Bosniak Classification, Version 2019

For a long time radiologists graded renal cysts on the original Bosniak scale using words like "thin," "minimally thickened," and "thick" with no agreed ruler. Two readers could look at the same septum, one calling it thin and the other thick, and both be following the rules. Version 2019 is the update that hands everyone the same ruler. It keeps the five familiar rungs, but every rung now has a definition you can measure with the calliper tool, and it spells out the attenuation shortcuts that let you close a case without contrast at all.

This page is the full table. The entity pages, simple versus complex cysts and the approach to the cystic renal mass, carry an abbreviated version and link back here when you need the fine print.

What it grades and why

The Bosniak classification answers one question: how likely is this cystic renal mass to be a cancer, and what should happen next? It was never meant to be a histology predictor. It is a management ladder. Each class maps to an action: leave it, watch it, refer it, or treat it.

The underlying biology is simple. A cyst is a water balloon; a cystic cancer is a water balloon with living tumor growing on the inside of the rubber. Living tissue has blood supply, blood supply carries contrast, and contrast shows up as enhancement. So the whole scale is a graded description of how much enhancing tissue is in the wall and septa, and what shape it takes. Smooth and thin means the lining is quiet. Thick, irregular, or lumpy means something is growing.

The 2019 revision fixed three things about the old system:

1. It gave the adjectives numbers. Thin, minimally thickened, and thick are now ≤2 mm, 3 mm, and ≥4 mm. Irregular is defined as convex protrusions ≤3 mm with obtuse margins. A nodule is a convex protrusion ≥4 mm with obtuse margins, or of any size with acute margins.
2. It made calcification almost irrelevant. Any type of calcification is allowed in class II. Calcium on its own does not raise the class; only enhancing tissue does.
3. It named the attenuation shortcuts explicitly so that a homogeneous hyperdense mass on unenhanced CT, or a homogeneous slightly-above-water mass on a portal venous scan, can be called class II without a dedicated study.

The system

Every measurement below comes from the version 2019 proposal. Where the class definition uses attenuation, the phase it applies to is part of the definition, not a detail; the same HU value means something different on an unenhanced and a contrast-enhanced scan.

Bosniak classification of cystic renal masses, version 2019 (CT criteria). Thin = ≤2 mm; minimally thickened = 3 mm; thick = ≥4 mm.

A few things worth reading twice:

• Class I versus class II subtype 4 look identical on paper (homogeneous −9 to 20 HU). The difference is context: class I is the textbook simple cyst on a study where you can see the thin wall and the absence of septa; subtype 4 is the shortcut that lets you call a homogeneous water-attenuation mass benign on an unenhanced scan without ever seeing it enhance. Both end the same way: no follow-up.
• "May enhance" appears in class I and class II. A thin wall or a thin septum with a normal blood supply lights up faintly. That is expected and does not promote the mass. What promotes the mass is thickening, multiplicity, irregularity, or a nodule in enhancing tissue.
• The ≥70 HU rule is for unenhanced CT only. On contrast-enhanced CT both a hyperdense cyst and a solid cancer can measure ≥70 HU, so the shortcut evaporates the moment contrast is on board.
• Class III versus class IV is a shape question. Both have enhancing tissue that is more than thin. In class III that tissue is a thick wall or septum (≥4 mm) or a low, obtuse-angled bump (≤3 mm). In class IV it is a nodule: taller (≥4 mm) with obtuse margins, or any size with acute margins. Picture a hill versus a mushroom. The hill is III; the mushroom is IV.

Figure · CT

The MRI side of the ladder

Version 2019 is a CT and MRI system. The wall, septal, and nodule definitions above carry over unchanged, and MRI is often the better tool for counting septa and judging subtle enhancement, particularly with subtraction images. The MRI-specific attenuation-equivalent shortcuts (the signal-intensity rules that let certain homogeneous T1- or T2-bright masses be called class II or IIF without contrast) I am not tabulating here rather than quote them from memory. The teaching point survives without them: on MRI, as on CT, a homogeneous mass with no enhancing tissue is benign, and the class rises only with enhancing thickening, irregularity, or a nodule.

Worked examples

Resolve each case against the table. The habit to build is: measure the attenuation, then measure the thickest enhancing wall or septum, then look for a nodule.

Case 1. The hyperdense surprise. A 2 cm well-circumscribed, homogeneous mass in the lower pole measures 78 HU on an unenhanced CT done for stones. No contrast was given.

Resolution: homogeneous and ≥70 HU on unenhanced CT is class II (subtype 2), a hyperdense cyst. No further evaluation. Do not order a contrast study to "confirm"; the rule exists precisely so you don't have to.

Case 2. The wall that grew up. On a renal-mass-protocol CT, a 4 cm cystic mass has a smooth wall measuring 3 mm that enhances between the unenhanced and nephrographic phases. No septa, no nodules.

Resolution: a smooth, minimally thickened (3 mm) enhancing wall is class IIF. Follow-up imaging. Had the wall measured ≥4 mm it would be thick and the mass would be class III.

Case 3. Two fat septa. A cystic mass contains two septa, both smooth, both enhancing, each measuring 5 mm across.

Resolution: enhancing septa ≥4 mm in width are thick; one or more of them makes the mass class III. Indeterminate, roughly half of these are malignant, and the report should recommend urologic referral. Note that the number of septa (two, which is "few") would have been fine on its own; the thickness is what carries it up the ladder.

Case 4. The mushroom. A cystic mass has a thin wall and a single thin septum, but on the post-contrast images a 5 mm enhancing convex protrusion with obtuse margins arises from the wall.

Resolution: an enhancing convex protrusion ≥4 mm with obtuse margins is a nodule; a nodule makes the mass class IV, roughly 90% malignant. Treatment for appropriate candidates. Had the same protrusion been 2 mm with obtuse margins it would count as irregular, class III. Had it been 2 mm with acute margins it would still be a nodule, class IV; acute margins make size irrelevant.

Where it breaks

Measurement lives at the edge of the ruler. The difference between IIF and III is one millimetre of wall (3 versus ≥4 mm), and the difference between II and IIF is the fourth thin septum. On thick slices with partial-volume blur, a 3 mm wall and a 4 mm wall can be honestly hard to separate. Thin slices, magnified views, and measuring perpendicular to the wall all help, and when the call is genuinely borderline, MRI often settles it.

Attenuation shortcuts have preconditions. Every HU-based subtype of class II says homogeneous. A mass that is ≥70 HU on unenhanced CT but heterogeneous is not a hyperdense cyst; it is a mass that needs pre- and post-contrast characterization. Likewise, the 21–30 HU rule is specific to the portal venous phase; the same number on a corticomedullary or delayed phase does not qualify.

Pseudoenhancement. Small cysts buried in avidly enhancing cortex can appear to gain a few Hounsfield units after contrast through a beam-hardening artifact rather than any real uptake. I'm not going to give you a specific HU cutoff for enhancement or for pseudoenhancement here; the concept is that an apparent small rise in a small intraparenchymal cyst should make you suspicious of artifact before you promote the mass, and MRI or ultrasound can arbitrate.

Inter-reader agreement. How consistently two radiologists assign the same class is exactly what v2019 was designed to improve, and the proposal paper itself calls for validation. I am not going to quote a kappa for version 2019's inter-reader reliability. What I can tell you is that the gray zone lives at IIF versus III, which is why the definitions there are so fussy.

The scale ranks risk; it does not diagnose histology. Class IV is roughly 90% malignant, which means roughly one in ten is not. Class III is roughly half. The classification tells the urologist how worried to be, and the pathologist tells everyone what it was.

The one-line version to carry out of here: measure the thickest enhancing wall or septum, look for a nodule, and read the number off the table. Version 2019 turned a vocabulary test into a measuring test, and measuring is something you can do the same way every time.

References

• Silverman SG, Pedrosa I, Ellis JH, et al. Bosniak Classification of Cystic Renal Masses, Version 2019: An Update Proposal and Needs Assessment. Radiology 2019;292(2):475–488. Used for: every class definition in "The system" (the ≤2 mm thin, 3 mm minimally thickened, and ≥4 mm thick wall and septal cutoffs; the 1–3 few and ≥4 many septa counts; the ≤3 mm irregular and ≥4 mm nodule protrusion definitions; the −9 to 20 HU, ≥70 HU unenhanced, >20 HU non-enhancing, and 21–30 HU portal venous attenuation subtypes of class II); the "roughly half" (class III) and "roughly 90%" (class IV) malignancy summaries; the follow-up, referral, and treatment actions per class; and the resolution of every case in "Worked examples" (the 78 HU, 3 mm wall, 5 mm septa, and 5 mm nodule examples) against those thresholds.
• Herts BR, Silverman SG, Hindman NM, et al. Management of the Incidental Renal Mass on CT: A White Paper of the ACR Incidental Findings Committee. J Am Coll Radiol 2018;15(2):264–273. Used for: the statement that the ≥70 HU hyperdense-cyst rule applies to unenhanced CT only because both RCC and hyperdense cysts may measure ≥70 HU after contrast (in "The system" and "Where it breaks").