VI-RADS: Vesical Imaging-Reporting and Data System
- VI-RADS is a five-point MRI score for one question: has a bladder tumor invaded the muscularis propria (the detrusor) or not?
- It reads three sequences, T2W, DWI, and DCE, and each one is asking whether the tumor respects the low-signal muscle line on T2W.
- Category 1 is a lesion under 1 cm with a continuous muscle line; category 2 is a larger lesion that still has a reassuring stalk or thickened inner layer; category 3 has lost those but keeps the line; category 4 has an interrupted line; category 5 is tumor out in the perivesical fat.
- When T2W and DWI disagree, DWI is the dominant sequence; if DWI is unusable, DCE takes over.
- It is a probability estimate of local muscle invasion, not a tissue diagnosis and not a full stage.
Every reporting system is born from the same frustration: ten radiologists looking at the same picture and writing ten different paragraphs. For the bladder, the paragraph everyone was writing badly was "I think it might be into the muscle." The urologist needed a yes or a no, or at least a number, because a tumor that stays above the detrusor can be shaved out through a scope, and one that has breached it usually means the bladder comes out. VI-RADS (Vesical Imaging-Reporting and Data System, "vesical" simply meaning "of the bladder") was published in 2018 to turn that paragraph into a single digit. This page is the full system; the entity-level story of the disease lives at bladder cancer and VI-RADS.
What it grades and why
VI-RADS grades exactly one thing: the likelihood that a bladder tumor has invaded the muscularis propria, on a multiparametric MRI performed before the urologist has resected the tumor. It does not grade the tumor's cell type, its grade under the microscope, whether there are nodes, or whether there is disease elsewhere. Think of it as a very well-trained dog that does one trick.
Why does one boundary matter that much? Because the bladder wall is layered like a mille-feuille, and the treatment forks at one specific layer. From the lumen outward: the urothelium, then the loose lamina propria (together the "inner layer"), then the thick muscularis propria, then the perivesical fat. Tumor confined to the inner layer is non-muscle-invasive and is managed with endoscopic resection and bladder instillations. Tumor into the muscularis propria is muscle-invasive, and the conversation changes to cystectomy or chemoradiation. The same lump, one layer deeper, is a different disease.
MRI can see those layers, which is the whole opportunity. On T2-weighted (T2W) images the muscularis propria is a crisp low-signal (dark) line, the inner layer above it is intermediate to high signal, and the fat outside it is bright. A tumor is intermediate signal, and it is either sitting on top of that dark line, or it isn't. Diffusion-weighted imaging (DWI) adds a second opinion: packed tumor cells restrict water motion, so the tumor is bright on high-b-value DWI and dark on the ADC map, while the muscle underneath is not. Dynamic contrast-enhanced (DCE) imaging adds a third: tumor and the inflamed inner layer enhance early, the muscularis propria enhances later, so early enhancement that reaches into the muscle is a sign the tumor has too.
Two morphologic clues do a lot of the work. A papillary tumor often grows on a stalk, a fibrovascular core that lifts the tumor off the wall; a visible stalk means the tumor is hanging off the wall rather than burrowing into it. A sessile (broad-based) tumor has no stalk, but if the inner layer beneath it is thickened and high-signal on T2W (edematous, inflamed, but not tumor), that thickened cushion is playing the same reassuring role: it is intact tissue between tumor and muscle.
The system
VI-RADS 2018 (Panebianco et al., Eur Urol 2018). Each sequence is scored on the same five-point ladder; the final category is assigned from the combination, with the dominant-sequence rule below.
| Category | T2W | DWI | DCE | Likelihood of muscle invasion |
|---|---|---|---|---|
| 1 | Lesion <1 cm with a continuous low-signal line of muscularis propria (with or without a stalk or a thickened inner layer) | Muscularis line intact | Muscularis line intact | Highly unlikely |
| 2 | Lesion >1 cm with a continuous low-signal muscularis line: papillary tumor with a stalk, or sessile tumor with a thickened high-signal inner layer | Muscularis line intact | Muscularis line intact | Unlikely |
| 3 | No stalk (papillary) or no high-signal thickened inner layer (sessile), but the low-signal muscularis line is still continuous | The reassuring stalk or thickened inner layer is absent, but the muscularis line is still continuous | The reassuring stalk or thickened inner layer is absent, but the muscularis line is still continuous | Equivocal |
| 4 | Interruption of the low-signal muscularis propria line | Focal extension of tumor into the muscularis propria | Focal extension of early enhancement into the muscularis propria | Likely |
| 5 | Tumor extends through the entire bladder wall into perivesical fat | Tumor extends through the entire wall into perivesical fat | Tumor extends through the entire wall into perivesical fat | Very likely, with invasion beyond the bladder |
The dominant-sequence rule. When T2W and DWI disagree, DWI is the dominant sequence and sets the category. If DWI is suboptimal (motion, susceptibility, gas in the bladder, a poor ADC map), DCE becomes the dominant sequence instead. In practice that makes T2W the anatomical map, DWI the referee, and DCE the substitute referee.
Notice what the ladder is really tracking. Categories 1 and 2 have an intact muscle line plus a reassuring buffer (small size, a stalk, or a thick inner layer). Category 3 has an intact line but no buffer. Category 4 has a broken line. Category 5 has no wall left at all. Once you see it as "buffer, line, fat," the whole table fits on the back of your hand.
The size cut between categories 1 and 2 is 1 cm, and it only applies while the muscle line is continuous on every sequence. A tiny lesion that has already broken through the line is a category 4, not a category 1 with an asterisk.
Worked examples
Each case is resolved by walking the three sequences and then applying the dominant-sequence rule. The lesion sizes are illustrative; the only sourced size in the system is the 1 cm cut.
Case 1: the small polyp
A patient with painless hematuria has a single lesion on the posterior wall that measures well under 1 cm. On T2W it sits on top of a continuous dark muscularis line; on DWI it is bright but the line beneath it is intact; on DCE it enhances early and the muscle beneath enhances later, line intact.
Resolution: lesion <1 cm with a continuous muscularis line on T2W, intact on DWI and DCE. VI-RADS 1. Muscle invasion highly unlikely. The urologist resects it through the scope with confidence.
Case 2: the big tumor on a stalk
A papillary tumor a few centimetres across projects into the lumen on a clearly visible stalk. On T2W the stalk is seen and the low-signal muscle line beneath the base is continuous. DWI shows the bright tumor stopping at an intact line. DCE shows early enhancement of the tumor and stalk with the muscle line intact.
Resolution: lesion >1 cm, papillary with a stalk, muscularis line continuous on T2W and intact on DWI and DCE. VI-RADS 2. Big is not the same as deep; the stalk is the reassurance.
Case 3: the sessile tumor with nothing to hold on to
A broad-based tumor on the lateral wall. On T2W there is no stalk and the inner layer beneath the tumor is not thickened or bright, but the dark muscularis line is still continuous under the whole base. DWI agrees: bright tumor, intact line, no reassuring inner layer. DCE agrees.
Resolution: no stalk and no high-signal thickened inner layer, but the muscularis line is still continuous. VI-RADS 3. Equivocal. This is the honest "I can't promise" category, and the report should say so rather than rounding it up or down.
Case 4: the sequences argue
A sessile tumor at the trigone. On T2W the low-signal muscle line beneath it looks interrupted, which on its own would be category 4. But DWI is of good quality and shows the bright tumor sitting on a continuous, intact muscle line, with no thickened inner layer and no stalk, which on DWI is a category 3 pattern. DCE also shows the line intact.
Resolution: T2W says 4, DWI says 3, and when T2W and DWI disagree, DWI is dominant. VI-RADS 3. Had the DWI been unreadable because of gas or motion, DCE would have become the dominant sequence instead, and here DCE agrees with DWI.
Case 5: through and through
A large sessile mass with T2W tumor signal replacing the muscle line and extending into the bright perivesical fat; DWI shows restricted tumor through the entire wall into the fat; DCE shows early enhancement through the wall and into the fat.
Resolution: tumor extends through the entire bladder wall into perivesical fat on T2W, DCE, and DWI. VI-RADS 5. Invasion of muscle and beyond the bladder very likely, and the report now moves on to nodes and adjacent organs, which VI-RADS itself does not score.
Where it breaks
After the scope, not before. VI-RADS was designed for tumors that have not yet been resected. A transurethral resection tears up the wall and leaves edema, granulation tissue, and hemorrhage that are intermediate on T2W, can restrict diffusion, and enhance early, in other words a convincing forgery of muscle invasion. Scan before resection whenever the pathway allows; if you must scan after, say so in the report and expect over-calling.
A bladder that isn't full enough. An underdistended bladder folds its wall, and a fold under a tumor looks like a thickened base and can blur the muscle line. Too full and the patient moves. Moderately distended is the target, and the report should note the distension.
When the muscle line was never crisp. Chronic outlet obstruction thickens and trabeculates the detrusor, diverticula have no muscularis propria at all, and prior radiation or instillation therapy inflames everything. In all of these the dark line is harder to trust, and category 3 gets used a lot.
DWI is only a referee if it is legible. Gas in the bladder, motion, and susceptibility from hip prostheses all wreck DWI, and that is precisely the situation the rule anticipates: DCE becomes the dominant sequence. Do not let a distorted DWI decide a category.
The inner layer can lie in both directions. A thickened high-signal inner layer is reassuring, but inflammation from the tumor itself or from infection can thicken it too, and a thin inner layer under a small lesion does not on its own mean invasion. The category comes from the muscle line, not from the cushion above it.
It scores the wrong question for some tumors. Flat carcinoma in situ has no mass to score. Tumors inside a diverticulum have no muscle to invade. Non-urothelial histologies (adenocarcinoma, squamous, urachal) were not what the system was built on. Use judgment and say what you are unsure about.
It is not staging. VI-RADS estimates local muscle invasion. Lymph nodes, the upper tracts (which need CT urography), the prostate and rectum, and distant disease are all outside its remit and have to be reported separately.
How well does it actually work? Meta-analyses of VI-RADS's sensitivity, specificity, and inter-reader agreement have been published, but I'm not quoting one here, so this page states none. Treat the category as a well-founded probability estimate, and remember that the pathologist's slide, not the MRI, gives the final answer.
The single most common way to misuse VI-RADS is to score the T2W image, glance at the rest, and move on. The rule is explicit: when T2W and DWI disagree, DWI wins. Read the DWI as carefully as the anatomy, and if the DWI is not readable, say so and let DCE decide.
VI-RADS is a five-rung ladder built around one dark line. Buffer intact and line intact: 1 or 2. Buffer gone, line intact: 3. Line broken: 4. Wall gone: 5. DWI is the referee, DCE the substitute. Learn the ladder, respect the referee, and the paragraph that used to say "might be into the muscle" becomes a number the urologist can act on.
References
- Panebianco V, Narumi Y, Altun E, et al. Multiparametric Magnetic Resonance Imaging for Bladder Cancer: Development of VI-RADS (Vesical Imaging-Reporting And Data System). Eur Urol 2018;74(3):294–306. Used for: the 2018 publication date, the VI-RADS 1 to 5 category definitions on T2W, DWI, and DCE (including the <1 cm and >1 cm size cut, the stalk and thickened high-signal inner layer criteria, the interrupted muscularis propria line, and extension into perivesical fat), the likelihood statements per category, and the dominant-sequence rule (DWI dominant over T2W; DCE dominant when DWI is suboptimal), in Key Points, "The system", "Worked examples", and "Where it breaks".