State of the art of radiomic analysis in the clinical management of prostate cancer: A systematic review
Graphical abstract
Introduction
Prostate cancer (PCa) affects over four million individuals and is the second most common cancer in men worldwide (World Health Organization, 2020). Early diagnosis is a key contributor to successful PCa management, with remarkable consequences on its relatively low mortality rate. Digital rectal examination (DRE), trans-rectal ultrasound (TRUS), and prostate-specific antigen (PSA) serum level are commonly used for PCa screening (Schröder et al., 2009). However, although sensitive, these tests are non-specific, and often result in unnecessary and expensive follow-up procedures (Schröder et al., 2009). Albeit biopsy is currently the gold standard for PCa diagnosis (European Association of Urology, 2020), it has some drawbacks including sampling error, the possibility to cause patients’ discomfort and pain and being not representative of the whole tumour heterogeneity (Guo et al., 2018; Kristiansen, 2012; Patel et al., 2014; Marusyk and Polyak, 2010).
Multi-parametric Magnetic Resonance Imaging (mp-MRI) has the potential to reliably detect PCa, reducing the number of unnecessary biopsies by guiding a more precise sampling (Carlaw and Woo, 2017; Gupta et al., 2013). The prostate imaging-reporting and data systems (PI-RADS) version 2 was released in 2015 for the standardization of the interpretation of mp-MRI scans (American College of Radiology Web site, 2015). Since its introduction it has reached widespread acceptance and it is now regularly implemented in patient care (Kasel-Seibert et al., 2016), where it has improved diagnostic accuracy of PCa up to 60–90 % (Jordan et al., 2017). However, PI-RADS v2 features are subjective and therefore susceptible to inter- and intra-observer variability (Rosenkrantz et al., 2016).
Radiomics has been defined as “high throughput extraction of quantitative features that results in the conversion of images into mineable data” (Gillies et al., 2016). More simply, radiomics consists in the extraction of quantitative information from medical images. This method uses features based on intensity, shape, size, volume, and texture to accurately describe the tumour phenotype. Radiomics potentially provides several quantitative, objective, imaging biomarkers that can aid in personalized PCa management (Gillies et al., 2016; Yip and Aerts, 2016). Another advantage of radiomics is that it makes use of different kinds of imaging modalities, raising the possibility to investigate multiple aspects of the tumour at the same time. Furthermore, as information is gathered from the entire tumour mass, this technique offers details on intra-tumoral heterogeneity, which is known to be closely related to cancer progression and therapeutic resistance (Marusyk and Polyak, 2010; Andor et al., 2016).
The field of radiomics is growing fast, yielding promising results, especially in oncology. The rapid progress of radiomics is mainly due to the ever increasing number of medical images that are collected every day and to the advances of machine learning (ML) and deep learning (DL) for their interpretation (Herrmann et al., 2019). The aim of the present review is to systematically report all the applications of radiomics to the clinical management of PCa, discussing the additional utility bore by radiomic analysis to PCa care. Methodological and practical challenges that still prevent radiomics to be applied in clinical settings as well as the adherence to existing guidelines of the included studies will also be discussed.
Section snippets
Materials and methods
The following online medical databases have been systematically searched in order to find publications of interest: PubMed, EMBASE, SCOPUS, MEDLINE, and Web of Science using these keywords: (radiomic or radiomics) AND (prostate cancer or prostate tumour or prostate tumor or prostate neoplasia). All publication until December 2020 have been included. After duplicates removal, abstracts were screened to identify and remove all the studies that did not meet the inclusion criteria. The PRISMA
Study selection
Studies were evaluated according to the exclusion/inclusion criteria and 76/571 were finally selected and included for qualitative analysis. The main focus of the selected works is reported in Table 1.
Prostate cancer detection
Histological examination is required for PCa diagnosis, but selecting good candidates for biopsy is problematic due to the high false-positive rate of screening methods (Schröder et al., 2009). However, mp-MRI have been proved effective in localizing PCa and is now commonly used in clinical
Discussion
Radiomics bears the potential to aid clinicians for the personalized management of PCa. Radiomic analysis is performed starting from medical images routinely collected in clinical practice and, therefore, is a non-invasive procedure that does not entail additional costs. Furthermore, since it exploits information collected throughout the whole organ it provides notions on tumour heterogeneity. Conversely, biopsy, the state-of-the-art technique to characterize PCa, presents multiple drawbacks
Conclusion
In conclusion, this work systematically reviewed all the applications of radiomic analysis for the clinical management of PCa, showing the great potential that it bears and the many limitations that still prevent it to enter clinical practice. Results gathered so far are encouraging, but often exploratory and non-comparable. Future studies with larger samples and external validation cohorts, including multimodal imaging RF and clinical factors are needed to assess with certainty the utility of
Ethical approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Declaration of Competing Interest
All the Authors have no conflicts of interest to disclose related to the present paper.
Acknowledgements
This work was supported by the Italian Association for Cancer Research (grant IG 2017 Id. 20571) and by the Italian Ministry of Health (PE-2016-02361273); EUDRACT number: 2018-001034-18.
Samuele Ghezzo, MSc.: Ph.D. student in Molecular Medicine at Vita-Salute San Raffaele University. His main interest is the application of machine/deep learning algorithms to medical imaging for improved clinical management of oncological diseases.
References (100)
- et al.
Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: recommendations of the RTOG-ASTRO Phoenix consensus Conference
Int J Radiat Oncol [Internet].
(2006) - et al.
Ultra-hypofractionated versus conventionally fractionated radiotherapy for prostate cancer: 5-year outcomes of the HYPO-RT-PC randomised, non-inferiority, phase 3 trial
. Lancet [Internet].
(2019) - et al.
Machine learning-based radiomic models to predict intensity-modulated radiation therapy response, Gleason score and stage in prostate cancer
Radiol Medica.
(2019) - et al.
MRI radiomic analysis of IMRT-Induced bladder wall changes in prostate Cancer patients: a relationship with radiation dose and toxicity
J. Med. Imaging Radiat. Sci.
(2019) - et al.
Machine learning for differentiating metastatic and completely responded sclerotic bone lesion in prostate cancer: a retrospective radiomics study
Br. J. Radiol.
(2019) - et al.
Radiomic features on MRI enable risk categorization of prostate cancer patients on active surveillance: preliminary findings
J. Magn. Reson. Imaging
(2018) - et al.
Combination of peri-tumoral and intra-tumoral radiomic features on bi-parametric mri accurately stratifies prostate cancer risk: A multi-site study
Cancers (Basel).
(2020) PI-RADS Prostate Imaging and Reporting and Data System: Version 2 [Internet]
(2015)- et al.
Pan-cancer analysis of the extent and consequences of intratumor heterogeneity
Nat. Med.
(2016) - et al.
Detection of dominant intra-prostatic lesions in patients with prostate Cancer Using an artificial neural network and MR multi-modal radiomics analysis
Front. Oncol.
(2019)
Comparison of machine learning algorithms to predict clinically significant prostate cancer of the peripheral zone with multiparametric MRI using clinical assessment categories and radiomic features
Eur. Radiol.
Multiparametric MRI and auto-fixed volume of interest-based radiomics signature for clinically significant peripheral zone prostate cancer
Eur. Radiol.
Radiomic machine learning for characterization of prostate lesions with MRI: comparison to ADC values
Radiology.
Cone-beam computed tomography-based radiomics in prostate cancer: a mono-institutional study
Strahlenther. Onkol.
MRI-derived radiomics to guide post-operative management for high-risk prostate cancer
Front. Oncol.
External validation of an MRI-derived radiomics model to predict biochemical recurrence after surgery for high-risk prostate cancer
Cancers (Basel).
Prostate gleason score detection and cancer treatment through real-time formal verification
IEEE Access
Radiomics for gleason score detection through deep learning
Sensors (Switzerland).
MAPS: A Quantitative Radiomics Approach for Prostate Cancer Detection
IEEE Trans. Biomed. Eng.
Evaluation of the changing landscape of prostate cancer diagnosis and management from 2005 to 2016
Prostate Int.
Predicting gleason score of prostate cancer patients using radiomic analysis
Front. Oncol.
Multimodal radiomic features for the predicting gleason score of prostate cancer
Cancers (Basel).
Deep radiomic analysis to predict gleason score in prostate Cancer
IEEE Access
Prostate Cancer differentiation and aggressiveness: assessment with a radiomic-based model vs. PI-RADS v2
J. Magn. Reson. Imaging
Prostate cancer detection via a quantitative radiomics-driven conditional random field framework
IEEE Access
Clinically significant prostate cancer detection on MRI: a radiomic shape features study
Eur. J. Radiol.
Machine learning-based analysis of [18F]DCFPyL PET radiomics for risk stratification in primary prostate cancer
Eur J Nucl Med Mol Imaging. European Journal of Nuclear Medicine and Molecular Imaging
Classification of suspicious lesions on prostate multiparametric MRI using machine learning
J. Med. Imaging Bellingham (Bellingham)
Biochemical recurrence prediction after radiotherapy for prostate cancer with T2w magnetic resonance imaging radiomic features
Phys Imaging Radiat Oncol.
Quantitative characterisation of clinically significant intra-prostatic cancer by prostate-specific membrane antigen (PSMA) expression and cell density on PSMA-11
Eur. Radiol.
Radiomics driven diffusion weighted imaging sensing strategies for zone-level prostate cancer sensing
Sensors (Switzerland).
EAU Guidelines [Internet]
Identification of the most significant magnetic resonance imaging (MRI) radiomic features in oncological patients with vertebral bone marrow metastatic disease: a feasibility study
Radiol Medica.
Risk of prostate cancer-specific mortality following biochemical recurrence after radical prostatectomy
J. Am. Med. Assoc.
Voxel-based supervised machine learning of peripheral zone prostate cancer using noncontrast multiparametric MRI
J. Appl. Clin. Med. Phys.
PI-RADS 3 lesions: role of prostate MRI texture analysis in the identification of prostate Cancer
Curr. Probl. Diagn. Radiol.
Correlation between SUVmax and CT radiomic analysis using lymph node density in PET/CT-based lymph node staging
J. Nucl. Med.
Radiomics: Images are more than pictures, they are data
Radiology.
Radiomic features for prostate cancer detection on MRI differ between the transition and peripheral zones: preliminary findings from a multi-institutional study
J. Magn. Reson. Imaging
Haralick textural features on T2-weighted MRI are associated with biochemical recurrence following radiotherapy for peripheral zone prostate cancer
J. Magn. Reson. Imaging
Noninvasive prediction of high-grade prostate Cancer via biparametric MRI radiomics
J. Magn. Reson. Imaging
MRI-based radiomics signature for localized prostate cancer: a new clinical tool for cancer aggressiveness prediction? Sub-study of prospective phase II trial on ultra-hypofractionated radiotherapy (AIRC IG-13218)
Eur. Radiol.
Multi-region proteome analysis quantifies spatial heterogeneity of prostate tissue biomarkers
Life Sci Alliance.
The state of prostate MRI in 2013
Oncol (United States)
Sub-differentiating equivocal PI-RADS-3 lesions in multiparametric magnetic resonance imaging of the prostate to improve cancer detection
Eur. J. Radiol.
Radiomics features measured with multiparametric magnetic resonance imaging predict prostate cancer aggressiveness
J. Urol.
Driving the future of nuclear medicine
J. Nucl. Med.
A radiomics machine learning-based redefining score robustly identifies clinically significant prostate cancer in equivocal PI-RADS score 3 lesions
Abdom. Radiol. (NY)
Evaluating the performance of PI-RADS v2 in the non-academic setting
Abdom. Radiol. (NY)
Assessment of PI-RADS v2 for the detection of prostate Cancer
Eur. J. Radiol.
Cited by (15)
News in magnetic resonance imaging use for radiation oncology
2022, Cancer/RadiotherapieCitation Excerpt :The current consensus is that quantitative imaging biomarkers should be included as much as possible in clinical studies to obtain more results, even negative. The same uncertainty quantification approach should be applied to radiomics [11]. Non-quantitative MRI exams introduce low reproducibility even for the same machine.
Nomogram to predict the presence of PSMA-negative but FDG-positive lesion in castration-resistant prostate cancer: a multicenter cohort study
2024, Therapeutic Advances in Medical OncologyBeyond diagnosis: is there a role for radiomics in prostate cancer management?
2023, European Radiology Experimental
Samuele Ghezzo, MSc.: Ph.D. student in Molecular Medicine at Vita-Salute San Raffaele University. His main interest is the application of machine/deep learning algorithms to medical imaging for improved clinical management of oncological diseases.
Carolina Bezzi, MSc.: Research fellow at the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute. Her activities are related to bioinformatics and data science, and her main interests are focused on radiomics and the generation of artificial intelligence-based clinical decision support technologies.
Luca Presotto, Ph.D.: Physicist and certified medical physics expert. He is a Research fellow at the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute. His research interests involve image processing and artificial intelligence.
Paola Mapelli, M.D., Ph.D.: Nuclear Medicine Physician at the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute and research fellow at Vita-Salute San Raffaele University. Her main research interests include the use of hybrid Molecular Imaging (PET/CT and PET/MRI) in oncology.
Valentino Bettinardi, MSc: Senior Physicist, staff of the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute. His research interests include use and development of Positron Emission Tomography (PET), Computed Tomography (CT), Magnetic Resonance Imaging (MRI), hybrid PET/CT and PET/MRI systems. Image processing, Image & data analysis, and PET Image reconstruction (2D and 3D)
Annarita Savi, MSc.: Physicist and certified medical physics expert, staff of the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute. Her activities are related to therapeutic and diagnostic applications of radionuclides, image analysis and dose patient monitoring.
Ilaria Neri, Biomedical engineer, MSc.: Research fellow at the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute. Her research activities are related to hybrid PET/MR image processing.
Erik Preza, M.D.: Nuclear Medicine physician at the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute. His main research interests are related to prostate cancer PET/CT and PET/MRI studies.
Ana Maria Samanes Gajate, M.D.: Nuclear Medicine physician at the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute, with more than 10 years of experience in the field of Nuclear Medicine. Main research interests include the use of hybrid Molecular Imaging (PET/CT and PET/MRI) in oncology, with special focus on genitourinary cancers.
Francesco De Cobelli, M.D.: Full professor of Radiology at Vita-Salute San Raffaele University; head physician of Clinical and Experimental Radiology Unit, IRCCS San Raffaele Scientific Institute. Director, Postgraduate School of Radiology and director of the PhD course in Molecular Medicine at Vita-Salute San Raffaele University. Responsible for the Experimental and Clinical Radiology Unit of the Experimental Imaging Centre, IRCCS San Raffaele Scientific Institute.
Paola Scifo, Electronic Engineer, PhD.: Senior researcher at the Nuclear Medicine Department at IRCCS San Raffaele Scientific Institute, expert in MR technology, advanced MR applications and image processing. Her actual interests are focused on PET/MRI.
Maria Picchio, M.D.: Associate Professor of Nuclear Medicine at Vita-Salute San Raffaele University; Head of Molecular imaging Unit, Clinical Research Unit of the Experimental Imaging Centre, IRCCS San Raffaele Scientific Institute; Nuclear Medicine physician at the Nuclear Medicine Department of IRCCS San Raffaele Scientific Institute.