What are the key selectors or domains in the Prostate Imaging-Reporting and Data System (PI-RADS) 2.1?

PI-RADS 2.1 Selectors and Domains

PI-RADS 2.1 utilizes three primary MRI sequence selectors (T2-weighted imaging, diffusion-weighted imaging with ADC maps, and dynamic contrast-enhanced MRI) that are evaluated in a zone-specific hierarchical manner to generate assessment categories ranging from 1 to 5. 1

The Three Core MRI Sequences (Selectors)

PI-RADS 2.1 assessment is fundamentally based on three distinct MRI sequences that serve as the primary selectors for lesion characterization 2:

• T2-weighted (T2W) imaging: Provides anatomical detail and is the dominant sequence for transition zone (TZ) lesion evaluation 1
• Diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) maps: Serves as the dominant sequence for peripheral zone (PZ) lesion assessment 1, 3
• Dynamic contrast-enhanced MRI (DCE-MRI): Functions in a clarification role, primarily for equivocal peripheral zone lesions (PI-RADS 3) 1, 4

Zone-Specific Hierarchical Assessment

The system employs a zone-dependent approach rather than summing individual sequence scores 3, 4:

Peripheral Zone Lesions

• Primary selector: DWI/ADC is the dominant sequence 1, 3
• Secondary role: DCE-MRI upgrades equivocal DWI score 3 lesions to overall PI-RADS category 4 if positive focal enhancement is present 1, 4
• Supporting role: T2W provides anatomical context 3

Transition Zone Lesions

• Primary selector: T2W imaging is the dominant sequence 1, 3
• Secondary role: DWI/ADC can upgrade equivocal T2W score 2 lesions (atypical BPH nodules with incomplete/absent capsule) to overall PI-RADS category 3 if marked diffusion restriction (DWI score 4-5) is present 1, 5, 6
• This creates the "2+1" TZ lesion pathway introduced in version 2.1 5

The Five Assessment Categories (Output Domain)

The selectors combine to generate five assessment categories that represent increasing probability of clinically significant prostate cancer 1:

• PI-RADS 1: Very low likelihood (clinically significant cancer is highly unlikely to be present)
• PI-RADS 2: Low likelihood (16% cancer probability) 2, 6
• PI-RADS 3: Intermediate/equivocal likelihood (14-33% cancer probability) 2, 7, 6
• PI-RADS 4: High likelihood (37-71% cancer probability) 2, 6
• PI-RADS 5: Very high likelihood (77-91% cancer probability) 2, 6

Key Modifications in Version 2.1

Version 2.1 introduced specific refinements to address interpretation challenges 1, 8:

• Atypical BPH nodule definition: TZ lesions with incomplete or absent capsule (T2 score 2) that can be upgraded to category 3 if marked diffusion restriction is present 5, 6
• "Markedly hypointense" ADC definition: More specific criteria provided, though this actually decreased interobserver agreement (κ = 0.26 in v2.1 vs κ = 0.52 in v2) 8
• "Linear/wedge-shaped" DWI descriptor: Replaced "indistinct hypointense" with improved agreement (κ = 0.52 vs κ = 0.18) and yielded 14 more true-negative versus 5 more false-negative interpretations 8

Important Clinical Caveats

The atypical BPH nodule upgrade pathway (2+1 TZ lesions) has shown limited clinical utility, with only 6% clinically significant cancer detection rate compared to 11% for conventional PI-RADS 3 TZ lesions (not statistically different, p=0.31) 5. One prospective study found upgraded PI-RADS 3 TZ lesions were actually less likely to harbor clinically significant cancer than non-upgraded counterparts (4% vs 20%, p=0.02) 6, suggesting this pathway may generate unnecessary biopsies.

Interobserver agreement improved with version 2.1 in the peripheral zone (κ = 0.64 vs 0.51) but remained similar in the transition zone (κ = 0.60 vs 0.64) 8. The modified TZ T2W category 2 descriptors in version 2.1 showed only fair agreement (κ = 0.21), with lower overall agreement for PI-RADS category 2 in the TZ compared to version 2 (κ = 0.31 vs 0.57) 8.