Multiparametric MRI: Standardizations Needed

The article by Gupta et al does a nice job of summarizing the role of multiparametric MRI (mpMRI) in the diagnosis and treatment of prostate cancer. Such an article is necessary because MRI of the prostate in 2013 is much different than it was even 5 or 10 years ago. For many years, the primary role of prostate MRI was to stage known disease-specifically, to determine whether carcinoma was confined to the gland in candidates for radical prostatectomy.

The potential role of prostate MRI has in recent years expanded significantly, and there is great hope that mpMRI can help address two major problems encountered when treating prostate cancer: overtreatment and understaging. Prostate-specific antigen (PSA) level, the currently used screening tool for prostate cancer, is nonspecific and often leads to prostate biopsy. Despite the fact that transrectal ultrasound (TRUS)-guided biopsy routinely misses and understages cancers, many patients are still unnecessarily treated for identified indolent cancers that will not kill them.

The goal of a multiparametric approach to prostate MRI is to increase the accuracy of the identification of tumors within the prostate. Many recent publications have confirmed that a multiparameter approach achieves this goal,[1-6] although with sensitivities ranging from 49% to 95%. The identification by mpMRI of tissue foci within the prostate suspicious for neoplasm has many potential benefits: it can identify lesions in patients with elevated PSA levels but negative prior TRUS biopsy, guide biopsies in patients who are candidates for active surveillance, provide serial information on the size of tumors under active surveillance, provide reassurance that low-risk patients do not harbor significant disease, and identify lesions for localized therapy. These uses should help us accurately stage and follow disease and reduce the incidence of unnecessary whole-gland treatment.

Recent studies have shown that MRI/US-guided biopsies diagnose significant cancers in patients with prior negative TRUS biopsy, independent of the number of prior negative biopsies.[7] In addition, the degree of suspicion for focal neoplasm on mpMRI correlates with the likelihood of obtaining a positive targeted biopsy[7-9] and the likelihood of Gleason 7 or greater disease,[8-10] while the size of a tumor on MRI correlates with its actual size on histopathology.[11] Finally, glands with no intermediate- or high-suspicion lesions on mpMRI are unlikely to harbor clinically significant disease.[12,13]

While the ability of mpMRI to aid in the treatment of prostate cancer has been described, it is premature to conclude that mpMRI has a key role in the treatment of prostate cancer. Much work remains to be done before the approach can be practically implemented. Although the sequences included in an mpMRI exam are available on most commercial scanners (ie, available outside of the research environment), there has been no standardization of the acquisition or interpretation/reporting of the images.

For example, in diffusion-weighted imaging (DWI), calculated apparent diffusion coefficient (ADC) values are dependent on the b-values chosen for the DWI acquisition. These values are not standardized and vary within the literature. Combining this variability with the known overlap in ADC values between benign and malignant tissue, it can be difficult for the practicing radiologist to know when to call DWI/ADC images abnormal. Some authors have developed a graded scoring system based on the average ADC value of the lesion. However, many authors simply define an abnormal focus as a region that is “dark” on the ADC map but bright on the high b-value DWI image, but they do not specify how the images are windowed or leveled to create this contrast.

Dynamic contrast-enhanced MRI (DCE-MRI) is less standardized than DWI. Spatial resolution, temporal resolution, contrast bolus rate, and interpretation strategy all affect the images, making it difficult to compare results between studies. In addition, the quantitative DCE-MRI analysis used by many authors is dependent on the model used and on accurate measurement of contrast concentration in the vasculature and in the tissues. Determining contrast concentration is technically challenging, and different techniques affect the model output (typically numeric Ktrans values). Some authors describe abnormal DCE-MRI as simply “early and intense enhancement,” but what does this mean? Is it subjective? Some authors use a graded scoring system based on mean lesion Ktrans values calculated using custom-designed model-fitting software, recognizing that this approach is not reproducible at other institutions. Other authors use commercial software that overlays a semi-quantitative Ktrans map on the T2 images, and then look for focal areas of asymmetry. There remains no standardization or consensus.

In addition to a lack of standardized interpretation (normal vs abnormal) of each individual sequence in the mpMRI exam, there is no standard way of weighting the findings in order to label a lesion as having a “low,” “intermediate,” or “high” suspicion of being cancer. Some authors who correlate mpMRI suspicion with the likelihood of a positive biopsy and Gleason 7 or greater disease simply sum the number of sequences where a region of tissue was abnormal. A focus with abnormal signal on only one sequence is deemed to have “low suspicion”; abnormal signal on two sequences, “intermediate suspicion”; and abnormal suspicion on three sequences, “high suspicion.” Other authors have created graded scoring systems for each sequence; the scores are summed, and receiver operating characteristic (ROC) analysis is used to find a cutoff most accurate for identifying neoplasm. Other authors simply score each lesion from 1 to 5 based on predetermined combinations of subjective and objective criteria for each sequence. Several groups have even created complicated models using linear discriminant analysis and logistic regression models to automatically assign a probability of neoplasm to suspicious foci.[14-16] The European Society of Uroradiology (ESUR) is currently seeking to address this issue of a lack of a standardized way to weigh the findings in an mpMRI exam so as to determine a lesion’s level of suspicion for cancer, and they are attempting to suggest standard criteria for interpreting and reporting mpMRI results.[17] The American College of Radiology is expected to release separate standardized interpretation and reporting guidelines in the near future. Subsequent research will hopefully adhere to these guidelines.

In summary, mpMRI is a promising tool for identifying cancer within the prostate. It has the potential to drastically change the way prostate cancer is staged and treated. However, work remains to make this technique reproducible and accessible to the community-based radiologist and urologist.

Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.

References:

REFERENCES

1. Delongchamps N, Rouanne M, Flam T, et al. Multiparametric magnetic resonance imaging for the detection and localization of prostate cancer: combination of T2-weighted, dynamic contrast-enhanced and diffusion-weighted imaging. BJU Int. 2010;107:1411-8.

2. Turkbey B, Mani H, Shah V, et al. Multiparametric 3T prostate magnetic resonance imaging to detect cancer: histopathological correlation using prostatectomy specimens processed in customized magnetic resonance imaging based molds. J Urol. 2011;186:1818-24.

3. Turkbey B, Pinto P, Mani H, et al. Prostate cancer: value of multiparametric MRI Imaging at 3T for detection – histopathologic correlation. Radiology. 2012;255:89-99.

4. Tamada T, Sone T, Higashi H, et al. Prostate cancer detection in patients with total serum prostate-specific antigen levels of 4-10 ng/mL: diagnostic efficacy of diffusion-weighed imaging, dynamic contrast enhanced MRI, and T2-weighted imaging. Am J Roentgenol. 2011;197:664-70.

5. Chen M, Dang J, Wang C, et al. Prostate cancer detection: comparison of T2-weighted imaging, diffusion-weighted imaging, proton magnetic resonance spectroscopic imaging, and the three techniques combined. Acta Radiol. 2008;5:602-10.

6. Isebaert S, Van den Bergh L, Haustermans K, et al. Multiparametric MRI for prostate cancer localization in correlation to whole-mount histopathology. JMRI. Nov 2012. [Epub ahead of print]

7. Vourganti S, Rastinehad A, Yerram NK, et al. Multiparametric magnetic resonance imaging and ultrasound fusion biopsy detect prostate cancer in patients with prior negative transrectal ultrasound biopsies. J Urol. 2012;188:2152-7.

8. Sonn G, Natarajan S, Margolis D, et al. Targeted biopsy in the detection of prostate cancer using an office based magnetic resonance ultrasound fusion device. J Urol. 2013;189:86-92.

9. Hadaschik B, Kuru T, Tulea C et al. A Novel stereotactic prostate biopsy system integrating pre-interventional magnetic resonance imaging and live ultrasound fusion. J Urol. 2011;186:2214-20.

10. Rosenkrantz A, Deng F, Kim S, et al. Prostate cancer: multiparametric MRI for index lesion localization-a multiple-reader study. Am J Roentgenol. 2012;199:830-7.

11. Turkbey B, Mani H, Aras O, et al. Correlation of magnetic resonance imaging tumor volume with histopathology. J Urol. 2012;188:1157-63.

12. Yerram N, Volkin D, Turkbey B. Low suspicion lesions on multiparametric magnetic resonance imaging predict for the absence of high-risk prostate cancer. BJU Int. 2012;110(11 Pt B):E783-8.

13. Delongchamps N, Beuvon F, Flam T, et al. Multiparametic MRI is helpful to predict tumr focality, stage, and size in patients diagnosed with unilateral low-risk prostate cancer. Prost Cancer and Prost Dis. 2011;14:232-7.

14. Langer D, van der Kwast T, Evans A. Prostate cancer detection with multi-parametric MRI: logistic regression analysis of quantitative T2, diffusion-weighted imaging, and dynamic contrast enhanced-MRI. JRMI. 2009;30:327-34.

15. Hambrock T, Vos P, Hulsbergen-van de Kaa CA, et al. Prostate cancer: computer-aided diagnosis with multiparametric 3-T MR imaging-effect on observer performance. Radiology. 2013; 266:521-30.

16. Peng Y, Jiang Y, Yang C, et al. Quantitative analysis of multiparametric prostate MR images: differentiation between prostate cancer and normal tissue and correlation with Gleason score-a computer-aided diagnosis development sudy. Radiology. 2013. [Epub before print]

17. Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR guidelines 2012. Eur Radiol. 2012;22:746-57.