A Urine-based Genomic Assay Improves Risk Stratification for Patients with High-risk Hematuria Stratified According to the American Urological Association Guidelines.

BACKGROUND: According to the recent American Urological Association (AUA) guideline on hematuria, patients are stratified into groups with low, intermediate, and high risk of urothelial carcinoma (UC). These risk groups are based on clinical factors and do not incorporate urine-based tumor markers. OBJECTIVE: To evaluate whether a urine-based genomic assay improves the redefined AUA risk stratification for hematuria. DESIGN, SETTING, AND PARTICIPANTS: We selected patients with complete biomarker status, as assessed on urinary DNA, from a previously collected prospective Dutch hematuria cohort (n = 838). Patients were stratified into the AUA risk categories on the basis of sex, age, and type of hematuria. Biomarker status included mutation status for the FGFR3, TERT, and HRAS genes, and methylation status for the OTX1, ONECUT2, and TWIST1 genes. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: The primary endpoint was the diagnostic model performance for different hematuria risk groups. Further analyses assessed the pretest and post-test UC probability in the hematuria subgroups using a Fagan nomogram. RESULTS AND LIMITATIONS: Overall, 65 patients (7.8%) were classified as low risk, 106 (12.6%) as intermediate risk, and 667 (79.6%) as high risk. The UC incidence differed significantly between the gross hematuria (21%, 98/457) and microscopic hematuria (4%, 14/381) groups (p < 0.001). All cancer cases were in the high-risk group, which had UC incidence of 16.8% (112/667). Application of the diagnostic model revealed robust performance among all risk groups (area under the receiver operating characteristic curve 0.929-0.971). Depending on the risk group evaluated, a negative urine assay was associated with post-test UC probability of 0.3-2%, whereas a positive urine assay was associated with post-test UC probability of 31-42%. CONCLUSIONS: This study shows the value that a urine-based genomic assay adds to the AUA guideline stratification for patients with hematuria. It seems justified to safely withhold cystoscopy for patients with AUA low risk who have a negative urine assay. In addition, evaluation should be expedited for patients with AUA intermediate or high risk and a positive urine assay. PATIENT SUMMARY: Patients who have blood in their urine (hematuria) can be classified as having low, intermediate, or high risk of having cancer in their urinary tract. We found that use of a urine-based genetic test improves the accuracy of predicting which patients are most likely to have cancer. Patients with a negative test may be able to avoid invasive tests, while further tests could be prioritized for patients with a positive test.

Assessment of Predictive Genomic Biomarkers for Response to Cisplatin-based Neoadjuvant Chemotherapy in Bladder Cancer.

Cisplatin-based neoadjuvant chemotherapy (NAC) followed by radical cystectomy is recommended for patients with muscle-invasive bladder cancer (MIBC). It has been shown that somatic deleterious mutations in ERCC2, gain-of-function mutations in ERBB2, and alterations in ATM, RB1, and FANCC are correlated with pathological response to NAC in MIBC. The objective of this study was to validate these genomic biomarkers in pretreatment transurethral resection material from an independent retrospective cohort of 165 patients with MIBC who subsequently underwent NAC and radical surgery. Patients with ypT0/Tis/Ta/T1N0 disease after surgery were defined as responders. Somatic deleterious mutations in ERCC2 were found in nine of 68 (13%) evaluable responders and two of 95 (2%) evaluable nonresponders (p = 0.009; FDR = 0.03). No correlation was observed between response and alterations in ERBB2 or in ATM, RB1, or FANCC alone or in combination. In an exploratory analysis, no additional genomic alterations discriminated between responders and nonresponders to NAC. No further associations were identified between the aforementioned biomarkers and pathological complete response (ypT0N0) after surgery. In conclusion, we observed a positive association between deleterious mutations in ERCC2 and pathological response to NAC, but not overall survival or recurrence-free survival. Other previously reported genomic biomarkers were not validated. PATIENT SUMMARY: It is currently unknown which patients will respond to chemotherapy before definitive surgery for bladder cancer. Previous studies described several gene mutations in bladder cancer that correlated with chemotherapy response. This study confirmed that patients with bladder cancer with a mutation in the ERCC2 gene often respond to chemotherapy.

An integrated multi-omics analysis identifies prognostic molecular subtypes of non-muscle-invasive bladder cancer.

The molecular landscape in non-muscle-invasive bladder cancer (NMIBC) is characterized by large biological heterogeneity with variable clinical outcomes. Here, we perform an integrative multi-omics analysis of patients diagnosed with NMIBC (n = 834). Transcriptomic analysis identifies four classes (1, 2a, 2b and 3) reflecting tumor biology and disease aggressiveness. Both transcriptome-based subtyping and the level of chromosomal instability provide independent prognostic value beyond established prognostic clinicopathological parameters. High chromosomal instability, p53-pathway disruption and APOBEC-related mutations are significantly associated with transcriptomic class 2a and poor outcome. RNA-derived immune cell infiltration is associated with chromosomally unstable tumors and enriched in class 2b. Spatial proteomics analysis confirms the higher infiltration of class 2b tumors and demonstrates an association between higher immune cell infiltration and lower recurrence rates. Finally, the independent prognostic value of the transcriptomic classes is documented in 1228 validation samples using a single sample classification tool. The classifier provides a framework for biomarker discovery and for optimizing treatment and surveillance in next-generation clinical trials.

Evaluation of the New American Urological Association Guidelines Risk Classification for Hematuria.

PURPOSE: Microhematuria is a prevalent condition and the American Urological Association has developed a new risk-stratified approach for the evaluation of patients with microhematuria. Our objective was to provide the first evaluation of this important guideline. MATERIALS AND METHODS: This multinational cohort study combines contemporary patients from 5 clinical trials and 2 prospective registries who underwent urological evaluation for hematuria. Patients were stratified into American Urological Association risk strata (low, intermediate or high risk) based on sex, age, degree of hematuria, and smoking history. The primary end point was the incidence of bladder cancer within each risk stratum. RESULTS: A total of 15,779 patients were included in the analysis. Overall, 727 patients (4.6%) were classified as low risk, 1,863 patients (11.8%) were classified as intermediate risk, and 13,189 patients (83.6%) were classified as high risk. The predominance of high risk patients was consistent across all cohorts. A total of 857 bladder cancers were diagnosed with a bladder cancer incidence of 5.4%. Bladder cancer was more prevalent in men, smokers, older patients and patients with gross hematuria. The cancer incidence for low, intermediate and high risk groups was 0.4% (3 patients), 1.0% (18 patients) and 6.3% (836 patients), respectively. CONCLUSIONS: The new risk stratification system separates hematuria patients into clinically meaningful categories with differing likelihoods of bladder cancer that would justify evaluating the low, intermediate and high risk groups with incremental intensity. Furthermore, it provides the relative incidence of bladder cancer in each risk group which should facilitate patient counseling regarding the risks and benefits of evaluation for bladder cancer.

A Urine Based Genomic Assay to Triage Patients with Hematuria for Cystoscopy.

PURPOSE: Current clinical guidelines recommend cystoscopy in patients who present with hematuria to rule out a bladder tumor. We evaluated whether our previously developed urine assay was able to triage patients with hematuria for cystoscopy in a large prospective cohort. MATERIALS AND METHODS: A urine sample was collected before cystoscopy and mutation/methylation status of 6 genes was determined on cellular DNA. The existing diagnostic model was validated on this cohort. Logistic regression was applied to investigate other potential variables. The primary end point was the model performance as indicated by the AUC. Secondary end points were sensitivity, specificity and negative predictive value. Clinical usefulness was determined by the net benefit approach. RESULTS: In 838 patients biomarker status could be determined for all genes. Urothelial cancer was observed in 112 patients (98 of 457 in the gross and 14 of 381 in the microscopic hematuria group). Validation of the existing model resulted in an AUC of 0.93. Logistic regression analysis identified type of hematuria as a significant additional variable. Adding type of hematuria resulted in an AUC of 0.95 (96% sensitivity, 73% specificity, 99% negative predictive value). The assay identified all upper tract tumors not visible by cystoscopy (in 6). Net benefit analysis showed that the urine assay should be preferred over current clinical practice. Implementing the urine assay as a triage tool could lead to a 53% reduction in cystoscopies. CONCLUSIONS: The urine assay detected urothelial cancer with a very high accuracy and can be used to triage patients presenting with hematuria for cystoscopy.

Long non-coding RNAs identify a subset of luminal muscle-invasive bladder cancer patients with favorable prognosis.

BACKGROUND: Muscle-invasive bladder cancer (MIBC) is a heterogeneous disease, and gene expression profiling has identified several molecular subtypes with distinct biological and clinicopathological characteristics. While MIBC subtyping has primarily been based on messenger RNA (mRNA), long non-coding RNAs (lncRNAs) may provide additional resolution. METHODS: LncRNA expression was quantified from microarray data of a MIBC cohort treated with neoadjuvant chemotherapy (NAC) and radical cystectomy (RC) (n = 223). Unsupervised consensus clustering of highly variant lncRNAs identified a four-cluster solution, which was characterized using a panel of MIBC biomarkers, regulon activity profiles, gene signatures, and survival analysis. The four-cluster solution was confirmed in The Cancer Genome Atlas (TCGA) cohort (n = 405). A single-sample genomic classifier (GC) was trained using ridge-penalized logistic regression and validated in two independent cohorts (n = 255 and n = 94). RESULTS: NAC and TCGA cohorts both contained an lncRNA cluster (LC3) with favorable prognosis that was enriched with tumors of the luminal-papillary (LP) subtype. In both cohorts, patients with LP tumors in LC3 (LPL-C3) were younger and had organ-confined, node-negative disease. The LPL-C3 tumors had enhanced FGFR3, SHH, and wild-type p53 pathway activity. In the TCGA cohort, LPL-C3 tumors were enriched for FGFR3 mutations and depleted for TP53 and RB1 mutations. A GC trained to identify these LPL-C3 patients showed robust performance in two validation cohorts. CONCLUSIONS: Using lncRNA expression profiles, we identified a biologically distinct subgroup of luminal-papillary MIBC with a favorable prognosis. These data suggest that lncRNAs provide additional information for higher-resolution subtyping, potentially improving precision patient management.

Urinary peptide panel for prognostic assessment of bladder cancer relapse.

Non-invasive tools stratifying bladder cancer (BC) patients according to the risk of relapse are urgently needed to guide clinical intervention. As a follow-up to the previously published study on CE-MS-based urinary biomarkers for BC detection and recurrence monitoring, we expanded the investigation towards BC patients with longitudinal data. Profiling datasets of BC patients with follow-up information regarding the relapse status were investigated. The peptidomics dataset (n = 98) was split into training and test set. Cox regression was utilized for feature selection in the training set. Investigation of the entire training set at the single peptide level revealed 36 peptides being strong independent prognostic markers of disease relapse. Those features were further integrated into a Random Forest-based model evaluating the risk of relapse for BC patients. Performance of the model was assessed in the test cohort, showing high significance in BC relapse prognosis [HR = 5.76, p-value = 0.0001, c-index = 0.64]. Urinary peptide profiles integrated into a prognostic model allow for quantitative risk assessment of BC relapse highlighting the need for its incorporation in prospective studies to establish its value in the clinical management of BC.

Recommendations for follow-up of muscle-invasive bladder cancer patients: A consensus by the international bladder cancer network.

RATIONALE: Several guidelines exist that address treatment of patients with nonmetastatic muscle-invasive bladder cancer (MIBC). However, most only briefly mention follow-up strategies for patients and hence the treating physician is often left to infer on what the preferred follow-up schema would be for an individual patient. Herein, we aim to synthesize recommendations for follow-up of patients with MIBC for easy reference. METHODS: A multidisciplinary MIBC expert panel from the International Bladder Cancer Network was assembled to critically assess currently available major guidelines on surveillance of MIBC patients. Recommendations for follow-up were extracted and critically evaluated. Important considerations for guideline assessment included both aspects of oncological and functional follow-up-frequency of visits, the use of different imaging modalities, the role of cytology and molecular markers, and the duration of follow-up. OUTCOME: An International Bladder Cancer Network expert consensus recommendation was constructed for the follow-up of patients with MIBC based on the currently available evidence-based data.

Molecular Markers Increase Precision of the European Association of Urology Non-Muscle-Invasive Bladder Cancer Progression Risk Groups.

Purpose: The European Association of Urology (EAU) guidelines for non-muscle-invasive bladder cancer (NMIBC) recommend risk stratification based on clinicopathologic parameters. Our aim was to investigate the added value of biomarkers to improve risk stratification of NMIBC.Experimental Design: We prospectively included 1,239 patients in follow-up for NMIBC in six European countries. Fresh-frozen tumor samples were analyzed for GATA2, TBX2, TBX3, and ZIC4 methylation and FGFR3, TERT, PIK3CA, and RAS mutation status. Cox regression analyses identified markers that were significantly associated with progression to muscle-invasive disease. The progression incidence rate (PIR = rate of progression per 100 patient-years) was calculated for subgroups.Results: In our cohort, 276 patients had a low, 273 an intermediate, and 555 a high risk of tumor progression based on the EAU NMIBC guideline. Fifty-seven patients (4.6%) progressed to muscle-invasive disease. The limited number of progressors in this large cohort compared with older studies is likely due to improved treatment in the past two decades. Overall, wild-type FGFR3 and methylation of GATA2 and TBX3 were significantly associated with progression (HR = 0.34, 2.53, and 2.64, respectively). The PIR for EAU high-risk patients was 4.25. On the basis of FGFR3 mutation status and methylation of GATA2, this cohort could be reclassified into a good class (PIR = 0.86, 26.2% of patients), a moderate class (PIR = 4.32, 49.7%), and a poor class (PIR = 7.66, 24.0%).Conclusions: We conclude that the addition of selected biomarkers to the EAU risk stratification increases its accuracy and identifies a subset of NMIBC patients with a very high risk of progression. Clin Cancer Res; 24(7); 1586-93. ©2018 AACR.

A reported 20-gene expression signature to predict lymph node-positive disease at radical cystectomy for muscle-invasive bladder cancer is clinically not applicable.

BACKGROUND: Neoadjuvant chemotherapy (NAC) for muscle-invasive bladder cancer (MIBC) provides a small but significant survival benefit. Nevertheless, controversies on applying NAC remain because the limited benefit must be weight against chemotherapy-related toxicity and the delay of definitive local treatment. Therefore, there is a clear clinical need for tools to guide treatment decisions on NAC in MIBC. Here, we aimed to validate a previously reported 20-gene expression signature that predicted lymph node-positive disease at radical cystectomy in clinically node-negative MIBC patients, which would be a justification for upfront chemotherapy. METHODS: We studied diagnostic transurethral resection of bladder tumors (dTURBT) of 150 MIBC patients (urothelial carcinoma) who were subsequently treated by radical cystectomy and pelvic lymph node dissection. RNA was isolated and the expression level of the 20 genes was determined on a qRT-PCR platform. Normalized Ct values were used to calculate a risk score to predict the presence of node-positive disease. The Cancer Genome Atlas (TCGA) RNA expression data was analyzed to subsequently validate the results. RESULTS: In a univariate regression analysis, none of the 20 genes significantly correlated with node-positive disease. The area under the curve of the risk score calculated by the 20-gene expression signature was 0.54 (95% Confidence Interval: 0.44-0.65) versus 0.67 for the model published by Smith et al. Node-negative patients had a significantly lower tumor grade at TURBT (p = 0.03), a lower pT stage (p<0.01) and less frequent lymphovascular invasion (13% versus 38%, p<0.01) at radical cystectomy than node-positive patients. In addition, in the TCGA data, none of the 20 genes was differentially expressed in node-negative versus node-positive patients. CONCLUSIONS: We conclude that a 20-gene expression signature developed for nodal staging of MIBC at radical cystectomy could not be validated on a qRT-PCR platform in a large cohort of dTURBT specimens.

Validation of a DNA Methylation-Mutation Urine Assay to Select Patients with Hematuria for Cystoscopy.

PURPOSE: Only 3% to 28% of patients referred to the urology clinic for hematuria are diagnosed with bladder cancer. Cystoscopy leads to high diagnostic costs and a high patient burden. Therefore, to improve the selection of patients for cystoscopy and reduce costs and over testing we aimed to validate a recently developed diagnostic urine assay. MATERIALS AND METHODS: Included in study were 200 patients from a total of 3 European countries who underwent cystoscopy for hematuria, including 97 with bladder cancer and 103 with nonmalignant findings. Voided urine samples were collected prior to cystoscopy. DNA was extracted and analyzed for mutations in FGFR3, TERT and HRAS, and methylation of OTX1, ONECUT2 and TWIST1. Logistic regression was used to analyze the association between predictor variables and bladder cancer. RESULTS: Combining the methylation and mutation markers with age led to an AUC of 0.96 (95% CI 0.92-0.99) with 93% sensitivity and 86% specificity, and an optimism corrected AUC of 0.95. The AUC was higher for T1 or greater tumors compared to Ta tumors (0.99 vs 0.93). The AUC was also higher for high grade tumors compared to low grade tumors (1.00 vs 0.93). Overall negative predictive value was 99% based on the 5% to 10% prevalence of bladder cancer in patients with hematuria. This would lead to a 77% reduction in diagnostic cystoscopy. CONCLUSIONS: Analyzing hematuria patients for the risk of bladder cancer using novel molecular markers may lead to a reduction in diagnostic cystoscopy. Combining methylation analysis (OTX1, ONECUT2 and TWIST1) with mutation analysis (FGFR3, TERT and HRAS) and patient age resulted in a validated accurate prediction model.

Elevated Derived Neutrophil-to-Lymphocyte Ratio Corresponds With Poor Outcome in Patients Undergoing Pre-Operative Chemotherapy in Muscle-Invasive Bladder Cancer.

Background: Platinum-based pre-operative chemotherapy (POC) for muscle-invasive bladder cancer (MIBC) increases the complete pathological response rate at cystectomy and improves overall survival. However, 60% of MIBC patients still has muscle-invasive disease at cystectomy despite POC. Therefore, accurate prediction of response to POC is an important clinical need. We hypothesized that an elevated neutrophil-to-lymphocyte ratio (NLR) corresponds with adverse outcome in patients undergoing POC and radical cystectomy. Objective: To explore the correlation between the NLR and outcome in MIBC patients treated by POC and radical cystectomy. Methods: In 123 MIBC patients (urothelial carcinoma) who were treated by platinum-based POC and radical cystectomy, the derived NLR (dNLR) was retrospectively calculated by dividing the neutrophil count by the difference between leukocytes and neutrophil counts, prior to the start of chemotherapy. The correlation of the dNLR with pathological response at cystectomy and survival was analyzed by logistic regression analysis or the Kaplan-Meier method. Results: The complete pathological response (ypT0N0Mx) rate was 28.5%, 8.9% obtained a partial response (ypTa/T1/TisN0Mx), and 62.6% were non-responders (stage ≥ ypT2 and/or N+). An elevated dNLR (>2.21) correlated with non-response to POC (OR 2.70, 95% confidence interval: 1.15-6.38, p = 0.02) but this effect was nullified when corrected for clinically node-positive disease and clinical T stage. Patients with an elevated dNLR had shorter progression-free and overall survival albeit non-significant (p = 0.42, and p = 0.45, respectively). Conclusions: An elevated dNLR corresponded with poor outcome in terms of survival and non-response to POC in MIBC patients undergoing radical surgery. However, after correction for well-known prognostic factors, such as positive lymph node status at diagnostic imaging and clinical T stage, the correlation for the dNLR was nullified. Therefore, we conclude that the dNLR is insufficient to predict response to POC in this heterogeneous patient population.

Comprehensive Transcriptional Analysis of Early-Stage Urothelial Carcinoma.

Non-muscle-invasive bladder cancer (NMIBC) is a heterogeneous disease with widely different outcomes. We performed a comprehensive transcriptional analysis of 460 early-stage urothelial carcinomas and showed that NMIBC can be subgrouped into three major classes with basal- and luminal-like characteristics and different clinical outcomes. Large differences in biological processes such as the cell cycle, epithelial-mesenchymal transition, and differentiation were observed. Analysis of transcript variants revealed frequent mutations in genes encoding proteins involved in chromatin organization and cytoskeletal functions. Furthermore, mutations in well-known cancer driver genes (e.g., TP53 and ERBB2) were primarily found in high-risk tumors, together with APOBEC-related mutational signatures. The identification of subclasses in NMIBC may offer better prognostication and treatment selection based on subclass assignment.

Evaluation of an Epigenetic Profile for the Detection of Bladder Cancer in Patients with Hematuria.

PURPOSE: Many patients enter the care cycle with gross or microscopic hematuria and undergo cystoscopy to rule out bladder cancer. Sensitivity of this invasive examination is limited, leaving many patients at risk for undetected cancer. To improve current clinical practice more sensitive and noninvasive screening methods should be applied. MATERIALS AND METHODS: A total of 154 urine samples were collected from patients with hematuria, including 80 without and 74 with bladder cancer. DNA from cells in the urine was epigenetically profiled using 2 independent assays. Methylation specific polymerase chain reaction was performed on TWIST1. SNaPshot™ methylation analysis was done for different loci of OTX1 and ONECUT2. Additionally all samples were analyzed for mutation status of TERT (telomerase reverse transcriptase), PIK3CA, FGFR3 (fibroblast growth factor receptor 3), HRAS, KRAS and NRAS. RESULTS: The combination of TWIST1, ONECUT2 (2 loci) and OTX1 resulted in the best overall performing panel. Logistic regression analysis on these methylation markers, mutation status of FGFR3, TERT and HRAS, and patient age resulted in an accurate model with 97% sensitivity, 83% specificity and an AUC of 0.93 (95% CI 0.88-0.98). Internal validation led to an optimism corrected AUC of 0.92. With an estimated bladder cancer prevalence of 5% to 10% in a hematuria cohort the assay resulted in a 99.6% to 99.9% negative predictive value. CONCLUSIONS: Epigenetic profiling using TWIST1, ONECUT2 and OTX1 results in a high sensitivity and specificity. Accurate risk prediction might result in less extensive and invasive examination of patients at low risk, thereby reducing unnecessary patient burden and health care costs.

Targeted therapies in bladder cancer: an overview of in vivo research.

Survival of patients with muscle-invasive bladder cancer is poor and new therapies are needed. Currently, none of the targeted agents that are approved for cancer therapy have been approved for the treatment of bladder cancer and the few clinical trials that have been performed had limited success, often owing to a lack of efficacy and toxic effects. However, many other novel targeted agents have been investigated in animal models of bladder cancer. EGFR, FGFR-3, VEGF, mTOR, STAT3, the androgen receptor and CD24 are molecular targets that could be efficiently inhibited, resulting in reduced tumour growth, and that have been investigated in multiple independent studies. Several other targets, for example COX-2, IL-12, Bcl-xL, livin and choline kinase α, have also been observed to inhibit tumour growth, but these findings have not been replicated to date. Limitations of several studies include the use of cell lines with mutations downstream of the target, providing resistance to the tested therapy. Furthermore, certain technologies, such as interfering RNAs, although effective in vitro, are not yet ready for clinical applications. Further preclinical research is needed to discover and evaluate other possible targets, but several validated targets are now available to be studied in clinical trials.

FGFR3 mutation analysis in voided urine samples to decrease cystoscopies and cost in nonmuscle invasive bladder cancer surveillance: a comparison of 3 strategies.

PURPOSE: We determined whether FGFR3 mutation analysis of voided urine samples would be cost-effective to partly replace cystoscopy in the surveillance of patients treated for nonmuscle invasive urothelial carcinoma. MATERIALS AND METHODS: In this decision analytical study we analyzed data on 70 Dutch patients with FGFR3 positive primary tumors and a median followup of 8.8 years. Surveillance strategies were compared in a Markov model. Modified surveillance consisted of FGFR3 mutation analysis of voided urine samples every 3 months, and cystoscopy at 3, 12 and 24 months. Standard surveillance was defined as cystoscopy every 3 months and minimal surveillance was defined as cystoscopy at 3, 12 and 24 months. Analysis was stratified for 3 risk profiles, including surveillance after 1) the primary tumor, 2) the first to third recurrence and 3) the fourth recurrence or more. Sensitivity analysis was performed to evaluate the impact of variations in cost, sensitivity and specificity. RESULTS: The probability of no recurrence after 2 years of surveillance after a primary tumor was higher for modified surveillance than for standard and minimal surveillance, eg after primary tumors (95.7% vs 95.0% and 93.9%, respectively). The total cost of surveillance after the primary tumor was lower for minimal and modified surveillance (€2,254 and €2,558, respectively) than for standard surveillance (€5,861). Results were robust to changing inputs over plausible ranges and consistent for each of the 3 risk profiles. CONCLUSIONS: Surveillance in which cystoscopy is partly replaced by FGFR3 mutation analysis of urine seems a safe, effective and cost-effective surveillance strategy. Further validation in larger cohorts is required.