By: Dan Sperling, MD

Multiparametric magnetic resonance imaging (mpMRI) is increasingly accepted as a standard of care for the detection and characterization of prostate cancer. The utilization of the two parameters considered to be the workhorse of prostate MRI, namely T2 weighted (T2 MRI) and diffusion weighted imaging (DWI) reveals detailed anatomic features coupled with functional tissue information that differentiates malignant lesions from normal gland tissue. Two other parameters, dynamic contrast enhancement and MR-spectroscopy, identify tumor blood flow and metabolic characteristics, respectively. When read by an experienced radiologist, mpMRI demonstrates high sensitivity and specificity for prostate cancer.[i] The high negative predictive value of mpMRI is useful in ruling the need for a biopsy in or out.

However, less experienced readers may find benign abnormalities such as benign prostatic hyperplasia (BPH) and prostatitis ambiguous, thus confounding their interpretations. One such condition is high grade prostatic intraepithelial neoplasia (PIN), which is considered a precancerous condition. High grade PIN does not cause symptoms, and cannot be detected by PSA or prostate ultrasound. It is usually found incidental to prostate biopsies, but is virtually impossible to detect and monitor by conventional means.

Improvements in MRI technology, correlation with prostatectomy pathology, and reader training are making it possible to identify and follow high grade PIN by means of mpMRI. Dwivedi et al. (2016)[ii] performed pre-biopsy mpMRI on 116 men with elevated PSA suspicious for prostate cancer. Nine were found to have high grade PIN by biopsy by no prostate cancer. Their pathology was compared with corresponding regions on mpMRI, and based on DWI and spectroscopy, the apparent diffusion coefficients (ADC) and metabolite ratios were calculated. The nine patients were followed to determine clinical outcomes, but over time only five remained available for follow up by mpMRI. Four of these patients were diagnosed with cancer on repeat biopsy. Their mpMRI at diagnosis showed further reduction in ADC – and the lower the ADC value, the greater the correlation with prostate cancer. There was no significant change in metabolite ratio. The authors concluded that the DWI imaging of high grade PIN and ADC values close to those of cancer “may help identify such men who may be candidates for close follow-up.”

Diffusion weighted imaging is the subject of a study by Litjens et al. (2016)[iii] using computer-extracted features of prostatectomy specimens. Their retrospective study involved the presurgery mpMRI scans of 70 men scheduled for prostatectomy. All scans were done on a 3T magnet and included T2-MRI, DWI-MRI and dynamic contrast enhanced imaging. Following prostatectomy, the specimen slides were digitized and annotated on software for cancer and noncancerous conditions. The annotated specimen images were coregistered (fused) with each patient’s mpMRI results using an interactive deformable coregistration scheme. According to the authors, “Computer-identified features for each of the noncancerous disease categories (e.g., benign prostatic hyperplasia [BPH], prostatic intraepithelial neoplasia [PIN], inflammation, and atrophy) and prostate cancer were extracted. Feature selection was performed to identify the features with the highest discriminatory power.” Their goal was to determine which features best distinguished between cancer and each benign condition, as well as their association with benign disease class and prostate cancer grade. With regard to high grade PIN, they found that the ADC results from DWI-MRI were most suitable for identifying PIN from cancer.

Both studies show that mpMRI, particularly diffusion weighted imaging, can be used to detect high grade PIN. More importantly, monitoring patients with high grade PIN is feasible, and reductions in the ADC values may signal a PIN progression into prostate cancer. This is a promising development that can reduce both the number of first biopsies and repeat biopsies.

[i] Margolis D. Multiparametric MRI for Localized Prostate Cancer: Lesion Detection and Staging. BioMed Research International. Volume 2014 (2014), Article ID 684127, 11 pages. http://dx.doi.org/10.1155/2014/684127

[ii] Dwivedi DK, Kumar R, Bora GS, Sharma S, Thulkar S, Gupta SD, Jagannathan NR (2016). Multiparametric MR can identify high grade prostatic intraepithelial neoplasia (HGPIN) lesions and predict future detection of prostate cancer in men with a negative initial prostate biopsy. Article in press, published online 19 May, 2016. DOI: http://dx.doi.org/10.1016/j.mri.2016.05.006

[iii] Litjens GJ, Elliott R, Shih NN, Feldman MD et al. Computer-extracted features can distinguish noncancerous confounding disease from prostatic adenocarcinoma at multiparametric MR imaging. Radiology. 2016 Jan;278(1):135-45. doi: 10.1148/radiol.2015142856.