Multiparametric Magnetic Resonance Imaging in Detection Prostate Cancer: Comparison
View Latest Version
Please note this is a comparison between Version 2 by Conner Chen and Version 1 by Daryoush Shahbazi-Gahrouei.

Prostate cancer is the second leading cause of cancer-related death in men. Its early and correct diagnosis is of particular importance to controlling and preventing the disease from spreading to other tissues. Multip-MRIarametric magnetic resonance imaging (mp-MRI) is recognised as the combination of conventional anatomical MRI and at least two functional magnet resonance sequences: diffusion-weighted imaging (DWI), dynamic contrast-enhanced MRI (DCE-MRI), and, optionally, MR spectroscopy (MRS).

  • prostate cancer
  • multiparametric MRI
  • machine learning

1. Introduction

Prostate cancer (PCa) is the most common cancer in men and the second leading cause of cancer-related death in them [1,2,3][1][2][3]. Various methods are used for PCa screening, though these methods are invasive or have low accuracies, such as digital rectal examination, prostate-specific antigen (PSA) tests, and transrectal ultrasound (TRUS)-guided prostate biopsy [4,5,6,7][4][5][6][7]. New biomarkers, named 8-hydroxy-2-deoxyguanosine (8-OHdG) and 8-iso-prostaglandin F2α (8-IsoF2α), have been reported. Increased levels of these biomarkers indicate prostate cancer in the patient and they are measured through urine tests. Of course, validating these urinary biomarkers in relation to prostate cancer still requires significant research [8]. Meanwhile, prostate MRI plays a crucial role before a biopsy in patients with raised PSA. Multiparametric magnetic resonance imaging (mp-MRI) is a commonly used imaging procedure for diagnosing PCa. Mp-MRI is recognised as the combination of conventional anatomical MRI and at least two functional magnet resonance sequences: diffusion-weighted imaging (DWI), dynamic contrast-enhanced MRI (DCE-MRI), and, optionally, MR spectroscopy (MRS) [9,10][9][10]. Various studies have noted that mp-MRI has good accuracy for diagnosing or determining the grade of prostate cancer [11,12][11][12]. Of course, it is more challenging to determine the aggressiveness of cancer using MRI than when it is detected by a physician with good reliability. Recently, various studies have used artificial intelligence and MRI images to diagnose or assess the characterization and severity of cancers, including prostate cancer, to reduce human error, increase the speed of diagnosis and classification, and improve overall efficiency and accuracy [13,14,15][13][14][15]. Indeed, artificial intelligence is beneficial in acquiring important clinical information that can help physicians to provide key and critical opinions about clinical prognosis, diagnosing diseases, and treatment outcomes [16,17][16][17].
Artificial Intelligence (AI) describes the capability of a computer to model intelligent behaviour, with minimal human intervention, and to reach a certain goal based on provided data. AI has multiple branches. One of these branches is machine learning (ML). ML describes algorithms used to incorporate intelligence into machines by automatically learning from data [16,18][16][18]. There are different types of ML. In general, ML types are branched into four groups: Unsupervised learning, Semi-supervised learning, Supervised learning, and Reinforcement learning [19,20,21][19][20][21]. In Supervised learning, an observer provides data to the machine and labels the data types. Input and output are specified and the machine attempts to learn a pattern from the input to the expected output [22,23][22][23]. In unsupervised learning, the computer finds connections between data and discovers patterns without the help of a trainer and without the use of labels that define the type of data [24,25][24][25]. Semi-supervised learning is a learning paradigm that studies how computers learn in the attendance of labeled and unlabeled data. During semi-supervised learning, the aim is to design algorithms using combinations of labeled and unlabeled data [26]. Reinforcement learning is conducted by encouraging desirable behaviour and punishing undesirable behaviour. In this way, the computer can understand and interpret various issues by trial and error, according to the feedback it receives as a result of its actions [27,28][27][28].
The most common categories of ML algorithms are classification and regression. Examples of supervised learning algorithms include linear and logistic regression, support vector machines (SVMs; classification), K nearest neighbours (KNN; classification and regression), naive Bayes (classification), decision tree and random forests (DT and RF, respectively; both classification), and deep learning techniques (classification) [16,25][16][25].

2. mp-MRI in the Detection PCa

Mp-MRI primarily contains at least three sequences: T2WI or T1WI, DWI, and DCE imaging [29]. T1WI is used to detect bleeding after a biopsy. T2-weighted images can detect the anatomical shape of the peripheral and transitional zones, where 70% and 30% of cancers are found, respectively [9]. DWI measures the Brownian movement of free water protons inside a tissue. Malignant tissue is denser than normal tissue, triggering restricted free water movement inside the cancerous tissue, thereby decreasing the diffusion of water [30,31,32][30][31][32]. DCE assesses the perfusion and vascular permeability throughout the prostate and within a cancerous tissue through the rapid administration of gadolinium chelates and the use of fast T1-weighted images. Unlike normal tissue, malignant tissue has more penetrable, heterogeneous, and disordered vessels due to neoangiogenesis [9,33][9][33]. Various studies have used mp-MRI to diagnose PCa and have noted its diagnostic performance. Di Campli et al. [35][34] conducted a study on mp-MRI to determine the diagnostic accuracy of PCa. A total of 85 patients underwent prostate MRI investigation at a 1.5 T MR system without an endorectal coil. In this study, the MR images were separately interpreted by three radiologists with 7 (reader 1), 3 (reader 2) and 1 year(s) (reader 3) of experience in prostate MRI, respectively (according to Prostate Imaging Reporting and Data System (PI-RADS) version 2). The sensitivity (CI 95%), specificity (CI 95%), area under the curve (AUC), and accuracy values for readers 1, 2, 3 were obtained (97.2% (90.3–99.7%), 88.9% (79.3–95.1%), 83.3% (72.7–91.1%)), (61.5% (31.6–86.1%), 23.1% (5–53.8%), 46.2% (19.2–74.9%)), (0.72, 0.70, 0.54), and 90.58, 78.82, and 77.64, respectively [35][34]. Kam et al. [36][35] assessed the accuracy of mpMRI to predict PCa pathology. In their work, 235 patients underwent mpMRI with a 1.5 T or 3 T MRI. The results of mpMRI were compared with the final radical prostatectomy specimen to analyze the performance of mpMRI for significant prostate cancer (sPCa) detection. They reported the accuracy of mpMRI for the prediction of sPCa. Overall, the sensitivity, specificity, and positive predictive value (PPV) of mpMRI for the detection of sPCa were 91%, 23%, and 95%, respectively. In 2020, Ippolito et al. [37][36] stated the multiparametric diagnostic accuracy of 201 patients for PCa detection. Patients underwent mp-MRI examination with a 3 T MR scanner and a body coil with sequences T2WI, DWI, and DCE. The sensitivity, specificity, and accuracy of PI-RADS for the detection of PCa were 65.1%, 54.9%, and 64.2% (55.1–72.7%), respectively. Consequently, in a study of systematic review and meta-analysis, Zhao et al. [38][37] reported the diagnostic performance of mp-MRI. The meta-analysis included 10 articles. At a per-patient level, the pooled sensitivity, specificity, and AUC values for mpMRI were 0.87 (0.83–0.91), 0.47 (0.23–0.71), and 0.84, respectively. At a per-lesion level, the pooled sensitivity, specificity, and AUC values were 0.63 (0.52–0.74), 0.88 (0.81–0.95), and 0.83, respectively.

References

  1. Chatterjee, A.; Gallan, A.J.; He, D.; Fan, X.; Mustafi, D.; Yousuf, A.; Antic, T.; Karczmar, G.S.; Oto, A. Revisiting quantitative multi-parametric MRI of benign prostatic hyperplasia and its differentiation from transition zone cancer. Abdom. Radiol. 2019, 44, 2233–2243.
  2. Bevacqua, E.; Ammirato, S.; Cione, E.; Curcio, R.; Dolce, V.; Tucci, P. The potential of microRNAs as non-invasive prostate cancer biomarkers: A systematic literature review based on a machine learning approach. Cancers 2022, 14, 5418.
  3. Crocetto, F.; Russo, G.; Di Zazzo, E.; Pisapia, P.; Mirto, B.F.; Palmieri, A.; Pepe, F.; Bellevicine, C.; Russo, A.; La Civita, E.; et al. Liquid Biopsy in Prostate Cancer Management—Current Challenges and Future Perspectives. Cancers 2022, 14, 3272.
  4. Welch, H.G.; Black, W.C. Overdiagnosis in cancer. J. Natl. Cancer Inst. 2010, 102, 605–613.
  5. Hegde, J.V.; Mulkern, R.V.; Panych, L.P.; Fennessy, F.M.; Fedorov, A.; Maier, S.E.; Tempany, C.M. Multiparametric MRI of prostate cancer: An update on state-of-the-art techniques and their performance in detecting and localizing prostate cancer. J. Magn. Reson. Imaging 2013, 37, 1035–1054.
  6. De Rooij, M.; Hamoen, E.H.J.; Fütterer, J.J.; Barentsz, J.O.; Rovers, M.M. Accuracy of multiparametric MRI for prostate cancer detection: A meta-analysis. Am. J. Roentgenol. 2014, 202, 343–351.
  7. Fütterer, J.J.; Briganti, A.; De Visschere, P.; Emberton, M.; Giannarini, G.; Kirkham, A.; Taneja, S.S.; Thoeny, H.; Villeirs, G.; Villers, A. Can clinically significant prostate cancer be detected with multiparametric magnetic resonance imaging? A systematic review of the literature. Eur. Urol. 2015, 68, 1045–1053.
  8. Di Minno, A.; Aveta, A.; Gelzo, M.; Tripodi, L.; Pandolfo, S.D.; Crocetto, F.; Imbimbo, C.; Castaldo, G. 8-Hydroxy-2-Deoxyguanosine and 8-Iso-prostaglandin F2α: Putative biomarkers to assess oxidative stress damage following robot-assisted radical prostatectomy (RARP). J. Clin. Med. 2022, 11, 6102.
  9. Johnson, L.M.; Turkbey, B.; Figg, W.D.; Choyke, P.L. Multiparametric MRI in prostate cancer management. Nat. Rev. Clin. Oncol. 2014, 11, 346–353.
  10. Barentsz, J.O.; Richenberg, J.; Clements, R.; Choyke, P.; Verma, S.; Villeirs, G.; Rouviere, O.; Logager, V.; Fütterer, J.J. ESUR prostate MR guidelines 2012. Eur. Radiol. 2012, 22, 746–757.
  11. Fusco, R.; Sansone, M.; Granata, V.; Setola, S.V.; Petrillo, A. A systematic review on multiparametric MR imaging in prostate cancer detection. Infect. Agents Cancer 2017, 12, 1–14.
  12. Zhu, G.; Luo, J.; Ouyang, Z.; Cheng, Z.; Deng, Y.; Guan, Y.; Du, G.; Zhao, F. The assessment of prostate cancer aggressiveness using a combination of quantitative diffusion-weighted imaging and dynamic contrast-enhanced magnetic resonance imaging. Cancer Manag. Res. 2021, 13, 5287.
  13. Bulten, W.; Kartasalo, K.; Chen, P.H.C.; Ström, P.; Pinckaers, H.; Nagpal, K.; Cai, Y.; Steiner, D.F.; van Boven, H.; Vink, R.; et al. Artificial intelligence for diagnosis and Gleason grading of prostate cancer: The PANDA challenge. Nat. Med. 2022, 28, 154–163.
  14. Sushentsev, N.; Moreira Da Silva, N.; Yeung, M.; Barrett, T.; Sala, E.; Roberts, M.; Rundo, L. Comparative performance of fully-automated and semi-automated artificial intelligence methods for the detection of clinically significant prostate cancer on MRI: A systematic review. Insights Imaging 2022, 13, 59.
  15. Van Booven, D.J.; Kuchakulla, M.; Pai, R.; Frech, F.S.; Ramasahayam, R.; Reddy, P.; Parmar, M.; Ramasamy, R.; Arora, H. A systematic review of artificial intelligence in prostate cancer. Res. Rep. Urol. 2021, 13, 31.
  16. Hajjo, R.; Sabbah, D.A.; Bardaweel, S.K.; Tropsha, A. Identification of tumor-specific MRI biomarkers using machine learning (ML). Diagnostic 2021, 11, 742.
  17. Trebeschi, S.; Drago, S.G.; Birkbak, N.J.; Kurilova, I.; Cǎlin, A.M.; Pizzi, A.D.; Lalezari, F.; Lambregts, D.M.J.; Rohaan, M.W.; Parmar, C.; et al. Predicting response to cancer immunotherapy using noninvasive radiomic biomarkers. Ann. Oncol. 2019, 30, 998–1004.
  18. Ding, W.; Abdel-Basset, M.; Hawash, H.; Ali, A.M. Explainability of artificial intelligence methods, applications and challenges: A comprehensive survey. Inf. Sci. 2022, 615, 238–292.
  19. Bagherzadeh, J.; Asil, H. A review of various semi-supervised learning models with a deep learning and memory approach. Iran J. Comput. Sci. 2019, 2, 65–80.
  20. Cierco Jimenez, R.; Lee, T.; Rosillo, N.; Cordova, R.; Cree, I.A.; Gonzalez, A.; Indave Ruiz, B.I. Machine learning computational tools to assist the performance of systematic reviews: A mapping review. BMC Med. Res. Methodol. 2022, 22, 322.
  21. Hazratifard, M.; Gebali, F.; Mamun, M. Using machine learning for dynamic authentication in telehealth: A tutorial. Sensors 2022, 22, 7655.
  22. Learned-Miller, E.G. Introduction to Supervised Learning; Department of Computer Science, University of Massachusetts: Amherst, MA, USA, 2014; p. 3.
  23. Castiglioni, I.; Rundo, L.; Codari, M.; Di Leo, G.; Salvatore, C.; Interlenghi, M.; Gallivanone, F.; Cozzi, A.; D’Amico, N.C.; Sardanelli, F. AI applications to medical images: From machine learning to deep learning. Phys. Med. 2021, 83, 9–24.
  24. Dike, H.U.; Zhou, Y.; Deveerasetty, K.K.; Wu, Q. Unsupervised learning based on artificial neural network: A review. In Proceedings of the 2018 IEEE International Conference on Cyborg and Bionic Systems (CBS), Shenzhen, China, 25–27 October 2018.
  25. Nasteski, V. An overview of the supervised machine learning methods. Horizons B 2017, 4, 51–62.
  26. Zhu, X.; Goldberg, A.B. Introduction to semi-supervised learning. Synth. Lect. Artif. Intell. Mach. Learn. 2009, 3, 1–130.
  27. Sutton, R.S.; Barto, A.G. Reinforcement learning. J. Cogn. Neurosci. 1999, 11, 126–134.
  28. Dayan, P.; Niv, Y. Reinforcement learning: The good, the bad and the ugly. Curr. Opin. Neurobiol. 2008, 18, 185–196.
  29. Nketiah, G.A.; Elschot, M.; Scheenen, T.W.; Maas, M.C.; Bathen, T.F.; Selnæs, K.M. Utility of T2-weighted MRI texture analysis in assessment of peripheral zone prostate cancer aggressiveness: A single-arm, multicenter study. Sci. Rep. 2021, 11, 2085.
  30. Shahbazi-Gahrouei, D.; Aminolroayaei, F.; Nematollahi, H.; Ghaderian, M.; Gahrouei, S.S. Advanced Magnetic Resonance Imaging Modalities for Breast Cancer Diagnosis: An Overview of Recent Findings and Perspectives. Diagnostics 2022, 12, 2741.
  31. Bonde, A.; Andreazza Dal Lago, E.; Foster, B.; Javadi, S.; Palmquist, S.; Bhosale, P. Utility of the Diffusion Weighted Sequence in Gynecological Imaging. Cancers 2022, 14, 4468.
  32. Manoharan, D.; Das, C.J.; Aggarwal, A.; Gupta, A.K. Diffusion weighted imaging in gynecological malignancies-present and future. World J. Radiol. 2016, 8, 288.
  33. Arledge, C.A.; Sankepalle, D.M.; Crowe, W.N.; Liu, Y.; Wang, L.; Zhao, D. Deep learning quantification of vascular pharmacokinetic parameters in mouse brain tumor models. Front. Biosci. Landmark 2022, 27, 99.
  34. Di Campli, E.; Pizzi, A.D.; Seccia, B.; Cianci, R.; d’Annibale, M.; Colasante, A.; Cinalli, S.; Castellan, P.; Navarra, R.; Iantorno, R.; et al. Diagnostic accuracy of biparametric vs multiparametric MRI in clinically significant prostate cancer: Comparison between readers with different experience. Eur. J. Radiol. 2018, 101, 17–23.
  35. Kam, J.; Yuminaga, Y.; Krelle, M.; Gavin, D.; Koschel, S.; Aluwihare, K.; Sutherland, T.; Skinner, S.; Brennan, J.; Wong, L.M.; et al. Evaluation of the accuracy of multiparametric MRI for predicting prostate cancer pathology and tumour staging in the real world: An multicentre study. BJU Int. 2019, 124, 297–301.
  36. Ippolito, D.; Querques, G.; Pecorelli, A.; Perugini, G.; Roscigno, M.; Da Pozzo, L.F.; Maino, C.; Sironi, S. Diagnostic accuracy of multiparametric magnetic resonance imaging combined with clinical parameters in the detection of clinically significant prostate cancer: A novel diagnostic model. Int. J. Urol. 2020, 27, 866–873.
  37. Zhao, Y.; Simpson, B.S.; Morka, N.; Freeman, A.; Kirkham, A.; Kelly, D.; Whitaker, H.C.; Emberton, M.; Norris, J.M. Comparison of multiparametric magnetic resonance imaging with prostate-specific membrane antigen positron-emission tomography imaging in primary prostate cancer diagnosis: A systematic review and meta-analysis. Cancers 2022, 14, 3497.
More
© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/)
ScholarVision Creations
Feedback