health Monitoring (ARHeMLab), Department of Electrical Engineering and Information Technology, University of Naples “Federico II”, 80100 Naples, Italy
Institute of Biostructures and Bioimaging of the National Research Council, 80131 Naples, Italy
Author to whom correspondence should be addressed.
Diagnostics 2022, 12(3), 578;
Received: 3 February 2022 / Revised: 19 February 2022 / Accepted: 21 February 2022 / Published: 24 February 2022


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.

1. Introduction

Adrenal glands can be affected by a great variety of different disorders, including, but not limited to, acute illnesses, functional abnormalities, infectious diseases, and oncological processes, with diagnostic imaging playing a relevant role in most cases [1,2,3]. Radiologists are currently required to guide the management of incidental adrenal masses, whose characterization and differential diagnosis is complex and often challenging [4,5]. Indeed, while most adrenal lesions are benign (i.e., adrenal adenomas), there are several less-common entities that cannot be neglected (e.g., pheochromocytomas, metastases, adrenocortical carcinoma, lymphoma) [6,7]. Furthermore, adrenal adenomas can occasionally present atypical features due to the paucity of fat content, making it difficult to qualitatively differentiate them from malignancies [8,9]. In the effort to overcome the limitations of conventional image assessment and further increase the value of diagnostic imaging, a complex multi-step post-processing approach defined as radiomics has been proposed [10]. The premise behind this rapidly growing field is that medical images can be mined to extract quantitative data possibly related to the pathophysiology of a biological tissue [11]. These quantitative parameters, known as radiomics features, can then be used either alone or in combination with clinical information to build machine learning-based decision support tools and predictive models to aid physicians in the management of patients [12,13]. However, radiomics requires a robust workflow based on standardized, rigorous methods which are required to ensure the reliability and generalizability of results [14,15,16,17]. Using the radiomics quality score (RQS), a scoring system proposed by Lambin and colleagues and designed to assess the crucial steps in radiomics pipelines, previous investigations have found that radiomics studies are heterogeneous in terms of quality [18]. The aim of this study was to systematically review radiomics studies focused on adrenal disorders applications and to assess their quality using the RQS.

2. Materials and Methods

The present review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [19] and the review protocol was registered on the International Prospective Register of Systematic Reviews (PROSPERO, registration number = CRD42021248536).

2.1. Search Strategy and Eligibility Criteria

A systematic search of published articles investigating radiomics applications to adrenal cross-sectional imaging was performed by two investigators on three major digital databases (PubMed, Scopus and Web of Science), with the end date set to 25 February 2021. The search strategy was based on the following terms and functions: “textural” OR “radiomics” OR “texture” OR “histogram” AND “adrenal” AND “computed tomography” OR “CT” OR “magnetic resonance” OR “MRI” OR “MR”. Duplicates were removed prior to screening titles and abstracts to identify original research dealing with the topic of interest, published in English and involving human subjects.

2.2. Data Collection and Study Evaluation

For a qualitative synthesis, the main characteristics of included studies were extracted from the full texts. In particular, first authorship, country and year of publication were registered. Then, details regarding study design (i.e., retrospective or prospective), general aim (i.e., differential diagnosis of adrenal masses, detection of adrenal diseases, evaluation of prognosis or characterization of adrenal disorders), mean goal (i.e., the specific purpose of the study), sample size (i.e., number of included patients and adrenal lesions) and imaging modality assessed were recorded. Thirdly, imaging analysis methods such as segmentation strategy, feature extraction software, extracted feature category and information about machine learning applications were noted. Finally, it was registered whether the study reported positive or negative results.
To assess the methodological quality of each radiomics study, two readers with previous experience in radiomics and the RQS (AS and RC) worked in consensus to evaluate the papers and thus calculate the score, with a third operator (VR) intervening to solve disagreements. The RQS consists of 16 items that explore crucial steps of a radiomics pipeline, including but not limited to imaging protocol, features extraction and handling, clinical relevance, and data openness [18]. Each of these has a different weight and can contribute positively or negatively in terms of points attributed, with −8 being the minimum and 36 being the maximum score that can be reached. The absolute score is then converted to a percentage value (with 36 = 100%). More details on each of the RQS items can be found in Table S1.

2.3. Statistical Analysis

Categorical data are presented as counts and percentages. The Shapiro-Wilk test was performed to verify whether the distribution of continuous variables was normal, and these variables are presented as either mean and standard deviation or median and interquartile range (IQR) according to the test results, as appropriate. The mode among evaluated studies was calculated for each RQS item. All analyses were conducted using the “stats” (v3.6.2) R package (v4.0.5).

3. Results

3.1. Literature Search

The PRISMA flowchart for this study can be found in Figure 1.
Briefly, 235 records were identified, of which 105 were duplicates. After applying the selection criteria during the title and abstract screening, an additional 105 were excluded; therefore, 25 eligible articles were finally included in the present systematic review.

3.2. Qualitative Synthesis of Included Studies

The main characteristics of the included articles are presented in Table 1.
The study design was retrospective in all cases. Out of 25, eight studies (32%) included less than 50 patients. Regarding study aim, most were focused on the differential diagnosis of adrenal masses (18/25, 72%) while only 3/25 (12%) [24,39,40] and 2/25 (8%) [21,22], respectively, investigated the role of radiomics in the evaluation of prognosis and characterization of adrenal disorders. A single article (4%) was aimed at the detection of adrenal disease (primary aldosteronism) [20] while one (4%) had multiple aims [36]. The number of included studies per publication year is illustrated in Figure 2.
CT was the most represented imaging modality (16/25, 64%), with more than half CT radiomics studies using contrast-enhanced images (11/16, 69%). A total of 4/25 (16%) MRI radiomics studies were identified, while a single article (4%) evaluated both MRI and CT images [26]. Finally, 4/25 (16%) were 2-[18F]FDG PET/CT studies. As for the segmentation strategy, regions of interest were predominantly annotated manually on medical images (19/25, 76%), being two-dimensional in most cases (13/25, 52%). Machine learning approaches to handle radiomics features were embraced in 12 out of 25 studies (48%). Table S2 illustrates the main characteristics of the machine learning algorithms listed in Table 1 is provided. Only one study (4%) reported negative results [40].

3.3. RQS Assessment

The results of the RQS assessment performed in this systematic review are reported in Table 2.
The median total score was 2 (IQR = −5–8), corresponding to a percentage of 6% (IQR = 0–22%), with the highest and lowest scores registered corresponding respectively to 12/36 (33%) and −5/36 (0%). Figure 3 presents the distribution of RQS percentage scores in the reviewed papers while Figure 4 shows its median value by year.
Swarm plots depicting the distribution of RQS according to radiomics feature type as well as imaging modality can be found respectively in Figures S1 and S2, while the mode for each single item can be found in Table S1. Among the evaluated articles, 5/25 (20%) did not provide details regarding the imaging protocol to allow reproducibility. Most studies (18/25, 72%) did not test radiomics features robustness to segmentation variabilities. Among the included studies, radiomics features robustness testing for scanner and temporal variabilities was never performed. Feature reduction or adjustment for multiple testing was performed in 16/25 (62%) studies. Non-radiomics features were considered in 4/25 (16%) studies. Validation was missing in more than half of the included studies (13/25, 52%). Only one study performed calibration statistics (4%) [42]. The formal assessment of radiomics models’ clinical utility was never presented; similarly, there were no studies in which a cost-effectiveness analysis was performed. Finally, only one point was assigned among all studies for item 16 (open science and data) [36].

4. Discussion

Beyond the growing interest in radiomics and its many promises, the clinical translation of actual real-life applications must be grounded in methodologically rigorous studies and high-quality evidence. Unfortunately, previous investigations have found that the overall quality of radiomics studies is heterogenous and lower than desirable, even when exclusively focusing on works published on high-ranking scientific journals [15,45]. In this systematic review, we confirm that radiomics applications on cross-sectional adrenal imaging do not represent an exception. Indeed, the average and highest RQS scores cannot be considered satisfactory and indicate a lack in terms of methodological quality. Furthermore, our findings do not reflect a trend of quality increase in time. While “how to” guides have been published and efforts towards standardization of practice in radiomics are being carried on, most are relatively recent and therefore it might take some time before researchers begin aligning more strictly to these recommendations [46,47,48]. The evidence gathered in our systematic review suggests that three main issues might be the root cause for the overall low RQS results, namely the lack of prospectively designed studies, the absence of proper validation and the poor data openness. Indeed, a hypothetical study validating a previously published radiomics signature in a prospective setting with data (i.e., images, segmentations and code) freely available could reach a provisional score of 15/36 (42%) with just three RQS items (thus without considering additional points to be assigned). On one hand, this highlights a certain degree of strictness for the RQS system, since about half of the points are linked to three of the sixteen items. On the other hand, it should be acknowledged that studies with those characteristics could indeed have the potential to significantly contribute to advancing the field and may be considered a requirement prior to clinical translation of these models. Nevertheless, the prospective validation of previously published models can only be possible if exploratory and pilot investigations (typically retrospective in design) have been previously carried out. Thus, the actual value of the papers assessed in this systematic review may go beyond the mere score, provided that they represent a steppingstone for more robust trials. Our findings indicate that we could be at a crossroads: research group efforts should either be directed towards the validation of these preliminary studies or increase the methodological quality of works proposing new radiomics models. The issue of open data in radiomics studies has also been highlighted in other settings [49,50]. A possible solution to the problem might be represented by publicly available datasets, which could increase reproducibility and transparency but should be subject to rigorous quality controls themselves since it has been found that they might not be perfect [51,52]. However, generalizability remains one of the main endpoints and all those strategies increasing repeatability and reproducibility in radiomics are worthy of consideration [11].
The total number of included studies in this systematic review appears lower when compared to similar investigations focused on more prevalent and clinically relevant disorders, such as radiomics applications on prostate or pancreatic cancer [53,54]. However, it is in line with those reported in the evaluation of less frequent diseases (e.g., osteosarcoma) [55]. This is probably due to the fact that larger imaging datasets are more frequently available for those abnormalities which require a higher workload for radiology departments. Furthermore, the efforts of research groups are obviously driven by the perceived clinical needs to be addressed and their relevance and urgency. Nevertheless, the number of articles retrieved in each of the last three years is higher compared to the number of studies published in both 2016 and 2017, suggesting a possible trend in the interest toward adrenal cross-sectional imaging radiomics that is coherent with the rising enthusiasm on radiomics in general. However, this trend is not as strong when compared to the field of radiomics taken as a whole, which has shown a yearly growth rate of over 170% [56]. Unsurprisingly, almost the entirety of studies was focused on oncologic disorders, with a single exception [20]. Indeed, the vast majority of adrenal imaging exams are requested to investigate adrenal masses and the main field of application for radiomics is oncologic imaging [57].
Nearly half of the included studies used radiomics features to train machine learning algorithms and build classifiers for the desired task. While artificial intelligence certainly represents a valid choice to handle massive amounts of data compared with more traditional approaches (e.g., traditional statistics), it can prove complex and inscrutable when it comes to gaining clear insights on what processes are behind a certain result [58,59]. If we also consider that the sample size in the included studies was often limited, it appears understandable and agreeable that many researchers did not choose to employ artificial intelligence algorithms in their pipelines. Accordingly, the RQS does not seem to penalize or favor an approach over the other.
Segmentation remains a challenging and crucial step in radiomics pipeline, mainly because of reproducibility concerns [11]. Ideally, fast automated segmentation of the whole volume of interest would represent the optimal solution to extract representative radiomics data without time-consuming hard manual labor. However, in this systematic review, we found that two-dimensional image annotations and manual segmentation strategies were slightly preferred. Studies regarding the potential impact of segmentation strategy in cross-sectional adrenal imaging could help guide future research and reduce heterogeneity in the literature.
It is interesting to note that merely one study among those included in this review presented negative findings [40]. Publishing negative results is of paramount importance, especially in an emerging field such as radiomics, but it is often difficult and our findings seem to confirm this well-known issue [56]; scientific rigor should be the main criteria for publishing, regardless of the outcome, and it could be speculated that we might be currently missing out on high quality research for a wrong reason [60].
The present study suffers from some limitations that deserve to be discussed. Firstly, only recognized databases were searched and therefore the grey literature data may be missing from our evaluation. While this might mean that some possibly relevant studies were not assessed, it should be considered that there are no standardized methods to search the grey literature and it is often difficult to verify the reliability of its findings [61]. Therefore, we believe that our search strategy allowed us to balance comprehensiveness and accuracy, while making it easier to reproduce our results. Another limitation is that the body of the literature is small and highly heterogeneous; thus, we were not able to offer a deeper formal insight performing sub-group analyses. Nevertheless, in future endeavors to update the current systematic review, it can be hypothesized that this issue will be solved.
In conclusion, 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 ensure that a valid and robust radiomics pipeline could in the future be translated into support decision tools to increase the value of adrenal imaging. The path is still long for research investigating adrenal radiomics applications, but it might be worth exploring.

Supplementary Materials

The following supporting information can be downloaded at:, Table S1: Radiomics quality score items and their respective scores as described by Lambin et al.; Table S2: Basic principles of the most adopted machine learning algorithms in adrenal imaging; Figure S1. Swarm plot of the RQS distribution according to imaging modality; Figure S2. Swarm plot of the RQS distribution according to feature type.

Author Contributions

Conceptualization, A.S., S.M. and R.C.; methodology, A.S. and R.G.; software A.S. and R.C.; validation, A.S., V.R. and R.G.; formal analysis, A.S.; investigation, R.C.; data curation, V.R. and F.V.; writing—original draft preparation, A.S.; writing—review and editing, A.B. and S.M.; visualization, F.V. and P.P.M.; supervision, R.C., A.B. and S.M. All authors have read and agreed to the published version of the manuscript.


This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are contained within the article or supplementary material.

Conflicts of Interest

The authors declare no conflict of interest.


  1. Udare, A.; Agarwal, M.; Siegelman, E.; Schieda, N. CT and MR imaging of acute adrenal disorders. Abdom. Radiol. 2021, 46, 290–302. [Google Scholar] [CrossRef] [PubMed]
  2. Karaosmanoglu, A.D.; Onder, O.; Leblebici, C.B.; Sokmensuer, C.; Akata, D.; Ozmen, M.N.; Karcaaltincaba, M. Cross-sectional imaging features of unusual adrenal lesions: A radiopathological correlation. Abdom. Radiol. 2021, 46, 3974–3994. [Google Scholar] [CrossRef] [PubMed]
  3. Alshahrani, M.A.; Bin Saeedan, M.; Alkhunaizan, T.; Aljohani, I.M.; Azzumeea, F.M. Bilateral adrenal abnormalities: Imaging review of different entities. Abdom. Radiol. 2019, 44, 154–179. [Google Scholar] [CrossRef] [PubMed]
  4. Glazer, D.I.; Mayo-Smith, W.W. Management of incidental adrenal masses: An update. Abdom. Radiol. 2020, 45, 892–900. [Google Scholar] [CrossRef] [PubMed]
  5. Elsayes, K.M.; Elmohr, M.M.; Javadi, S.; Menias, C.O.; Remer, E.M.; Morani, A.C.; Shaaban, A.M. Mimics, pitfalls, and misdiagnoses of adrenal masses on CT and MRI. Abdom. Radiol. 2020, 45, 982–1000. [Google Scholar] [CrossRef]
  6. Albano, D.; Agnello, F.; Midiri, F.; Pecoraro, G.; Bruno, A.; Alongi, P.; Toia, P.; Di Buono, G.; Agrusa, A.; Sconfienza, L.M.; et al. Imaging features of adrenal masses. Insights Imaging 2019, 10, 1. [Google Scholar] [CrossRef][Green Version]
  7. Hanafy, A.K.; Mujtaba, B.; Roman-Colon, A.M.; Elsayes, K.M.; Harrison, D.; Ramani, N.S.; Waguespack, S.G.; Morani, A.C. Imaging features of adrenal gland masses in the pediatric population. Abdom. Radiol. 2020, 45, 964–981. [Google Scholar] [CrossRef]
  8. Elbanan, M.G.; Javadi, S.; Ganeshan, D.; Habra, M.A.; Rao Korivi, B.; Faria, S.C.; Elsayes, K.M. Adrenal cortical adenoma: Current update, imaging features, atypical findings, and mimics. Abdom. Radiol. 2020, 45, 905–916. [Google Scholar] [CrossRef]
  9. Romeo, V.; Maurea, S.; Guarino, S.; Mainenti, P.P.; Liuzzi, R.; Petretta, M.; Cozzolino, I.; Klain, M.; Brunetti, A. The role of dynamic post-contrast T1-w MRI sequence to characterize lipid-rich and lipid-poor adrenal adenomas in comparison to non-adenoma lesions: Preliminary results. Abdom. Radiol. 2018, 43, 2119–2129. [Google Scholar] [CrossRef]
  10. Gillies, R.J.; Kinahan, P.E.; Hricak, H. Radiomics: Images Are More than Pictures, They Are Data. Radiology 2016, 278, 563–577. [Google Scholar] [CrossRef][Green Version]
  11. Rizzo, S.; Botta, F.; Raimondi, S.; Origgi, D.; Fanciullo, C.; Morganti, A.G.; Bellomi, M. Radiomics: The facts and the challenges of image analysis. Eur. Radiol. Exp. 2018, 2, 36. [Google Scholar] [CrossRef] [PubMed]
  12. Cuocolo, R.; Caruso, M.; Perillo, T.; Ugga, L.; Petretta, M. Machine Learning in oncology: A clinical appraisal. Cancer Lett. 2020, 481, 55–62. [Google Scholar] [CrossRef] [PubMed]
  13. Stanzione, A.; Cuocolo, R.; Verde, F.; Galatola, R.; Romeo, V.; Mainenti, P.P.; Aprea, G.; Guadagno, E.; Del Basso De Caro, M.; Maurea, S. Handcrafted MRI radiomics and machine learning: Classification of indeterminate solid adrenal lesions. Magn. Reson. Imaging 2021, 79, 52–58. [Google Scholar] [CrossRef] [PubMed]
  14. Van Timmeren, J.E.; Cester, D.; Tanadini-Lang, S.; Alkadhi, H.; Baessler, B. Radiomics in medical imaging—“How-to” guide and critical reflection. Insights Imaging 2020, 11, 91. [Google Scholar] [CrossRef] [PubMed]
  15. Park, J.E.; Kim, D.; Kim, H.S.; Park, S.Y.; Kim, J.Y.; Cho, S.J.; Shin, J.H.; Kim, J.H. Quality of science and reporting of radiomics in oncologic studies: Room for improvement according to radiomics quality score and TRIPOD statement. Eur. Radiol. 2020, 30, 523–536. [Google Scholar] [CrossRef] [PubMed]
  16. Spadarella, G.; Calareso, G.; Garanzini, E.; Ugga, L.; Cuocolo, A.; Cuocolo, R. MRI based radiomics in nasopharyngeal cancer: Systematic review and perspectives using radiomic quality score (RQS) assessment. Eur. J. Radiol. 2021, 140, 109744. [Google Scholar] [CrossRef]
  17. Ugga, L.; Perillo, T.; Cuocolo, R.; Stanzione, A.; Romeo, V.; Green, R.; Cantoni, V.; Brunetti, A. Meningioma MRI radiomics and machine learning: Systematic review, quality score assessment, and meta-analysis. Neuroradiology 2021, 63, 1293–1304. [Google Scholar] [CrossRef]
  18. Lambin, P.; Leijenaar, R.T.H.; Deist, T.M.; Peerlings, J.; de Jong, E.E.C.; van Timmeren, J.; Sanduleanu, S.; Larue, R.T.H.M.; Even, A.J.G.; Jochems, A.; et al. Radiomics: The bridge between medical imaging and personalized medicine. Nat. Rev. Clin. Oncol. 2017, 14, 749–762. [Google Scholar] [CrossRef]
  19. Moher, D.; Liberati, A.; Tetzlaff, J.; Altman, D.G. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. BMJ 2009, 339, b2535. [Google Scholar] [CrossRef][Green Version]
  20. Akai, H.; Yasaka, K.; Kunimatsu, A.; Ohtomo, K.; Abe, O.; Kiryu, S. Application of CT texture analysis to assess the localization of primary aldosteronism. Sci. Rep. 2020, 10, 472. [Google Scholar] [CrossRef][Green Version]
  21. Ahmed, A.A.; Elmohr, M.M.; Fuentes, D.; Habra, M.A.; Fisher, S.B.; Perrier, N.D.; Zhang, M.; Elsayes, K.M. Radiomic mapping model for prediction of Ki-67 expression in adrenocortical carcinoma. Clin. Radiol. 2020, 75, 479.e17–479.e22. [Google Scholar] [CrossRef] [PubMed]
  22. Ansquer, C.; Drui, D.; Mirallié, E.; Renaudin-Autain, K.; Denis, A.; Gimenez-Roqueplo, A.-P.; Leux, C.; Toulgoat, F.; Kraeber-Bodéré, F.; Carlier, T. Usefulness of FDG-PET/CT-Based Radiomics for the Characterization and Genetic Orientation of Pheochromocytomas Before Surgery. Cancers 2020, 12, 2424. [Google Scholar] [CrossRef] [PubMed]
  23. Chen, N.; Ng, C.; Hobbs, B.P. Bayesian classifiers of solid lesions with dynamic CT: Integrating enhancement density with washout density and delay interval. In Proceedings of the 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018), Washington, DC, USA, 4–7 April 2018; IEEE: Piscataway, NJ, USA, 2018; pp. 236–239. [Google Scholar]
  24. Daye, D.; Staziaki, P.V.; Furtado, V.F.; Tabari, A.; Fintelmann, F.J.; Frenk, N.E.; Shyn, P.; Tuncali, K.; Silverman, S.; Arellano, R.; et al. CT Texture Analysis and Machine Learning Improve Post-ablation Prognostication in Patients with Adrenal Metastases: A Proof of Concept. Cardiovasc. Interv. Radiol. 2019, 42, 1771–1776. [Google Scholar] [CrossRef] [PubMed]
  25. Elmohr, M.M.; Fuentes, D.; Habra, M.A.; Bhosale, P.R.; Qayyum, A.A.; Gates, E.; Morshid, A.I.; Hazle, J.D.; Elsayes, K.M. Machine learning-based texture analysis for differentiation of large adrenal cortical tumours on CT. Clin. Radiol. 2019, 74, 818.e1–818.e7. [Google Scholar] [CrossRef] [PubMed]
  26. Ho, L.M.; Samei, E.; Mazurowski, M.A.; Zheng, Y.; Allen, B.C.; Nelson, R.C.; Marin, D. Can Texture Analysis Be Used to Distinguish Benign From Malignant Adrenal Nodules on Unenhanced CT, Contrast-Enhanced CT, or In-Phase and Opposed-Phase MRI? Am. J. Roentgenol. 2019, 212, 554–561. [Google Scholar] [CrossRef] [PubMed]
  27. Koyuncu, H.; Ceylan, R.; Asoglu, S.; Cebeci, H.; Koplay, M. An extensive study for binary characterisation of adrenal tumours. Med. Biol. Eng. Comput. 2019, 57, 849–862. [Google Scholar] [CrossRef] [PubMed]
  28. Li, X.; Guindani, M.; Ng, C.S.; Hobbs, B.P. Classification of adrenal lesions through spatial Bayesian modeling of GLCM. In Proceedings of the 2017 IEEE 14th International Symposium on Biomedical Imaging (ISBI 2017), Melbourne, Australia, 18–21 April 2017; IEEE: Piscataway, NJ, USA, 2017; pp. 147–151. [Google Scholar]
  29. Li, X.; Guindani, M.; Ng, C.S.; Hobbs, B.P. Spatial Bayesian modeling of GLCM with application to malignant lesion characterization. J. Appl. Stat. 2019, 46, 230–246. [Google Scholar] [CrossRef]
  30. Li, X.; Guindani, M.; Ng, C.S.; Hobbs, B.P. A Bayesian nonparametric model for textural pattern heterogeneity. J. R. Stat. Soc. Ser. C Appl. Stat. 2021, 70, 459–480. [Google Scholar] [CrossRef]
  31. Liu, J.; Xue, K.; Li, S.; Zhang, Y.; Cheng, J. Combined Diagnosis of Whole-Lesion Histogram Analysis of T1- and T2-Weighted Imaging for Differentiating Adrenal Adenoma and Pheochromocytoma: A Support Vector Machine-Based Study. Can. Assoc. Radiol. J. 2021, 72, 452–459. [Google Scholar] [CrossRef][Green Version]
  32. Nakajo, M.; Jinguji, M.; Nakajo, M.; Shinaji, T.; Nakabeppu, Y.; Fukukura, Y.; Yoshiura, T. Texture analysis of FDG PET/CT for differentiating between FDG-avid benign and metastatic adrenal tumors: Efficacy of combining SUV and texture parameters. Abdom. Radiol. 2017, 42, 2882–2889. [Google Scholar] [CrossRef]
  33. Romeo, V.; Maurea, S.; Cuocolo, R.; Petretta, M.; Mainenti, P.P.; Verde, F.; Coppola, M.; Dell’Aversana, S.; Brunetti, A. Characterization of Adrenal Lesions on Unenhanced MRI Using Texture Analysis: A Machine-Learning Approach. J. Magn. Reson. Imaging 2018, 48, 198–204. [Google Scholar] [CrossRef] [PubMed]
  34. Shi, B.; Zhang, G.-M.-Y.; Xu, M.; Jin, Z.-Y.; Sun, H. Distinguishing metastases from benign adrenal masses: What can CT texture analysis do? Acta Radiol. 2019, 60, 1553–1561. [Google Scholar] [CrossRef] [PubMed]
  35. Schieda, N.; Krishna, S.; McInnes, M.D.F.; Moosavi, B.; Alrashed, A.; Moreland, R.; Siegelman, E.S. Utility of MRI to Differentiate Clear Cell Renal Cell Carcinoma Adrenal Metastases From Adrenal Adenomas. Am. J. Roentgenol. 2017, 209, W152–W159. [Google Scholar] [CrossRef] [PubMed]
  36. Shoemaker, K.; Hobbs, B.P.; Bharath, K.; Ng, C.S.; Baladandayuthapani, V. Tree-based Methods for Characterizing Tumor Density Heterogeneity. Pac. Symp. Biocomput. 2018, 23, 216–227. [Google Scholar]
  37. Tu, W.; Verma, R.; Krishna, S.; McInnes, M.D.F.; Flood, T.A.; Schieda, N. Can Adrenal Adenomas Be Differentiated From Adrenal Metastases at Single-Phase Contrast-Enhanced CT? Am. J. Roentgenol. 2018, 211, 1044–1050. [Google Scholar] [CrossRef]
  38. Umanodan, T.; Fukukura, Y.; Kumagae, Y.; Shindo, T.; Nakajo, M.; Takumi, K.; Nakajo, M.; Hakamada, H.; Umanodan, A.; Yoshiura, T. ADC histogram analysis for adrenal tumor histogram analysis of apparent diffusion coefficient in differentiating adrenal adenoma from pheochromocytoma. J. Magn. Reson. Imaging 2017, 45, 1195–1203. [Google Scholar] [CrossRef]
  39. Wang, M.; Xu, H.; Xiao, L.; Song, W.; Zhu, S.; Ma, X. Prognostic Value of Functional Parameters of 18 F-FDG-PET Images in Patients with Primary Renal/Adrenal Lymphoma. Contrast Media Mol. Imaging 2019, 2019, 2641627. [Google Scholar] [CrossRef][Green Version]
  40. Werner, R.A.; Kroiss, M.; Nakajo, M.; Mügge, D.O.; Hahner, S.; Fassnacht, M.; Schirbel, A.; Bluemel, C.; Higuchi, T.; Papp, L.; et al. Assessment of tumor heterogeneity in treatment-naïve adrenocortical cancer patients using 18F-FDG positron emission tomography. Endocrine 2016, 53, 791–800. [Google Scholar] [CrossRef]
  41. Yi, X.; Guan, X.; Chen, C.; Zhang, Y.; Zhang, Z.; Li, M.; Liu, P.; Yu, A.; Long, X.; Liu, L.; et al. Adrenal incidentaloma: Machine learning-based quantitative texture analysis of unenhanced CT can effectively differentiate sPHEO from lipid-poor adrenal adenoma. J. Cancer 2018, 9, 3577–3582. [Google Scholar] [CrossRef][Green Version]
  42. Yi, X.; Guan, X.; Zhang, Y.; Liu, L.; Long, X.; Yin, H.; Wang, Z.; Li, X.; Liao, W.; Chen, B.T.; et al. Radiomics improves efficiency for differentiating subclinical pheochromocytoma from lipid-poor adenoma: A predictive, preventive and personalized medical approach in adrenal incidentalomas. EPMA J. 2018, 9, 421–429. [Google Scholar] [CrossRef]
  43. Yu, H.; Parakh, A.; Blake, M.; McDermott, S. Texture Analysis as a Radiomic Marker for Differentiating Benign From Malignant Adrenal Tumors. J. Comput. Assist. Tomogr. 2020, 44, 766–771. [Google Scholar] [CrossRef] [PubMed]
  44. Zhang, G.-M.-Y.; Shi, B.; Sun, H.; Jin, Z.-Y.; Xue, H.-D. Differentiating pheochromocytoma from lipid-poor adrenocortical adenoma by CT texture analysis: Feasibility study. Abdom. Radiol. 2017, 42, 2305–2313. [Google Scholar] [CrossRef] [PubMed]
  45. Sanduleanu, S.; Woodruff, H.C.; de Jong, E.E.C.; van Timmeren, J.E.; Jochems, A.; Dubois, L.; Lambin, P. Tracking tumor biology with radiomics: A systematic review utilizing a radiomics quality score. Radiother. Oncol. 2018, 127, 349–360. [Google Scholar] [CrossRef] [PubMed]
  46. Vallières, M.; Zwanenburg, A.; Badic, B.; Cheze Le Rest, C.; Visvikis, D.; Hatt, M. Responsible Radiomics Research for Faster Clinical Translation. J. Nucl. Med. 2018, 59, 189–193. [Google Scholar] [CrossRef]
  47. Zwanenburg, A.; Vallières, M.; Abdalah, M.A.; Aerts, H.J.W.L.; Andrearczyk, V.; Apte, A.; Ashrafinia, S.; Bakas, S.; Beukinga, R.J.; Boellaard, R.; et al. The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping. Radiology 2020, 295, 328–338. [Google Scholar] [CrossRef] [PubMed][Green Version]
  48. Shur, J.D.; Doran, S.J.; Kumar, S.; Ap Dafydd, D.; Downey, K.; O’Connor, J.P.B.; Papanikolaou, N.; Messiou, C.; Koh, D.-M.; Orton, M.R. Radiomics in Oncology: A Practical Guide. RadioGraphics 2021, 41, 1717–1732. [Google Scholar] [CrossRef] [PubMed]
  49. Chetan, M.R.; Gleeson, F.V. Radiomics in predicting treatment response in non-small-cell lung cancer: Current status, challenges and future perspectives. Eur. Radiol. 2021, 31, 1049–1058. [Google Scholar] [CrossRef]
  50. Wakabayashi, T.; Ouhmich, F.; Gonzalez-Cabrera, C.; Felli, E.; Saviano, A.; Agnus, V.; Savadjiev, P.; Baumert, T.F.; Pessaux, P.; Marescaux, J.; et al. Radiomics in hepatocellular carcinoma: A quantitative review. Hepatol. Int. 2019, 13, 546–559. [Google Scholar] [CrossRef][Green Version]
  51. Cuocolo, R.; Stanzione, A.; Castaldo, A.; De Lucia, D.R.; Imbriaco, M. Quality control and whole-gland, zonal and lesion annotations for the PROSTATEx challenge public dataset. Eur. J. Radiol. 2021, 138, 109647. [Google Scholar] [CrossRef]
  52. Oakden-Rayner, L. Exploring Large-scale Public Medical Image Datasets. Acad. Radiol. 2020, 27, 106–112. [Google Scholar] [CrossRef][Green Version]
  53. Abunahel, B.M.; Pontre, B.; Kumar, H.; Petrov, M.S. Pancreas image mining: A systematic review of radiomics. Eur. Radiol. 2021, 31, 3447–3467. [Google Scholar] [CrossRef] [PubMed]
  54. Stanzione, A.; Gambardella, M.; Cuocolo, R.; Ponsiglione, A.; Romeo, V.; Imbriaco, M. Prostate MRI radiomics: A systematic review and radiomic quality score assessment. Eur. J. Radiol. 2020, 129, 109095. [Google Scholar] [CrossRef] [PubMed]
  55. Zhong, J.; Hu, Y.; Si, L.; Jia, G.; Xing, Y.; Zhang, H.; Yao, W. A systematic review of radiomics in osteosarcoma: Utilizing radiomics quality score as a tool promoting clinical translation. Eur. Radiol. 2021, 31, 1526–1535. [Google Scholar] [CrossRef] [PubMed]
  56. Song, J.; Yin, Y.; Wang, H.; Chang, Z.; Liu, Z.; Cui, L. A review of original articles published in the emerging field of radiomics. Eur. J. Radiol. 2020, 127, 108991. [Google Scholar] [CrossRef]
  57. Scapicchio, C.; Gabelloni, M.; Barucci, A.; Cioni, D.; Saba, L.; Neri, E. A deep look into radiomics. Radiol. Med. 2021, 126, 1296–1311. [Google Scholar] [CrossRef]
  58. Kocak, B.; Durmaz, E.S.; Ates, E.; Kilickesmez, O. Radiomics with artificial intelligence: A practical guide for beginners. Diagn. Interv. Radiol. 2019, 25, 485–495. [Google Scholar] [CrossRef]
  59. Chartrand, G.; Cheng, P.M.; Vorontsov, E.; Drozdzal, M.; Turcotte, S.; Pal, C.J.; Kadoury, S.; Tang, A. Deep Learning: A Primer for Radiologists. RadioGraphics 2017, 37, 2113–2131. [Google Scholar] [CrossRef][Green Version]
  60. Buvat, I.; Orlhac, F. The Dark Side of Radiomics: On the Paramount Importance of Publishing Negative Results. J. Nucl. Med. 2019, 60, 1543–1544. [Google Scholar] [CrossRef]
  61. Adams, J.; Hillier-Brown, F.C.; Moore, H.J.; Lake, A.A.; Araujo-Soares, V.; White, M.; Summerbell, C. Searching and synthesising ‘grey literature’ and ‘grey information’ in public health: Critical reflections on three case studies. Syst. Rev. 2016, 5, 164. [Google Scholar] [CrossRef][Green Version]
Figure 1. Literature search and study selection flow-chart.
Figure 1. Literature search and study selection flow-chart.
Diagnostics 12 00578 g001
Figure 2. Count plot showing studies published per year.
Figure 2. Count plot showing studies published per year.
Diagnostics 12 00578 g002
Figure 3. Distribution of RQS percentage score of the papers included in our review. This is presented both as a histogram (bars) and its corresponding density function (line).
Figure 3. Distribution of RQS percentage score of the papers included in our review. This is presented both as a histogram (bars) and its corresponding density function (line).
Diagnostics 12 00578 g003
Figure 4. Line plot depicting the median RQS percentage score in relation to publication year.
Figure 4. Line plot depicting the median RQS percentage score in relation to publication year.
Diagnostics 12 00578 g004
Table 1. Main characteristics of included articles.
Table 1. Main characteristics of included articles.
Study IDYearCountryAimMean GoalStudy DesignPatient Population (Number of Lesions)Imaging ModalitySegmentation Method (Software/Algorithm)Feature ExtractionMLFeatures Type
Akai et al. [20]2020JapanDetectionLocalization of primary aldosteronismRetrospective82 (82)Unenhanced CTSemi-automatic, 2D (TexRAD)TexRAD;NoFirst-order
Amhed et al. [21]2020USACharacterizationPrediction of Ki-67 expression in ACCRetrospective53 (53)Contrast-enhanced CTManual, 3D (AMIRA)PyRadiomicsNoShape-based, first- and higher-order
Ansquer et al. [22]2020FranceCharacterizationBiological and genetic profiling of pheoRetrospective49 (52)2-[18F]FDG PET/CTAutomatic, 3D (STAPLE)Image Biomarker Standardization InitiativeNoHigher-order
Chen et al. [23]2018USADDBenign vs. malignantRetrospective222 (222)Unenhanced and contrast-enhanced CTNRNRYes (Bayesian classifier)NR
Daye et al. [24]2019USAPrognosisLocal progression and survival in ablated adrenal metastasisRetrospective21 (21)Contrast-enhanced CTManual, 3D (NR)MATLABYes (support vector machine)Higher-order
Elmohr et al. [25]2019USADDAdenoma vs. ACCRetrospective54 (54)Unenhanced and contrast-enhanced CTManual, 3D (AMIRA)PyRadiomicsYes (random forest)Shape-based, first- and higher-order
Ho et al. [26]2019USADDBenign vs. malignantRetrospective20 (23)Unenhanced and contrast-enhanced CT and MRI 3T or 1,5T T1 IN-OUTManual 3D (Seg3D)Image Biomarker Standardization InitiativeNoFirst- and higher-order
Koyuncu et al. [27]2018TurkeyDDBenign vs. malignantRetrospectiveNR (114) *Unenhanced and contrast-enhanced CTSemi-automatic, 2D
MATLABYes (neural network)First- and higher-order
Li et al. [28]2017USADDBenign vs. malignantRetrospective223 (230)Unenhanced CTManual, 2D (NR)NRYes (Bayesian)Higher-order
Li et al. [29]2018USADDBenign vs. malignantRetrospective204 (210)Unenhanced and contrast-enhanced CTManual, 2D (NR)NRYes (Bayesian)Higher-order
Li et al. [30]2020USADDBenign vs. malignantRetrospective204 (210)Unenhanced and contrast-enhanced CTManual, 2D (NR)NRYes (Bayesian)Higher-order
Liu et al. [31]2020ChinaDDAdenoma vs. pheoRetrospective58 (60)MRI 3T
T1 IN-OUT, T2w
Manual, 3D
MaZdaYes (support vector machine)First-order
Nakajo et al. [32]2017JapanDDBenign vs. malignantRetrospective31 (35)2-[18F]FDG PET/CTSemi-automatic, 3D (Advantage Windows Workstation)Python #NoFirst-order
Romeo et al. [33]2018ItalyDDAdenoma vs. non-adenomaRetrospective60 (60)MRI 3T T1 IN-OUT, T2wManual, 3D (3D Slicer)3D SlicerYes (decision tree)First- and higher-order
Shi et al. [34]2019ChinaDDBenign vs. malignantRetrospective225 (265)Unenhanced and contrast-enhanced CTManual, 2D (TexRAD)TexRAD;Yes
(support vector machine)
Schieda et al. [35]2017CanadaDDAdenoma vs. RCC metastasisRetrospective39 (44)MRI 3T or 1.5T
T1 IN-OUT, T2w, GRE T1w pre- and post-contrast
Manual, 2D (Image J)Image JNoFirst-order
Shoemaker et al. [36]2018USADD/CharacterizationBenign vs. malignant; calcified vs. non calcified; functioning vs. non functioningRetrospective356 (379)Unenhanced CTNRNRYes
(decision tree)
First- and higher-order
Tu et al. [37]2018CanadaDDAdenoma vs. lung cancer metastasisRetrospective61 (76)Contrast-enhanced CTManual, 2D
Umanodan et al. [38]2016JapanDDAdenoma vs. pheoRetrospective47 (52)MRI 3T
Manual, 2D
(Synapse Vincent)
Synapse VincentNoFirst-order
Wang et al. [39]2019ChinaPrognosisSurvival in primary adrenal non-Hodgkin’s lymphomaRetrospective19 (19) §2-[18F]FDG PET/CTManual, 3D
(LifeX package)
LifeX packageNoFirst- and higher-order
Werner et al. [40]2016GermanyPrognosisDisease progression and survival in ACCRetrospective10 (10)2-[18F]FDG PET/CTManual, 3D
(Interview Fusion
Interview Fusion
NoFirst- and higher-order
Yi et al. [41]2018ChinaDDAdenoma vs. pheoRetrospective108 (110)Unenhanced CTManual, 2D
(logistic regression)
First- and higher-order
Yi et al. [42]2018ChinaDDAdenoma vs. pheoRetrospective265 (265)Unenhanced and contrast-enhanced CTManual, 2D
MaZdaNoFirst- and higher-order
Yu et al. [43]2020USADDBenign vs. malignantRetrospective125 (125)Contrast-enhanced CTManual, 2D (TexRAD)TexRAD,NoFirst-order
Zhang et al. [44]2017ChinaDDAdenoma vs. pheoRetrospective155 (164)Unenhanced and contrast-enhanced CTManual, 2D (TexRAD)TexRAD,NoFirst-order
* reported number of images. # formulas reported in the article. § patients with adrenal and/or kidney lymphoma included. ACC: adrenocortical carcinoma, DD: differential diagnosis of adrenal masses, Pheo: pheochromocytoma, NR: not reported.
Table 2. Radiomics Quality Score (RQS) assessment for all included articles.
Table 2. Radiomics Quality Score (RQS) assessment for all included articles.
Study IDItem 1Item 2Item 3Item 4Item 5Item 6Item 7Item 8Item 9Item 10Item 11Item 12Item 13Item 14Item 15Item 16Total(%)
Akai et al. [20]10003001100−5000013%
Amhed et al. [21]1000−3001100−50000−50
Ansquer et al. [22]11003111200200001233%
Chen et al. [23]0000300010020000617%
Daye et al. [24]1000−30002002000026%
Elmohr et al. [25]11003000200220001131%
Ho et al. [26]10003000100−5000000%
Koyuncu et al. [27]0000300010020000617%
Li et al. [28]0000300010020000617%
Li et al. [29]1000300010020000719%
Li et al. [30]00003000100−50000−10
Liu et al. [31]1100300110020000925%
Nakajo et al. [32]1000−3001100−50000−50
Romeo et al. [33]10003000200220001028%
Shi et al. [34]1100300020020000925%
Schieda et al. [35]1000−3100100−50000−50
Shoemaker et al. [36]0000300020020001822%
Tu et al. [37]1000−3001100−50000−50
Umanodan et al. [38]1100−3001100−50000−40
Wang et al. [39]10003101100−5000026%
Werner et al. [40]1000−3100100−50000−50
Yi et al. [41]11003000100−5000013%
Yi et al. [42]11003001110200001028%
Yu et al. [43]1000−3001100−50000−50
Zhang et al. [44]1000−3001100−50000−50
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Stanzione, A.; Galatola, R.; Cuocolo, R.; Romeo, V.; Verde, F.; Mainenti, P.P.; Brunetti, A.; Maurea, S. Radiomics in Cross-Sectional Adrenal Imaging: A Systematic Review and Quality Assessment Study. Diagnostics 2022, 12, 578.

AMA Style

Stanzione A, Galatola R, Cuocolo R, Romeo V, Verde F, Mainenti PP, Brunetti A, Maurea S. Radiomics in Cross-Sectional Adrenal Imaging: A Systematic Review and Quality Assessment Study. Diagnostics. 2022; 12(3):578.

Chicago/Turabian Style

Stanzione, Arnaldo, Roberta Galatola, Renato Cuocolo, Valeria Romeo, Francesco Verde, Pier Paolo Mainenti, Arturo Brunetti, and Simone Maurea. 2022. "Radiomics in Cross-Sectional Adrenal Imaging: A Systematic Review and Quality Assessment Study" Diagnostics 12, no. 3: 578.

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop