You are on page 1of 10

Note: This copy is for your personal non-commercial use only.

To order presentation-ready
copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Original Research  n  Genitourinary


Follow-up for Bosniak Category
2F Cystic Renal Lesions1

Nicole M. Hindman, MD
Purpose: To determine percentage of Bosniak category 2F complex
Elizabeth M. Hecht, MD

Imaging
cystic renal masses that progress to malignancy based
Morton A. Bosniak, MD
on serial follow-up studies,and to determine if there are
demographic and/or imaging features associated with
progression.

Materials and This retrospective study was institutional review board–


Methods: approved with waiver of informed consent. Hospital da-
tabase system was searched from January 1, 1996, to
May 1, 2011, for category 2F cysts studied with contrast
agent–enhanced computed tomography or magnetic reso-
nance imaging and followed with serial contrast-enhanced
imaging. Demographics of patients and imaging features
of lesions that progressed were compared with those that
did not. The relationship of these features to progression
or stability was assessed by using x2, Fisher exact, or Co-
chran Armitage trend tests.

Results: Identified in 144 patients (98 men, 46 women; age range,


31–83 years; average, 63 years) were 156 category 2F le-
sions. Follow-up studies were from 6 months to 13 years
(median, 3.6 years; average, 4.2 years). Nineteen of 156
lesions progressed to category 3 or 4 in 6 months to 3.2
years; 17 lesions (89.5% of those that progressed and
10.9% of initial 2F lesions) were malignant and two were
benign. To date, no patients had recurrent or metastatic
disease. Men had significant risk for progression to ma-
lignancy (P = .003). Of 17 category 2F lesions that pro-
gressed to malignancy, 12 were endophytic (P = .02). Cat-
egory 2F lesions with minimally irregular septa (nine of
17; P= .001) or wall (seven of 17; P = .016), and lesions
with indistinct parenchymal interface (nine of 17; P ,
.001) were associated with progression to cancer. A mul-
tilobulated border was not associated with progression (P
= .999).

Conclusion: Based on this study, 10.9% (17 of 156) Bosniak category


2F cystic lesions progress to malignancy, and progression
occurs within 6 months to 3.2 years.

1
 From the Department of Radiology, NYU School of Med-  RSNA, 2014
q

icine, 550 1st Ave, HW 202, New York, NY 10016 (N.M.H.,


M.A.B.); and Department of Radiology, Columbia University, Online supplemental material is available for this article.
New York, NY (E.M.H.). Received December 31, 2012;
revision requested February 12, 2013; revision received
August 11; accepted September 19; final version accepted
March 6, 2014. Address correspondence to N.H. (e-mail:
Nicole.Hindman@nyumc.org).

q
 RSNA, 2014

Radiology: Volume 272: Number 3—September 2014  n  radiology.rsna.org 757


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

C
omplex renal cysts pose a sub- category 3 in appearance. The 2F cate- examinations) were excluded for the
stantial diagnostic and manage- gory has allowed for the safe nonopera- following reasons: 423 examinations
ment dilemma. The Bosniak tive monitoring of complex cystic renal no longer had images available from
classification, introduced in 1986, es- lesions, but the risk of malignancy and archive, 117 examinations had no fol-
tablished an imaging framework for how long lesions should be followed low-up imaging (of which 12 exami-
differentiation of benign and malignant remains unclear in this category. The nations proceeded directly to surgery
cystic renal lesions. It serves as a clini- purpose of this study was to determine with a ratio of 50-to-50 malignant-to-
cally useful tool that communicates im- the percentage of Bosniak category 2F benign pathologic analysis), 27 exami-
aging findings to referring clinicians in a complex cystic renal masses that pro- nations had a follow-up shorter than 6
practical way that aids in management gress to malignancy based on serial months with contrast-enhanced CT or
(1). By using the cyst morphologic and follow-up studies and to determine if MR imaging (these 27 cysts were sta-
enhancement characteristics, cysts can there are demographic and/or imaging ble on this very short-term follow-up),
be categorized into one of five updat- features associated with progression. and 13 were of insufficient quality to
ed groups (categories 1, 2, 2F, 3, and include for analysis. Thus, 225 of 2435
4), with suggested recommendations renal cysts were evaluated. Among
for management of each group (1–4). Materials and Methods these 225 renal cysts, eight patients
The 2F category (“F” for “follow-up”) This Health Insurance Portability and with 10 lesions were previously report-
is composed of lesions that are thought Accountability Act–compliant retro- ed (6). For the follow-up examinations,
most likely to be benign, but still must spective study was institutional re- if the patient had multiple follow-up
be proven to be so by demonstrating view board–approved with a waiver imaging studies, the last obtained fol-
stability over serial imaging examina- of informed consent. A radiology da- low-up study was used for stable Bos-
tions. Category 2F lesions are complex tabase was retrospectively searched niak category 2F cysts, and the first
cystic lesions that have complexity be- from January 1, 1996, through May 1, follow-up study that demonstrated re-
tween category 2 lesions (clearly be- 2011, for all contrast agent–enhanced gression or progression was used for
nign lesions that contain a few very computed tomographic (CT) or mag- changed lesions.
thin and smooth septa or walls with netic resonance (MR) examinations in
possible minimal smooth calcifications which the terms “renal or kidney” and Imaging Technique
and without obvious enhancement) and “cyst,” “cystic,” “complex,” “mass,” “le- CT scans were obtained on one of the
category 3 lesions (these contain thick- sion,” “indeterminate,” or “Bosniak,” following four clinical scanners: a single
ened irregular clearly enhancing septa “2,” “2F,” “Bosniak 3,” or “renal mass” detector row scanner (CT Hi Speed Ad-
or walls). Category 3 lesions are benign appeared in the radiologist’s report or vantage; GE Medical Systems, Milwau-
approximately 40% of the time and ma- impression. Only studies that included kee, Wis), a four-section multi–detector
lignant approximately 60% of the time contrast material administration were row scanner (Somatom Volume Zoom,
(4–8), and they usually require surgery. included because contrast enhance- Siemens AG, München, Germany), a
Category 2F lesions contain a range of ment is a fundamental component of 16-section multi–detector row scan-
appearances; some are closer to cat- the classification system. All prospec- ner (Sensation 16; Siemens AG), or a
egory 2 in complexity, while others tive readings that documented a Bos- 64-section multi–detector row scanner
are more worrisome and are closer to niak category 2F renal cyst (or, for (Sensation 64; Siemens AG). CT pa-
example, “complex cyst” and “lesion”) rameters, contrast agent, and contrast
Advances in Knowledge were collected. The initial database
search yielded 2435 examinations, of
nn There were 10.9% (17 of 156) which 1630 examinations were ex-
Published online before print
Bosniak category 2F cystic renal 10.1148/radiol.14122908  Content code:
cluded because they were performed
lesions that progressed to malig- with noncontrast technique and in- Radiology 2014; 272:757–766
nancy within 6 months to 3.2 evitably read as an indeterminate or Abbreviation:
years on retrospective review and complex cyst. The remainder (580 RCC = renal cell carcinoma
in 6 months to 3.8 years for ini-
tial radiologist interpretation. Author contributions:

nn Based on this study, the recom- Implication for Patient Care Guarantor of integrity of entire study, N.M.H.; study
concepts/study design or data acquisition or data analysis/
mended length of follow-up for nn Dedicated follow-up examina- interpretation, all authors; manuscript drafting or manu-
the majority of Bosniak 2F le- tions, up to 4 years in the ma- script revision for important intellectual content, all authors;
sions is 4 years; however, the jority of Bosniak category 2F le- approval of final version of submitted manuscript, all
length of follow-up may vary sions, are required to establish authors; literature research, N.M.H., M.A.B.; clinical studies,
depending on the complexity of benignity, and longer or shorter all authors; experimental studies, N.M.H.; statistical
analysis, N.M.H.; and manuscript editing, all authors
the lesion, the age, and/or any follow up periods may be appro-
comorbidities of the patient. priate for some lesions. Conflicts of interest are listed at the end of this article.

758 radiology.rsna.org  n Radiology: Volume 272: Number 3—September 2014


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

material injection rate and dose varied. Figure 1


Details of the ranges are given in Table
E1 (online).
MR examinations were performed
on one of three clinical 1.5-T imagers
(Magnetom Avanto, Sonata, or Sym-
phony; Siemens Medical Solutions,
Erlangen, Germany) by using a torso
phased-array coil (six-element ante-
rior and posterior coil arrays for the
Avanto imager; four-element anterior
and posterior coil arrays for the So-
nata and Symphony imagers). Relevant
sequences, parameters, and contrast
agent administration details are sum-
marized in Table E2 (online).

Data Collection
One radiologist (N.M.H., 7 years of
experience in genitourinary radiology)
performed the initial database search,
medical chart review, and retrieval of
images 6 months before analysis of the
examinations (Fig 1). Data recorded at
this time included sex, age, indication
for imaging, imaging protocol, presence
of an additional renal mass, history of
renal or other malignancy, the side of
the lesion (ie, right or left), and the lo-
cation of the lesion (ie, image and series
number). Renal lesions were classified
as benign or malignant based on patho-
logic or follow-up imaging reports.
Before image interpretation, the
three radiologists agreed on imaging Figure 1:  Study flowchart. B2F = Bosniak category 2F, B/W = between.
criteria for a scoring system to be used
for the cyst classifications, and listed
criteria as described as in Table 1. Ad- completely intrarenal lesions greater size)/years between measures], were
ditionally, a training set of 20 unique than 3.0 cm or any intrarenal or par- calculated from the measured cyst size.
Bosniak category 2F, 2, and 3 cysts were tially intrarenal lesions with an indis- The lesion was considered to have pro-
reviewed independently by two radiolo- tinct interface with the adjacent renal gressed if there was a change in Bosniak
gists (E.M.H., 11 years of experience, parenchyma were also considered to category to a 3 or 4 (ie, interval devel-
and N.M.H.) and then in consensus be category 2F cysts. opment of enhancement in a previously
with M.A.B. (52 years of experience) If a lesion was unchanged from the nonenhancing septation, nodule, or
before data collection. These cysts were initial study at follow-up (eg, still had wall; or if there was increased number
not included in the study. appearances consistent with a 2F le- or thickness or irregularity of an en-
A lesion was considered to be sion), it would be called a 2F on fol- hancing septation, nodule, or wall), or
category 2F if it contained any of the low-up. Lesions that regressed on fol- if an enhanced nodule appeared within
following features, typically in isolation: low-up were downgraded to a cyst that the lesion (1–4).
multiple thin septations, septations or was category 1 or 2. Additional features
cyst wall with minimal thickening (ie, of the cysts were evaluated according to Image Review
smooth or minimally irregular) and the outline in Table 1. Change in lesion Six months after the initial image eval-
perceived enhancement, and calcifica- size, found by using the equation (lesion uation, N.M.H. and a second genitouri-
tions that may be chunky or nodular size during last study 2 lesion size dur- nary radiologist (E.M.H.) independently
without associated enhanced soft tis- ing the first study), and the equation for evaluated all of the acquired imaging se-
sue. High-attenuation or high-intensity growth rate, [(follow-up size 2 initial quences of the contrast-enhanced MR

Radiology: Volume 272: Number 3—September 2014  n  radiology.rsna.org 759


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

Table 1
k Analysis of Lesion Characteristics Evaluated By Two Independent Reviewers
Initial Study Follow-up Study
Parameter k Statistic 95% Confidence Interval k Statistic 95% Confidence Interval

Quality of study 0.907 0.803, 1.000 NA NA


No. of septations 0.973 0.956, 0.990 0.950 0.921, 0.978
Thickness of septations 0.881 0.823, 0.939 0.958 0.925, 0.991
Morphology of septations 0.856 0.785, 0.928 0.947 0.902, 0.993
Presence of septal enhancement 0.941 0.894, 0.987 0.951 0.909, 0.993
Wall thickness 0.972 0.934, 1.000 0.956 0.908, 1.000
Morphology of wall 0.915 0.834, 0.997 0.944 0.866, 1.000
Presence of enhancement of wall 0.888 0.812, 0.963 0.951 0.896, 1.000
Presence of multilobulated contour (eg, grape-like lobulated contour) 1.000 NA 0.977 0.932, 1.000
Calcification (CT scans only) 0.928 0.875, 0.980 0.989 0.969, 1.000
Presence or absence of nonenhancing soft tissue nodule 1.000 NA 1.000 NA
Presence or absence of hemorrhage or protein* 1.000 NA 0.987 0.961, 1.000
Interface with parenchyma 1.000 NA 0.937 0.851, 1.000

Note.—NA = not applicable.


* On CT: hemorrhage defined as 50 HU on noncontrast CT. On CT: protein defined as internal Hounsfield units equal or greater than background renal parenchyma on noncontrast CT. On MR imaging:
hemorrhage or protein defined as high signal on both T1 and T1 fat saturated sequences.

or CT images and assigned a Bosniak the 225 cysts on the initial examina- study (median, 6 months; average, 6.9
cyst category to each complex renal tion (Fig 1). After the third reader months); 20 patients with 21 lesions
cyst. For discrepancies between the two reviewed them, nine of 20 cysts were who had follow-up 1–2 years after the
readers, a third radiologist (M.A.B.) ultimately included as category 2F. The original study (median, 1.5 years; aver-
analyzed the cases and resolved any dif- third reader made decisions based only age, 1.5 years); and 110 patients with
ferences of opinion. on the index study, without knowledge 121 lesions who had long-term fol-
of the subsequent examinations. The low-up of 2 years or more (median, 4.5
Final Cohort and Follow-up Intervals third reader excluded 11 of 20 cases in years; average, 5.0 years; range, 2–13.1
Among the 225 cysts that were eval- which nine lesions were downgraded years). Of the 121 lesions with long-
uated, 52 of 225 lesions (15 on CT to a category 2 (all of which were sta- term follow-up longer than 2 years, 27
imaging; 37 on MR imaging) were ex- ble or regressed on follow-up of 1–4 lesions had follow-up examinations at 4
cluded because they were overcalled years [mean 2.4 years]), and two le- years or more (median, 6 years; aver-
as a complex, indeterminate, or 2F sions were upgraded to a category 3 age, 6.7 years; range, 4.2–13.1 years).
cyst and were downgraded to a Bos- (both of which progressed on follow-up
niak category 2 when they were re- imaging within 12 months and were Imaging Study Details
viewed again by both reviewers. Six cystic RCC when they were resected). Of the 156 cysts that were category 2F,
of 225 (four on CT imaging; two on Thus, 69 lesions were excluded from the initial imaging modalities were 77
MR imaging) lesions were initially de- the initial dataset of 225 lesions after CT examinations and 79 MR examina-
scribed as a complex, indeterminate, the three radiologists reviewed them. tions. Fifty-two CT examinations were
or 2F cyst and were upgraded to a None of the patients had von Hippel performed with a renal mass protocol
Bosniak category 3 when they were re- Lindau syndrome. (ie, noncontrast evaluation of the kid-
viewed again by both radiologists (cor- The final cohort included 144 pa- neys followed by nephrographic phase
responding pathologic analysis for these tients with 156 lesions that were Bos- evaluation of the kidneys) and 25 CT
six lesions that were Bosniak category niak category 2F (12 patients each had examinations were performed with
3: two multilocular cystic renal cell car- two lesions) who had follow-up con- an abdominal pain or routine proto-
cinomas [RCCs], one papillary RCC, trast-enhanced CT or MR examinations col without a noncontrast baseline.
three clear-cell RCCs with multilocular 6 months or longer after the original Of the 25 CTs performed after con-
cystic and solid components). study (median, 3.6 years; average, 4.2 trast agent administration without a
Initially, there were 20 disagree- years). Of these, there were 12 patients noncontrast baseline, six patients who
ments between the two radiologists with 14 lesions who had follow-up imag- underwent CT imaging also under-
regarding the Bosniak category of ing 6 months to 1 year after the original went contrast-enhanced MR imaging

760 radiology.rsna.org  n Radiology: Volume 272: Number 3—September 2014


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

Figure 2

Figure 2:  Images in a 60-year-old man with a history of a previous left upper pole papillary RCC, with a Bosniak category 2F cyst that progressed to a Bosniak cat-
egory 3 cystic lesion in 2 years. (a) Postcontrast CT image demonstrates a nonenhancing (confirmed on precontrast images) hemorrhagic (69 HU) category 2F cyst in
the lower pole of the right kidney. (b) Follow-up MR image 6 months later shows a stable appearance of the Bosniak 2F cyst with minimal irregularity of the posterior
margin on axial contrast-enhanced three-dimensional T1 subtraction gradient-echo images. (c) Follow-up MR image 2 years after the initial CT demonstrates internal
enhanced septations on axial contrast-enhanced three-dimensional T1 subtraction gradient-echo images compatible with a category 3 lesion. Partial nephrectomy
demonstrated a papillary RCC.

performed within 2 months of the CT, by using x2 test (for categorical data the patients had a concomitant re-
while an additional six patients who with cell counts .5), Fisher exact nal cancer or history of a previously
underwent CT imaging also underwent test (for categorical data with any treated renal cancer. Average cyst size
contrast-enhanced MR imaging within cell counts ,5), or Cochran Armitage was 2.8 cm (range, 0.8–13.9 cm).
the year. Therefore, the presence or trend tests for ordinal data (SAS 9.2;
absence of perceived enhancement SAS Institute, Cary, NC). A statisti- Lesions That Progressed
was confirmed by follow-up examina- cally significant result was indicated Of the 156 lesions, 17 (10.9%) le-
tion. Thus, 13 CT examinations did by a P value less than .05. sions progressed to category 3 or 4
not have a noncontrast control, which Of the 144 patients, 132 patients and proved to represent renal neo-
limited assessment for true enhance- had one lesion (132 of 144; 91.7%) plasms (four cystic clear-cell RCC,
ment. For the final follow-up studies, and 12 patients had two lesions (12 of five clear-cell RCC with multilocular
there were 54 CT examinations and 144; 8.3%). Therefore, the clustering growth patterns, seven papillary RCC,
102 MR examinations. For the pro- effect was minimal, and we chose not and one tubulocystic carcinoma) and
gressed lesions, the initial radiologist to account for this effect. two were benign (one fibrohyaline
reads and initial timeline to document cyst with blood and one multilocu-
progression were documented by one lar cystic nephroma). Overall time
of the readers (N.M.H.) after the Results to progression based on retrospec-
blinded analysis was completed. tive review for the category 2F cysts
Initial Imaging Evaluation ranged from 6 months to 3.2 years
Statistical Analysis k statistics between the two readers for (average, 19.2 months) (Figs 2–4;
Analysis of k statistic for interob- imaging features ranged from 0.856 to Figs E1–E5 [online]). However, when
server agreement between the two 1.000, with a mean of 0.943 (very good initially interpreted by the radiologist,
radiologists was performed. The to excellent) (Table 1). The k statistic the overall time to progression ranged
agreement of intraclass correlation for overall assessment of category 2F from 6 months to 3.8 years (average,
coefficient and k statistic was rated was 0.930 (95% confidence interval: 29.5 months).
as follows: 0–0.4, fair agreement; 0.794, 1.000). Eight of the progressed lesions
0.41–0.6, moderate agreement; 0.61– occurred in the subset of 14 lesions
0.8, substantial agreement; and 0.81– Characteristics of Bosniak Category 2F with follow-up imaging of 6 months
1.00, excellent agreement (9). Imag- Cysts to 1 year (the two progressed lesions
ing features, change in cyst size and We identified 156 Bosniak category that were benign when surgically re-
growth rate imaging modality, and pa- 2F lesions in 144 patients. The age of sected were in this subset); seven
tient demographics of the lesions that the 98 men (age range, 31–83 years; lesions with progression occurred
progressed were compared with those average age, 63.1 years) was similar in the subset of 21 lesions with fol-
that remained unchanged over time. to that of the 46 women (age range, low-up imaging of 1–2 years; and the
The relationship of these features to 28–85 years; average age, 64.7 years; remainder (4) of the progressed le-
progression or stability was assessed P = .488). Thirty of 156 (19.2%) of sions came from the subset of lesions

Radiology: Volume 272: Number 3—September 2014  n  radiology.rsna.org 761


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

Figure 3

Figure 3:  Images in a 56-year-old man with a right midpole Bosniak category 2F lesion that progressed to a Bosniak category 4 in 3 years. (a) Postcontrast CT
scan demonstrates a high-attenuation (39 HU) right renal cyst (3.5 cm) in the midpole without internal enhancement. (b) Follow-up contrast-enhanced MR image 2
years later demonstrates thin internal septations on axial single-shot fast spin echo. (c) Image from an axial subtraction T1 three-dimensional gradient-recalled-echo
MR examination 3 years after the initial CT demonstrates increase in number and thickness of internal septations with a new peripheral nodule of enhancement that
indicates a Bosniak category 4 lesion.

Figure 4 on clinical follow-up (follow-up, 1–10


years; average, 3.4 years).
For the 156 lesions, the Bos-
niak categories at final follow-up
were as follows: regression to cat-
egory 1 (often secondary to col-
lapse of the cyst; Fig E4 [online]),
seven lesions (4.5%); regression to
category 2 (often secondary to in-
crease in cyst size and increase in
internal fluid content within the cyst,
with associated stretching out of any
internal septae), 17 lesions (10.9%);
category 2F (stable), 113 lesions
(72.4%); progression to category 3,
eight lesions (5.1%); and progression
to category 4, 11 lesions (7.0%).
Men had statistically significant
progression to cancer (P = .003). Age
between 50 and 59 years and history
Figure 4:  Images in a 69-year-old man with left midpole Bosniak category of solid renal cancer were not statisti-
2F lesion that progressed to a Bosniak category 4 cystic lesion in 3.2 years. cally significant but approached signifi-
(a) Contrast-enhanced CT image demonstrates an anterior left midpole cystic cance (P = .07 and .053, respectively).
lesion with a thin internal septation and minimally thickened wall and a poorly Of cysts with blood within the cyst,
defined margin with the parenchyma. (b) Subtraction contrast-enhanced three- nine of 50 (18%) were malignant com-
dimensional T1 gradient-recalled-echo image 3.2 years after the initial CT pared with eight of 106 (7.5%) that
scan demonstrates interval development of an internal enhanced nodule that did not have blood within the cyst (P
indicates a Bosniak category 4 tumor. Partial nephrectomy revealed a cystic = .051). There was a range of imaging
clear-cell RCC. findings in category 2F lesions: some
cysts were more complex and closer
(121 lesions) with long-term follow-up not progressed), and the remainder of to category 3 (typically based on the
of 2 years or more. Eight of 14 le- these lesions (six of 14) had no fur- cyst that demonstrated more than one
sions with follow-up imaging of 6–12 ther imaging studies. of the variables described in Table 1),
months proceeded directly to surgical None of the patients with lesions while others were less complex and
resection after the follow-up study that progressed had recurrent or met- closer to category 2. Initial imaging
(even though four of these cysts had astatic disease after treatment, based characteristics that were associated

762 radiology.rsna.org  n Radiology: Volume 272: Number 3—September 2014


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

Table 2 with progression to malignancy in-


cluded minimally irregular thickness of
Association of Progression to Malignancy with Patient and Lesion Characteristics septae or wall (P , .001 and P = .016,
No. of Progression to No Progression to respectively) and an indistinct inter-
Parameter Patients Malignancy (%) Malignancy (%) P Value face of the enhanced lesion with the
adjacent renal parenchyma (P , .001,
Sex* ,.003 Table 2). Of the category 2F lesions
 Male 98 15 (15) 85 (83)
that progressed to malignancy, 12 of
 Female 46 0 (0) 100 (46)
17 were endophytic (P = .02).
Age* ,.069
Of the 156 category 2F lesions,
  ,50 years 19 0 (0) 100 (19)
145 had stable imaging features at 1
  50–59 years 31 26 (8) 74 (23)
year and, of these, 14 (9.6%) lesions
  60–69 years 37 11 (4) 89 (33)
  .70 years 57 5 (3) 95 (54)
progressed to a category 3 or 4 on
First study type* ,.657 follow-up imaging; therefore, there
 CT 75 9 (7) 91 (68) was no association of 1-year stabil-
  MR imaging 69 12 (8) 88 (61) ity with future stability of a cyst (P
No. of lesions* ,.113 = .999). Six of the 156 lesions were
 1 132 9 (12) 91 (120) downgraded at 1 year and of these,
 2 12 25 (3) 75 (9) 0% had future progression. Twelve
No. of cysts* ,.691 percent (19 of 156) of lesions showed
 1 35 9 (3) 91 (32) progression based on imaging crite-
 2–4 56 11 (6) 89 (50) ria; 10.9% (17 of 156) showed pro-
 5+ 53 11 (6) 89 (47) gression to malignancy.
Bilateral cysts* ,.455
 Yes 83 12 (10) 88 (73) Lesions That Did Not Progress
 No 61 8 (5) 92 (56) Initial cyst size, change in lesion size
History of solid renal cancer* .053 (increase or decrease), and growth
 Yes 9 33 (3) 67 (6) rate [growth rate = (follow-up size 2
 No 135 9 (12) 91 (123) initial size)/years between measures]
History of any prior renal ,.457
were not found to correlate with
  cancer (cystic or solid)*
progression (P values of .843, .188,
 Yes 26 15 (4) 85 (22)
and .163, respectively). There were
 No 118 9 (11) 91 (107)
22 small cysts (,1.5 cm) of the 156
Pole of kidney .011
 Upper 48 2 (1) 98 (47)
category 2F cysts, and only three of
 Mid 71 11 (8) 89 (63)
22 (14%) of these small category 2F
 Lower 37 22 (8) 78 (29) cysts demonstrated progression (all of
Laterality .400 which were cancer on resection, and
 Right 70 9 (6) 91 (64) all of which were in men with a history
 Left 86 13 (11) 87 (75) of cancer) (Fig E5 [online]). Of these
Cyst size .721 22 small cysts, 14 showed growth at
  ,1.5 22 14 (3) 86 (19) follow-up (average growth, 0.68 cm;
 1.5–2.9 57 11 (6) 89 (51) range, 0.1–1.9 cm), and 11 of those
 3+ 77 10 (8) 90 (69) were downgraded to a Bosniak cate-
Location of tumor in the renal .020 gory 2 after growth demonstrated in-
parenchyma creased internal fluid, with decreased
 Endophytic 69 17 (12) 83 (57) internal soft tissue (eg, initial wall
 Exophytic 87 6 (5) 94 (82) thickening or thickened septations
Number of septations .957 became thinner as the fluid contents
 0 48 10 (5) 90 (43) in the cyst increased). In other words,
 1–3 77 12 (9) 88 (68) some lesions grew in our series, but
 4+ 31 10 (3) 90 (28)
they had no soft-tissue (ie, septae or
Thickness of septations .170
wall) growth.
  No septations 48 10 (5) 90 (43)
A multilobulated border of the le-
  Pencil thin 45 4 (2) 96 (43)
sion was not found to correlate with
 Measureable 63 16 (10) 84 (53)
progression (P = .999). Similarly,
Table 2 (continues)
none of the lesions with calcification

Radiology: Volume 272: Number 3—September 2014  n  radiology.rsna.org 763


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

Table 2 (continued) are those that are probably benign but


are somewhat worrisome and can only
Association of Progression to Malignancy with Patient and Lesion Characteristics be proven to be benign by demonstra-
tion of stability in follow-up studies,
No. of Progression to No Progression to
while those that show progression are
Parameter Patients Malignancy (%) Malignancy (%) P Value
considered to be surgical lesions (unless
Morphologic structure of septations ,.001
there are clinical contraindications to
  No septations 48 10 (5) 90 (43) surgery). Few series of follow-up of cat-
 Smooth 76 4 (3) 96 (73) egory 2F lesions are available, and none
  Minimally Irregular 32 28 (9) 72 (23) include a large number of patients who
Septal enhancement .120 underwent MR imaging (6–8,14). Previ-
  No septations 48 11 (5) 89 (42) ous series showed progression rates of
 Yes 71 15 (11) 85 (60) 2F lesions of nine of 69 (13%) lesions
 No 37 3 (1) 97 (37)
(of which eight were surgically resected
Wall thickness .956
and four of those eight were malignant)
  Pencil thin 100 11 (11) 89 (89)
(7), 12 of 81 (15%) lesions (6), and 14
 Measureable 56 11 (6) 89 (50)
of 201 (7.0%) lesions (14). Our retro-
Morphologic structure .016
spective review showed progression to
 Smooth 128 8 (10) 92 (118)
  Minimally irregular 28 25 (7) 75 (21)
malignancy in 17 of 156 (10.9%) lesions.
Wall enhancement .031 Imaging features that determine
 Yes 55 18 (10) 82 (45) whether a category 2F cyst has pro-
 No 101 7 (7) 93 (94) gressed are based on whether enhanced
Multilobulated .999 solid portions of a tumor appeared
 Yes 6 0 (0) 100 (6) or whether there was an increase in
 No 150 11 (17) 89 (133) thickness, number or irregularity of
Calcification .050 enhanced septa, or thickness of the en-
  NA because of MR imaging 82 13 (11) 87 (71) hanced wall of the lesion. Increase in
 Yes 33 0 (0) 100 (33) the size of the lesion alone is not asso-
 No 41 15 (6) 85 (35) ciated with progression (15), as shown
Nonenhancing soft-tissue nodule .032 in our results where change in cyst size
 Yes 3 67 (2) 33 (1) and growth rate had no correlation with
 No 153 10 (15) 90 (138) progression. Some lesions grew in our
Hemorrhagic or proteinaceous fluid .051 series, but had no growth of soft tissue
 Yes 50 18 (9) 82 (41) (ie, septae or wall). Two of the 17 cate-
 No 106 8 (8) 92 (98)
gory 2F cysts that progressed to malig-
Interface with parenchyma ,.001
nancy were smaller on follow-up imag-
 Sharp 126 6 (8) 94 (118)
ing, but demonstrated the development
 Indistinct 30 30 (9) 70 (21)
of enhancing internal soft tissue. Some
Note.—Data in parentheses are number of lesions unless otherwise indicated. NA = not applicable. of the small 2F lesions (,1.5 cm) that
* Data in parentheses are number of patients.. initially looked worrisome grew, and
they became more benign in appearance
when they were imaged because the
in this group (n = 33) progressed (P vs 14% [14 of 102], respectively; fluid within the cyst increased in volume
= .05); however, this observation is P = .456; Table 3). and the septae became stretched out. In
limited by the high number of lesions our practice, if a lesion is indeterminate
evaluated with MR imaging, and the or has indeterminate enhancement (eg,
Discussion between 10–20 HU) by using CT imag-
presence or absence of calcification
could not be determined. Accurate diagnosis of complex cystic le- ing, then an MR examination is often
sions of the kidney remains one of the obtained (16,17).
Association of Imaging Modality and more difficult problems in diagnostic ra- Although Bosniak criteria are a CT-
Lesion Progression diology (2,4,9–11). Imaging findings for defined classification system, our data
The difference in percentage with complex cystic lesions can also be seen suggest that CT and MR imaging may
progression to malignancy was not in either malignant or benign lesions be used interchangeably when following
significant for CT imaging at fol- (12,13) and that often causes removal these lesions; however, more research
low-up compared with MR imag- of benign lesions because malignancy is needed to confirm or refute this
ing at follow-up (6% [three of 54] cannot be ruled out. Category 2F lesions conclusion.

764 radiology.rsna.org  n Radiology: Volume 272: Number 3—September 2014


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

Table 3
Association between Progression of Category 2F Cysts and Imaging Modality
Modality of First Study and Latest Follow-up Study*
First study, CT; First study, CT; First study, MR; First study, MR;
Parameter latest follow-up, CT latest follow-up, MR latest follow-up, CT latest follow-up, MR P Value

No. of cysts 43 37 11 65
Bosniak category at follow-up .237
  Regression to category 1 2 (5) 2 (5) 0 (0) 3 (4)
  Regression to category 2 20 (47) 12 (32) 4 (36) 29 (45)
  2F (stable appearance) 19 (44) 17 (46) 3 (27) 23 (35)
  Progression to category 3 1 (2) 2 (5) 0 (0) 5 (8)
  Progression to category 4 1 (2) 4 (11) 1 (9) 5 (8)
Any progression 2 (5) 6 (16) 1 (9) 10 (15) .313
Progression to malignancy 2 (5) 5 (14) 1 (9) 9 (14) .456

Note.—Data in parentheses are percentages.


* First follow-up study was used if the lesion was stable. If not, first follow-up study that demonstrated progression or regression was used.

There are several limitations to our than 2 years cannot be fully assessed cystic lesions tend to progress slowly.
study. This is a retrospective study from based on our data. Another limitation of We are aware that there will be an oc-
a single institution performed by geni- this study is the small number of lesions casional case in which a lesion may ap-
tourinary imaging specialists who were that progressed to cancer (17 of 156 pear dormant for years, but progress
all trained by a single person (M.A.B.), [10.9%]), which limits by power con- later and prove to be malignant. Cur-
with an inherent inclusion bias and a clusions about features associated with rent literature that investigates the ori-
probably nonreplicable interobserver progression. For example, as shown in gin of malignancy in cystic renal tumors
agreement. Additionally, there was a Table 2, patients with a history of a solid supports that malignancy can arise
change in the MR and CT technology renal cancer (seen in three of the 15 in what appears to be a benign cyst
over the 11-year period of evaluation. men who had lesions that progressed (19–22). While the length of follow-up
For this reason, 13 patients had CT to malignancy) were 24% more likely to needed in an individual case cannot be
scans obtained without a noncontrast have progression (33% [three of nine] in determined with certainty, the results
control, which limited our assessment those with history of solid renal cancer of this study would suggest that at least
for measurement of enhancement (18). vs 9% [12 of 135] in those without), but 4 years would be a safe follow-up pe-
To have expert readers evaluate these the difference was not statistically signif- riod for the majority of patients. Some
cystic lesions may improve the standard- icant (P = .053). patients with lesions that appear less
ization of 2F cysts, however, it may not We did not demonstrate a statically worrisome might need a shorter fol-
be reflective of how 2F cysts are clas- significant difference between MR im- low-up, and others, particularly young
sified in practice. Additionally, we were aging and CT in follow-up of Bosniak patients, might need a longer follow-up.
able to document progression within the category 2F lesions. However, this re- Larger studies of category 2F cysts are
upper limit of 3.2 years in all cases, but sult may not reflect clinical practice be- needed to corroborate our findings.
one of the 19 cases of progression was cause more MR imaging cases than CT In conclusion, based on the expe-
not prospectively documented until 3.8 cases that were initially called category rience at our institution, 10.9% (17 of
years after the initial examination. An- 2F cysts (15 lesions on CT, 37 lesions 156) Bosniak category 2F cystic lesions
other limitation is our inclusion criteria. MR imaging) were downgraded to cat- progressed to malignancy. Therefore,
We included 14 lesions with follow-up egory 2 and were therefore excluded. the recommended length of follow-up
imaging of only 6 months to 1 year. While 10.9% (17 of 156) category for the majority of Bosniak category 2F
Eight of these patients went on to sur- 2F cases in our series progressed to ma- lesions is 4 years, although longer or
gery, but six patients had no further fol- lignancy in 6 months to 3.2 years (with shorter follow ups will be appropriate
low-up. These six patients were included an upper limit of 4 years when prospec- in some cases.
secondary to uniform application of the tive reads were taken into account),
inclusion criteria to capture the eight one cannot be certain that those cases Disclosures of Conflicts of Interest: N.M.H.
No relevant conflicts of interest to disclose.
patients with a follow-up examination that have been followed for less than 4
E.M.H. No relevant conflicts of interest to dis-
that showed progression before surgery. years will not progress in a longer time close. M.A.B. No relevant conflicts of interest
Therefore, follow-up for lesions for less period, particularly considering that to disclose.

Radiology: Volume 272: Number 3—September 2014  n  radiology.rsna.org 765


GENITOURINARY IMAGING: Follow-up for Bosniak Category 2F Cystic Renal Lesions Hindman et al

References kidney (Bosniak category IIF). AJR Am J magnetic resonance imaging for solid com-
Roentgenol 2003;181(3):627–633. ponents in cystic renal neoplasms. Urology
1. Bosniak MA. The current radiological ap-
2008;71(1):47–51.
proach to renal cysts. Radiology 1986;158(1): 9. Shrout PE. Measurement reliability and
1–10. agreement in psychiatry. Stat Methods Med 17. Israel GM, Bosniak MA. MR imaging of cys-
Res 1998;7(3):301–317. tic renal masses. Magn Reson Imaging Clin
2. Bosniak MA. Problems in the radiologic di- N Am 2004;12(3):403–412, v.
agnosis of renal parenchymal tumors. Urol 10. Silverman SG, Israel GM, Herts BR, Rich-
Clin North Am 1993;20(2):217–230. ie JP. Management of the incidental renal 18. Curry NS, Cochran ST, Bissada NK. Cystic
mass. Radiology 2008;249(1):16–31. renal masses: accurate Bosniak classification
3. Bosniak MA. Diagnosis and management requires adequate renal CT. AJR Am J
of patients with complicated cystic lesions
11. Warren KS, McFarlane J. The Bosniak Roentgenol 2000;175(2):339–342.
of the kidney. AJR Am J Roentgenol 1997; classification of renal cystic masses. BJU Int
169(3):819–821. 2005;95(7):939–942. 19. Bielsa O, Lloreta J, Gelabert-Mas A. Cystic
renal cell carcinoma: pathological features,
4.
Bosniak MA. The Bosniak renal cyst 12. Eble JN, Bonsib SM. Extensively cystic renal survival and implications for treatment. Br J
classification: 25 years later. Radiology 2012; neoplasms: cystic nephroma, cystic partially Urol 1998;82(1):16–20.
262(3):781–785. differentiated nephroblastoma, multilocular
cystic renal cell carcinoma, and cystic ham- 20. Hartman DS, Davis CJ Jr, Johns T, Gold-
5. Harisinghani MG, Maher MM, Gervais DA, artoma of renal pelvis. Semin Diagn Pathol man SM. Cystic renal cell carcinoma. Urol-
et al. Incidence of malignancy in complex 1998;15(1):2–20. ogy 1986;28(2):145–153.
cystic renal masses (Bosniak category III):
should imaging-guided biopsy precede sur- 13. Freire M, Remer EM. Clinical and radiologic 21. Smith AY, Feddersen RM, Gardner KD Jr,
gery? AJR Am J Roentgenol 2003;180(3): features of cystic renal masses. AJR Am J Davis CJ Jr. Cystic renal cell carcinoma
755–758. Roentgenol 2009;192(5):1367–1372. and acquired renal cystic disease associ-
ated with consumption of chaparral tea:
6. O’Malley RL, Godoy G, Hecht EM, Stifelman 14. Hwang JH, Lee CK, Yu HS, Cho KS, Choi
a case report. J Urol 1994;152(6 Pt 1):
MD, Taneja SS. Bosniak category IIF desig- YD, Ham WS. Clinical outcomes of Bosniak 2089–2091.
nation and surgery for complex renal cysts. category IIF complex renal cysts in Korean
patients. Korean J Urol 2012;53(6):386–390. 22. Thoma CR, Frew IJ, Krek W. The VHL tu-
J Urol 2009;182(3):1091–1095.
mor suppressor: riding tandem with GSK-
7. Smith AD, Remer EM, Cox KL, et al. Bos- 15. Weibl P, Klatte T, Waldert M, Remzi M.
3beta in primary cilium maintenance. Cell
niak category IIF and III cystic renal le- Complex renal cystic masses: current stan- Cycle 2007;6(15):1809–1813.
sions: outcomes and associations. Radiology dards and controversies. Int Urol Nephrol
2012;262(1):152–160. 2012;44(1):13–18.

8. Israel GM, Bosniak MA. Follow-up CT of 16. Adey GS, Pedrosa I, Rofsky NM, Sanda MG,
moderately complex cystic lesions of the DeWolf WC. Lower limits of detection using

766 radiology.rsna.org  n Radiology: Volume 272: Number 3—September 2014