Investigational biomarkers and the evaluation of acute kidney injury

INTRODUCTION — Acute kidney injury (AKI) is a common clinical problem [1-7]. Although measurement of the serum creatinine concentration is widely used for the detection of AKI, it does not permit early diagnosis of acute tubular necrosis (ATN), since tubular injury precedes a significant rise in serum creatinine. Investigational biomarkers have been evaluated in patients with possible ATN in an attempt to detect tubular injury at an earlier stage. The US Food and Drug Administration (FDA) has approved use of the first platform measuring tissue inhibitor of metalloproteinases-2/insulin-like growth factor-binding protein 7 (TIMP-2/IGFBP7) to assess for the development of AKI. Other novel biomarkers such as neutrophil gelatinase-associated lipocalin (NGAL) and liver-type fatty acid-binding protein (L-FABP) are approved in Europe and Japan, respectively.

We discuss here biomarkers that are being studied for the diagnosis of ATN. The pathophysiology, etiology, clinical presentation, and evaluation and diagnosis of prerenal disease and ATN are discussed elsewhere. (See "Etiology and diagnosis of prerenal disease and acute tubular necrosis in acute kidney injury in adults".)

The diagnostic approach to patients with acute or chronic kidney disease (CKD), the possible prevention and management of ATN, and kidney and patient outcomes after ATN are also discussed elsewhere:

●(See "Diagnostic approach to adult patients with subacute kidney injury in an outpatient setting".)

●(See "Possible prevention and therapy of ischemic acute tubular necrosis".)

●(See "Kidney and patient outcomes after acute kidney injury in adults".)

OVERVIEW — The loss of kidney function in AKI is most easily detected by measurement of the serum creatinine, which is used to estimate the glomerular filtration rate (GFR). Although the serum creatinine is widely used in diagnosing the presence of AKI, it is a suboptimal biomarker. It is a lagging marker of change in kidney function; therefore, it has poor sensitivity for the early diagnosis of AKI, and, as a marker of glomerular filtration, it is unable to differentiate among the various causes of AKI [8]. As an example, the rise in serum creatinine is slow following the onset of AKI. By the time a change is observed in the serum creatinine, a critical therapeutic window may have been missed, particularly among those with ATN. A number of factors may contribute to a lag in rise of serum creatinine after AKI, including dilutional effect of fluid administration [9] and decrease in creatinine generation [10]. (See "Assessment of kidney function".)

Thus, different urinary and serum proteins have been intensively investigated as possible biomarkers for the early diagnosis of ATN. There are promising candidate biomarkers that report on kidney and tubule function, detect an early and graded increase in tubular epithelial cell injury, and distinguish prerenal disease from ATN [8,11-13]. These novel biomarkers have the potential to reflect physiologic and pathophysiologic processes of the injured kidney. Some biomarkers are detected in the urine of patients without a diagnostic increase in serum creatinine, which defines a group of patients with "subclinical AKI" who are at risk for adverse outcomes [14]. Biomarkers are used in clinical investigation to facilitate early randomization to different treatment arms [15]. These vanguard studies using biomarkers may lead to the identification of new therapies and the practical use of biomarkers in routine patient care [8,11,16-18]. Future studies are also needed to investigate whether biomarker profiles reflect unique injuries and identify the site of injury. For instance, sepsis-associated AKI may have a biomarker profile that is distinctly different from that of nephrotoxin-associated AKI. The National Institutes of Health (NIH)/National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) initiated the Kidney Precision Medicine Project (KPMP). The objectives of KPMP are to ethically obtain human kidney biopsy samples from participants with AKI and chronic kidney disease (CKD), with a goal of identifying critical cells, pathways, and targets for therapy [19]. Tissue-driven definitions of AKI will permit identification of AKI subphenotypes that may be responsive to specific, targeted therapies.

Many candidate biomarkers have been identified. However, the following steps are necessary before they are used clinically:

●Validation in different settings of AKI (cardiac surgery, sepsis, contrast-induced nephropathy, emergency and pediatric medicine) and in different clinical centers.

●Development and testing of rapid assays.

●Development of a panel of tubular biomarkers that can be used in combination with clinical (eg, fluid overload) and/or functional biomarkers (eg, estimated GFR [eGFR] kinetics). It is unlikely that a single biomarker will suffice; rather, a panel of biomarkers will be necessary [18].

Additional barriers for implementing novel biomarkers exist [20]. For example, the optimal method of reporting biomarker excretion has not been determined. Urinary biomarker excretion may be reported as an absolute concentration or normalized to creatinine excretion. One study that compared these methods showed that the biomarker normalized to creatinine best predicted death, dialysis, or the development of AKI, although the absolute concentration best diagnosed AKI on admission [21].

DIAGNOSTIC BIOMARKERS

Urinary tubular enzymes — Urinary tubular enzymes consist of proximal renal tubular epithelial antigen (HRTE-1), alpha-glutathione S-transferase (alpha-GST), pi-glutathione S-transferase (pi-GST), gamma-glutamyltranspeptidase (gamma-GT), alanine aminopeptidase (AAP), lactate dehydrogenase (LDH), N-acetyl-beta-glucosaminidase (NAG), and alkaline phosphatase (ALP). Most of these are released from proximal tubular epithelial cells within 12 hours and four days earlier than a detectable rise in serum creatinine [11,22]. No validated cut-off points currently exist to help distinguish prerenal disease from ATN.

Urinary low-molecular-weight proteins — Alpha1-microglobulin (alpha1-m), beta2-microglobulin (beta2-m), retinol-binding protein (RBP), adenosine deaminase-binding protein (ABP), and urinary cystatin C are urinary low-molecular-weight proteins. They are produced at different sites, filtered at the glomerulus, and reabsorbed at the proximal tubule with no secretion. Although promising prognostically and to help distinguish prerenal disease from ATN, increased levels may be observed after reversible and mild dysfunction and may not necessarily be associated with persistent or irreversible damage [11,22,23].

Neutrophil gelatinase-associated lipocalin (NGAL) — NGAL is markedly upregulated and abundantly expressed in the kidney after kidney ischemia [24]. In this setting, NGAL may function to dampen toxicity by reducing apoptosis and increasing the normal proliferation of kidney tubule cells. In addition, by enhancing the delivery of iron, NGAL upregulates heme oxygenase-1, thereby helping protect kidney tubule cells [24-28]. NGAL is now approved for use as a biomarker of AKI in some countries. (See "Pathogenesis and etiology of ischemic acute tubular necrosis".)

NGAL shows promise as a biomarker to help diagnose early ATN [14,29-38]. Furthermore, NGAL can differentiate prerenal disease from ATN. Using an NGAL reporter mouse, the NGAL-Luc2-mC reporter responds to endogenous signals that illuminate sites of injury (NGAL expression) in vivo and in real time. They found that, following ischemia reperfusion, as evidenced by a rise in creatinine, the kidneys illuminate, indicating NGAL production at the site of injury. Interestingly, following maneuvers that lead to significant prerenal disease associated with hypernatremia, there was no NGAL illumination, indicating that prerenal disease does not induce NGAL expression. Thus, NGAL may potentially be useful in differentiating prerenal disease from ATN [39].

NGAL has been tested in multiple studies of patients at risk for AKI due to sepsis, cardiac surgery, exposure to contrast media, or after kidney transplantation. In these studies, the average sensitivity and specificity of NGAL measured one to three days prior to AKI diagnosis was 76 and 77 percent, respectively, for cardiac surgery patients and 73 and 80 percent, respectively, for patients admitted to the intensive care unit (ICU) [38]. In a meta-analysis that examined the performance of a variety of novel AKI biomarkers in 110 studies with over 38,000 patients, biomarkers utilizing urinary and/or serum NGAL were the most accurate at predicting AKI [40].

Examples of specific studies are as follows:

●Urinary NGAL levels were determined in 196 children following cardiopulmonary bypass [32]. Ninety-nine patients sustained AKI, defined as a >50 percent increase in serum creatinine. Urinary NGAL levels obtained at two hours following procedure correlated with severity and duration of AKI, length of stay, requirement for dialysis, and death.

●Urinary NGAL levels were determined in 635 consecutive patients assessed in the emergency department [33]. Thirty patients had AKI, defined as the new onset of a 50 percent increase in the serum creatinine level (compared with historic baseline) or a 25 percent decrease in estimated glomerular filtration rate (eGFR). The mean urinary NGAL level was significantly elevated in patients with AKI compared with those with normal kidney function, chronic kidney disease (CKD), or prerenal azotemia.

●Urinary NGAL among other biomarkers was measured in patients coming to the emergency department and at later time points in the ICU [41]. Urinary NGAL diagnosed AKI up to 48 hours. Testing later did not affect the AKI diagnostic performance.

●Urinary NGAL, monomeric NGAL (mNGAL), interleukin (IL)-18, and other conventional urinary biomarkers for AKI (albumin, beta-2 microglobulin, fractional excretion of sodium) were measured at diagnosis and on days 3, 7, and 14 in 320 patients with AKI and decompensated cirrhosis. Among all biomarkers, urinary NGAL measured at day 3 had the greatest accuracy for differentiating ATN from all other types of AKI [42].

Urinary kidney injury molecule-1 — Kidney injury molecule-1 (KIM-1), also known as T cell immunoglobulin mucin domain-1 (TIM-1) [43], is a type 1 transmembrane glycoprotein that is low in normal kidneys but in high levels in the proximal tubule cells of kidneys with ischemic or toxic injury [44,45]. KIM-1 is a receptor for phosphatidylserine, an "eat-me" signal, expressed on cell surfaces of apoptotic cells that identifies them for removal by efferocytosis [46].

The ectodomain of KIM-1 is a soluble fragment that can be measured in urine by immunoassay. KIM-1 has been tested in a number of cohorts serving as a sensitive and specific biomarker for AKI [35,37,45,47,48].

●In a prospective study that included 123 adults undergoing cardiac surgery, urinary KIM-1, NGAL, cystatin C, hepatocyte growth factor, pi-GST, alpha-GST, and fractional excretion of sodium and urea were measured preoperatively, postoperatively, and at the time of the clinical diagnosis of AKI [37]. At various postoperative time points, cystatin C, NGAL, KIM-1, alpha-GST, and pi-GST all demonstrated ability to diagnose stage 3 AKI. Preoperative KIM-1 and alpha-GST predicted the development of stage 1 and stage 3 AKI, possibly reflecting subclinical proximal tubular injury present at this time.

●In one study of 38 patients, KIM-1 differentiated ATN from other forms of AKI and CKD [45]. The normalized urinary KIM-1 levels were significantly higher in patients with ischemic ATN (2.92) compared with levels in patients with other forms of AKI (0.63) or CKD (0.72). After adjustment for age, sex, time between the initial insult, and sampling of the urine, a one-unit increase in normalized KIM-1 was associated with a greater than 12-fold (odds ratio [OR] 12.4, 95% CI 1.2-119) risk for the presence of ATN.

●KIM-1 was tested in a case-control study of 20 children who underwent cardiopulmonary bypass surgery both with and without complicating AKI [48]. Urinary KIM-1 was increased 6 to 12 hours following cardiopulmonary bypass and remained elevated up to 48 hours in patients who sustained a >50 percent increase in serum creatinine within the first 48 hours but not in children who had normal kidney function. The increase in KIM-1 was paralleled by that of NGAL.

Urinary interleukin-18 — Urinary interleukin (IL)-18 has been shown to be elevated in patients with ATN compared with patients with prerenal azotemia, urinary tract infection (UTI), or CKD [49-51]. Its predictive utility for AKI following cardiac surgery could not be demonstrated in one prospective, observational study of 100 adult patients [52]. In addition, in a large study population of 1439 critically ill patients, IL-18 had a poor-to-moderate ability to predict AKI, kidney replacement therapy, or 90-day mortality [53]. Further studies are needed to clarify the role of urinary IL-18 as a biomarker.

Urinary liver-type fatty acid-binding protein in acute kidney injury — Urinary excretion of urinary liver-type fatty acid-binding protein (L-FABP) reflects stress of proximal tubular epithelial cells and correlates with severity of ischemic tubular injury [54]. A meta-analysis of 15 prospective cohort studies demonstrated that, in hospital–based cohorts of patients at risk of AKI, L-FABP can discriminate for the diagnosis of AKI and predict the need for dialysis and in-hospital mortality [55]. Further validation is needed in large cohort studies.

Combination of biomarkers — Several biomarkers are now investigated in panels [35,56,57], some including clinically available markers such as a urinary microscopy score [58]. In a prospective study of 90 adults with cardiac surgery, KIM-1, NAG, and NGAL were evaluated for early detection of AKI in the postoperative phase. The combined analysis of these three markers enhanced the sensitivity of early detection of postoperative AKI [35].

●In another prospective, observational study, NGAL, cystatin C, and conventional markers like creatinine and urea were measured at different time points before and after cardiac surgery in 100 adult patients. NGAL and cystatin C (in patients with initial eGFR >60 mL/min/1.73 m²) were independent predictors of AKI [57]. Similar findings were reported in another study of pediatric patients [59].

●In a multicenter, prospective, cohort study, five biomarkers were measured in 1635 unselected emergency department patients at time of hospital admission. Urinary NGAL and urinary KIM-1 predicted a composite outcome of dialysis initiation or death during hospitalization. These biomarkers could detect a subpopulation with low serum creatinine on admissions that was at risk of adverse events [60].

●Comparing five different biomarkers in a prospective, observational study of 529 patients in two ICUs, performance of biomarkers could be improved by stratification for time of collection with respect to kidney insult and for baseline kidney function before injury [61].

●In a multicenter, prospective cohort study of patients with cirrhosis and AKI, multiple biomarkers were used to assess the differential diagnosis of AKI [62]. Patients with ATN had significantly higher values of NGAL, IL-18, KIM-1, L-FABP, and albumin.

While new biomarkers are being validated, some older functional biomarkers (serum creatinine, serum urea, and urine output) continue to be utilized in the diagnosis and staging of AKI. In the future, the use of existing or newer damage biomarkers (eg, KIM-1) will likely be combined with the use of new and old functional biomarkers to enhance the ability of RIFLE, Acute Kidney Injury Network (AKIN), or Kidney Disease: Improving Global Outcomes (KDIGO) criteria to define AKI and to delineate a more precise AKI phenotype [13,63]. (See "Definition and staging criteria of acute kidney injury in adults".)

These functional and damage biomarkers can also be tested along with other systemic and organ-specific biomarkers to assess the patient's AKI phenotype, which may then guide tailored therapeutic intervention.

Some investigators have studied the combination of biomarker data with clinical data to develop prediction models for AKI and other adverse outcomes. A three-variable model including age, cirrhosis, and soluble tumor necrosis factor (TNF) receptor-1 concentrations was tested in internal and external validation cohorts and had a high negative predictive value (0.94 to 0.95) for occurrence of severe AKI within 72 hours of ICU admission [64]. In another study, four distinct phenotypes of AKI were identified utilizing clinical data and biomarkers including KIM-1, NGAL, B-type natriuretic peptide (BNP), troponin T levels, uromodulin, and vascular inflammation markers (KL-40 and MCP-1). These AKI phenotypes were associated with longitudinal outcomes such as risk of CKD, cardiovascular events, and death [65].

Other areas of research for biomarkers of AKI

●Proteomics – The role of protein profiling with laser mass spectrometry in urinary samples of patients with AKI is being investigated. This method allows identification of new and early markers of AKI [66]. In a derivation cohort of cardiac bypass surgery patients, a panel of 204 urinary peptides related to hemolysis, inflammation, immune cells, cell growth, and survival was identified as a potential predictor of AKI [67]. In two validation cohorts, this novel peptide signature outperformed several single biomarkers and clinical scores.

●Metabolomics – Metabolomics is the study of small-molecule metabolites that are produced by the body and provide insight into physiological and pathophysiological conditions. Metabolomic analysis can be readily performed in biofluids such as blood and urine, and, because there are fewer metabolites than there are genes, mRNA, and proteins, analyses are simpler. This method may allow for identification of new markers in AKI [68,69].

●Extracellular vesicles (EVs) – EVs are a heterogenous population of small submicron membrane fragments released from multivesicular bodies (exosomes, <100 nm) or shed from various cell types into different body fluids (microvesicles, 100 to 1000 nm). They carry markers of their parent cells that are utilized to identify their origin. In particular, urinary EVs contain proteins from various nephron segments providing nephron-specific information [70]. In addition, EVs represent an important mode of intercellular communication by serving as vehicles for transfer between cells of membrane and cytosolic proteins, lipids, and genetic information. Therefore, these EVs have important implications with regard to biomarkers and mechanisms of disease [71].

●MicroRNAs – Serum microRNAs are being explored in patients with AKI [72]. MicroRNAs were profiled in 77 patients with AKI and 18 critically ill patients with acute myocardial infarction [73]. Circulating microRNAs were altered in patients with AKI. MicroRNA-210 predicted mortality in this patient cohort.

Other pilot studies are indicating other sets of microRNAs to be altered days before increase in serum creatinine [74,75].

●Markers of inflammation – Plasma IL-6 and IL-10 have been measured in adults undergoing cardiac surgery. IL-6 and IL-10 were elevated after surgery and associated with higher risk for AKI [76]. In a substudy of the Translational Research Investigating Biomarker Endpoints in AKI (TRIBE-AKI) cohort, plasma IL-6 and IL-10 were measured in 106 children undergoing cardiopulmonary bypass. Preoperative plasma IL-6 concentrations were associated with the development of stage 2 and 3 AKI [77].

●Others – Additional novel markers that have been evaluated include serum cystatin C levels, isoform 3 of the sodium-hydrogen exchanger (NHE3), perforin and granzyme B, CXCR3-binding chemokines, endothelin, ProANP (1Y 98), tryptophan glycoconjugate and cysteine-rich protein 61 (CYR61), fatty acid-binding protein, TNF receptor-I, plasminogen activator inhibitor-1, netrin-1 [78], activating transcription factor 3 (ATF3), and MCP-1 [8,11,79-82]. Prospective studies need to be done to evaluate the utility of these biomarkers.

PROGNOSTIC BIOMARKERS — Most of the biomarkers described above allow early detection of AKI but do not predict severe AKI.

Soluble urokinase plasminogen activator receptor (suPAR) — suPAR is the circulating form of the membrane-bound uPAR, a glycosyl-phosphatidylinositol-anchored protein normally expressed on endothelial cells, podocytes, and, with induced expression, on monocytes and lymphocytes [83,84]. It can be measured in plasma using commercially available assays. In the setting of inflammation, uPAR is shed by immune cells following proteolytic cleavage and travels as the soluble form, suPAR. suPAR is thought to represent a biomarker of immune activation and inflammation [85]. suPAR is thought to interact with cells in the kidney (eg, podocytes, glomerular cells, and tubular epithelial cells) to trigger mitochondrial superoxide generation and a heightened energy demand, ultimately leading to kidney injury [86-89].

As a biomarker, suPAR was predictive of progressive decline in kidney function in a number of studies that included healthy participants and patients with CKD of various etiologies [86,90-94]. Systemic levels of suPAR have also been shown to correlate with markers of organ dysfunction and may enable prognostication in critically ill patients [85].

suPAR was also predictive of AKI in multiple cohorts [87,95,96]. In one study, for example, high levels of suPAR were associated with risk of mild to moderate AKI within the first seven days after cardiac surgery, coronary angiography, or ICU admission [87]. However, the relationship of suPAR with AKI is confounded by its association with conditions such as infection, sepsis, cardiovascular disease, diabetes, cancer, and liver failure, which themselves are risk factors for AKI [85,88].

Dickkopf-3 — The preoperative urinary concentration of dickkopf-3 (DKK3), a urinary cytokine and tubular stress biomarker, has been used to identify surgical patients at high risk for AKI [97]. In over 700 patients who were scheduled to undergo cardiac surgery, elevated urinary DKK3 to creatinine ratios were associated with an increased risk of postoperative AKI, independent of baseline kidney function. Furthermore, compared with clinical and other laboratory data, urinary DKK3 to creatinine ratios improved AKI prediction.

Uromodulin (Tamm-Horsfall protein) — Uromodulin is a 95-kD glycoprotein produced by the thick ascending limb and the distal convoluted tubule [98] and is thought to bind pathogenic bacteria and prevent stone formation [99,100]. Mice deficient of uromodulin are more susceptible to ischemia-reperfusion kidney injury [101]. In a post-hoc analysis of a prospective cohort study of 218 adult patients undergoing cardiac surgery, lower urinary uromodulin/creatinine ratio was associated with higher odds for AKI [102]. In children undergoing cardiopulmonary bypass surgery, preoperative urinary uromodulin also has been assessed as a predictive biomarker [103]. Children in the lowest quartile of urinary uromodulin levels had a markedly increased risk of postoperative AKI compared with those in the highest quartile.

Plasma NGAL — In a prospective, cohort study in which 616 patients admitted from the emergency department were classified by clinical criteria as AKI, transient azotemia, stable chronic kidney disease (CKD), or normal kidney function, plasma NGAL discriminated AKI from normal function and from transient azotemia [104]. Higher concentrations of NGAL were associated with more severe AKI.

Urinary insulin-like growth factor-binding protein 7 (IGFBP7) and tissue inhibitor of metalloproteinases-2 (TIMP2) — Sepsis- and ischemia-induced cell injury and repair are associated with cell cycle regulation [105,106]. IGFBP7 and TIMP-2, two urinary biomarkers identified in a discovery study, are expressed in epithelial cells and act in an autocrine and paracrine manner to arrest cell cycle in AKI.

In the Sapphire validation study of over 700 critically ill patients, the primary endpoint was moderate to severe AKI (Kidney Disease: Improving Global Outcomes [KDIGO] stage 2 to 3) within 12 hours of sample collection. These markers performed well in patients with sepsis (with area under the receiver operating characteristics curve [AUC] of 0.82) and postsurgery (AUC 0.85) in comparison with traditional biomarkers and improved risk stratification for AKI well ahead of clinical manifestations (azotemia and oliguria) [107]. An analysis of the Sapphire validation study showed that the use of TIMP-2/IGFBP7 identified patients at risk for death or dialysis during the ensuing nine months after study enrollment.

A point-of-care device measuring those two biomarkers was approved by the US Food and Drug Administration (FDA) in 2014, though defining the appropriate timing and frequency of biomarker measurement and interpreting these results in individual patients remains the focus of future studies [108]. The positive predictive value to diagnose CKD stages 2 and 3 (ie, estimated glomerular filtration rate [eGFR] 30 to 89 mL/min/1.73 m²) was 49 percent, and the negative predictive value was 97 percent [109,110]. The clinical value of this point-of-care device is being assessed in different clinical settings of AKI [111]. (See "Definition and staging of chronic kidney disease in adults", section on 'GFR'.)

Two randomized trials have utilized postoperative urinary TIMP-2/IGFBP7 values to identify patients at high risk of AKI after cardiac surgery [112,113].

Furosemide stress test — The furosemide stress test (FST) was developed as a clinical tool to assess the risk for AKI progression [114]. FST entails administration of intravenous furosemide (1 mg/kg in furosemide-naïve and 1.5 mg/kg in non-naïve) to euvolemic patients with stage 1 or 2 AKI. A urine output of >200 mL over two hours after administration of furosemide indicates FST responsiveness. In observational studies, FST-unresponsiveness was predictive of progression to stage 3 AKI, need for kidney replacement therapy, and higher inpatient mortality [114-117].

SUMMARY

●Investigational biomarkers of acute kidney injury – Acute kidney injury (AKI) is a common clinical problem. Although measurement of the serum creatinine concentration is widely used for the detection of AKI, it does not permit early diagnosis of acute tubular necrosis (ATN), since tubular injury precedes a significant rise in serum creatinine. Investigational biomarkers have been evaluated in patients with possible ATN in an attempt to detect tubular injury at an earlier stage. (See 'Introduction' above.)

●Steps necessary for clinical use – Various urinary and serum proteins have been intensively investigated as possible biomarkers for the early diagnosis of ATN. Before any such proteins are used clinically, validation in different settings of AKI and the development and testing of rapid assays are necessary. In addition, it needs to be shown if there is an association between levels of biomarkers and outcome. (See 'Overview' above.)

●Diagnostic biomarkers – Promising biomarkers for the diagnosis of AKI include neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule-1 (KIM-1), urinary interleukin (IL)-18, and liver-type fatty acid-binding protein (L-FABP), among others. (See 'Diagnostic biomarkers' above.)

●Prognostic biomarkers – The development of biomarkers that predict the occurrence and/or severity of AKI would allow for personalized interventions to prevent or minimize AKI. Assays proposed to identify patients at high risk of AKI, such as the product of urinary insulin-like growth factor-binding protein 7 (IGFBP7) and tissue inhibitor of metalloproteinases-2 (TIMP-2), are the focus of ongoing studies. (See 'Prognostic biomarkers' above.)