Roles of mpMRI in my contemporary practice
- Identify “targets” for biopsy to increase pick up rate of clinically significant cancers (reduce false negative biopsies)
- Reduce the number of biopsies performed unnecessarily (avoid negative biopsies)
- Local staging of prostate cancer (nodes and estimate of T stage)
Additional uses:
- Identify lesions “missed” on an initial negative biopsy
- Aid in active surveillance (both candidates for, and ongoing monitoring)
- Aid in determining eligibility for focal therapy
- Investigation of other prostate diseases (abscess, sarcoma)
As per EAU guidelines, I :
- Perform MRI prostate before biopsy
- Use MRI as a tool for biopsy indication
Evidence for use
PROMIS:
- 93 % sensitivity for detecting prostate cancer (cf 48 % TRUS biopsy)
- 89 % NPV for csPCa (cf 74 % TRUS biopsy)
PRECISION:
- 38 % csPCa found in MRI targeted biopsy vs 26 % in systematic biopsy
- 9 % gleason 6 prostate cancer in MRI group (neg MRI no biopsy) vs 22 % in systematic biopsy without MRI
- 28 % negative MRI and therefore no biopsy
In summary – mpMRI before biopsy enhances detection of clinically significant prostate cancer (accepting 10 – 15 % false negative rate) and can reduce the number of unnecessary biopsies and reduce diagnosis of Gleason 6 disease.
Hardware and interpretation
Multiparametric MRI should be performed in centres with high volume and be read by experienced radiologists.
mpMRI provides anatomical and functional information.
A 3 Tesla scanner is preferred for better images (“higher signal to noise ratio”)
Endorectal coils were traditionally used for 1.5 T machines and are not standard practice now where I refer. Buscopan is often used to reduce bowel motion artefact, and the rectum is preferentially empty.
Traditional mpMRI sequences:
- T2 weighted
- T1 weighted
- Diffusion weighted imaging (DWI)
- Dynamic contrast enhancement (DCE)
The T1 images generally show the whole pelvis, while T2 and DWI images are more focussed on the prostate for better signal to noise ratios.
Normal anatomy on MRI
On a T2 scan:
- Peripheral zone has higher signal. Can be compressed and thin with BPH.
- Linear or wedge shaped hypointense/low signal areas are generally benign
- Transition zone enlarged with “organised chaos” appearance of BPH.
- “Smudged charcoal” more suspicious
- On T2, water is white.
- SVs – homogenous high signal fluid within thin low signal walls.
- Anterior FMS – low signal, continuous with bladder neck.
Tumour on a T2 scan is generally dark (low signal). This should be correlated with T1 – because haemorrhage post-biopsy can also appear dark on T2, but will be bright white high signal on T1.
Assessing nodes:
- Generally seen on the T1 images with largest field of view
- Concerning features – > 8 mm, rounded shape, loss of fatty hilum, low signal similar to tumour on T2, irregular margins
Bone mets generally are seen as low intensity (dark) spots on the T1 images (and T2). Benign bone lesions are common and need to be considered in context and compared with other imaging.
Diffusion weighted imaging
The cornerstone of modern MRI prostate – it is based on the diffusion of water molecules within tissue against a high magnetic field gradient.
- More tightly packed/denser cells – cancer cells – exhibit less water diffusion (“restricted diffusion”)
- The difference between water diffusion in cancer and normal cells becomes more marked with higher magnetic gradients (b values)
- As the magnetic gradient and b values increase, the tumours become more obvious, but the signal to noise ratio becomes worse
Tumour is bright / high signal on the DWI images and becomes brighter on higher b value images.
The average diffusion coefficient (ADC) map is produced from an average of 3 DWI images (b50 – b1000).
Tumour is dark / low signal on ADC map.
Higher grade tumours tend to have more densely packed cells, and therefore lower ADC values / brighter high signal areas on high b value.
ADC value < 900 generally suspicious, and the lower the more likely to be high grade cancer.
Dynamic contrast enhancement
Prostate cancer is vascular with angiogenesis, and more aggressive tumours have higher vessel microdensity – therefore cancer has early and more prominent contrast enhancement.
Studies underway to see if DCE can feasibly be omitted (making MRI “bi-parametric”)
Post contrast scans for at least 2 minutes, with 15 second runs.
Most suspicious is early focal enhancement within 10 seconds of contrast in femoral arteries.
Interpretation of prostate MRI – PIRADs
Use of the standardised PIRADs v2.1 reporting system is strongly encouraged.
Different scoring systems and methods for peripheral or transition zone lesions.
Peripheral zone – almost exclusively scored on DWI images, with DCE “tie-breaker” for 3’s.
| DWI image interpretation | |
| 1 | No abnormality |
| 2 | Linear/wedge-shaped areas (hypointense on ADC, hyperintense on high b value) |
| 3 | Focal area hypointense on ADC/hyperintense on high b value, or markedly hypo/hyperintense on one but not the other
If +ve contrast enhancement can be upgraded to a 4 |
| 4 | Focal area markedly hypointense on ADC / markedly hyperintense on high b val ; < 15 mm |
| 5 | As above, > 15 mm or definite concern for EPE |
Transition zone – usually scored on the T2 images, with occasional upgrades if DWI very concerning.
| T2 image interpretation | |
| 1 | Normal / completely encapsulated nodule |
| 2 | Mostly encapsulated nodule, homogenous circumscribed nodule, homogenous mildly hypointense area between nodules |
| 3 | Heterogenous signal intensity with obscured margins, not fitting other PIRADs score |
| 4 | Lenticular or non-circumscribed, homogenous, moderately hypointense < 15 mm |
| 5 | As above, > 15 mm or concern for EPE |
*DWI scores 2 higher than the T2 score, allows upgrade in total of 1 point.
What is the likelihood of a clinically significant prostate cancer being found based on MRI?=
| 2 | 5 – 10 % |
| 3 | > 20 % |
| 4 | > 60 % |
| 5 | > 75 % |
Note MRI is not sensitive for picking up Gleason 6 disease.
Excellent article from Rosenkrantz and Taneja re: MRI pitfalls here
