Plasma KIM-1 for preoperative prediction of renal cell carcinoma versus benign renal masses, and association with clinical outcomes

Abstract

Purpose:

Both clear cell and papillary renal cell carcinomas overexpress KIM-1 (kidney injury molecule-1). We investigated whether plasma KIM-1 (pKIM-1) may be a useful risk stratification tool among patients with suspicious renal masses.

Methods:

Pre-nephrectomy pKIM-1 was measured in two independent cohorts of patients with renal masses. Cohort 1, from the prospective K2 trial, included 162 patients found to have ccRCC (cases) and 162 patients with benign renal masses (controls). Cohort 2 included 247 patients with small (cT1a) renal masses from an academic biorepository, of whom 184 had RCC. We assessed the relationship between pKIM-1, surgical pathology, and clinical outcomes.

Results:

In Cohort 1, pKIM-1 distinguished RCC versus benign masses with AUC-ROC 0.81 (95% CI: 0.76–0.86). In Cohort 2 (cT1a only), pKIM-1 distinguished RCC versus benign masses (AUC-ROC 0.74; 95% CI 0.67–0.80) and the addition of pKIM-1 to an established nomogram for predicting malignancy improved the model AUC-ROC (0.65 [0.57–0.74] vs 0.78 [0.72–0.85]). A pKIM-1 cut-point identified using Cohort 2 demonstrated sensitivity of 92.5% and specificity of 60% for identifying RCC in Cohort 1.

In long-term follow-up of RCC cases (Cohort 1), higher pre-nephrectomy pKIM-1 was associated with worse metastasis free survival (multivariable MFS HR 1.29 per unit increase in log pKIM-1, 95% CI 1.10–1.53) and overall survival (multivariable OS HR 1.31 per unit increase in log pKIM-1, 95% CI 1.10–1.54). In long-term follow-up of Cohort 2 no metastatic events occurred, consistent with the favorable prognosis of resected cT1a RCC.

Conclusions:

Among patients with renal masses, pKIM-1 is associated with malignant pathology, worse metastasis free survival, and risk of death. pKIM-1 may be useful for selecting patients with renal masses for intervention versus surveillance.

INTRODUCTION

There are over 76,000 cases of renal cell carcinoma (RCC) in the United States per year, and RCC causes over 13,000 deaths annually [1]. Clear cell RCC and papillary RCC are the most common RCC subtypes and together account for approximately 90% of all cases [2,3]. Suspicious renal masses, representing possible RCC, are often found incidentally on imaging studies and represent a common clinical conundrum. While renal masses are often managed by surgical resection [4], up to one-third of partial nephrectomies for renal masses result in benign findings only [5,6]. Percutaneous renal mass biopsy is invasive and has only modest negative predictive value, limiting its routine use [4]. There is significant overtreatment of benign renal masses and clinically indolent RCCs, and better risk stratification is needed to select appropriate patients for active surveillance [7,8].

Blood-based biomarkers for kidney cancer could fulfill this unmet need. KIM-1 (kidney injury molecule-1) is a transmembrane glycoprotein that is shed from kidney cancer cells and is a candidate circulating biomarker for clear cell and papillary RCC [9–12]. We previously found that elevated plasma KIM-1 (pKIM-1) after nephrectomy for RCC predicts worse disease-free survival and overall survival [13], and that patients with new diagnosis of RCC have elevated KIM-1 in banked plasma samples up to 5 years prior to diagnosis [9]. No prior studies, to our knowledge, have addressed the question of whether pKIM-1 can help to risk stratify patients prior to surgery. We therefore investigated whether pre-nephrectomy pKIM-1 can distinguish RCC versus benign renal masses, and whether higher pre-nephrectomy pKIM-1 is associated with worse cancer related outcomes.

METHODS

Patient population

We analyzed pre-nephrectomy plasma and clinical data from two independent cohorts. Cohort 1 consists of adults with renal masses from the WHO/IARC K2 multinational prospective study [14], who presented to the N.N. Blokhin Russian Medical Research Center for Oncology in Moscow between 2007–2012. Within this cohort, 162 patients were subsequently found to have benign renal masses (controls), and the WHO/IARC study statistical team selected an equal number of 162 patients found to have ccRCC (cases). Cases and controls were matched for age (within 5 years) and sex. Sample size was calculated using estimated population means of 59 pg/mL in patients without RCC and 122 pg/mL among patients with T1a RCC, which would give >80% power to distinguish non-malignant vs malignant histologies (standard error of the mean 15.6 pg/mL and two-tailed α rate 5%).

Clinical data were collected for all patients including baseline patient demographics, surgery type, and tumor histology determined after resection. Prospective follow-up for clinical outcomes was performed as part of the IARC K2 study protocol. Patients received annual follow-up visits by phone to evaluate for clinical outcomes including survival and metastasis; these data were verified with medical records from the treating cancer center and with the Moscow Cancer Registry. Additional clinical data including serum creatinine levels and radiographic measurements of tumor size prior to nephrectomy were abstracted from the patients’ medical records.

Cohort 2 is from the Johns Hopkins Brady Urological Institute Biorepository [15]. We obtained plasma from 247 patients who were diagnosed with renal masses ≤4cm and had plasma collected as part of the biorepository banking protocol prior to surgical resection. Clinical characteristics including baseline patient demographics, surgical characteristics, and tumor histology were collected by prospective follow-up and retrospective chart review [7,15].

To establish baseline pKIM-1 levels among patients without renal masses, we additionally obtained plasma from 48 healthy volunteers from the Brigham and Women’s Hospital biorepository. Participants in all cohorts provided informed consent prior to participation, and consent for exploratory biomarker analysis was embedded within the respective consent forms for the IARC K2 trial and the Johns Hopkins Biorepository. Approval was granted by the Institutional Review Board at each participating study center. This study was conducted in accordance with the Declaration of Helsinki.

Measurement of pKIM-1

We measured pKIM-1 levels using a microbead-based assay in Cohort 1 and among healthy volunteers [12], and a custom MSD assay in Cohort 2 [16,17]. Both assays were described and validated in previous studies [12,16–18]. Additionally, a subset of 80 samples from Cohort 1 were re-measured for pKIM-1 using the MSD assay to confirm assay concordance. Plasma KIM-1 measurements were conducted in duplicate, and investigators were blinded to the identity of individual plasma samples during pKIM-1 analysis. For the microbead-based assay, each sample was diluted 10-fold in sample diluent buffer (0.1M HEPES, 0.1M NaCl, 0.1% Tween-20, and 1% BS; pH 7.4; filter sterilized). Thirty microliters (30 μL) of diluted sample, recombinant standards, and internal controls were incubated with ~6000 microbeads coupled with KIM-1 capture antibody for 1 hour (R&D Systems, Cat # AF1750). Beads were then washed 3x with PBST and incubated with detection antibody for 45 min (R&D Systems, Cat # BAF1750). Beads were washed 3x with PBS-Tween and incubated with Streptavidin-PE (Invitrogen) for 15 min. The signal from the fluorochrome was captured using Bio-Plex 200 system (Bio-Rad). Data were generated and interpreted using a five parametric logistic regression analysis.

For the MSD assay, pKIM-1 was assessed with a custom sandwich immunoassay using the R-PLEX platform from Meso Scale Discovery (MSD, Rockville, MD). Human TIM-1/KIM-1/HAVCR Antibody (R&D systems, Minneapolis, MN), EZ-Link Sulfo-NHS-LC-Biotin (Thermo Fisher, Waltham, MA) and MSD GOLD SULFO-TAG NHS-ESTER (MSD, Rockville, MD) were used to prepare the biotin conjugated antibody and detection antibody. Human KIM-1 recombinant protein (R&D systems, Minneapolis, MN) was used to prepare the calibration curve (lowest calibration point, 4.88 pg/mL). KIM-1 assays were performed by Metabolism and Mitochondrial Research Core (Beth Israel Deaconess Medical Center, Boston, MA). The assay plates were read using a MESO QUICKPLEX SQ 120 instrument and data were analyzed using Discovery workbench 4.0 software.

Statistical analysis

Classification of renal masses

AUC-ROC analyses and univariable logistic models were used to characterize the value of pKIM-1 in identifying benign masses versus ccRCC. To evaluate for potential confounders, we assessed the relationships between pKIM-1 and tumor size, and between pKIM-1 and serum creatinine. Multivariable logistic regression models were used to assess the added value of pKIM-1 to a historical nomogram that featured age, sex, smoking status, and tumor size to identify benign vs malignant masses [8]. Correlation statistics between continuous variables were calculated using the Spearman method, and Wilcoxon tests were used to evaluate continuous variables between groups.

Time to event analysis

In Cohort 1, patients with ccRCC were followed prospectively as part of the K2 trial. Univariable and multivariable Cox survival models were used to determine whether pre-nephrectomy pKIM-1 was prognostic for metastasis free survival (MFS) or overall survival (OS). Patients who had known metastasis at time of surgery (n=28) were excluded from the both the univariable MFS and OS analyses. Metastasis free survival was defined as months elapsed from time of surgery to first distant metastatic event or death. Overall survival time was defined as months elapsed from time of surgery to death. Patients without events were censored at the date of last follow-up. Multivariable survival models were adjusted for age, sex, cancer stage, and surgery type (partial versus radical nephrectomy). Patients with metastases (n=28) at baseline were included in the multivariable models. Time to event analysis was not performed in Cohort 2 as there were no disease recurrence or metastasis events detected during long term follow-up, consistent with the expected low recurrence rate among patients with surgically resected cT1a RCC.

RESULTS

Patient characteristics

Baseline characteristics of our patient cohorts are shown in Table 1 and 2. Cohort 1 (K2 study) consisted of 162 patients with ccRCC (cases) and 162 patients with benign renal masses (controls). Demographics including age and gender distribution were similar among patients with malignant versus benign renal masses. The proportion of males (39%) was lower than expected, which was a result of gender matching since the benign renal mass samples with plasma available were predominantly female. The majority of patients underwent radical nephrectomy (78%), and the remainder underwent partial nephrectomy or surgery type was unknown. Among patients with radiographic tumor measurements, the majority (67%) had renal masses ≤7 cm in maximal diameter, and 39% had masses ≤4 cm. Among patients with benign masses, the majority had oncocytomas (63) or angiolipomas (59) while the remainder had renal cysts (29), angiomyolipomas (8), or angioleiomyomas (3). During the follow-up period (median follow-up 119 months), 83 patients died. Among the patients who died, 82 (99%) had ccRCC, and the cause of death was attributed to ccRCC in 66 (80%) of those patients.

Table 1: Baseline patient characteristics.

Tumor size represents maximal dimension measured on pre-surgical imaging. Histology was determined from pathologic analysis of resected tumor.

		Number of patients (%) or median (IQR)
	Cohort 1 (K2 study)		Cohort 2 (Hopkins)
	Benign	Malignant	Benign	Malignant
Histology
Oncocytoma	63 (39)		34 (54)
Renal cyst	29 (18)		11 (17)
Angiomyolipoma/angiolipoma	67 (41)		18 (29)
Angioleiomyoma	3 (2)
Clear cell RCC		162 (100)		112 (61)
Papillary RCC				44 (24)
Chromophobe RCC				28 (15)
Tumor size (cm)
0–4	80 (49)	47 (29)	63 (100)	184 (100)
4–7	46 (28)	43 (27)
7–10	16 (10)	33 (20)
>10	16 (10)	35 (22)
Unknown	4 (2)	4 (2)
Age	55.2 (43.4–62.4)	57.9 (49.7–64.2)	62.6 (53.2–70.0)	61.9 (53.3–69.2)
Sex
Male	56 (35)	69 (43)	37 (59)	114 (62)
	162 (50)	162 (50)	63 (26)	184 (74)
Total	324 (100)		247 (100)

Table 2:

Characteristics of patients found to have renal cell carcinoma.

	Number of patients (%)
	Cohort 1	Cohort 2
Total	162 (100)	184 (100)
Nodal status
N0	138 (85)	184 (100)
N1	24 (15)
Metastases at presentation
No	134 (83)	184 (100)
Yes	28 (17)
Overall cancer stage
I	76 (47)	184 (100)
II	16 (10)
III	39 (24)
IV	31 (19)

Cohort 2 (Hopkins) included 247 patients, all of whom had renal masses ≤4 cm. The majority of patients (79%) underwent partial nephrectomy, while others had radical nephrectomy (12%) or surgery type was unknown (10%). 184 patients were found to have RCC including ccRCC (112), papillary RCC (44) and chromophobe RCC (28). Among benign masses, the most common were oncocytomas (34) followed by angiomyolipomas/angiolipomas (18) and renal cysts (11). During the follow-up period (median follow-up 28 months), 4 patients died and no deaths were attributed to RCC.

Assay characteristics

We compared the results of pKIM-1 measurement using the microbead and MSD based assays among 80 samples from Cohort 1 (40 cases and 40 controls) in which additional plasma was available. The two assays were closely correlated (Pearson correlation coefficient 0.95, eFigure 1 in the Supplement).

Since KIM-1 is shed directly from the ccRCC cell membrane, we investigated whether tumor size and pKIM-1 are correlated. pKIM-1 was correlated with tumor size among patients with ccRCC but not among patients with benign masses (eFigure 2 in the Supplement).

Serum creatinine data was available from patients in Cohort 1. We investigated whether impaired renal function is a confounder for pKIM-1. No association was seen between serum creatinine and pKIM-1 among patients with either benign or malignant masses (eFigure 3A in the Supplement). The discriminatory value of pKIM-1 as a biomarker for RCC was similar among patients with high versus low baseline creatinine (eFigure 3B in the Supplement).

Pre-nephrectomy pKIM-1 in patients with RCC versus benign masses

We investigated pKIM-1 levels among patients with benign versus malignant renal masses. The distribution of pKIM-1 among patients in each cohort is shown in Figure 1A. In both Cohort 1 and Cohort 2, univariable logistic regression models showed that pKIM-1 levels were higher in patients with RCC compared to those with benign renal masses (Cohort 1, OR 1.63 per log increase in pKIM-1, 95% CI 1.43–1.85, p < 0.001; Cohort 2, OR 3.11 per log increase in pKIM-1, 95% CI 2.02–4.79, p < 0.001).

Figure 1:

Plasma KIM-1 in patients with clear cell renal cell carcinoma versus benign renal masses.

Figure 1A: Plasma KIM-1 values among patients with renal masses (Cohort 1 and Cohort 2) and healthy volunteers.

Healthy volunteer plasma samples were taken from the Brigham and Women’s Hospital biorepository. The lower limit of detection is 1.0 pg/mL for the microbead based assay (Cohort 1 and healthy volunteers) and 11.7 pg/mL for the MSD assay (Cohort 2).

We evaluated the performance of pKIM-1 as a diagnostic test using receiver operating curve analysis (Figure 1B). In Cohort 1, pKIM-1 was able to distinguish ccRCC versus benign masses with area under the receiver operating curve (AUC-ROC) 0.81 (95% CI: 0.76–0.86). Plasma KIM-1 was found to be elevated in both large and small ccRCCs, but discrimination was improved among patients with T2 tumors (>7cm) compared to T1 tumors (≤ 7cm) (Figure 1C). Among small (cT1a) renal masses 0–4cm, pKIM-1 was associated with AUC-ROC 0.73 (95% CI: 0.64–0.82) for distinguishing benign versus malignant tumors. As an exploratory analysis, excluding patients with known metastases in Cohort 1 resulted in a similar AUC-ROC of 0.79 (0.73–0.84, eFigure 4 in the Supplement).

Figure 1B:

Receiver operating curve analysis of plasma KIM-1 to distinguish clear cell renal cell carcinoma versus benign renal masses. Cohort 1, all patients

Figure 1C:

Receiver operating curve analysis, Cohort 1, patients stratified by tumor size

We next validated the performance of pKIM-1 in Cohort 2. In these patients with cT1a renal masses, pKIM-1 had AUC-ROC 0.74 (95% CI: 0.67–0.80, Figure 1D) for distinguishing all RCC subtypes versus benign tumors. This improved slightly (AUC-ROC 0.75, 95% CI 0.68–0.82) when limiting the comparison to papillary and clear cell RCC versus benign tumors. The overall prevalence of RCC was 97% among patients with the highest quartile of pKIM-1, and 55% among patients with the lowest quartile of pKIM-1. Using the Lane clinical nomogram for likelihood of malignancy in renal masses which includes age, gender, tumor size, and smoking history, we found no evidence of correlation between pKIM-1 values and nomogram probability estimates (Figure 1E). The addition of pKIM-1 to the Lane nomogram was able to better identify patients who had malignant masses despite a low Lane score and improved the model AIC and AUC-ROC (AIC 227 vs 202, AUC-ROC 0.65 [0.57–0.74] vs 0.78 [0.72–0.85]) [8].

Figure 1D:

Receiver operating curve analysis, Cohort 2 (cT1a renal masses)

Figure 1E:

Scatter plot of Lane nomogram score versus plasma KIM-1. The dashed lines represent median values of the Lane score and plasma KIM-1 value, 0.79 and 4.81 pg/mL respectively.

As an exploratory analysis, we evaluated the performance of pKIM-1 among the 44 patients in Cohort 2 who had papillary RCC. Patients with papillary RCC had higher pKIM-1 compared to those with benign masses (p<.001). Among patients with cT1a renal masses, pKIM-1 distinguished patients with papillary RCC vs benign histologies with AUC-ROC 0.77 (95% CI: 0.68–0.82).

We performed a cutoff analysis using only samples tested with the MSD assay. This included 247 patients from Cohort 2 (“development set”) and 80 patients from Cohort 1 (“test set”). In the development set, an optimal cut point (pKIM-1 ≥ 105 pg/mL) was chosen based on shortest distance to the upper left corner of the receiver operating curve. When this cutoff was evaluated in the test set, this resulted in a sensitivity of 92.5% and a specificity of 60% for identifying RCC (eFigure 5 in the Supplement).

pKIM-1 as a predictor of metastasis free survival among patients with RCC

In Cohort 1, we assessed whether pre-nephrectomy pKIM-1 is associated with metastasis free survival among patients with localized RCC. Each increase in pKIM-1 quartile was associated with progressively higher risk of subsequent metastatic event or death (Figure 2a). A univariable Cox proportional hazards model for metastasis free survival showed a hazard ratio of 1.36 (95% CI: 1.19–1.56, p<0.001) per log increase in pKIM-1. This association remained significant after multivariable adjustment for age, sex, surgery type, and cancer stage (hazard ratio 1.29 per log increase in pKIM-1, 95% CI 1.10–1.53, p=0.003, Table 3).

Figure 2a:

Kaplan-Meier curves for metastasis free survival (MFS) by pre-nephrectomy plasma KIM-1 quartile, Cohort 1.

Table 3: Multivariable HR estimates for plasma KIM-1 and metastasis free survival, Cohort 1.

Hazard ratios are calculated relative to a reference group with female sex, partial nephrectomy, and stage I renal cell carcinoma.

	Hazard Ratio	95% CI	P-value
Log pKIM-1	1.29	1.10–1.53	0.0025
Age	1.01	0.98–1.03	0.6551
Male sex	1.26	0.71–2.26	0.4317
Radical nephrectomy	1.34	0.60–2.97	0.4716
Cancer stage
II	1.17	0.47–2.91	0.7300
III	2.03	1.03–3.98	0.0400
IV	4.15	1.15–15.02	0.0299

Time to event analyses were not performed in Cohort 2, since no metastasis events occurred during the follow-up period which reflects the favorable prognosis of resected pT1a RCC.

pKIM-1 as a predictor of overall survival among patients with RCC

In Cohort 1, we assessed whether pKIM-1 prior to nephrectomy were associated with subsequent risk of death among patients with localized RCC. Higher pKIM-1 quartiles were associated with progressively worse overall survival (Figure 2b). Among the 66 patients who had death attributed to RCC, all except one (98.5%) had pre-nephrectomy pKIM-1 higher than the median.

Figure 2b:

Kaplan-Meier curves for overall survival (OS) by pre-nephrectomy plasma KIM-1 quartile, Cohort 1.

In a univariable Cox model, pKIM-1 was associated with increased risk for death (HR 1.37 per log increase in log pKIM-1, 95% CI 1.20–1.58, p<0.001) and this remained significant after multivariable adjustment for age, sex, surgery type, and tumor stage (HR 1.31 per log increase in log pKIM-1, 95% CI 1.10–1.54, p=0.002) (Table 4).

Table 4: Multivariable HR estimates for plasma KIM-1 and overall survival, Cohort 1.

Hazard ratios are calculated relative to a reference group with female sex, partial nephrectomy, and stage I renal cell carcinoma.

	Hazard Ratio	95% CI	P-value
Log pKIM-1	1.31	1.10–1.54	0.0019
Age	1.01	0.98–1.03	0.6672
Male sex	1.21	0.67–2.18	0.5285
Radical nephrectomy	1.32	0.60–2.93	0.4908
Cancer stage
II	1.17	0.47–2.92	0.7313
III	2.18	1.10–4.31	0.0259
IV	4.49	1.22–16.51	0.0237

As a sensitivity analysis, further adjustment of our multivariable models for tumor size did not substantively alter the predictive value of pKIM-1 and MFS or OS (eTable 1. in the Supplement).

DISCUSSION

In this study, we analyzed pKIM-1 levels among patients presenting with a renal mass suspicious for malignancy. We demonstrate in two independent cohorts that elevated pre-nephrectomy pKIM-1 is associated with increased risk of finding RCC at time of surgery. Among patients with RCC, those with higher pre-surgery pKIM-1 have worse metastasis free survival and worse overall survival. We find that pKIM-1 is, to our knowledge, the only known circulating biomarker that is specific for both clear cell and papillary RCC, the two most common histologies of kidney cancer. This data supports a potential role for pKIM-1 as a diagnostic and prognostic biomarker in patients with suspicious renal masses.

There are currently no circulating biomarkers in clinical use for RCC. While prior studies have investigated RCC detection using several different platforms to identify circulating cytokines, tumor associated proteins, circulating tumor cells, cell-free tumor DNA and RNA, and cell-free DNA methylation patterns, none of these techniques are currently used in routine practice [19–25]. Recently, a novel imaging biomarker ⁸⁹Zr-DFO-girentuximab PET demonstrated sensitivity of 86% and specificity of 87% for detecting ccRCC among patients with renal masses, but is not designed to detect non-clear cell RCCs which comprise 30% of all renal malignancies [26]. ^99mTc-sestamibi SPECT/CT has demonstrated 87.5% sensitivity and 95% specificity for distinguishing benign oncocytic tumors from other renal masses, but its diagnostic utility is limited by identifying only a subset of benign masses [27]. Compared to these other approaches, pKIM-1 has several properties that make it an attractive candidate biomarker for RCC. The assays for measuring pKIM-1 are inexpensive, easily scalable and require only small volumes of plasma (30 μL) [28]. pKIM-1 is stable to multiple freeze/thaw cycles and can therefore be measured in banked plasma [18,29]. KIM-1 is already qualified by the United States Food and Drug Administration (FDA) as a biomarker for nephrotoxicity, which would simplify its implementation if found to have clinical value in kidney cancer [30].

The identification of circulating biomarkers is especially challenging for patients with early stage disease. Tumor products must be released into the systemic circulation in order to be detectable in blood samples. On the one hand, this is a clinical challenge since many early stage tumors may have limited access to the systemic circulation, and there may be an inherent sensitivity limit for detection of such tumors from peripheral blood tests. On the other hand, tumors that have access to the circulation might also have greater metastatic potential, which is supported by our finding that patients with low pKIM-1 have better MFS and OS even when RCC is present.

In a prior study, we showed that elevated pKIM-1 after nephrectomy is associated with recurrence risk in RCC [13]. The current manuscript investigates a different clinical question, e.g. the use of pKIM-1 to elucidate cancer risk prior to surgery. Our results represent an important proof of concept that blood-based risk stratification of kidney cancer is feasible even for patients with early stage disease.

Strengths of this study include its large sample size and the use of two independent cohorts, including a large prospective study (K2) and a mixed prospective/retrospective study (Hopkins). The majority of our patients had small renal masses, which reflects a real-world clinical conundrum where there is often equipoise as to whether surgery is necessary. Our cohorts include a wide spectrum of benign and malignant renal mass histologies which also reflects real-world practice. Currently, the management of renal masses relies on clinical risk stratification, with some patients proceeding to surgery or other interventions while others are managed with imaging surveillance [8,31]. In the DISSRM registry, among 224 patients undergoing active surveillance for small renal masses zero metastatic events occurred, suggesting that current surveillance protocols might be too conservative [32]. In the future, pKIM-1 could be a helpful tool to identify more candidates for active surveillance.

There are several limitations to this study. We did not have data on the presence of necrosis on preoperative imaging, limiting our ability to evaluate an alternative preoperative nomogram [33,34]. Cohort 1 was a matched case-control study while Cohort 2 was a retrospective cohort, therefore the baseline characteristics of our cohorts do not represent the global population of individuals with renal masses. There is overlap in pKIM-1 between benign and malignant masses, making this biomarker less specific among the lower end of the pKIM-1 distribution and among patients with smaller renal masses (Figure 1A and Figure 1E). pKIM-1 is also a known biomarker for renal injury including medication related toxicity, and we were not able to assess the effects of concurrent medications on pKIM-1 levels within our cohorts [28].

Our study design did not directly evaluate active surveillance as an alternative strategy to resection. Since all patients in our study underwent excision of their kidney masses, we do not know how patients with low pKIM-1 would have fared on surveillance alone. Among patients with renal masses, high pKIM-1 strongly suggested the presence of cancer and higher metastatic risk. Although we found that a subset of small RCCs do not release KIM-1 into the circulation, they also appear to have low metastatic potential. These findings, in our opinion, support future prospective studies of active surveillance in patients with small renal masses and low pKIM-1. Future studies could determine whether serial measurement of pKIM-1 is a useful adjunct for surveillance of a renal mass (similar to the use of serial PSA in active surveillance for prostate cancer) [35], and whether combining pKIM-1 with other biomarkers may further improve assay performance.

CONCLUSIONS

Among patients with renal masses, those with elevated pre-nephrectomy pKIM-1 have higher risk of malignant pathology, worse metastasis free survival, and higher risk of death. Plasma KIM-1 may be a useful minimally invasive biomarker for risk stratification of patients with suspicious renal masses.

Context Summary.

Key Objective:

Does plasma KIM-1 predict higher chance of malignancy and worse cancer outcomes among patients with suspicious renal masses?

Knowledge Generated:

In two independent international cohorts, pre-nephrectomy plasma KIM-1 distinguishes benign masses from renal cell carcinoma and improves performance when added to an established clinical nomogram. Among patients with renal cell carcinoma, those with higher pre-nephrectomy KIM-1 have worse risk of metastasis and risk of death.

Relevance statement:

This evaluation of KIM-1 in individuals with renal masses is an important proof of concept that blood-based risk stratification of kidney cancer is feasible even for patients with early stage disease. Patients with renal cell cancer had higher values of KIM-1 when compared to those with benign tumors or cysts -intriguing enough yet requires subsequent validation and clinical trial testing.

Relevance statement written by Dr. Carducci