Prognostic role of carcinoembryonic antigen and carbohydrate antigen 19-9 in metastatic colorectal cancer: a BRAF-mutant subset with high CA 19-9 level and poor outcome.

BACKGROUND: Mutation status of RAS and BRAF, as well as serum levels of carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA 19-9), are biomarkers used in clinical management of patients with gastrointestinal cancers. This study aimed to examine the prognostic role of these biomarkers in a patient population that started first-line chemotherapy for unresectable metastatic colorectal cancer (mCRC) in the NORDIC-VII study. METHODS: CEA and CA 19-9 were measured in serum samples from 545 patients obtained before the start of chemotherapy. Four hundred and ninety-four patients had detectable levels of carbohydrate antigen 19-9 (CA 19-9). RAS (exons 2-4) and BRAF (V600E) mutation status were available from 440 patients. Overall survival (OS) was estimated in patient groups defined by serum CEA or CA 19-9 levels using cut-off values of 5 µg/L and 35 kU/L, respectively, in the total population and in subgroups according to RAS and BRAF mutation status. RESULTS: For both CEA and CA 19-9, elevated serum levels were associated with reduced OS in adjusted analyses which included RAS and BRAF mutation status, baseline World Health Organization performance status, and levels of alkaline phosphatase and C-reactive protein. The negative prognostic information provided by an elevated CA 19-9 level was particularly marked in patients with BRAF mutation (hazard ratio = 4.35, interaction P = 0.003, in an adjusted model for OS). CONCLUSIONS: High baseline serum concentrations of CEA and CA 19-9 provide independent information of impaired prognosis in mCRC. In patients with BRAF-mutant tumours, elevated serum CA 19-9 may identify a subgroup with highly aggressive disease and could contribute to improving therapeutic decisions.

Lowered reference limits for hCG improve follow-up of patients with hCG-producing tumors.

BACKGROUND: Human Chorionic Gonadotropin (hCG) is produced by germ cell tumors, but can also be elevated in benign conditions such as primary hypogonadism, where hCG is produced by the pituitary gland. In our experience, the reference limits for hCG (Elecsys hCG+ß-assay, Roche Diagnostics), were unnecessarily high and did not reflect levels encountered in clinical practice. We wanted to establish new reference limits to increase the clinical utility of the hCG-assay. METHODS: We analysed hCG in serum samples from a healthy adult population and in a cohort of testicular cancer survivors. The gonadotropins LH and FSH were measured in the cohort and in a selection of the reference population to assess gonadal function. RESULTS: We found low hCG levels for all men and women <45years (97.5 percentiles 0.1 and 0.2IU/L, respectively) from the healthy population (n=795) having normal FSH and LH. Due to assay limitations, we suggest a common reference limit of <0.3IU/L. For the age group ≥45, the 97.5 percentiles in the healthy population were 0.5IU/L for men and 6.0IU/L for women. In all subjects from both the reference population and the cohort (n=732), hCG levels exceeding the reference limit could be fully explained by reduced gonadal function indicated by elevated LH and FSH levels. CONCLUSION: The Elecsys hCG+ß-assay should have lower reference limits than recommended by the manufacturer, with important implications for tumor follow-up. Elevated hCG is rare with intact gonadal function, both in a normal population and among survivors of testicular cancer, and should lead to further investigations when encountered in clinical practice.

HE4 as an Early Detection Biomarker of Epithelial Ovarian Cancer: Investigations in Prediagnostic Specimens From the Janus Serumbank.

OBJECTIVES: Epithelial ovarian cancer is characterized by nonspecific signs and clinical symptoms arising at late stages. Early detection is therefore important and may significantly improve the survival rate. Cancer antigen 125 (CA125) has been the most extensively studied serum biomarker in epithelial ovarian cancer, but low specificity limits its usefulness. A relatively novel biomarker, human epididymis protein 4 (HE4), has shown promise in early detection of the disease. The aim of this study was to investigate how early the tumor marker increases before diagnosis. METHODS/MATERIALS: A nested case-control design was used to evaluate the performance of HE4 and CA125 in prediagnostic serum samples from the Janus Serumbank. Serial specimens from 120 women with invasive epithelial ovarian cancer were compared with healthy controls. Serum level of CA125, HE4, and cotinine was measured. Spearman correlation and multiple linear regression analyses were used to investigate impact of smoking, age, storage time, and lag time (time from sampling until date of diagnosis). RESULTS: Spearman correlation showed a strong positive correlation between HE4 and smoking in both cases and controls. Multiple linear regression analyses for pairwise differences between case and control showed that serum level of HE4 and CA125 was significantly increased (P = 0.002 and P < 0.001, respectively) 2 years before diagnosis and that CA125 also was significantly increased up to 4 years before diagnosis (P = 0.002). CONCLUSIONS: The present study showed that a difference between cases and controls in serum concentration of HE4 seemed to be increased 2 years before diagnosis and that CA125 was increased until 4 years before diagnosis.

Heterophilic antibody interference in immunometric assays.

Immunometric assays are inherently vulnerable to interference from heterophilic antibodies, endogenous antibodies that bind assay antibodies. The consequences of such interference can be devastating. In this review, we discuss strategies that reduce the damage caused by heterophilic antibodies. Clinicians should only order blood tests that are indicated for the patient and clinical setting at hand, and have the confidence to question laboratory results discordant with the clinical picture. Laboratorians should familiarize themselves with the vulnerability of the assays they offer, and be able to perform and interpret adequate confirmatory measures correctly. When designing immunoassays, the immunoassay industry should invest the necessary resources in specific protective measures against heterophilic antibody interference. Examples include using antibody fragments and the addition of effective blockers to assay reagents. The increasing use of modified monoclonal mouse antibodies both in therapy and diagnostics could present a particular challenge in the future.

Human epididymis protein 4 reference limits and natural variation in a Nordic reference population.

The objectives of this study are to establish reference limits for human epididymis protein 4, HE4, and investigate factors influencing HE4 levels in healthy subjects. HE4 was measured in 1,591 samples from the Nordic Reference Interval Project Bio-bank and Database biobank, using the manual HE4 EIA (Fujirebio) for 802 samples and the Architect HE4 (Abbott) for 792 samples. Reference limits were calculated using the statistical software R. The influence of donor characteristics such as age, sex, body mass index, smoking habits, and creatinine on HE4 levels was investigated using a multivariate model. The study showed that age is the main determinant of HE4 in healthy subjects, corresponding to 2% higher HE4 levels at 30 years (compared to 20 years), 9% at 40 years, 20% at 50 years, 37% at 60 years, 63% at 70 years, and 101% at 80 years. HE4 levels are 29% higher in smokers than in nonsmokers. In conclusion, HE4 levels in healthy subjects are associated with age and smoking status. Age-dependent reference limits are suggested.

Heterophilic antibody interference in commercial immunoassays; a screening study using paired native and pre-blocked sera.

BACKGROUND: Heterophilic antibodies are still an important source of interference in immunoassays. We have conducted a screening study for interference in a panel of commercially available assays using two sera known to contain high titer Fc-reactive heterophilic antibodies. METHODS: The sera were distributed to laboratories participating in the Nordic External Quality Assessment cooperation (EQANord). Duplicate samples pre-blocked with aggregated murine monoclonal MAK33 were also supplied. Discrepancies (>50%) between the results for native and blocked samples were used to classify the tested assays as susceptible to interference. A total of 170 different assay kits covering 91 analytes were tested. RESULTS: We found that 21 assays, covering 19 different analytes, were susceptible to interference from the heterophilic antibodies in the two sera. Many of these are clinically and commercially important assays. Some of the false results were grossly elevated and could have been detrimental to patient care in a clinical setting. CONCLUSIONS: Heterophilic antibodies with Fc-reactivity remain a threat. A more widespread use of antibody fragments and aggregated immunoglobulin could potentially improve the heterophilic antibody resistance of assays intended for clinical use.

Splenic marginal zone lymphoma with VH1-02 gene rearrangement expresses poly- and self-reactive antibodies with similar reactivity.

One-third of all splenic marginal zone lymphomas (SMZL) use the IgH VH1-02 gene. These cases are usually not associated with hepatitis C virus infection. Of interest, the rearranged VH1-02 genes display similar complementarity determining regions 3, a finding confirmed by our study. The latter suggests that these SMZL may produce antibodies with similar reactivity. We produced recombinant antibodies from 5 SMZL cases with VH1-02 gene rearrangement to study the binding reactivity of these antibodies. Surprisingly, the recombinant antibodies demonstrated poly- and self-reactivity as demonstrated by their reactivity with nuclear, cytoplasmic, as well as membranous antigens expressed by human cells and by reactivity with human serum. This polyreactivity was specific as demonstrated by ELISA. The antibodies did not react with proteins on the cell surface that are induced by apoptosis as shown for antibodies produced by chronic lymphatic leukemia with VH1-02 gene rearrangement. The results indicate that a common subset of SMZL arises from polyreactive B cells, a subset of marginal zone B cells that are important in the immunologic defense against infection.

Differences in total human chorionic gonadotropin immunoassay analytical specificity and ability to measure human chorionic gonadotropin in gestational trophoblastic disease and germ cell tumors.

OBJECTIVE: To determine the ability of several radioimmunoassays and commercial two-site immunoassays to detect the first World Health Organization International Reference Reagents (IRRs) for 6 defined human chorionic gonadotropin (hCG) variants and to compare their performance in measuring hCG in sera from patients with gestational trophoblastic disease (GTD) and germ cell tumors (GCTs) of the testis or ovary. STUDY DESIGN: The reactivity of the different assays with the 6 IRRs together with the current fourth International Standard (IS, 75/589) was tested using 5 commercial two-site assays as well as 2 competitive polyclonal radioimmunoassays (RIAs) and a competitive monoclonal immunoassay. Individual samples from 41 patients (19 GCT and 22 GTD) with high circulating levels of hCG (range, 718-6,055,000 IU/L) were diluted and measured using the various immunoassays. RESULTS: The results of 4 GCT patient samples varied markedly among the assays, including 1 sample that was grossly underestimated by 3 of the commercial assays. CONCLUSION: Comparison of each assay's reactivity to the variant isoforms revealed that recognition of the isoforms was highly variable, particularly for hCGbeta and hCGbeta core fragment (hCGbetacf).

Exploring the complementary selectivity of immunocapture and MS detection for the differentiation between hCG isoforms in clinically relevant samples.

Whereas numerous immunoassays have been developed to ensure detection of the entire spectrum of isoforms displayed by the human chorionic gonadotropin (hCG) molecule, significant variation has been demonstrated in how these isoforms are recognized by the antibodies in different immunoassays. The aim of this study was to establish a method using the dual selectivity of the immunoextraction and the mass spectrometry detection for the differentiation between various hCG isoforms in clinically relevant samples. Immunoextraction of endogenous hCG isoforms using a monoclonal antibody (E27) on a 96-well microtitier plate, followed by in-well tryptic digestion, and SIM monitoring of the selected signature peptides, resulted in the qualitative differentiation between several hCG isoforms in serum or urine. We conclude that the orthogonal selectivity conferred by the combination of immunoaffinity extraction and LC-MS analysis offers valuable complementary information to the conventional immunoassays.

Expression and epitope characterization of a recombinant CA 125 repeat: fourth report from the ISOBM TD-1 workshop.

BACKGROUND: CA 125 antigenic domains appear to reside within a region containing 156-amino acid sequence repeats. Surprisingly, anti-CA 125 antibodies can be classified into three families (groups A, B and C) indicating limited epitope diversity. In this study we describe the heterologous expression of a CA 125 repeat unit (R11) and an analysis of its epitope topography. METHODS: R11 was expressed using a baculovirus approach and purified from culture supernatants by sequential ion exchange chromatography. Monoclonal antibody binding was assessed using antigen capture and cross-inhibition methods. RESULTS: The recombinant repeat was purified to 2.5 x 10(7) U/mg. Although a number of group A and B monoclonal antibodies were found to bind R11, the prototype antibody OC125 (group A) showed little reactivity. However, the prior binding of some group B monoclonal antibodies dramatically enhanced subsequent OC125 binding. Low monoclonal antibody reactivity to R11 correlated well with poor binding to SDS-denatured human ascites CA 125. CONCLUSION: The ability to 'activate' R11 epitopes indicates that some may not be displayed optimally on isolated repeats. This observation, together with the concordance between monoclonal antibody binding to R11 and denatured CA 125, suggests that a number of epitopes are preferentially displayed only when contained within multiple repeat domains.

Progastrin-releasing peptide: stability in plasma/serum and upper reference limit.

BACKGROUND: Progastrin-releasing peptide (proGRP) is a promising tumor marker for small cell lung cancer (SCLC). Here we study the stability of proGRP in serum and plasma, as well as proGRP levels in healthy individuals, to provide a framework for clinical studies. METHODS: Serum, with and without protease inhibitors, and plasma from SCLC patients and healthy individuals were assayed for proGRP immediately after collection and following various storage conditions. RESULTS: No degradation was observed in serum or plasma after storage for 4 weeks at -30 degrees C. Serum proGRP levels were stable for up to 3 days at 4 degrees C, but decreased at room temperature. Addition of protease inhibitors to patient serum did not markedly improve stability. In EDTA plasma, proGRP concentrations increased upon storage in some samples at room temperature and 4 degrees C. When assayed immediately after collection, no significant variations in proGRP concentrations were observed between serum and EDTA plasma (n = 171). A 97.5-percentile reference limit of 58.9 ng/l was calculated from data from 806 individuals. However, proGRP levels were significantly correlated with age, sex, creatinine concentrations, body mass index and smoking. CONCLUSION: Serum is the preferred material for measuring proGRP, provided it is stored at 4 degrees C and assayed within 3 days.

Reference intervals for carcinoembryonic antigen (CEA), CA125, MUC1, Alfa-foeto-protein (AFP), neuron-specific enolase (NSE) and CA19.9 from the NORIP study.

OBJECTIVE: Adhering to current IFCC recommendations, we calculated upper 97.5 % reference limits for serum tumor markers. MATERIAL AND METHODS: Serum samples from 498 healthy individuals from the Nordic reference interval project (NORIP) were investigated for carcinoembryonic antigen (CEA), CA125 and MUC1 (episialin, CA15.3) using in-house immunofluorometric assays and, for alpha-foetoprotein (AFP), a PerkinElmer Life Sciences assay, neuron-specific enolase (NSE) using an in-house immunoradiometric assay and CA19.9 using a Beckman Access assay. All assays participate in external quality assessment programs. RESULTS: CEA concentrations increased with age and smoking. Upper reference limits for non-smokers were 3.59 microg/L at 50 years and 4.12 microg/L at 70 years. CA125 concentrations were age-independent and the upper reference limit was 35.8 kU/L. MUC1 increased with age and body mass index (BMI). Upper reference limits were 31.7 kU/L at 40 years and BMI 24, 37.5 kU/L at 70 years and BMI 24, and 33.7 kU/L at 40 years and BMI 30. AFP increases with age, and the upper reference limits were 3.82 kU/L at 20 years and 8.70 kU/L at 60 years. An upper reference limit for NSE was 8.91 microg/L in non-smokers; smokers exhibited significantly lower levels. The upper reference limit for individuals expressing CA19.9 was 28.3 kU/L. CONCLUSIONS: For AFP, CA125 and CA19.9, the reference levels obtained were close to previously reported reference ranges. Smoking and age were confirmed as covariates for CEA. The associations between MUC1 with age and BMI and between NSE and smoking have not been reported previously.

Automated time-resolved immunofluorometric assay for progastrin-releasing peptide.

BACKGROUND: Small cell lung cancer accounts for approximately 20% of new cases of lung cancer, and advanced disease is prevalent at the time of diagnosis. Neuron-specific enolase (NSE) has been the primary tumor marker in small cell lung cancer but it has relatively low sensitivity in early-stage disease. Progastrin-releasing peptide (proGRP) is a promising alternative or complementary marker for NSE. We have previously described a time-resolved immunofluorometric assay (TR-IFMA) for proGRP that lacked the necessary sensitivity and robustness for use in the routine clinical laboratory. Herein we describe the development of an improved assay using a novel monoclonal antibody pair. METHODS: Mice were immunized with different conjugated proGRP peptides, including residues 31-98, 1-98, and preproGRP(-23-125). Pair combinations of the resulting monoclonal antibodies (mAb) were tested. The improved TR-IFMA was compared with the only other available proGRP assay, the proGRP ELISA (IBL). RESULTS: A panel of 12 high-affinity mAbs was produced. The best assay combination was between our original E146 mAb as solid-phase antibody and the new mAb M16 as tracer. The new TR-IFMA had a linear dose-response curve, a wide dynamic range (13-13 500 ng/L), and a limit of detection of 2.8 ng/L. Total CV was <5.6% over the whole measuring range. Bland-Altman difference analysis indicated a significant positive bias between the IFMA and the ELISA. CONCLUSIONS: We describe a sensitive and robust mAb-based TR-IFMA for proGRP. The assay is fully automated and displays high quality performance.

Human heterophilic antibodies display specificity for murine IgG subclasses.

OBJECTIVES: The study investigated heterophilic antibodies: the human immunoglobulin classes involved and their specificity for different murine IgG subclasses. DESIGN AND METHODS: Using immunofluorometric assays for human IgA, IgM and IgG binding murine IgG1, we analyzed 173 samples displaying positive interference and 97 negative control samples from a previous study. We also set up assays for heterophilic antibody interference using Mabs from different murine IgG subclasses. Three Mabs each of murine IgG1, IgG2a and IgG2b subclasses, one murine IgG3 Mab and one rat Mab were used. RESULTS: Elevated levels of human murine IgG1-binding immunoglobulins of IgM class only were found in 40% of interference-positive samples, human IgG only in 1.7%, and human IgA only in 2.3% of the samples. Both elevated human IgG and IgM classes were found in 3.5% of the samples, IgA and IgM in 4.0%, and finally, all three immunoglobulin classes in 1.7% of the samples. Eighty percent of interference positive samples showed heterophilic assay interference for at least one murine IgG1 Mab, 35% for IgG2a, 66% for IgG2b, 52% for IgG3a and 17% for the rat Mab. CONCLUSIONS: Heterophilic antibody interference is mainly caused by IgM class human antibodies with a marked murine IgG subclass specificity. Combining assay antibodies from different murine IgG subclasses may reduce interference in immunoassays.

Use of an in vivo biotinylated single-chain antibody as capture reagent in an immunometric assay to decrease the incidence of interference from heterophilic antibodies.

BACKGROUND: Heterophilic antibodies are a common source of interference in immunometric assays. We tested the hypothesis that the incidence of such interference could be decreased by use of a recombinant in vivo-biotinylated single-chain antibody (scFv) as the capture reagent. METHODS: We established three assays for carcinoembryonic antigen (CEA) with the capture antibody either chemically biotinylated whole monoclonal T84.66 immunoglobulin, a corresponding F(ab')2 fragment, or a site-specifically biotinylated T84.66-derived single-chain antibody (scFv). Antibodies were attached to streptavidin-coated microplates. A common europium-labeled anti-CEA tracer monoclonal antibody was used. The F(ab')2 assay used a buffer that contained bovine immunoglobulin and aggregated irrelevant monoclonal antibody MAK33 as blocking agents. The whole T84.66 immunoglobulin and scFv assays were performed without addition of blocking agents. From a previous study of 11 261 sera, we tested 390 samples that had displayed heterophilic antibody interference and 179 samples that had not. RESULTS: After correction for bias and analytical variation [2.56 x SD (from the precision profile)], 383 samples displayed significantly different values (>1 microg/L) in the whole T84.66-based assay and the F(ab')2 assay. In contrast, only nine samples showed falsely high CEA concentrations in the scFv assay. After blocking agents were added to the assay buffer, eight of the nine samples displayed results equivalent to those of the F(ab')2 assay, and sample dilution produced equivalent results for the remaining sample. CONCLUSION: Their ability to be site-specifically biotinylated and their relative resistance to heterophilic antibody interference indicate that single-chain antibodies may be useful solid-phase reagents in immunometric assays.

Immunofluorometric assay for the metastasis-related protein S100A4: release of S100A4 from normal blood cells prohibits the use of S100A4 as a tumor marker in plasma and serum.

The metastasis-related protein S100A4 is released from tumor cells, and since it is highly expressed in colorectal cancer (CRC), it could be a potential tumor marker in plasma or serum. Monoclonal antibodies (MAbs) were raised against human recombinant S100A4 and shown to detect native and recombinant antigen with high sensitivity and specificity. Using two MAbs, an immunofluorometric assay (IFMA) was established to detect S100A4 in clinical samples with high sensitivity and precision. S100A4 in plasma and serum from patients with CRC was highly influenced by sample hemolysis. Both red blood cells and mononuclear cells were found to contain S100A4, possibly contributing to the measured levels in serum and plasma. Since even very low-level hemolysis influenced the results, a potential contribution from an S100A4-expressing tumor could not be discerned, indicating that S100A4 is not suitable as a plasma or serum tumor marker for CRC. The antibodies and the IFMA may still be useful for research purposes.

Testing and validating a homogeneous immunometric assay for interference.

We analyzed 95 sera, demonstrating interference in a previous study, with the Kryptor homogeneous time-resolved fluorescence resonance energy transfer carcinoembryonic antigen (CEA) immunoassay (Brahms AG, Berlin, Germany). Only one serum differed, i.e., 6.0 microg/l for Kryptor vs. 13.3 microg/l for a microtiter plate in-house immunofluorometric assay (IFMA), using both aggregated mouse immunoglobulins as blocker and capture monoclonal antibody (Mab) F(ab')2-fragments to reduce interference. Sera were further analyzed with experimental IFMAs with intact capture Mabs and without blocker. The Kryptor-CEA assay Mab GFR44 (capture) had elevated results in 71% of sera with the Kryptor Mab G15 (tracer) or in 81% with our tracer Mab. G15 (capture) had 65% elevated results with GFR44 (tracer) or 99% with our tracer Mab. The latter was reduced from 99% to 62% by adding Kryptor blocker to the buffer or to 30% by adding our blocker. Finally, combining both assay Mabs and assay buffer from Kryptor still gave interference in 16% of sera. Kryptor-CEA assay results thus agreed with our in-house CEA assay results, showing no interference. As the Kryptor-CEA assay antibodies were sensitive to interference and the Kryptor-CEA assay buffer did not reduce interference as efficiently as our in-house assay buffer, the Kryptor-CEA assay format was crucial for the absence of interference.

Immunometric assay interference: incidence and prevention.

BACKGROUND: The primary aim of the study was to reduce interference in an in-house two-site, two-step immunometric assay. METHODS: In the running laboratory routine, 11 261 samples were tested with a carcinoembryonic antigen (CEA) assay with bovine immunoglobulin but no murine immunoglobulins in the buffer, in parallel to our routine CEA assay, using 15 mg/L heat-treated nonspecific murine immunoglobulin (MAK33) in the buffer and with the Fc fragments removed from the capture antibody. RESULTS: The frequency of interference was estimated to be 4.0% (95% confidence interval, 3.3-4.7%). The addition of 15 mg/L native MAK33 had little effect (frequency, 3.9%; 95% confidence interval, 3.2-4.6%), whereas adding 15 mg/L heat-treated MAK33 reduced interference to 0.86% (0.61-1.12%), and adding 50 mg/L reduced it further to 0.06% (0-0.13%). Removing the Fc fragments by itself reduced interference to 0.10% (0.02-0.19%). There were no statistically significant differences for age (P <0.23) or gender (P <0.40) between patients with interference (n = 210) and a randomly selected interference-negative control group (n = 186). Interference was not constant in patients: 15 of 25 individuals positive for interference and with four or more samples screened for interference had an interference-negative sample either before or after the peak of interference. CONCLUSIONS: In a two-site, two-step immunometric assay using mouse monoclonal antibodies, use of heat-treated nonspecific murine immunoglobulin in the buffer or removal of the Fc fragment from the capture antibody could improve performance.