Secondary Logo

Journal Logo

How are we going to train a generation of radiologists (and urologists) to read prostate MRI?

Puech, Philippea,b,c; Randazzo, Marcod; Ouzzane, Adilb,c,e; Gaillard, Vianneya,b; Rastinehad, Ardeshirf; Lemaitre, Laurenta,b; Villers, Arnauldb,c,e

doi: 10.1097/MOU.0000000000000217
IMAGING IN PROSTATE CANCER – IMPLICATIONS FOR THE UROLOGIST: Edited by Mark Emberton
Free

Purpose of review Multiparametric MRI has gained tremendous importance in the daily practice for patients at risk or diagnosed with prostate cancer. Interpretation of multiparametric-MRI is a complex task, supposedly restricted to experienced radiologists. The purpose of this review is to analyze fundamentals of multiparametric-MRI interpretation and to describe how multiparametric-MRI training could be organized.

Recent findings Recently, professional guidelines have been published to provide technical and interpretation frameworks and harmonize multiparametric-MRI practice, but the question of physicians training in prostate multiparametric-MRI reading is still pending. What kind of education, practice, and training makes a radiologist able to reliably interpret a prostate multiparametric-MRI? How can findings be reported to be easily understood? How much experience is needed? How can we train urologists and other physicians to review the examinations they request? Is double-reading necessary?

Summary An institutional-based competency certification process for prostate multiparametric-MRI interpretation may encourage nonspecialized radiologists to qualify for prostate imaging in a standardized and reproducible way, exactly as urologists need it.

aCHU Lille, Department of Radiology, F59000 Lille

bUniv. Lille, U1189 - ONCO-THAI - Thérapies Lasers Assistées par l’Image pour l’Oncologie, F59000 Lille

cInserm, U1189, F-59000 Lille, France

dUniversity Hospital Zürich, Department of Urology, Zürich, Switzerland

eCHU Lille, Department of Urology, F59000 Lille, France

fIcahn School of Medicine at Mount Sinai New York, New York, USA

Correspondence to Pr Philippe Puech, MD, PhD, Service d’imagerie génito-urinaire et ORL Hôpital Claude Huriez CHRU de Lille 1, rue Michel Polonovski 59037, Lille, Cedex, France. Tel: +33 3 20 44 59 83; e-mail: puech@dicomworks.com

Back to Top | Article Outline

INTRODUCTION

Over the last decade, the developments in MRI technology have established the role of multiparametric prostate MRI for the detection, staging, surveillance, and treatment planning of prostate cancer. In experienced hands, this examination allows distinction of significant cancers requiring a potentially invasive treatment, from indolent cancers that only require active surveillance.

Physicians ordering prostate MRI, and especially urologists, need to be confident regarding the MRI report they receive, because the decision-making process (e.g., if or not to treat a patient with prostate cancer) depends among other things on the count, location, and radiographic stage of lesions given by the radiologist. Today, unless they personally know the radiologist who performed the examination, they have no rational clue on how reliable the final report is, and still, many of them believe that multiparametric-MRI should be left to the experts.

In an effort to codify prostate MRI protocols, interpretation and reporting, to provide indicators on MRI quality, as well as to standardize the practices, professional guidelines such as European Society of Uro-Radiology's (ESUR) and American College of Radiology's (ACR) Prostate Imaging-Reporting and Data System (PI-RADS) have recently been published [1,2]. It is a real step forward for the adoption of imaging in prostate cancer management. Nevertheless, even when the examination is technically optimized, a significant interobserver variability remains in reporting prostate MRI findings. Basically, a reader accuracy depends on three skills: the ability to identify a clinically significant lesion within the parenchyma (to distinguish it from normal or inflammatory tissue); to rule out a number of false-positive images and preserve multiparametric-MRI's high negative predictive value; to assign a reliable PI-RADS and staging score for each lesion.

Currently, it is to be established what makes a radiologist (residents, nonspecialized practitioners, and others) efficient and autonomous in these three basic interpretation skills, what comes from knowledge, and what comes from experience. Some authors have shown that young residents, having a simple training, could yield similar accuracy as experts [3,4▪▪]. In a recently published study, there was a significant improvement in diagnostic accuracy after a 2-year study period with almost doubling of the cancer detection rates.

In this review, we will consider the training requirements for urologists and radiologists on how to read MRI.

Box 1

Box 1

Back to Top | Article Outline

WHAT DO UROLOGISTS (AND CLINICIANS) EXPECT FROM PROSTATE MULTIPARAMETRIC-MRI?

Urologists and other clinicians (radiation therapists, oncologists, pathologists, etc.) can not be specialized in imaging, MRI, and prostate image interpretation, but they should be able to recognize lesions from report without assistance (for instance by using corresponding slice number(s) on the anatomical T2-w series) as they will be in first line to show suspicious lesions to the patient, to coregister magnetic resonance and ultrasound data in case of fusion biopsy or image-guided treatment in the operating room (high intensity focalized ultrasound or other focal therapy techniques, etc.) [5]. They need landmarks on protocol quality and conformance to guidelines to check that the examinations they request are correctly performed.

According to recent guidelines, physicians appreciate standardized MRI prostate reporting of lesions along with a prose report [6–8]. Several items on what urologists (and clinicians) expect from prostate multiparametric-MRI data, and from its standardized reporting are detailed in Table 1.

Table 1

Table 1

Back to Top | Article Outline

BASIC STEPS OF MULTIPARAMETRIC-MRI INTERPRETATION AND REPORTING

Localization of suspicious lesions on multiparametric imaging

The first step of the interpretation process is to detect foci suspicious for clinically significant cancer within the gland. This requires a knowledge of prostate cancer's natural history, gland anatomy and radioanatomy, cancer and benign tissue semiology. In a recent article [5], we encouraged a systematic and independent review of each of the three main compartments of the gland: peripheral zone, transition zone, and then anterior fibromuscular stroma, as each one has a slightly different semiology. Such analysis allows faster reading, and review of the entire gland, without omission.

Back to Top | Article Outline

Scoring and staging of suspicious images

Once detected, the second step is to provide a degree of suspicion of malignancy for all significant images. This classification can be qualitative or quantitative (with a score). Classifications are always based on semiology criteria, but these latter are numerous (up to 14 per lesion), often subjective, sequence-specific, and complicated by multiples exceptions (size, location, etc.) [4▪▪,5,9]. Thus, the main complaint of urologists is the subjectivity, complexity, and low reproducibility of multiparametric-MRI [10,11].

In 2012, following a consensus conference [6], ESUR published professional recommendations [1] in an effort to harmonize practices.

They introduced the PI-RADS score to help classify multiparametric-MRI lesions. An updated version of this score is available [2]. Several studies have shown good accuracies of the PI-RADS for classification of suspicious images [4▪▪,12▪,13▪,14], and moderate to good inter-reader agreement between readers. Two studies suggest that the PI-RADS score could be used as a triage test for selecting patients with suspicious images requiring target biopsies or conversely to defer biopsy procedure, depending on high (4 and 5 out of 5) or low (1 or 2 out of 5) PI-RADS score, respectively [15,16]. However, PI-RADS has few drawbacks:

  1. Several studies have shown that PI-RADS accuracy and reproducibility for prostate cancer detection was slightly lower than pre-existing subjective scoring techniques [4▪▪,14,17], suggesting that PI-RADS criteria do not completely reflect, at the moment, components of an expert's judgement.
  2. A large proportion of significant MRI images remain ‘equivocal’ (score of 3 out of 5) and cannot clearly be classified. Strict application of semiology criteria could overcome this drawback [13▪], such as use of computer-aided diagnosis (CAD) systems either using atlases, computerized decision systems, or automatic analysis [18,19,20▪,21–24], new MRI sequences [25,26], biomarkers [27] or new postprocessing algorithms that may be included in next PI-RADS versions.
Back to Top | Article Outline

Reporting of significant images

Format, content, and terminology

There is great variability in the way multiparametric-MRI is reported (conventional free-text, semistructured report, description of lesions with or without suspicion score, use of PI-RADS or different score, addition or not of schematic prostate map or key images, etc.) [5,8,28–32,33▪]. It has been proven that referring physicians prefer structured reports, and that structured reports are evidence of the reader's training and knowledge in the domain [34].

Prostate MRI reports should include all relevant information to allow use of multiparametric-MRI data under all circumstances (consultation with the patient, patient himself, multidisciplinary meeting, further comparison, follow-up, double-reading, etc.), and easy data collection or comparison. Therefore, according to most authors [1,2,5–7,30,31,33▪], its format has to be structured, and its content to be in a standardized and understandable terminology. We describe a framework and the fundamental content of a standardized prostate multiparametric-MRI report in Table 2.

Table 2

Table 2

Table 2

Table 2

ESUR guidelines include a dedicated section describing key elements for the reporting of this examination [1]. The primary objective of this approach is to harmonize practices in Europe, but thanks to a virtuous circle [35], structured reporting allows wider promotion of guidelines, quicker adoption in radiology and urology, and in the case of PI-RADS an extension outside Europe, to the rest of the world [2]. The recent ACR/ESUR joint work on PI-RADS version 2 includes a detailed glossary of terms, in an additional effort to harmonize the terminology [2]. Recently, Silveira et al.[33▪] an improvement prostate multiparametric-MRI quality reports by using a structured template and informatics tools to generate it in routine.

Back to Top | Article Outline

Report appendices

In addition to the structured reports, a copy of a standardized prostate map, as described in current guidelines [1], with manual or electronic drawing of lesions (Fig. 1), position, and size relative to the schematic gland, as well as key images for the index lesion should be included in the final report, to help physicians localize, recognize the most suspicious lesions [5,30,31,36,37▪]. Lesions having a suspicion of extracapsular extension should clearly have margins outside the contours of the schematic prostate slice [37▪]. It is unclear whether a 16, 27 [38], or 39 [2] sectors prostate map is optimal for the localization of lesions; however, all schematic representations of the prostate divide the gland in three craniocaudal sections (base, mid-gland, and apex), with similar functionality [1,2,5,31,37▪,38,39].

FIGURE 1

FIGURE 1

Back to Top | Article Outline

Tools for interpretation

Several information technology tools are required to ensure optimal and comfortable image interpretation and reporting:

  1. First, the most important is a dedicated interpretation software, able to display multiple series simultaneously, and sync them in three-dimention. Depending on the context (fast review, standard or extended multiparametric-MRI protocol), up to 20 series of images have to be synchronized in real-time, as shown in Table 3.
  2. Second, a database dedicated to prostate imaging allowing the tracking of activity, data collection and sharing, patient follow-up, disease management (pending results, correlation, etc.), key image collection, teaching.
  3. Third, CAD software that may help radiologists detect or characterize suspicious images [18,19,20▪,21–24].
  4. Last, new computer-assisted reporting tools are promising additions to radiology information system, or prostate-specific databases, for data collection and sharing, and rapid building of standardized reports [33▪,37▪,40▪,41].
Table 3

Table 3

Back to Top | Article Outline

LEVELS OF COMPETENCE IN MULTIPARAMETRIC-MRI READING

  1. Image-review is a simplified set of technical and reading skills, sufficient to meet the requirements of physicians prescribing and using multiparametric-MRI routinely (urologists, radiation therapists, medical oncologists, pathologists). Image review basically consists of recognizing and localizing the lesion(s) described on the report, to put it in the clinical context of the patient, and take full advantage of imaging for patient management.
  2. Image interpretation is the ability to search, select, and describe relevant information on the images, in order to answer a clinical question rose by a referring physician. Interpretation leads to the reduction of a report, and engages the radiologist's responsibility. Only appropriately trained physicians should carry out this duty. In addition to technical skills (MRI physics, acquisition of images), interpretation requires knowledge of prostate diseases, radioanatomy, semiology and its variants, and a capacity to transcribe these findings into an intelligible report that will be useful in the management of the patient. All of this is part of the curriculum of radiology trainees, and partially available in CME programs or congress sessions.

Table 4 summarizes differences between image interpretation and image review.

Table 4

Table 4

Back to Top | Article Outline

Reader's experience

Both image interpretation and image review benefit from the reader's experience [3,42,43], but there is no established indicator of that experience in practice, and no study has investigated the time necessary to finish the so-called ‘learning curve’ [44,45]. Radiologists fulfill it at different speeds, depending on their abilities, their investment, and the number of cases they have analyzed. There are several classes of ‘experience’ for image interpretation in the literature, usually expressed by the years in practice: nonexperienced; ‘experienced’ (2–3 years), and ‘of high experience’ (>5 years). Readers having about 2 years of experience in genito-urinary imaging are usually considered as experienced. Depending on center size, this represents between 400 and 1200 cases (5–15 cases a week) with full analysis of indications, clinical settings, multiparametric-MRI images, histology feedback, and patient outcome. Yet, there is no rational relationship between the years or number of cases of experience, and an established ‘experience’ or ‘high experience’ status. Two recent articles, including a consensus conference agreed on a minimal number of 50 patient cases [31,46] to be considered as sufficiently trained to read multiparametric-MRI. Additionally, the quality of initial training, the ability to easily review, share, or discuss difficult cases with colleagues and to get high-quality disorder feedback can certainly balance this purely quantitative approach of what is called ‘experience’.

After initial education (1–2 days seminary dedicated to prostate multiparametric-MRI reading), a radiologist should be considered as able to read the examinations but under a second reading performed by an experienced reader. After an intermediate period of 50–100 cases modulated by the mutual appreciation of the interest of this process, they should be considered as ‘experienced’ and work autonomously. Quantification of first and secondary readings concordances in time may help determine the optimal duration of this intermediate phase. In centers having only one radiologist dedicated to prostate MRI, double-reading could be achieved with help of remote tele-expertise services.

Table 5 summarizes the main elements that may be required to ensure experience and qualification in prostate multiparametric-MRI interpretation.

Table 5

Table 5

Back to Top | Article Outline

Double-reading

Multiparametric-MRI interpretation is always a challenging task, especially for the less experienced readers. Double-reading is commonly used in original research studies to improve multiparametric-MRI reading robustness, to reduce the risk of missing significant images due to human mistake [4▪▪,12▪,15,47]. Double-reading may be of interest depending on initial findings:

  1. To confirm negativeness of multiparametric-MRI and consolidate its negative predictive value (NPV) in cases with PI-RADS 1 or 2;
  2. To improve classification and selectivity in cases with PI-RADS 3;
  3. To improve local staging accuracy in cases with high risk PI-RADS 4 or 5.

However, there is no evidence to date, that this process can improve prostate multiparametric-MRI accuracy in routine, as it does for breast imaging, whose classification is also based on a standardized score [48–50]. Further studies, based on prospective evaluation of large routine series of prostate multiparametric-MRI are still expected.

In our experience, double-reading is especially interesting in the startup-phase of prostate imaging training, as readers with low experience tend to describe typically false-positive images, or still need methodological assistance. Further studies will be required to answer this question in the future.

Back to Top | Article Outline

Standard levels of competence

On the basis of previous observations, and similarly to what other disciplines did [51], it is possible to propose three standardized levels of competence in prostate multiparametric-MRI reading:

  1. Level I: experience on selecting the appropriate diagnostic modality, reviewing images, and using the results. This level does not include performing the technique.
  2. Level II: having achieved initial training, with practical experience in image interpretation but with cover of a double-reading.
  3. Level III: able to independently perform and interpret the examination under all circumstances, with confidence, without double-reading.
Back to Top | Article Outline

TRAINING

General aspects of training in prostate multiparametric-MRI

The European Society of Radiology has published a description of three levels of training, consistent with a forementioned levels of competence. Levels I and II correspond to years 1–3 and 4–5 of residency training periods, respectively. Level III corresponds to a subspecialization beyond the fifth year of training [52]. Level I training in uroradiology includes ‘understanding of imaging features and differential diagnoses of pathologies of the prostate, seminal vesicles, and testes/scrotum’. Level II training includes ‘description of zonal anatomy of the prostate’ and ‘description of imaging features of prostatic zones with ultrasound and MRI’, as well as ‘inflammatory and tumoral disorders of the prostate’.

In order to specialize in prostate MRI (and reach a level III competence), radiologists need to acquire advanced clinical and interpretation skills specific to prostate cancer. They benefit of a level III baseline technical background. Inversely, urologists and other physicians have a strong (level III) baseline clinical background that radiologists do not, but usually no education in MRI physics (level I), and only basic (level I or undergraduate) reading skills. Consequently, physicians with different specialties and levels of knowledge will share a common base of technical, clinical, and interpretation skills, and will have to deepen each of them, depending on their background, in order to review or interpret prostate multiparametric-MRI, and use it at its best.

Initial education is the cornerstone of prostate multiparametric-MRI reading. In 2009, Akin et al.[3] showed significant increase in diagnostic accuracy with area under the ROC curve climbing from 0.52 to 0.66 after a single didactic education, and remaining stable after a 200 cases training curriculum. In a series of 31 cases read by five readers, Garcia-Reyes et al.[53▪] showed that a single dedicated education allows significant increase of accuracy (from 74.2 to 87.7% for index lesion detection and from 54.8 to 73.5% for Gleason assessment) and reader confidence.

Usually, dedicated education in prostate MRI is part of the genitourinary imaging curriculum, or performed during a period of fellowship. Nonspecialized radiologists will be able to get it through various CME programs that are frequently opened to urologists and other physicians.

Back to Top | Article Outline

Available material

Currently, there are numerous teaching courses on prostate MRI provided all over the world. As coaching is increasingly important, online, web-based, and even tablet-based prostate MRI analyzing tools have been developed to support radiologists [4▪▪]. In Table 6, we collected a selection of teaching material or dedicated courses available online, or regularly repeated. Many are live CME courses focused on image interpretation (with ‘hands-on’ sessions), and consequently aimed at radiologists. However, provided that basic training in technique is acquired, many of them will be accessible to urologists. CME courses usually include didactic lectures and interactive practical sessions tailored to meeting the level of knowledge of the audience. In Table 7, we present the program of a theoretical CME course typically designed for radiologists and other physicians, with technical, clinical, and interpretation training pathways, and corresponding requirements in knowledge.

Table 6

Table 6

Table 7

Table 7

Table 7

Table 7

Back to Top | Article Outline

Certification

There is currently no certification in prostate multiparametric-MRI interpretation. Certificates of attendance for masterclasses or live CME courses are available to testify involvement into a level I competence process. An experience of prostate multiparametric-MRI reading in a specialized imaging department during a traineeship, residency or assistantship, with an experience of about 50–100 cases may testify of a good experience, but what makes a radiologist completely autonomous and experienced in prostate imaging remains to be delineated. No trainee assessment program had been published to date. A scalable process, from level I to level III of competence should be organized and validated by the institutions or third party organizations [54].

Back to Top | Article Outline

CONCLUSION

Because of its increasing importance in prostate cancer management, a structured training is required for radiologists willing to interpret prostate multiparametric-MRI. Urologists, as well as other specialists also need to learn how to review this examination, as they need it daily for clinical decisions, treatment planning, or transrectal ultrasound/MRI fusion procedure purposes. Education includes technical, clinical, and image interpretation fundamentals. Currently, there is plenty of teaching material available for this purpose, but initial education must be consolidated with a practical experience in image interpretation (or review). Prior to being considered as experienced readers, we believe radiologists could benefit of a temporary intermediate competency certification process based on the experience of 50–100 cases and supervised by a systematic double-reading. This may encourage nonspecialized radiologists to involve in prostate multiparametric-MRI, and ensure they can perform, interpret and report this examination in a standardized and reproducible way, exactly as urologists need it.

Back to Top | Article Outline

Acknowledgements

None.

Back to Top | Article Outline

Financial support and sponsorship

None.

Back to Top | Article Outline

Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

  • ▪ of special interest
  • ▪▪ of outstanding interest
Back to Top | Article Outline

REFERENCES

1. Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR guidelines 2012. Eur Radiol 2012; 22:746–757.
2. 2015ACR ISCoEa. PI-RADS v2. Prostate Imaging and Report and Data System: Version 2: European Society of UroRadiology (ESUR); 2014. http://www.esur.org/fileadmin/content/user_upload/_v2_20141223.pdf. [Accessed 25 July 2015]
3. Akin O, Riedl CC, Ishill NM, et al. Interactive dedicated training curriculum improves accuracy in the interpretation of MR imaging of prostate cancer. Eur Radiol 2010; 20:995–1002.
4▪▪. Vache T, Bratan F, Mege-Lechevallier F, et al. Characterization of prostate lesions as benign or malignant at multiparametric MR imaging: comparison of three scoring systems in patients treated with radical prostatectomy. Radiology 2014; 272:446–455.

This article shows robust methodology. It shows that classical scoring systems (Likert-like) perform better than version 1.0 of the PI-RADS, suggesting that its criteria do not perfectly reflect prostate cancer MRI semiology, and opening way for improvements in the new version of PI-RADS.

5. Puech P, Villers A, Ouzzane A, Lemaitre L. Prostate cancer: diagnosis, parametric imaging and standardized report. Diagn Interv Imaging 2014; 95:743–752.
6. Dickinson L, Ahmed HU, Allen C, et al. Magnetic resonance imaging for the detection, localisation, and characterisation of prostate cancer: recommendations from a European consensus meeting. Eur Urol 2011; 59:477–494.
7. Dickinson L, Ahmed HU, Allen C, et al. Scoring systems used for the interpretation and reporting of multiparametric MRI for prostate cancer detection, localization, and characterization: could standardization lead to improved utilization of imaging within the diagnostic pathway? J Magn Reson Imaging 2012; 37:48–58.
8. Moore CM, Kasivisvanathan V, Eggener S, et al. Standards of reporting for MRI-targeted biopsy studies (START) of the prostate: recommendations from an International Working Group. Eur Urol 2013; 64:544–552.
9. Puech P, Ouzzane A, Gaillard V, et al. Multiparametric MRI-targeted TRUS prostate biopsies using visual registration. Biomed Res Int 2014; 2014:819360.
10. Mueller-Lisse U, Mueller-Lisse U, Scheidler J, et al. Reproducibility of image interpretation in MRI of the prostate: application of the sextant framework by two different radiologists. Eur Radiol 2005; 15:1826–1833.
11. Dickinson L, Ahmed HU, Allen C, et al. Clinical applications of multiparametric MRI within the prostate cancer diagnostic pathway. Urol Oncol 2013; 31:281–284.
12▪. Muller BG, Shih JH, Sankineni S, et al. Prostate cancer: interobserver agreement and accuracy with the revised prostate imaging reporting and data system at multiparametric MR imaging. Radiology 2015; 142818.

This original research is first to assess accuracy of the new PI-RADS version 2.0 score for detection of clinically significant prostate cancer. However, it concludes to moderate reproducibility.

13▪. Hamoen EH, de Rooij M, Witjes JA, et al. Use of the Prostate Imaging Reporting and Data System (PI-RADS) for prostate cancer detection with multiparametric magnetic resonance imaging: a diagnostic meta-analysis. Eur Urol 2015; 67:1112–1121.

This review article pooled data of 14 studies in the literature about PI-RADS scoring, and confirmed reliable sensitivity, specificity, and NPV of prostate multiparametric-MRI for identifying clinically significant prostate cancer.

14. Rosenkrantz AB, Lim RP, Haghighi M, et al. Comparison of interreader reproducibility of the prostate imaging reporting and data system and likert scales for evaluation of multiparametric prostate MRI. AJR Am J Roentgenol 2013; 201:W612–W618.
15. Renard-Penna R, Mozer P, Cornud F, et al. Prostate imaging reporting and data system and likert scoring system: multiparametric MR imaging validation study to screen patients for initial biopsy. Radiology 2015; 275:458–468.
16. Grey AD, Chana MS, Popert R, et al. Diagnostic accuracy of magnetic resonance imaging (MRI) prostate imaging reporting and data system (PI-RADS) scoring in a transperineal prostate biopsy setting. BJU Int 2015; 115:728–735.
17. Rosenkrantz AB, Kim S, Lim RP, et al. Prostate cancer localization using multiparametric MR imaging: comparison of Prostate Imaging Reporting and Data System (PI-RADS) and Likert scales. Radiology 2013; 269:482–492.
18. Betrouni N, Makni N, Lakroum S, et al. Computer-aided analysis of prostate multiparametric MR images: an unsupervised fusion-based approach. Int J Comput Assist Radiol Surg 2015.
19. Lemaitre G, Marti R, Freixenet J, et al. Computer-aided detection and diagnosis for prostate cancer based on mono and multiparametric MRI: a review. Comput Biol Med 2015; 60:8–31.
20▪. Wang S, Burtt K, Turkbey B, et al. Computer aided-diagnosis of prostate cancer on multiparametric MRI: a technical review of current research. Biomed Res Int 2014; 2014:789561.

This article is a nice and detailed review of current literature on CAD analysis of multiparametric prostate MRI data.

21. Zhao K, Wang C, Hu J, et al. Prostate cancer identification: quantitative analysis of T2-weighted MR images based on a back propagation artificial neural network model. Sci China Life Sci 2015; 58:666–673.
22. Regge D, Halligan S. CAD: how it works, how to use it, performance. Eur J Radiol 2013; 82:1171–1176.
23. Niaf E, Rouviere O, Mege-Lechevallier F, et al. Computer-aided diagnosis of prostate cancer in the peripheral zone using multiparametric MRI. Phys Med Biol 2012; 57:3833–3851.
24. Puech P, Betrouni N, Makni N, et al. Computer-assisted diagnosis of prostate cancer using DCE-MRI data: design, implementation and preliminary results. Int J Comput Assist Radiol Surg 2009; 4:1–10.
25. Rosenkrantz AB, Geppert C, Grimm R, et al. Dynamic contrast-enhanced MRI of the prostate with high spatiotemporal resolution using compressed sensing, parallel imaging, and continuous golden-angle radial sampling: preliminary experience. J Magn Reson Imaging 2015; 41:1365–1373.
26. Rakow-Penner RA, White NS, Parsons JK, et al. Novel technique for characterizing prostate cancer utilizing MRI restriction spectrum imaging: proof of principle and initial clinical experience with extraprostatic extension. Prostate Cancer Prostatic Dis 2015; 18:81–85.
27. Dikaios N, Alkalbani J, Sidhu HS, et al. Logistic regression model for diagnosis of transition zone prostate cancer on multiparametric MRI. Eur Radiol 2015; 25:523–532.
28. Wibmer A, Vargas HA, Sosa R, et al. Value of a standardized lexicon for reporting levels of diagnostic certainty in prostate MRI. AJR Am J Roentgenol 2014; 203:W651–W657.
29. Westphalen AC, Rosenkrantz AB. Prostate imaging reporting and data system (PI-RADS): reflections on early experience with a standardized interpretation scheme for multiparametric prostate MRI. AJR Am J Roentgenol 2014; 202:121–123.
30. Rothke M, Blondin D, Schlemmer HP, Franiel T. PI-RADS classification: structured reporting for MRI of the prostate. Rofo 2013; 185:253–261.
31. Kirkham AP, Haslam P, Keanie JY, et al. Prostate MRI: who, when, and how? Report from a UK consensus meeting. Clin Radiol 2013; 68:1016–1023.
32. Quentin M, Blondin D, Klasen J, et al. Evaluation of a structured report of functional prostate magnetic resonance imaging in patients with suspicion for prostate cancer or under active surveillance. Urol Int 2012; 89:25–29.
33▪. Silveira PC, Dunne R, Sainani NI, et al. Impact of an information technology-enabled initiative on the quality of prostate multiparametric MRI reports. Acad Radiol 2015; 22:827–833.

Implementation of a structured report template and CAD tool significantly improves the quality of prostate multiparametric-MRI reports.

34. Plumb AA, Grieve FM, Khan SH. Survey of hospital clinicians’ preferences regarding the format of radiology reports. Clin Radiol 2009; 64:386-394395-386.
35. Kahn CE Jr, Heilbrun ME, Applegate KE. From guidelines to practice: how reporting templates promote the use of radiology practice guidelines. J Am College Radiol 2013; 10:268–273.
36. El-Shater Bosaily A, Parker C, Brown LC, et al. PROMIS - Prostate MR imaging study: a paired validating cohort study evaluating the role of multiparametric MRI in men with clinical suspicion of prostate cancer. Contemp Clin Trials 2015; 42:26–40.
37▪. Puech P. Prostate MRI structured report wizard. http://www.pcih.fr/mpmri. [Accessed 25 July 2015]

This is the URL to a free online web-based computer assisted reporting tool, designed to easily draw multiparametric-MRI findings on a standardized prostate map.

38. Haffner J, Lemaitre L, Puech P, et al. Role of magnetic resonance imaging before initial biopsy: comparison of magnetic resonance imaging-targeted and systematic biopsy for significant prostate cancer detection. BJU Int 2011; 108:E171–E178.
39. Barentsz J, Villers A, Schouten M. ESUR prostate MR guidelines. Author reply. Eur Radiol 2013; 23:2322–2323.
40▪. Litjens GJ, Barentsz JO, Karssemeijer N, Huisman HJ. Clinical evaluation of a computer-aided diagnosis system for determining cancer aggressiveness in prostate MRI. Eur Radiol 2015; doi 10.1007/s00330-015-3743-y. [Epub ahead of print].

This article shows that PI-RADS assessment benefits of the addition of objective criteria, based on computer assisted image analysis.

41. Mata C, Walker PM, Oliver A, et al. ProstateAnalyzer: web-based medical application for the management of prostate cancer using multiparametric MR imaging. Inform Health Soc Care 2015; 1–21.
42. Futterer JJ, Engelbrecht MR, Huisman HJ, et al. Staging prostate cancer with dynamic contrast-enhanced endorectal MR imaging prior to radical prostatectomy: experienced versus less experienced readers. Radiology 2005; 237:541–549.
43. Mullerad M, Hricak H, Wang L, et al. Prostate cancer: detection of extracapsular extension by genitourinary and general body radiologists at MR imaging. Radiology 2004; 232:140–146.
44. Latchamsetty KC, Borden LS Jr, Porter CR, et al. Experience improves staging accuracy of endorectal magnetic resonance imaging in prostate cancer: what is the learning curve? Can J Urol 2007; 14:3429–3434.
45. Harris RD, Schned AR, Heaney JA. Staging of prostate cancer with endorectal MR imaging: lessons from a learning curve. Radiographics 1995; 15:813–829.
46. Muller BG, Futterer JJ, Gupta RT, et al. The role of magnetic resonance imaging (MRI) in focal therapy for prostate cancer: recommendations from a consensus panel. BJU Int 2014; 113:218–227.
47. Trimboli RM, Verardi N, Cartia F, et al. Breast cancer detection using double reading of unenhanced MRI including T1-weighted, T2-weighted STIR, and diffusion-weighted imaging: a proof of concept study. AJR Am J Roentgenol 2014; 203:674–681.
48. Klompenhouwer EG, Duijm LE, Voogd AC, et al. Variations in screening outcome among pairs of screening radiologists at nonblinded double reading of screening mammograms: a population-based study. Eur Radiol 2014; 24:1097–1104.
49. Klompenhouwer EG, Voogd AC, den Heeten GJ, et al. Blinded double reading yields a higher programme sensitivity than nonblinded double reading at digital screening mammography: a prospected population based study in the south of The Netherlands. Eur J Cancer 2015; 51:391–399.
50. Klompenhouwer EG, Weber RJ, Voogd AC, et al. Arbitration of discrepant BI-RADS 0 recalls by a third reader at screening mammography lowers recall rate but not the cancer detection rate and sensitivity at blinded and nonblinded double reading. Breast 2015; doi: 10.1016/j.breast.2015.06.004. [Epub ahead of print].
51. Popescu BA, Andrade MJ, Badano LP, et al. European Association of Echocardiography recommendations for training, competence, and quality improvement in echocardiography. Eur J Echocardiogr 2009; 10:893–905.
52. Ertl-Wagner B. European curriculum for further education in radiology. Der Radiologe 2014; 54:11061107-1110.
53▪. Garcia-Reyes K, Passoni NM, Palmeri ML, et al. Detection of prostate cancer with multiparametric MRI (mpMRI): effect of dedicated reader education on accuracy and confidence of index and anterior cancer diagnosis. Abdominal Imaging 2015; 40:134–142.

This article shows that simple training (two dedicated lectures) for prostate multiparametric-MRI reading significantly increases readers accuracy (especially for anterior tumors detection), and reduces interobserver variability, with a yield of +13.5%.

54. Howard SA, Krajewski KM, Weissman BN, et al. Cancer imaging training in the 21st century: an overview of where we are, and where we need to be. J Am Coll Radiol 2015; 12:714–720.
Keywords:

education; guidelines; interpretation; MRI; prostate; reporting

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.