European Radiology

, Volume 22, Issue 2, pp 484–492

Arterial spin labeling MR imaging for characterisation of renal masses in patients with impaired renal function: initial experience

Authors

    • Department of RadiologyBeth Israel Deaconess Medical Center and Harvard Medical School
    • Department of RadiologyUT Southwestern Medical Center
  • Khashayar Rafatzand
    • Department of RadiologyBeth Israel Deaconess Medical Center and Harvard Medical School
  • Philip Robson
    • Department of RadiologyBeth Israel Deaconess Medical Center and Harvard Medical School
  • Andrew A. Wagner
    • Surgery, Division of UrologyBeth Israel Deaconess Medical Center and Harvard Medical School
  • Michael B. Atkins
    • Hematology/OncologyBeth Israel Deaconess Medical Center and Harvard Medical School
  • Neil M. Rofsky
    • Departments of RadiologyUniversity of Texas Southwestern Medical Center
  • David C. Alsop
    • Department of RadiologyBeth Israel Deaconess Medical Center and Harvard Medical School
Urogenital

DOI: 10.1007/s00330-011-2250-z

Cite this article as:
Pedrosa, I., Rafatzand, K., Robson, P. et al. Eur Radiol (2012) 22: 484. doi:10.1007/s00330-011-2250-z

Abstract

Objectives

To retrospectively evaluate the feasibility of arterial spin labeling (ASL) magnetic resonance imaging (MRI) for the assessment of vascularity of renal masses in patients with impaired renal function.

Methods

Between May 2007 and November 2008, 11/67 consecutive patients referred for MRI evaluation of a renal mass underwent unenhanced ASL-MRI due to moderate-to-severe chronic or acute renal failure. Mean blood flow in vascularised and non-vascularised lesions and the relation between blood flow and final diagnosis of malignancy were correlated with a 2-sided homogeneous variance t-test and the Fisher Exact Test, respectively. A p value <0.05 was considered statistically significant.

Results

Seventeen renal lesions were evaluated in 11 patients (8 male; mean age = 70 years) (range 57–86). The median eGFR was 24 mL/min/1.73 m2 (range 7–39). The average blood flow of 11 renal masses interpreted as ASL-positive (134 +/− 85.7 mL/100 g/min) was higher than that of 6 renal masses interpreted as ASL-negative (20.5 +/− 8.1 mL/100 g/min)(p = 0.015). ASL-positivity correlated with malignancy (n = 3) or epithelial atypia (n = 1) at histopathology or progression at follow up (n = 7).

Conclusions

ASL detection of vascularity in renal masses in patients with impaired renal function is feasible and seems to indicate neoplasia although the technique requires further evaluation.

Key Points

  • Arterial spin labeling may help to characterise renal masses in patients with renal failure

  • Detection of blood flow on ASL in a renal mass supports the presence of a neoplasm

  • Renal masses with high blood-flow levels on ASL seem to progress rapidly

Keywords

Magnetic resonance imagingKidney neoplasmsPerfusionRenal insufficiencyNephrogenic systemic fibrosis

Introduction

Lesion vascularity, the most reliable feature to characterise a renal mass as a neoplasm [1, 2], is typically evaluated with contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI). The known nephrotoxicity of iodinated contrast media and the recently identified association of gadolinium-based contrast agents (GBCAs) with nephrogenic systemic fibrosis (NSF) [3, 4] have limited their use in patients with renal failure. While the risk of NSF may be drastically decreased with the use of a macrocyclic GBCA, current recommendations include contraindication of certain GBCA in high-risk patients (i.e. severe chronic or acute renal failure) [5, 6]. For these reasons, there is an increasing need for alternative methods that allow evaluation of patients with renal masses without the need of administering intravenous contrast medium.

With MR imaging, arterial spin labeling (ASL) can be used to measure blood flow (i.e. perfusion) into tissue by magnetically ‘labeling’ the nuclear spins of the endogenous water in arterial blood [7, 8]. Qualitative and quantitative images of blood flow can be generated without intravenous contrast material. ASL has shown promise for assessment of renal perfusion [911] and in monitoring tumour blood flow in patients undergoing antiangiogenic therapy for metastatic renal cell carcinoma (RCC) [12, 13]. Furthermore, there is excellent correlation between tumour blood flow on ASL in human RCC xenografts and the presence of vascularised, viable tumour at histopathology [14]. Moreover, areas of tumour necrosis induced by anti-angiogenic therapy at histopathology demonstrate absence of detectable blood flow on ASL [14].

Based on these preliminary data, we have been using ASL as a complementary tool for assessment of renal masses in patients with moderate/severe impairment of the renal function in whom avoiding GBCAs is desirable. Our aim is to retrospectively examine this clinical practice to determine the feasibility of ASL in the assessment of renal masses in patients with impaired renal function.

Material and methods

Study patients

This retrospective IRB-approved, HIPAA-compliant study was conducted with waiver of informed consent. Between May 1st 2007 and November 25th, 2008 sixty-seven consecutive patients were referred for MRI evaluation of one or more suspected renal masses. Following standard procedures, renal function was screened in all patients prior to MRI and those with an estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m2 underwent 1.5 T MRI without receiving GBCA. eGFR was calculated using the modification of diet in renal disease (MDRD) equation. Ten patients with an eGFR below this threshold and an additional patient with adult polycystic kidney disease (APKD) and worsening moderate renal failure (eGFR = 39 mL/min/1.73 m2) underwent unenhanced MRI using ASL imaging. Therefore, a total of eleven patients (16%) of the 67 referrals represent our study group (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00330-011-2250-z/MediaObjects/330_2011_2250_Fig1_HTML.gif
Fig. 1

Patient selection flowchart and lesion categories. RCC Renal Cell Carcinoma; TCC Transitional Cell Carcinoma

MRI protocol

All studies were performed at 1.5 T (Excite-HDxt, GE Healthcare, Waukesha, WI, USA). Preliminary unenhanced sequences included axial 2D T1-weighted in-phase/opposed-phase gradient echo (TR = 180–205 ms, TE = 2.2–2.7 / 4.5–5.2 ms, flip angle = 80°, slice thickness = 6–8 mm, gap 1 mm, matrix = 160 × 256, FOV = 35 cm, NEX = 1), coronal and axial T2-weighted single shot fast spin echo (SSFSE) (TR = 1,100 ms, TE = 60 ms, slice thickness = 4 mm, 1 mm gap, matrix 192 × 256, flip angle 130°, bandwidth ± 62 kHz, FOV 35–40 cm), and coronal 3D fat-saturated T1-weighted spoiled gradient echo (LAVA) (TR = 3.8–4.5 ms, TE = 1.8–2.0 ms, FA = 12°, slice thickness = 3–4 mm, pre-interpolation) sequences.

ASL-MRI

ASL slices were acquired through the centre of the mass in the axial, coronal, or sagittal plane depending on the location of the mass. A single ASL slice was acquired because of time constraints. The ASL sequence was developed at the authors’ institution and is not commercially available. Separate ASL acquisitions were obtained in patients with more than one suspicious mass. Adequate anatomical prescription of the ASL slices was confirmed by the monitoring radiologist. Perfusion imaging was achieved with a pseudo-continuous labeling [15], optimised background suppression, and SSFSE acquisition [16]. For control images, SSFSE images with a FOV of 40 cm, a 128 × 128 matrix and a slice thickness of 8–10 mm were acquired. Labeling was performed in the axial plane 8–10 cm superior to the centre of the kidneys for 1,500 ms followed by a 1,500 ms post-labeling delay [16]. A repetition time of 6 s was used to allow for recovery of blood signal and for the subjects to breath in during the quiet period between acquisitions; a very short (i.e. <2 s) breath-hold is then performed repeatedly at each TR for image acquisition. Sixteen averages of label and control were acquired for a total acquisition time of 3.5 min.

Image reconstruction

Measurement of the fully relaxed magnetisation signal, M0, and the tissue T1 (based on two inversion recovery images) were obtained to quantify tumour blood flow [16]. A custom reconstruction was performed using the IDL programming language (ITT Visual Information Solutions, Boulder, CO, USA). ASL label-control pairs were subtracted and averaged in complex k-space prior to image reconstruction. Each image acquisition was reconstructed to generate: 1) a proton-density weighted “reference image”; 2) a “difference ASL image” (labeled minus control image); and 3) a “quantitative perfusion image”, which assumes the labeled water spends most of the time after labeling within the blood of arteries and microvasculature [16, 17]. Blood flow measurements were not based on tissue T1 determinations because of the propensity of misregistration between images at different inversion times in the abdomen [16].

Imaging analysis

Qualitative analysis

MRI examinations were prospectively evaluated for the presence of blood flow on ASL in the renal lesions the radiologist covering the body MRI service. A retrospective review was then performed to correlate the initial MRI interpretations with available pathological, clinical, and/or imaging follow-up, when available. For patients with more than one MRI, only the first ASL examination was included in the analysis. Lesions with subjective focal or diffuse high signal intensity (SI) relative to background on the ASL difference images were interpreted as vascularised masses (i.e. neoplastic) whereas lesions without appreciable increased signal on ASL were considered non-vascularised (i.e. non-neoplastic). Non-vascularised lesions with homogeneous low SI on T1-weighted images and high SI on T2-weighted images, without internal architecture, and thin walls were interpreted as simple cysts [18]. Cysts with thin linear filling defects but without nodularity and/or increased SI on ASL-MRI were considered as benign septated cysts [18, 19]. Non-vascularised lesions with high SI on un-enhanced fat-saturated T1-weighted images and variable SI on T2-weighted images were interpreted as haemorrhagic/proteinaceous cysts [18, 20].

Quantitative analysis

Tumour blood flow was measured by region-of-interest (ROI) analysis using the open-source DICOM viewer Osirix (Osirix X, v3.1 32-bit) on a Mac Pro platform (OS X; Apple Computer, Cupertino, CA, USA) by a single reviewer (I.P.) with 10 years of experience interpreting abdominal MRI. ROIs were initially drawn around the outer anatomical contour of the target lesions on the proton density-weighted “reference image”. The mean ROI area was 105 mm2 ± 13 mm2. To obtain measurements of SI and blood flow, ROIs were then copied and pasted on both the “difference ASL image” (“ASL SI-lesion”) and “perfusion image” (averaged blood flow “ABF”) obtained at the same level, respectively. ROI values obtained from perfusion images represent blood flow in mL/100 g/min [16]. The same ROI was then copied and pasted outside the body on the same difference ASL images to measure the background SI (“ASL SI-background”). For tumours exhibiting blood flow on ASL images, a second ROI was placed in the area within the region of the tumour that demonstrated the most blood flow by visual assessment (peak blood flow “PBF”).

An offset resulting from noise in the difference ASL images in combination with magnitude imaging may cause a small positive blood flow value even in the absence of blood flow. Though the use of homodyne reconstruction and phased coil combination reduces this effect relative to straight sum of squares combination [21], a consistent noise-induced offset remains. To estimate this offset, we took advantage of the linearity of blood flow quantification, i.e. (Blood Flow = calibration constant x difference ASL signal) where the calibration constant depends on the SI in the reference image and other parameters unrelated to the ASL difference signal [16]. Since noise is uniformly distributed in the difference ASL image, the “ASL-SI background” value should reflect the signal that would be measured in the lesion ROI in the absence of blood flow. To convert the “ASL-SI background” mean to an estimated blood flow offset (“Perf. Noise”), we used the same calibration constant as for the lesions as follows:
$$ \left( {{\text{Perf}}.\;{\text{Noise}}} \right) = \left( {{\text{Perf}}.\;{\text{Lesion}}} \right) \times \left( {{\text{Background}}\;{\text{Difference}}\;{\text{Signal}}} \right)/\left( {{\text{Lesion}}\;{\text{Difference}}\;{\text{Signal}}} \right), $$
where “Perf. Noise” is the estimated blood flow offset and “Perf. Lesion” is the measured perfusion in the lesion. This value was averaged across lesions to estimate the contribution to blood flow from noise.

Reference standard

There were 17 suspicious renal masses in 11 patients. Histopathological evaluation (when available) and a combination of presence/lack of growth on serial imaging with unenhanced MRI features were regarded as reference standards (Fig. 1). Because some patients had contrast-enhanced cross-sectional imaging prior to obtaining the ASL examination, the estimated follow up time was considered from the initial available imaging examination demonstrating the renal lesion to the last examination available for each patient.

Histopathological evaluation was available in 4 masses (Patients 3, 4, 9, 10) through biopsies of the renal lesions (n = 2), lymph node at the ipsilateral renal hilum (n = 1) and a liver metastasis (n = 1). A composite parameter of growth on serial imaging studies and unenhanced MRI features was regarded as a reference standard for 10 masses (Table 2). In one patient, 3 lesions were imaged only at one time-point, and considered simple cysts based on their classic appearance on T2-weighted imaging and lack of perfusion on ASL. No further follow up was available for this patient.

Statistical analysis

All statistical tests were performed by lesions and by patient. In by patient analyses, ROI values were averaged across lesions of the subject and treated as a single measure. In binary comparisons of outcome or radiologic assessment of blood flow present in the lesion, all lesions successfully scanned had identical findings and one representative lesion was used per patient.

The difference of mean blood flow between lesions identified as vascularised (ASL-positive) by the radiologist vs. those non-vascularised (ASL negative) was tested for significance with a 2-sided homogenous variance t-test. Contingency tables of the relationship between the presence of blood flow and clinical or imaging indications of malignancy/progression were tested with the Fisher Exact Test. All statistical tests were considered significant if p < 0.05.

Results

Seventeen suspicious renal masses were evaluated in 11 patients (8 males, 3 females), mean age 70 years (range 57–86) (Table 1). All patients had moderate-severe renal failure with a mean eGFR of 23 mL/min/1.73 m2 and median eGFR of 24 mL/min/1.73 m2 (range 7–39). Renal masses included in our series were previously seen on CT (10 masses in 5 patients) and contrast-enhanced MRI (7 masses in 7 patients). The latter 7 patients with prior contrast enhanced MRI examinations were followed with ASL because of worsening of renal function.
Table 1

Patient demographics, lesion distribution and renal status

Patient

Age/gender

Lesions

eGFRa

Creatinineb

Kidneys

1

68/F

1

30

1.8

Single left kidney

Right nephrectomy for RCC

2

77/M

1

14

4.4

Single left kidney

Right nephrectomy for RCC

3

66/M

1

24

2.7

Single right kidney

Left nephrectomy for RCC

4

75/M

1

7

7.6

Functional single right kidney

Hydronephrotic L kidney (staghorn calculus)

5

73/M

3

29

2.2

Two kidneys

Chronic renal failure, Diabetes

6

57/M

2

39

1.8

Two kidneys

Acute on chronic renal failure

Renal Transplant, APKD

7

75/M

1

28

2.3

Two kidneys

Chronic renal failure, Diabetes

8

86/F

2

19

2.4

Functional single left kidney

Atrophic right kidney

9

79/M

4

18

3.5

Two kidneys

Chronic renal failure, Diabetes

10

53/M

1

16

4.0

Two kidneys

Chronic renal failure, IgA nephropathy

11

64/F

1

28

2.3

Two kidneys

Chronic renal failure, Diabetes

amL/min/1.73 m2

bmg/dL

APKD Adult polycystic kidney disease

Qualitative analysis

Eleven masses were characterised as having blood flow on ASL (i.e. neoplastic) (Table 2). Histological evaluation was performed in 4 of these (Patients 3, 4, 9, 10) with confirmation of malignancy in 3 and epithelial atypia in one. Percutaneous biopsy was non-diagnostic in a complex cystic lesion with a solid nodule with blood flow on ASL (patient 11). Previous contrast-enhanced CT and MR examinations demonstrated enhancement of this nodule. Overall, an increase in size on serial imaging was confirmed in 10/11 masses classified as vascularised on ASL.
Table 2

Lesion distribution and characteristics, flow on qualitative ASL-MRI, reference standard and patient’s clinical outcome. L = Left, R = Right, UP = Upper Pole, IP = Interpolar, LP = Lower Pole. SRT = stereotactic radiation therapy

Patient

Lesion location

Size (cm)

Flow ASL-MRI

Histo-pathology

MRI characterisation

Size on serial imaging

Clinical outcome

Imaging follow-up (months)

1

R, LP

4

Yes

N/A (1)

Solid mass

Increased

↓ size and absent blood flow post SRT

13

2

L, UPJ (2)

5.7

No

N/A

Hemorrhagic cyst

No change

Hydronephrosis resolved

11

3

R, IP

7.2

Yes

Clear cell RCC (3)

Solid mass

Increased

Progression: IVC tumour thrombus

3

4

R, LP

10.1

Yes

Urothelial carcinoma (4)

Solid mass

Increased

Progression: new liver metastases

N/A

5

R, LP

1.6

No

N/A

Hemorrhagic cyst

N/A

No clinical evidence of progression.

36(5)

L, IP

4.2

No

N/A

Septated cyst

L, UP

1.2

No

N/A

Septated cyst

6

R, LP

2.7

No

N/A

Hemorrhagic cyst

No change

No enhancement on CE-MRI

14

L, IP

2

No

N/A

Hemorrhagic cyst

7

R, LP

2.3

Yes

N/A

Cystic neoplasm

Increased

Clinical & MRI surveillance

19

8

L, IP

2.9

Yes

N/A

Solid mass

Increased

Expired 1 month after ASL-MRI

1

L, LP

1.2

Yes

N/A

Solid mass

9

R, LP

6.2

Yes

Clear cell RCC (6)

Solid mass

Increased

Succumbed to metastatic disease

17

L, UP

3

Yes

N/A

Cystic neoplasm

Increased

L, IP

3.8

Yes

N/A

Cystic neoplasm

10

L, IP

2.4

Yes

Epithelial atypia

Solid mass

Increased

↓ size but persistent blood flow post SRT

8

11

R, UP

2

Yes

Inconclusive

Cystic neoplasm

Decreaseda

↓ size but persistent blood flow post biopsy

41

aRepresents post-biopsy reduced size of cystic component, with unchanged size of enhancing mural nodule

(1) Not available, (2) UPJ Ureteropelvic junction, (3) From biopsy of renal mass, (4) From biopsy of liver metastasis, (5) Clinical follow up only, no imaging available, (6) From biopsy of retroperitoneal lymph node adjacent to ipsilateral renal hilum

One lesion (patient 1) demonstrated progressive increase in tumour vascularity in ASL over time, which preceded increase in size of the mass (Fig. 2). Three lesions (patients 3, 4 and 9) with high levels of blood flow on ASL had evidence of rapid progression. Pathologic confirmation was not available in three other vascularised lesions (patients 7 and 8) with demonstrated growth on serial imaging.
https://static-content.springer.com/image/art%3A10.1007%2Fs00330-011-2250-z/MediaObjects/330_2011_2250_Fig2_HTML.gif
Fig. 2

Axial single-shot fast spin echo T2-weighted (TR = 1,100 ms, TE = 60 ms, slice thickness = 4 mm, 1 mm gap, matrix 192 × 256) (top row) and arterial spin labeled perfusion (TR = 6.0 s, TE = 62.1 ms, matrix 128 × 128, thickness = 10 mm) (bottom row) images in a patient with prior right nephrectomy for renal cell carcinoma. A biopsy-proven ccRCC (arrow) in the left kidney was treated with stereotactic radiation therapy (SRT) in 2005, and followed with gadolinium-enhanced MRI confirming persistent enhancement (not shown). Because of decrease in renal function, the vascularity in the mass was monitored after 2007 with ASL MRI. The mass had low blood flow on 8/2007 and showed minimal growth but substantial increase in measured blood flow on 1/2008. Note continued slow growth and persistent high perfusion in 4/2008. This patient underwent a second therapy with SRT (June 2008) that resulted in a substantial increase in size but a decrease in blood flow 1 month later (7/2008). These findings were consistent with post-radiation changes and an effective antiangiogenic effect of SRT. ASL on 10/2008 confirmed decrease in size of the tumour and marked decreased blood flow in the mass

The ASL imaging plane was incorrectly prescribed through the purely cystic component of a cystic mass, missing a small mural nodule, in a patient with 3 other vascularised masses on ASL and subsequent diagnosed with metastastic clear cell RCC (patient 9). This lesion was excluded from further analysis.

Six lesions were classified as non-neoplastic due to lack of detectable blood flow on ASL (Table 2). One haemorrhagic lesion decreased in size over time (Fig. 3). Two other lesions (Patient 6) were evaluated by gadolinium-enhanced MRI 12 months later, allowable due to improved renal function, demonstrating absence of enhancement. Imaging follow up or pathologic confirmation was not available for three other non-vascularised lesions (patient 5) with features consistent with haemorrhagic/proteinaceous cyst (1) and septated cysts (2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00330-011-2250-z/MediaObjects/330_2011_2250_Fig3_HTML.gif
Fig. 3

Coronal T2-weighted single-shot fast spin echo (SSFSE)(TR = 1,100 ms, TE = 60 ms, slice thickness = 4 mm, 1 mm gap, matrix 192 × 256) (a) and axial T1-weighted gradient echo (TR = 180, TE = 4.5 ms, flip angle = 80°, slice thickness = 6 mm, gap 1 mm, matrix = 160 × 256) (b) images in a 77-year-old male with remote history of complete remission after IL-2 therapy for metastatic right renal cell carcinoma presenting with obstruction of his left kidney due to a complex hemorrhagic mass (asterisk). The patient was scheduled for a left nephrectomy. Note the heterogeneous signal intensity within the mass with high signal intensity on the T1-weighted image and the central hypointense area (asterisk in a, b) suggestive of clot within a hemorrhagic cyst. Reference proton density (c, top) and arterial spin labeled (c, bottom) perfusion image (TR = 6.0 s, TE = 62.1 ms, matrix 128 × 128, thickness = 10 mm) confirmed lack of blood flow in the mass (arrow) and an adjacent simple cyst (arrowhead). The nephrectomy was cancelled based on these imaging findings. Follow-up coronal T2-weighted SSFSE (TR = 1,100 ms, TE = 60 ms, slice thickness = 4 mm, 1 mm gap, matrix 192 × 256)(d) and axial T1-weighted gradient echo (TR = 1,100 ms, TE = 60 ms, slice thickness = 4 mm, 1 mm gap, matrix 192 × 256)(E) images obtained 3 months later confirm the decrease in the overall size of the cyst (arrowheads) and the clot (asterisk)

Quantitative analysis

The results of the quantitative analysis are shown in Table 3. The averaged and peak blood flow values of the renal masses interpreted by the radiologists as ASL-positive were 134 ml/100 g/min (+/− 85.7 ml/100 g/min) and 265 ml/100 g/min (+/− 262 ml/100 g/min), respectively. The measured averaged blood flow of masses interpreted by the radiologists as ASL-negative was 20.5 ml/100 g/min (+/− 8.1 ml/100 g/min). The difference of mean flow between ASL positive and negative lesions was significant when tested by lesion (p = 0.015) and by patient (p = 0.006).
Table 3

Results of quantitative analysis. Peak blood flow was not measured in renal masses characterised as ASL-negative by subjective assessment by the radiologist

Patient #

Renal mass location

Average blood flow (mL/100 g/min)

Peak blood flow (mL/100 g/min)

1

R, LP

91

146

2

L, UPJ(1)

11

3

R, IP

277

729

4

R, LP

163

291

5

R, LP

31

L, IP

20

L, UP

12

6

R, LP

21

L, IP

28

7

R, LP

71

110

8

L, IP

164

222

L, LP

54

77

9

R, LP

255

811

L, UP

47

72

L, IP

28

63

10

L, IP

207

260

11

R. UP

117

134

The average noise contribution to perfusion was 11.9 ml/100 g/min ± 5.0. The mean ASL-negative lesion flow of 20.5 is only 2 standard deviations above this level, supporting the absence of clearly detectable perfusion. With current techniques, reliable identification of perfusion below 30 may be difficult as the lowest mean perfusion identified by the radiologist as high signal in an ASL positive lesion was 28 ml/100 g/min.

An analysis of the relationship between positive ASL findings and malignancy was performed. The 8 subjects and 11 lesions with histopathological confirmation, occurrence of new lesion, or increase size in serial imaging were all found to be ASL positive. The two patients and three lesions with no size change or decreased size on follow up imaging were ASL negative. The contingency table of this relationship was found to be significant (p = 0.003 by lesion, p = 0.022 by patient). One patient with three lesions was excluded from this analysis due to lack of reference standard.

Discussion

The recently reported association between gadolinium-based contrast agents (GBCAs) and the systemic condition, nephrogrenic systemic fibrosis (NSF), has constrained the use of GBCAs in the clinical assessment of patients with renal dysfunction. Sixteen percent of patients referred for MRI evaluation of renal masses in our series did not receive gadolinium due to decreased renal function supporting the need for alternative methods in this clinical context.

In our study, non-vascularised renal masses (i.e. presumptive simple and hemorrhagic cysts) displayed very low levels of signal intensity on ASL MRI compared to vascularised masses, which demonstrated largely high signal intensity on these acquisitions. Blood flow in hypervascular renal masses is thus readily recognised on ASL images based on these large differences in signal intensity.

The average noise contribution to measured perfusion in our cohort was 11.9 ml/100 g/min ± 5.0 while the lowest mean and peak perfusion identified by the radiologist during subjective assessment of ASL images in vascularised masses were 28 ml/100 g/min and 63 mL/100 g/min, respectively. Therefore, detection of blood flow in hypovasular masses may not be possible with the proposed ASL strategy when perfusion levels are below three standard deviations of the noise contribution, at approximately 30 mL/100 g/min. This may indicate a limited sensitivity of ASL MRI to detect blood flow in some specific RCC subtypes, like papillary RCC, which is characterised by very low levels of enhancement on contrast-enhanced MRI [22]. Furthermore, the lower limits of renal lesion size detectable when using ASL need to be determined and compared with those documented when using GBCAs for renal lesions [23].

ASL was useful in guiding the clinical management in our series. In 2 patients, high levels of blood flow on ASL-MRI correlated with an aggressive biologic behavior of these tumours, which demonstrated IVC and nodal involvement respectively. Importantly, all 11 lesions that demonstrated subjective vascularity on ASL were proven to be neoplastic by either histopathology or imaging follow up.

Our study has several limitations. First, although the renal masses were prospectively evaluated by ASL, the study design was retrospective. The small number of renal masses included in this study limit the ability to draw definitive conclusions about the sensitivity of ASL MRI for renal mass characterisation. Clinical limitations precluded access to definite pathological proof in all cases. Instead, we have used a combination of lesion growth and suspicious features on unenhanced MRI as an indicator of a neoplastic process. Second, the current 2D technique may be prone to sample errors dependent on the prescription plane. Indeed, data from one renal mass in one patient was excluded from further analysis in our study upon confirmation of the incorrect prescription of the ASL acquisition, which did not include the solid component of a cystic mass. This lesion was deemed neoplastic based on its internal growth and cystic-solid morphology although was not included in our analysis as our intention was to assess the feasibility of ASL to detect differences in neoplastic versus benign renal masses. However, it should be considered a false negative for ASL from the diagnostic accuracy point of view. The potential low sensitivity of ASL to detect low levels of blood flow and the possibility of sample error could decrease its negative predictive value. Moreover, ASL may be less useful dealing with hypovascular neoplasms as opposed to clear cell RCC, which is characteristically hypervascular [24].

In summary, our initial experience suggests that ASL MRI is feasible for the characterisation of renal masses in patients with impaired renal function in whom administration of GBCA is contraindicated. The detection of blood flow in a renal mass by ASL MRI seems to be indicative of neoplasm. However, further studies are necessary to assess the diagnostic performance of ASL in the characterisation of renal masses, and to determine the lower levels of tumour blood flow and size detectable by ASL.

Copyright information

© European Society of Radiology 2011