Renal mass sampling: An enlightened perspective

Authors


Steven C Campbell md phd, Department of Urology, Glickman Urological and Kidney Institute, Cleveland Clinic Foundation, Cleveland, OH 44195, USA. Email: campbes3@ccf.org

Abstract

Renal mass sampling (RMS) can be carried out by core biopsy or fine needle aspiration with each presenting potential advantages and limitations. The literature about RMS is confounded by a lack of standardized techniques, ambiguous terminology, imprecise definitions of accuracy, substantial rates of non-informative biopsies, and recurrent diagnostic challenges with respect to eosinophilic neoplasms. Despite these concerns, RMS has an expanding role in the evaluation and treatment of renal masses, in order to stratify biological aggressiveness and guide management that can range from surgery to active surveillance. Non-informative biopsies can be managed with surgical excision or repeat biopsy, with the latter showing encouraging results in recent studies. We propose a new classification in which all biopsies are categorized as non-informative versus informative, with the latter being subclassified as confirmed accurate, presumed accurate or confirmed inaccurate. This terminology will facilitate the comparison of results from various studies and stimulate progress. Incorporation of novel biomarkers and molecular fingerprinting into RMS protocols will likely allow for more rational management of patients with renal masses in the near future.

Introduction

The incidence of kidney cancer worldwide has been steadily increasing for the past two decades, now representing well over 200 000 new cancer diagnoses per year across the globe.1 Many of these tumors are incidentally detected, and it is estimated that 13–27% of abdominal imaging studies identify a renal lesion.2,3 As more masses are detected, the incidence of both renal cell carcinoma (RCC)4,5 and benign renal lesions6 has increased, with a concurrent decrease in tumor size at the time of detection.7,8 These smaller masses are associated with improved survival; however, a proportion of them prove to be potentially aggressive.9 Some renal masses can be diagnosed radiographically, such as simple cysts and angiomyolipomas (AML);10 however, many renal tumors, including oncoctyomas, “lipid-poor” AML11–13 and most RCC,14,15 are indistinguishable on imaging and histological analysis is required for a definitive diagnosis.16–21

Traditionally, all localized solid renal masses have been considered potentially malignant and treated with surgical excision, most often radical nephrectomy, in an effort to minimize the risk of metastatic dissemination. This dogma stems from the recognition that systemic medical therapy, even in the era of targeted molecular treatments, is rarely curative for kidney cancer.22 However, despite this aggressive surgical stance, mortality rates for kidney cancer have not declined.8 This discordance, namely aggressive treatment of early stage lesions without a decrease in mortality, has raised questions regarding the current paradigm of management.

Kidney cancer is now recognized as a heterogeneous disease process, with a number of distinct histopathological subtypes, having substantial variance in biological aggressiveness.23,24 For solid, enhancing, clinical T1 (<7.0 cm) renal masses, 20% will prove to be benign, 60% represent indolent RCC and only 20% have potentially aggressive features, such as high nuclear grade or locally invasive phenotype.25–28 Reflecting this, active surveillance is now an option for appropriately selected patients,8,29 and ablative therapies including radiofrequency ablation and cryoablation have entered the urologist's armamentarium. However, these treatment options risk tumor progression and metastasis. Tissue diagnosis through renal mass sampling (RMS) may help define tumor subtype and stratify aggressive potential, allowing for a more rational treatment protocol for this challenging disease.30 The past decade has been notable for a renaissance of RMS, with recent publications allowing for a more enlightened perspective.27

In the present article, we review the current approaches to RMS, including technical considerations. Encouraging results published in recent years with respect to accuracy and morbidity for both initial and repeat biopsies are summarized. We also discuss the evolving indications for RMS, particularly risk stratification for patients with small renal masses (SRM). The limitations of the current literature are reviewed, and a perspective of what will be required to advance this field is outlined. Finally, we will highlight the promising role of RMS combined with biomarkers and molecular profiling to stratify tumor aggressiveness, because this will likely become standard of care within the next decade.

Technical considerations

Imaging for renal mass sampling

RMS is most often carried out using computed tomography (CT), ultrasound (US), or a combination of the two, with each modality presenting distinct advantages related to tumor location, body habitus and other important considerations (Table 1). The primary advantage of US is real-time, multiplanar imaging,41,42 and potential disadvantages include the inability to discern isoechoic renal masses from normal renal parenchyma, distinguishing adjacent pleural folds and bowel, as well as difficulty imaging in the obese population.43 CT typically provides a higher resolution image and thereby facilitates the avoidance of adjacent vital structures and necrotic areas at the time of sampling.43,37 Magnetic resonance imaging (MRI) can also be used to guide RMS, but is only rarely used, typically for masses that cannot be seen by US or CT, or before MRI guided thermal therapy.44 Although each modality has merits and limitations, the choice should ultimately be based on operator preference, taking into account patient and tumor-related characteristics. Recently, US-guided RMS using advanced software and instrumentation to fuse prior CT or MR imaging to real-time US images has been evaluated in animals and humans.45 This might combine the benefits of each approach and overcome some of the difficulties, but more studies and improved instrumentation are needed.

Table 1.  Studies of renal mass sampling 2006–2010: biopsy characteristics
ReferenceNo. tumorsNo. pathologically confirmed (%)No. on radiographic surveillance (%)Length of radiographic surveillance (months)No. cancer (%)BiopsiesGuidanceComplications
Needle type (gauge)No. samplesMinorMajor§
  • Radiographic surveillance was most commonly carried out with computed tomography (CT) or ultrasound (US).

  • Minor complications primarily included perirenal or retroperitoneal hematomas, which were managed conservatively; that is, not requiring transfusion or other forms if intervention (n = 8). Other minor complications included flank pain (n = 1), small pneumothoraces managed conservatively (n = 2) and wound infection (n = 3).

  • §

    Major complications included postprocedural hemorrhage requiring transfusion (n = 3) and pseudoaneurysm requiring endovascular embolization (n = 1). In addition, one intrarenal hematoma was misdiagnosed as a local recurrence and erroneously excised at partial nephrectomy rather than the tumor; the patient subsequently underwent completion radical nephrectomy. FNA, fine needle aspiration; NA, not applicable.

Vasudevan et al. (2006)3110048 (48%)43 (43%)Not mentioned51 (51%)16-GNACT or US01
Beland et al. (2007)32588 (13.8%)45 (77.6%)1–8639 (67.2%)VariedNACT (55) or US (3)00
Lebret et al. (2007)3311964 (53.8%)55 (46.2%)21–4682 (68.9%)17-G coaxial cannula, 18-G core biopsy1–4CT00
Maturen et al. (2007)3415287 (57.2%)NANA87 (57.2%)18-G3–4CT (76) or US (76)12
Reichelt et al. (2007)353024 (80.0%)5 (16.7%)Not mentioned20 (66.7%)18-G2–3US10
Somani et al. (2007)367040 (57.1%)30 (42.9%)3–7248 (68.6%)16-G2–4US (67) or CT (3)10
Schmidbauer et al. (2008)377878 (100%)0NA65 (83.3%)18-G and FNA2–3CT (76) or US (76)10
Shannon et al. (2008)15235108 (50.0%)115 (48.9%)1–91146 (62.1%)18-G1–4US or CT02
Volpe et al. (2008)3810021 (21%)79 (79%)Not mentioned67 (67%)17-G coaxial cannula, 22-G FNA 18-G core biopsyNAUS (45), CT (11), both (44)30
Masoom et al. (2009)393131 (100%)0NA28 (90.3%)FNANANANANA
Wang et al. (2009)4011036 (55.4%)88 (80%)0–5668 (61.8%)17-G coaxial cannula, 18-G core biopsy>2CT (66) or US (44)80

Technique for RMS

Even in 2010, it is interesting to note that there are no particularly informative direct comparative in vivo studies assessing outcomes with various approaches to RMS. Both fine needle aspiration (FNA) and core renal mass biopsy (RMB) can present distinct advantages in certain situations (Fig. 1). FNA is safer in cases where the biopsy tract might potentially traverse bowel or vascular organs, and is processed quickly, allowing for the determination of an adequate sampling within minutes.46 Although no specialized instrumentation or personnel are required, RMB requires fixation and pathological processing, which may take up to 24 h.46 A frozen section can be considered, but requires dedicated support, and carries distinct limitations with respect to accuracy.46 Regardless of the technique, RMS should be carried out through a coaxial cannula,47 because this minimizes the risk of needle tract seeding, improves patient comfort by avoiding multiple needle sticks, and improves radiographic visualization as a result of the larger diameter of the cannula.48

Figure 1.

(a) Fine needle aspiration (FNA) of a renal oncocytoma shows abundant single or small nests of tumor cells with abundant eosinophilic granular cytoplasm, well-demarcated cell borders and round nuclei with small-sized nucleoli (Diff-Quick stain; figure supplied courtesy of Dr Longwen Chen, Cleveland Clinic). (b) Renal mass biopsy (RMB) of the same tumor shows that the tumor cells have abundant granular cytoplasm and uniform nuclei, confirming the diagnosis. (c) FNA of a clear cell renal cell carcinoma shows a cohesive cell group. The cells have abundant wispy cytoplasm with ill-defined edges and a large, round, eccentrically placed nucleus. Cytoplasmic vacuoles are also seen (figure supplied courtesy of Dr Longwen Chen, Cleveland Clinic). (d) RMB shows a tumor with clear cells which are separated by thin and arborizing vessels.

Fine needle aspiration

FNA is carried out using negative pressure and mechanical disruption (repetitive advancement of the needle throughout the substance of the tumor), which may produce more cellular samples than RMB.46 Samples can be used to prepare cell blocks, which may be useful in distinguishing between chromophobe RCC and oncocytoma,46 immunohistochemistry, fluorescence in situ hybridization (FISH), and cytometric techniques.43 Rapid handling is imperative to avoid clot formation, which can compromise diagnosis, and a cytopathologist must be present to assess sample quality.43

Although advocated as easier to carry out, the diagnostic accuracy of FNA appears to be somewhat variable.49,50 Accuracy is highly dependent on the skills of the cytologist, and is generally perceived to be lower than RMB.46 However, the processing technique may influence accuracy rates,51 and recent studies have shown that the accuracy of FNA can approach 100% for malignancy and 92% for histological subtype,50,52,53 with excellent concordance with final surgical pathology.39

Core biopsy

RMB is most often carried out using an 18-gauge needle to sample various areas of the mass, and larger core needles are theoretically associated with an increased risk of bleeding.46 Procoagulants or Gelfoam pledgets can be injected along the needle tract to promote hemostasis at completion of the procedure. Because central necrosis is often found in larger tumors (>4 cm), most authors advocate two to three peripheral biopsies as standard protocol.54 For tumors <4 cm, the standard recommendation is for one or more biopsies from both the central and peripheral regions.54 Cores should be visually assessed for size and quality, and those that are fragmented or particularly small (<1 cm) should prompt additional biopsy passes.55 Samples can be used to extract genetic material for genomic analyses, such as complementary DNA expression microarrays or comparative genomic hybridization.43

Most authors' experiences suggest that the use of an 18-gauge needle improves biopsy yield relative to FNA,56–58 although there is no randomized prospective data to confirm this, and FNA continues to play an important role in certain clinical settings.39,38 For instance, Breda et al. carried out ex vivo RMB and found that an 18-gauge needle (or larger) was most accurate in determining histological diagnosis,59 and Schmidbauer et al. reported similar findings related to their in vivo study. In their experience, RMS was initially carried out using both RMB and FNA; however, as the study proceeded, the authors felt that FNA had a prohibitively high rate of non-informative samples and a lower diagnostic yield for tumor subtype and grade. These authors also felt that the need for an experienced cytopathologist rendered FNA less attractive, which led them to eventually eliminate FNA from their RMS protocol.37

Some studies suggest that RMB and FNA may provide complementary results and have incorporated both into their RMS approach. For example, Wood et al. found a 93% sensitivity and 95% accuracy for malignancy using a combination of FNA and RMB,60 better than either technique in isolation.61,62 Volpe et al. confirmed these results, finding that FNA increased the diagnostic yield from RMB, providing a diagnosis in an additional 22.2% of cases.38 Incorporation of FNA has also improved the assessment of tumor subtype or grade in some studies.38

In summary, although RMB appears to provide more robust information in most studies, FNA can complement RMB and also presents distinct advantages in certain clinical settings. At our center, FNA is primarily used to evaluate for metastatic renal lesions in patients with an extrarenal malignancy, because differentiation of such lesions from primary RCC is readily accomplished, and this technique appears to be associated with less discomfort and lower morbidity. In contrast, we have routinely used RMB in most other circumstances, taking advantage of the increased tissue yield that it provides.

Accuracy and false-negative results

The “diagnostic accuracy” of RMS for malignancy continues to increase;27 before 2001 averaging 82%, increasing to 90% between 2001 and 2006,27 and currently >95% (Table 2). However, to some extent this is misleading, given the highly variable definitions of “accuracy” used by many of the studies in the literature. This is one of the most challenging aspects of the RMS literature and represents a major source of confusion. In Table 2, similar to most series in the literature, accuracy is defined as the percentage of informative biopsies for which the pathological diagnosis appeared to be correct; that is, not a false-positive or a false-negative, based on final surgical pathology or radiographic and clinical surveillance. This definition is to some extent artifactual, because it presumes that radiographic and clinical surveillance is reliable and it ignores one of the major limitations of RMS in this era, namely non-informative biopsies.

Table 2.  Studies of renal mass sampling 2006–2010: results
ReferenceNo. biopsy failure (%)‡,¶No. indeterminate (%)§,¶No. biopsy non-informative (%)††No. false-negatives (%)No. false-positives (%)No. accurate/Total No. (%)Sensitivity for malignancy‡‡
Malignant vs benignHistologyGrade
  • Accuracy for malignancy was defined as the percentage of informative biopsies for which the diagnosis was accurate; that is, not false-negative or false-positive.

  • ‡Biopsy failure was defined as any procedure for which the tumor could not be targeted or tumor tissue could not be obtained.

  • §

    §Indeterminate biopsies included cases in which tumor tissue was obtained, but the pathologist could not make a definitive diagnosis.

  • ¶Percentage of indeterminate, false-negative and false-positive biopsies were based on the total number of biopsies.

  • ††

    ††Non-informative biopsies included biopsy failures and indeterminate biopsies.

  • ‡‡

    ‡‡Sensitivity was defined as the number of definitively positive biopsies for cancer relative to the total number of cancers in the study. NA, not applicable.

Vasudevan et al. (2006)63NANA29 (29%)0071/71 (100%)44/44 (100%)NA43/51 (84.3%)
Beland et al. (2007)323 (5.2%)3 (5.2%)6 (10.4%)1 (1.9%)051/52 (98%)NANA38/39 (97.4%)
Lebret et al. (2007)3312 (10.1%)13 (10.9%)25 (21%)0094/94 (100%)55/64 (86%)29/64 (46%) 47/64 (73.4%) when classified as low-grade (Furhman 1 & 2) and high-grade (Furhman 3 & 4)70/82 (85.3%)
Maturen et al. (2007)34NANA6 (4%)00146/146 (100%)NANA85/87 (97.7%)
Reichelt et al. (2007)351 (3.3%)4 (13.3%)5 (16.6%)0025/25 (100%)NANA17/20 (85%)
Somani et al. (2007)36NANA9 (13%)0061/61 (100%)NANA44/48 (91.7%)
Schmidbauer, et al. (2008)3702 (3%)2 (3%)3 (3.8%)073/76 (96.1%)59/60 (98.3%)44/58 (76%)60/65 (92.3%)
Shannon et al. (2008)1539 (16.6%)12 (5.1%)51 (21.7%)00184/184 (100%)106/108 (98%)NA138/146 (94.5%)
Volpe et al. (2008)388 (8%)8 (8%)16 (16%)0084/84 (100%)56/60 (93%) for RCCs41/60 (68%)66/67 (98.5%)
Masoom et al. (2009)3900001 (3.2%)30/31 (96.7%)28/29 (96.6%)NA28/28 (100%)
Wang et al. (2009)40NANA10 (9.1%)00100/100 (100%)NANA65/68 (95.6%)

In reality, all biopsies fall into four basic categories, non-informative or informative, with the latter including those that are confirmed accurate, presumed accurate or confirmed inaccurate.

Non-informative

This group is comprised of failed and indeterminate biopsies, with a “failed” biopsy being one in which the tumor could not be targeted or tumor tissue was not obtained, and an “indeterminate” biopsy being one in which tumor tissue was obtained, but the pathologist could not provide a definitive diagnosis. In studies since 2006, the rate of non-informative biopsies has averaged 10–20% (Table 2). In the past, many non-informative biopsies were inappropriately considered “false-negative”, representing a major criticism of RMS, in that missed malignancies would potentially remain untreated. However, a recognized non-informative biopsy can be managed accordingly, either with proactive surgical extirpation or repeat RMS. Recent data about repeat RMS is encouraging and will be reviewed later in this article. In summary, non-informative biopsies are not uncommon and remain a limitation of conventional RMS, but with appropriate insight they are readily manageable.

Informative and confirmed accurate

These are the true-positive and true-negative biopsies, which have been confirmed by surgical extirpation. They are the optimal outcomes of RMS, at present representing the majority of cases (50–60%; Table 2).

Informative but proven inaccurate

These are biopsies in which a diagnosis was rendered at RMS, but was divergent from the final surgical pathology. This group represents the truly inaccurate biopsies, including false-negative and false-positive results, and is the most concerning for clinicians. Fortunately, these represent only a small proportion of cases encountered in recent studies, approximately 1% (Table 2).

False-negatives are the most disturbing outcome, because they would typically lead to observation of a malignancy with the potential for metastatic dissemination.

Sampling error and tumor heterogeneity are inherent to any biopsy procedure and are the most common reasons for a false-negative biopsy.56 Small masses are more difficult to target at biopsy, as illustrated by a lower biopsy sensitivity56 and a higher rate of biopsy failure.57 However, sampling error is also present in larger tumors as a result of the presence of central necrosis.56,57 Complicating the situation, a degree of tumor heterogeneity is present within most renal masses, with grade variability present in up to 25%.64 Concurrent histological subtypes32 and gene expression heterogeneity are also common.65,66 Finally, intrinsic to any biopsy technique is a degree of interobserver and intraobserver variability, as shown in a series of repeat biopsy experiments by Kummerlin et al.53,61,67

Informative and presumed true

This group reflects another important limitation of RMS in the current literature. These are biopsies that are informative, in that they yielded a diagnosis of benign or malignant, but pathological confirmation has not been obtained, representing approximately 20–35% of cases in recent series of RMS (Table 1). Almost all of these are biopsies that yielded a benign diagnosis, leading to active surveillance. Typical follow-up combines clinical and radiographic inputs, with many studies presuming confirmation of benign histology as long as substantial local or metastatic progression is not observed. However, without surgical extirpation (or autopsy), this presumed diagnosis cannot be proven true, and follow-up in most series is rather limited (Table 1). One source of confusion is that many series count these cases as accurate, although they are not pathologically confirmed.

Obtaining a clear understanding of the reliability of this group of RMS diagnoses would require surgery, even for patients with a benign diagnosis from RMS. As one could imagine, this is not routinely carried out, and very few recent studies have pursued this course. In the series by Schmidbauer et al., all masses were excised, including 21 with a benign diagnosis at RMS.37 Of these 3 (14%) were found to be malignant, representing false-negative diagnoses. Notably, this included a grade 2 clear cell RCC and two hybrid tumors, with features of both chromophobe RCC and oncocytoma.37 The latter tumors are often classified as “oncocytic neoplasms” on RMS and are typically considered to be benign at many centers.

The Schmidbauer study37 shows that the presumption that a benign diagnosis at RMS is truly benign is false in a small but substantial number of cases. Although it is likely that most of these tumors are indolent, such false-negatives are concerning. The overwhelming majority of presumed benign tumors in recent series have been stable on radiographic and clinical surveillance, suggesting that most of these tumors are indeed not aggressive (Table 2). In the end, final management of a benign biopsy result will depend on radiographic appearance, technical biopsy related factors, patient age, comorbidities and patient treatment preference.

The term “oncocytic neoplasm” is commonly encountered and can be descriptive of oncocytoma as well as eosinophilic RCC such as chromophobe RCC, granular variants of clear cell RCC, eosinophilic variants of papillary RCC, and epithelioid AML. This differentiation remains a major diagnostic dilemma in the field as a result of overlapping cytomorphology, particularly with the limited pathological material obtained from RMS. An excellent example of this is the study by Liu and Fanning, which analyzed 18 tumors with FNA, yielding a diagnosis of oncocytic neoplasm in the majority of cases (n = 10, 56%).68 Of these, the final pathology was oncocytoma in eight (80%), and eosinophilic papillary RCC and chromophobe RCC in one each.68 These data combined with the Schmidbauer study37 suggest that 10–20% of “oncocytic neoplasms” or benign diagnoses from RMS might in fact be malignant if subjected to more rigorous pathological analysis. Immunohistochemical stains that are routinely used appear to have limited success in tumor diagnosis and classification in this setting.69 Ancillary studies, such as FISH and electron microscopy, can improve our diagnostic accuracy for “oncocytic tumors”, but have not been routinely incorporated into most RMS studies to date.68,70

Although the accuracy quoted in many recent studies is artifactually high (Table 2), for reasons discussed earlier, other performance characteristics of RMS provide a more realistic indication of the utility and value of this procedure. The sensitivity of contemporary RMS series, defined as the number of cancers identified by RMS relative to the total number of cancers, ranges from 84.3% to 100% (Table 2). This is somewhat lower than the figures quoted for “accuracy”, because it takes into account non-informative biopsies that failed to diagnose the malignancy. Positive predictive value is extremely high in these series as reflected by the very low false-positive rate. Hence, a diagnosis of cancer can be relied on and will almost always be confirmed as malignant on final surgical pathology. Data about negative predictive value must be considered limited, because most patients with a benign diagnosis on RMS do not proceed to surgery; however, data from the Schmidbauer37 and Liu68 studies suggest that this will be approximately 80–90%. Specificity is often discussed in the RMS literature, but again, conclusions about this should be restricted as a result of the lack of surgical confirmation in most studies.

In summary, the RMS literature to date has been compromised by ambiguous terminology, as well as unclear and at times inappropriate definitions. Quoted “accuracy” rates may be unrealistically high, as RMS is still challenged with major concerns with respect to sampling error, tumor heterogeneity, and differentiation of various eosinophilic neoplasms as illustrated by the conundrum of the “oncocytic neoplasm”. Nevertheless, sensitivity and positive predictive value remain very high overall, and most importantly, false-negative rates are almost negligible, which has great clinical relevance. Non-informative biopsies still represent 10–20% of all RMS, although with appropriate recognition, they can be managed in a sensible manner.

Accuracy of tumor grade and subtype

As experience with RMS has increased, so has interest in the grading of biopsies, because this may have implications for tumor management by stratifying tumor risk. Concordance rates between biopsy specimens and final surgical pathology range from 46% to 94% for Fuhrman nuclear grade,43,37,50,54,55,57,71–73 although most discordant cases are only one grade off. This divergence may be a result of both interobserver variability71 and tumor heterogeneity.64 Notably, concordance was increased in the Lebret study to 76% by compressing the classification to “low” and “high” grade, which is likely to have clinical relevance.55,33 Other studies have confirmed this, with concordance rates ranging from 76% to 100% when this simplified grading system was used.55,57,72 The significance of this becomes apparent with the increasing use of active surveillance and ablative therapies, which generally should not be utilized in the setting of a high-grade cancer, regardless of size.46

Historically, RMS has been shown to have an accuracy of 87–100% for identifying histological subtype,27 with RMB being more reliable than FNA.61,62 In studies since 2006, accuracy for histological subtype was assessed in six studies, substantially adding to this literature. Accuracy rates ranged from 86% to 100%, averaging approximately 94% (Table 2). Clear cell RCC, in particular, tends to be more aggressive, and RMS could help stratify patients with this histology for tailored counseling and management.74,75

Non-informative results and the role of repeat RMS

Prior to 2006, up to 36% of FNA and 7.7% of RMB specimens were considered non-informative, comprised of “failed” and “indeterminant” biopsies.27 As outlined earlier, the rate of non-informative biopsies is now down to 10–20% in studies since 2006 (Table 2).

Several recent studies have reported on repeat RMS for appropriately selected tumors with initially non-informative biopsies, generally with encouraging results, as summarized in Table 3. In these studies a definitive diagnosis was obtained in 83–100% of repeat RMS cases, most of which revealed a malignancy. However, similar to first-time biopsies, not all repeat biopsies have been subjected to surgical extirpation, again representing a potential source of inaccuracy. These important and novel studies establish repeat RMS as a viable option in the management of non-informative biopsies.

Table 3.  Outcome of repeat biopsy of non-informative renal mass sampling
ReferenceNo. tumorsNo. non-informative biopsiesNo. tumors that underwent repeat biopsyDiagnostic yield (% definitive)Bx result
  • Diagnostic yield was defined as the number of informative repeat biopsies relative to the number of repeat biopsies performed. AML, angiomyolipoma; RCC, renal cell carcinoma.

Wood et al. (1999)607966100%5 malignant, 1 benign
Lechevallier et al. (2000)577315475%3 RCC
Vasudevan et al. (2006)3110029888%4 RCC, 3 oncocytoma
Lebret et al. (2007)331192513100%11 RCC, 1 pyelonephritis, 1 oncocytoma
Somani et al. (2007)367091100%1 RCC
Shannon et al. (2008)15235511283%5 RCC, 1 TCC, 4 oncocytoma
Volpe et al. (2008)38100162100%1 RCC, 1 oncocytoma
Wang et al. (2009)40110101100%1 AML

Complications

RMS has proven to be a safe procedure with low and manageable morbidity. Contemporary series show a minor complication rate of <5%, with serious complications being exceedingly rare.27 We found that of 1083 RMS carried out since 2006, there were 15 (1.4%) minor and five (0.46%) major complications (Table 1). Most complications were related to bleeding and did not require a transfusion. The risk of this can be minimized through the introduction of Gelfoam pledgets through the coaxial cannula,38 although persistent bleeding should prompt suspicion for an arteriovenous fistula.43 Small pneumothoraces may also occur, particularly after a posterior approach to an upper pole renal mass, but these are rare and generally managed conservatively.43,37,38,34

Major complications, such as tumor seeding along the needle tract, have been estimated to occur in 0.01% of RMS,64 with the last reported case of needle tract seeding being in 1992.76–78 Urothelial carcinomas are at higher risk for tract seeding than RCC and therefore percutaneous biopsy should be carried out with extreme caution for centrally located infiltrative renal masses.64 When urothelial carcinoma is suspected, an endoscopic approach is generally preferred. A coaxial sheath helps prevent tumor seeding, and there have been no reported cases of seeding using this technique.77,78 Other variables that might correlate with complication rates include needle size,79,80 operator experience, and tumor location.46 Major complications in recent series include post-procedural hemorrhage requiring transfusion (n = 3), pseudoaneurysm requiring endovascular embolization (n = 1), and an intrarenal hematoma that was erroneously excised at partial nephrectomy rather than the tumor, for which the patient subsequently underwent completion radical nephrectomy. With this exception, RMS has not been shown to negatively impact subsequent surgery.55,57,60 Therefore, the risk of complications from RMS needs to be weighed against the risks of managing a patient with suboptimal information. Many will be managed too aggressively, and perhaps some too conservatively placing the patient at risk of cancer progression.

Indication for RMS in 2010

Traditionally, RMS has been utilized in the diagnosis of lymphoma, metastatic carcinoma, infection/abscess or concurrent with ablative therapies (Table 4). For patients with a known extrarenal primary cancer, biopsy is 90% sensitive for the detection of malignancy, and half of such lesions will prove to be new primary renal tumors, illustrating the utility of RMS in this setting.56 With the expanding body of knowledge regarding kidney cancer, particularly with regard to cytogenetic and molecular factors, as well as expanding therapeutic options, there has been a renewed interest in RMS as a diagnostic tool.8,53 The role of RMS has expanded to include the evaluation of complex cystic lesions, SRM <4 cm, and to determine tumor subtype (Table 4).81,82 In these settings, RMS can be used to tailor treatment to each individual patient and optimize oncological efficacy.

Table 4.  Indications for renal mass sampling
  1. UTI, urinary tract infection.

Absolute indications:
 Renal mass and known extrarenal malignancy
 Renal mass and febrile UTI, possible abscess
 Suspected lymphoma
 Concomitant with thermal ablation
Relative indications:
 Mass in a solitary kidney or bilateral renal masses
 Renal mass with imaging features suggestive of unresectable renal cancer
 Medically unfit
Emerging Indications:
 Small enhancing renal masses
 Indeterminate cystic lesions
 Determination of tumor subtype in metastatic setting

RMS of small renal masses

For the past two decades, the incidence of SRM, has disproportionately increased, now comprising up to 66% of all renal tumors.9,83 These SRM have a higher likelihood of being benign than other renal tumors, because larger tumor size and symptoms at presentation are important predictors of malignant histology.84–88 For instance, a review of 2770 solid renal mass resections found that 25% of masses <3 cm, 30% of masses <2 cm, and 44% of masses <1 cm were benign at extirpation.26 Similarly, our institution reported that 20% of SRM treated with partial nephrectomy were benign and only 20% were cancers with potentially aggressive histopathological characteristics.28 In addition, radiographic surveillance has shown relatively slow growth rates and a low risk of metastatic progression for selected SRM.26,29,77,89–92

RMS can provide valuable information that might allow for the stratification of oncological risk, and should be considered for select patients with SRM in whom it might impact clinical management. Young, healthy patients who will not accept the ongoing uncertainty and low level risk of RMS should be managed proactively, preferably by partial nephrectomy, which essentially represents an excisional biopsy that is both diagnostic and therapeutic. In older or frail patients who are not candidates for proactive treatment, RMS will not change the management paradigm. RMS should be considered for most other patients with SRM, who would be potential candidates for a variety of treatment options, ranging from surgery to active surveillance, particularly those with tumors that might require radical nephrectomy and its associated increased risk of chronic kidney disease.93 In this setting, the information from RMS, while not infallible, will allow for risk stratification and help guide management.

Ablative therapies

As ablative therapies have gained more traction in the field, the sampling of masses before treatment is paramount to both identify benign lesions and determine response to treatment.28,29,86–88,90,92 These therapies are still relatively new and long-term oncological efficacy data remains limited.94 RMS, either before or concomitant with ablation, should be carried out routinely, representing an absolute indication in this era. Some have advocated RMS as a separate procedure before ablation to spare patients with benign masses from potential complications,95 and to channel patients with potentially aggressive tumors towards more rigorous treatments.46 After ablative therapy, RMS is indicated when imaging findings are atypical, to rule out persistent or recurrent disease.96

Indeterminate cystic lesions

Although simple renal cysts can be reliably diagnosed on the basis of radiographic findings, complex cystic masses present a spectrum of malignant potential. Bosniak III and IV cystic masses have traditionally been managed surgically, although their risk of malignancy is highly variable, ranging from 31% to 100%.97,98 RMS is being investigated in this population as well as select Bosniak IIF lesions that have lower, but not insignificant, risk of malignancy.

As many cysts have both solid and liquid components, sampling error and false-negative results are major concerns; however, the perspective on this is beginning to change. Richter et al. found that the combination of FNA and RMB was able to histologically classify 90% of 227 Bosniak II-III lesions, as confirmed at surgical extirpation.99 Lang et al. had similar results, making a definitive diagnosis in 88% of Bosniak IIF and III renal cysts using FNA and RMB.100 The data supporting the role of RMS in complex cystic lesions is limited, but as experience increases and reliable immunohistochemical markers such as carbonic anhydrase-IX (CA-IX) emerge,101 RMS in this cohort might play an increasing role. Tumor spillage has not been reported in recent studies, but remains an ongoing concern, as does the alteration in radiographic appearance that might occur after cyst sampling, which might complicate follow-up imaging.

Metastatic RCC

RMS of renal primary lesions in patients with metastatic RCC is likely to play an increasing role in the future, because it could identify the relevant molecular pathways that are pathogenic and allow for more precise targeted systemic treatment.46,102–104 However, RMS in this setting appears to be more challenging, likely reflecting increased tumor heterogeneity, occasional divergent pathologies, predominance of tumor necrosis and interest in a more precise diagnosis than “cancer” versus “benign”.56,105 In our experience, an increased number of biopsy specimens are often required in this setting.

Pathological advances

Exciting advances in immunohistochemical and cytogenetic techniques, and the emergence of reliable markers for identifying specific renal neoplasms, hold great promise for RMS (Table 5).102–104 These analyses have the potential for reducing the incidence of non-informative biopsies and providing increased differentiation of “oncocytic neoplasms” (Fig. 2). One example of this is the study from Beland et al., who analyzed RMS that were non-informative by conventional hematoxylin–eosin staining alone, and reported a definitive diagnosis in 89% of cases with the addition of immunohistochemistry and other ancillary techniques.32 Others have corroborated this, although studies in this domain remain limited.106,107

Table 5.  Histological, immunophenotypic and genetic characteristics of renal tumors frequently encountered in renal mass sampling
DiagnosisHistologyImmunophenotypeGenetics
  1. RCC, renal cell carcinoma.

Clear cell RCCNests of tumor cells separated by thin interconnecting vasculature and dilated sinusoidal space. Tumor cells with clear cytoplasm.(+): CA-IX, CD10, vimentin, EMA
(−): Cytokeratin 7, AMACR, E-cadherin
• Deletion of chromosome 3p
• Mutation in VHL gene
• Hypermethylation of VHL promoter
Papillary RCCTumor cells forming papillary structures with foamy histiocytes and hemosiderin.(+): Cytokeratin 7, AMACR, CD10
(−): CA-IX and E-cadherin
• Gain of chromosomes 7 and 17
• Loss of Y chromosome
Chromophobe RCCLarge and polygonal tumor cells with finely reticulated cytoplasm, prominent cell border resembling plant cells and irregular wrinkled nuclei with perinuclear clearing.(+): Hale's colloidal iron stain, Cytokeratin 7, E-cadherin.
(−): CA-IX
• Extensive chromosomal loss, most commonly involving chromosomes 1, 2, 6, 10, 13, 17 and 21
OncocytomaNests of tumor cells with abundant eosinophilic granular cytoplasm and uniform nuclei.(+): Hale's colloidal iron in a luminal fashion, Cytokeratin 7 in single cells or small clusters, E-cadherin.
(−): Vimentin
• Mixed population of cells with normal and abnormal karyotypes.
• Loss of chromosomes 1, 14, and t(5; 11) is sometimes observed
AngiomyolipomaClassical cases with namesake components: fat, abnormally formed vessels and smooth muscle cells. Tumors with predominantly one element can occur.(+): Melanocytic markers: HMB-45 and Melan-A
(−): Epithelial markers
• None
Figure 2.

(a) Computed tomography (CT) image of a low-fat angiomyolipoma (AML). (b) Renal mass biopsy (RMB) shows a spindle cell neoplasm with bland, elongated spindle cells with abundant eosinophilic cytoplasm. No adipose or vascular elements are present, and a definitive diagnosis of AML cannot be made. (c,d) Immunohistochemical staining shows tumor cells staining positively for (c) the melanocytic marker HMB-45 and (d) Melan-A, confirming the diagnosis of a fat-poor AML. (e) CT image of a clear cell renal cell carcinoma (RCC). (f) Hematoxylin–eosin image of the corresponding renal mass biopsy sample shows clear cell RCC at the tip of the core. Tumor cells are separated into small nests by a thin, interconnecting vascular network. (g) Immunohistochemical staining shows positive staining for CA-IX consistent with clear cell histology. (h) FISH also confirms the diagnosis, showing deletion of chromosome 3p in tumor cells (green signals) and preservation of chromosome 3 centromeres (red signals; figure supplied courtesy of Dr Liang Cheng, Indiana University School of Medicine).

One renal neoplasm for which immunohistochemical markers have proven particularly useful is AML. AML are generally diagnosed using imaging alone, with the exception of low-fat AML, which is often radiographically indistinguishable from RCC.12,108–110 HMB-45,111 desmin, actin,111,112 cytokeratin and epithelial membrane antigen82 have emerged as reliable markers to confirm the diagnosis and exclude other tumors that mimic AML (Table 5). Similar markers and electron microscopy are available for differentiating oncocytoma from RCC (Table 5), which if applied to RMS would allow for reliable differentiation of the two most common benign renal histological subtypes. However, the technologies that are required to pin down these diagnoses may be difficult to apply to the limited tissue obtained from RMS, and much research is needed in this area.

Enhanced RMS

As outlined earlier, the future of the field clearly is enhanced RMS, in which molecular analyses are utilized to provide increased diagnostic capacity and risk stratification. One example of this are the ex vivo RMS studies from Barocas et al., which improved the accuracy of RCC subtyping by adding polymerase chain reaction (PCR) and FISH for the most common genetic aberrations in RCC.49,106 Roh et al. also recently applied FISH to FNA to improve the differentiation of histological subtype.113 Li et al. used PCR to amplify CA-IX in FNA specimens and reported enhanced sensitivity (53 to 68%), specificity (71 to 100%) and positive predictive value (88 to 100%) for clear cell RCC, as compared with FNA alone.114 Similarly, CA-IX testing of aspirated fluid from renal cysts has been shown to be useful in distinguishing between malignant and benign lesions.101

There is increasing evidence that molecular profiling from proteomics or gene expression analyses can provide molecular fingerprints that may correlate with tumor histological subtype and aggressive potential.65,107,115,116 For instance, gene expression microarrays have been used to classify patterns of gene expression characteristic of aggressive variants of RCC,65 and these gene expression profiles have been shown to be concordant with final surgical pathology.117 It is likely that information from molecular profiling of RMS will be combined with clinical factors (age, sex, symptoms), tumor related factors (size), and radiographic parameters (degree of enhancement and heterogeneity) to provide refined information to risk-stratify patients with renal tumors for counseling and management.

Conclusions

The RMS literature continues to be confounded by a lack of standardized techniques, ambiguous terminology, unclear and inappropriate definitions of accuracy, and recurrent diagnostic dilemmas with respect to non-informative biopsies and “oncocytic neoplasms”. We propose a new terminology in which all biopsies are classified as non-informative versus informative, with the latter being subclassified as confirmed accurate, presumed accurate or confirmed inaccurate. Despite these limitations, RMS has a definite and expanding role in the evaluation and treatment of renal masses. One of the most significant areas that RMS is being used is in the evaluation of SRM, in order to stratify biological aggressiveness. As our experience continues to expand, future studies will focus on the role of repeat biopsy, and the use of biomarkers and molecular fingerprinting in order to facilitate a more rational approach to the management of renal masses.22

Ancillary