Pre-analytical stability of the CEA, CYFRA 21.1, NSE, CA125 and HE4 tumor markers

BACKGROUND

For lung cancer, circulating tumor markers (TM) are available to guide clinical treatment decisions. To ensure adequate accuracy, pre-analytical instabilities need to be known and addressed in the pre-analytical laboratory protocols.

OBJECTIVE

This study investigates the pre-analytical stability of CA125, CEA, CYFRA 21.1, HE4 and NSE for the following pre-analytical variables and procedures; i) whole blood stability, ii) serum freeze-thaw cycles, iii) electric vibration mixing and iv) serum storage at different temperatures.

METHODS

Left-over patient samples were used and for every investigated variable six patient samples were used and analysed in duplicate. Acceptance criteria were based on analytical performance specifications based on biological variation and significant differences with baseline.

RESULTS

Whole blood was stable for at least 6 hours for all TM except for NSE. Two freeze-thaw cycles were acceptable for all TM except CYFRA 21.1. Electric vibration mixing was allowed for all TM except for CYFRA 21.1. Serum stability at 4°C was 7 days for CEA, CA125, CYFRA 21.1 and HE4 and 4 hours for NSE.

CONCLUSIONS

Critical pre-analytical processing step conditions were identified that, if not taken into account, will result in reporting of erroneous TM results.

Abstract 3358: PLCRC-PROVENC3: assessing the prognostic value of post-surgery liquid biopsy cell-free circulating tumor DNA in stage III colon cancer patients

Introduction: Surgery followed by adjuvant chemotherapy (ACT) is standard of care in stage III colon cancer. However, 50% of the patients would be cured by surgery alone and are being overtreated, while 30-35% will experience a recurrence despite adjuvant treatment, resulting in only 15-20% of the patients benefitting from ACT. Therefore, there is a need for prognostic biomarkers to better stratify this group of patients for ACT decisions. Recent observational and interventional studies in non-metastatic colon cancer have shown that detection of cell-free circulating tumor DNA (ctDNA) in blood after surgery is highly prognostic for development of recurrence. Hence, ctDNA analysis is a promising approach to guide treatment decisions in stage III colon cancer, but studies with large well-defined patient cohorts are needed to prove clinical utility.

Aim: Determine the prognostic value of ctDNA in stage III colon cancer patients treated with ACT to reduce futile treatment.

Methods: 241 stage III colon cancer patients treated with ACT were included in the prospective observational study “PROVENC3” (PROgnostic Value of Early Notification by Ctdna in Colon Cancer stage 3), a substudy of the Prospective Dutch Colorectal Cancer cohort (PLCRC). The PLCRC infrastructure accrued patients with colorectal cancer in 23 participating hospitals in the Netherlands. After informed consent, blood was collected pre-surgery, post-surgery, post-ACT and every six months up to 3 years. Tumor-informed detection of ctDNA was performed through integrated whole genome sequencing (WGS) analyses of formalin-fixed paraffin-embedded tumor tissue DNA (80x), germline DNA (40x), and plasma cell-free DNA (30x).

Results: Patient accrual was completed in 2021, with a median follow-up of 35.6 months. In total, 1090 blood samples have been collected to date. Analytical studies demonstrated a limit of detection of the test of 0.005% ctDNA utilizing contrived reference models derived from six independent cell lines, with a specificity of 99.6% across 119 noncancerous donor plasma specimens. From the PROVENC3 study, ctDNA analyses are ongoing for pre-surgery (n=68) and post-surgery (n=241 patients) blood samples. Preliminary results demonstrated a ctDNA detection rate of 93.4% pre-surgery and 17.1% post-surgery, which was associated with disease recurrence. Final analysis will enable determination of: 1) the proportion of ctDNA-positive/negative patients after surgery and the corresponding recurrence rates; 2) the prognostic value of post-surgery ctDNA; and 3) the lead time between post-surgery ctDNA detection and recurrence.

Future Perspective: Ultimately, the results of this study will be used to model and design a ctDNA-guided interventional trial in stage III colon cancer patients, to reduce futile ACT and its associated side-effects.

Citation Format: Carmen Rubio-Alarcon, Steven L. Ketelaars, Ingrid A. Franken, Sietske C. van Nassau, Dave E. van der Kruijssen, Suzanna J. Schraa, Theodora C. Linders, Pien Delis-van Diemen, Maartje Alkemade, Anne Bolijn, Marianne Tijssen, Margriet Lemmens, Miranda van Dongen, Mirthe Lanfermeijer, Annegien Broeks, Lana Meiqari, Linda J. Bosch, Victor E. Velculescu, Amy Greer, Samuel V. Angiuoli, Andrew Georgiadis, David Riley, James R. White, Christopher Greco, Liam Cox, Daan van den Broek, Cornelis J. Punt, Veerle M. Coupé, Miriam Koopman, Jeanine Roodhart, Gerrit A. Meijer, Mark Sausen, Geraldine R. Vink, Remond J. Fijneman. PLCRC-PROVENC3: assessing the prognostic value of post-surgery liquid biopsy cell-free circulating tumor DNA in stage III colon cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3358.

Combining variant detection and fragment length analysis improves detection of minimal residual disease in post‐surgery circulating tumour DNA of stage II‐IIIA NSCLC patients

Stage II–IIIA nonsmall cell lung cancer (NSCLC) patients receive adjuvant chemotherapy after surgery as standard‐of‐care treatment, even though only approximately 5.8% of patients will benefit. Identifying patients with minimal residual disease (MRD) after surgery using tissue‐informed testing of postoperative plasma circulating cell‐free tumour DNA (ctDNA) may allow adjuvant therapy to be withheld from patients without MRD. However, the detection of MRD in the postoperative setting is challenging, and more sensitive methods are urgently needed. We developed a method that combines variant calling and a novel ctDNA fragment length analysis using hybrid capture sequencing data. Among 36 stage II–IIIA NSCLC patients, this method distinguished patients with and without recurrence of disease in a 20 times repeated 10‐fold cross validation with 75% accuracy (P = 0.0029). In contrast, using only variant calling or only fragment length analysis, no signification distinction between patients was shown (P = 0.24 and P = 0.074 respectively). In addition, a variant‐level fragmentation score was developed that was able to classify variants detected in plasma cfDNA into tumour‐derived or white‐blood‐cell‐derived variants with 84% accuracy. The findings in this study may help drive the integration of various types of information from the same data, eventually leading to cheaper and more sensitive techniques to be used in this challenging clinical setting.

An Automated Correction Algorithm (ALPACA) for ddPCR Data Using Adaptive Limit of Blank and Correction of False Positive Events Improves Specificity of Mutation Detection

BACKGROUND

Bio-Rad droplet-digital PCR is a highly sensitive method that can be used to detect tumor mutations in circulating cell-free DNA (cfDNA) of patients with cancer. Correct interpretation of ddPCR results is important for optimal sensitivity and specificity. Despite its widespread use, no standardized method to interpret ddPCR data is available, nor have technical artifacts affecting ddPCR results been widely studied.

METHODS

False positive rates were determined for 6 ddPCR assays at variable amounts of input DNA, revealing polymerase induced false positive events (PIFs) and other false positives. An in silico correction algorithm, known as the adaptive LoB and PIFs: an automated correction algorithm (ALPACA), was developed to remove PIFs and apply an adaptive limit of blank (LoB) to individual samples. Performance of ALPACA was compared to a standard strategy (no PIF correction and static LoB = 3) using data from commercial reference DNA, healthy volunteer cfDNA, and cfDNA from a real-life cohort of 209 patients with stage IV nonsmall cell lung cancer (NSCLC) whose tumor and cfDNA had been molecularly profiled.

RESULTS

Applying ALPACA reduced false positive results in healthy cfDNA compared to the standard strategy (specificity 98 vs 88%, P = 10-5) and stage IV NSCLC patient cfDNA (99 vs 93%, P = 10-11), while not affecting sensitivity in commercial reference DNA (70 vs 68% P = 0.77) or patient cfDNA (82 vs 88%, P = 0.13). Overall accuracy in patient samples was improved (98 vs 92%, P = 10-7).

CONCLUSIONS

Correction of PIFs and application of an adaptive LoB increases specificity without a loss of sensitivity in ddPCR, leading to a higher accuracy in a real-life cohort of patients with stage IV NSCLC.

Performance of four platforms for KRAS mutation detection in plasma cell-free DNA: ddPCR, Idylla, COBAS z480 and BEAMing

Multiple platforms are commercially available for the detection of circulating cell-free tumour DNA (ctDNA) from liquid biopsies. Since platforms have different input and output variables, deciding what platform to use for a given clinical or research question can be daunting. This study aimed to provide insight in platform selection criteria by comparing four commercial platforms that detect KRAS ctDNA hotspot mutations: Bio-Rad droplet digital PCR (ddPCR), BioCartis Idylla, Roche COBAS z480 and Sysmex BEAMing. Platform sensitivities were determined using plasma samples from metastatic colorectal cancer (mCRC) patients and synthetic reference samples, thereby eliminating variability in amount of plasma analysed and ctDNA isolation methods. The prevalence of KRAS nucleotide alterations was set against platform-specific breadth of target. Platform comparisons revealed that ddPCR and BEAMing detect more KRAS mutations amongst mCRC patients than Idylla and COBAS z480. Maximum sample throughput was highest for ddPCR and COBAS z480. Total annual costs were highest for BEAMing and lowest for Idylla and ddPCR. In conclusion, when selecting a platform for detection of ctDNA hotspot mutations the desired test sensitivity, breadth of target, maximum sample throughput, and total annual costs are critical factors that should be taken into consideration. Based on the results of this study, laboratories will be able to select the optimal platform for their needs.

Abstract 2276: Performance and cost comparison of circulating tumor DNA detection platforms

Purpose

The field of circulating cell free DNA (cfDNA) testing is quickly developing. Multiple platforms to detect hotspot mutations in ctDNA are available, including Bio-Rad droplet digital PCR (ddPCR), BioCartis Idylla, Roche COBAS z480 and Sysmex BEAMing. These platforms vary in the amount of plasma required, the method of ctDNA isolation, the number of hotspots analyzed, and the costs per sample. These factors can impact the applicability of a specific platform in the analysis of ctDNA.

In this study we compared these platforms in terms of sensitivity and total costs per sample.

Methods

The platform comparison was performed as follows:

1. Plasma from six metastatic colorectal cancer (mCRC) patients with known tissue KRAS mutation status was analyzed according to the manufacturers’ protocols.

2. Sensitivity was tested using six constructed reference samples with seven KRAS hotspot mutations at various total input (10ng, 50ng) and mutant allele frequencies (0%, 0.04%, 0.50%); each in four replicates.

3. Twelve mCRC patients were analyzed using equal amounts of ctDNA for each platform.

4. Total costs per sample were evaluated, including costs for consumables, technician hands-on time, equipment and maintenance.

Results

1. In six mCRC samples BEAMing (3ml plasma) detected 5/6 mutations, ddPCR and Idylla (both 1ml plasma) 4/6 mutations and COBAS z480 (2ml plasma) 3/6 mutations.

2. In constructed reference samples ddPCR (65%) and BEAMing (46%) yielded the highest sensitivity. With 10ng input BEAMing and COBAS z480 produced “Too little DNA” errors in 60% and 100% of cases, respectively.

3. In twelve mCRC samples, eight had more than 10ng ctDNA and four had less. All platforms were fully concordant for samples with more than 10ng input. Detection rate across twelve samples: Idylla detected 7/11 detectable mutations, COBAS z480 and BEAMing both detected 5/11, ddPCR detected 4/10.

4. BEAMing has the highest cost per sample (€486-€821) whereas ddPCR has the lowest cost per sample (€39-€298).

Conclusions

A direct comparison of ctDNA mutation detection platforms is complex and should take into account the differences in input and output specifications of the platforms.

Factors such as complexity of analysis (Idylla is a low complexity platform, whereas BEAMing requires more specialized training), total costs (varying from ddPCR to BEAMing), sensitivity (ddPCR and BEAMing yielded the highest sensitivity) and the number of mutations evaluated vary greatly between the platforms. All these factors influence ctDNA analysis and will have to be considered when choosing a specific platform for a specific (clinical) question, or when comparing results between studies.

Our data provide insight in the comparative performance of four commercial ctDNA analysis platforms, allowing future users to make an informed decision regarding a platform.

Acknowledgement

Powered by Health~Holland, Top Sector Life Sciences & Health, grant LSHM16047-H005.

Citation Format: Daan C. Vessies, Marjolein J. Greuter, Karlijn L. van Rooijen, Theodora C. Linders, Mirthe Lanfermeijer, Kalpana L. Ramkisoensing, Flore E. Grijseels, Boris van Doorn, Gerrit A. Meijer, Miriam Koopman, Veerle M. Coupé, Geraldine R. Vink, Remond J. Fijneman, Daan van den Broek. Performance and cost comparison of circulating tumor DNA detection platforms [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2276.

Abstract 1381: Technical validation of Roche AVENIO sequencing platform for liquid biopsies

Purpose

Circulating cell-free tumor DNA (ctDNA) is rapidly gaining attention as a valuable biosource for the detection and diagnosis of multiple types of cancer. ctDNA can be easily obtained through a minimally invasive liquid biopsy. Multiple NGS-based methods have been developed specifically for application in liquid biopsies.

Before such methods can be implemented in clinical care, they need to be technically validated through measurement of independent reference standards. In the current project we technically validated the Roche AVENIO ctDNA Assay, which is based on the CAPP-seq technology [Newman AM, et al. Nat Biotechnol. 2016;34:547-55]. Its Targeted Kit interrogates 17 genes for Single Nucleotide Variants (SNVs), Insertions and Deletions (InDels), Copy Number Variations (CNVs) and gene fusions in a single workflow.

Methods

The detection of SNVs, CNVs, Indels and gene fusions was evaluated as follows:

- SeraCare Seraseq ctDNA Complete was analyzed in four replicates at 2.5%, 0.5% and 0.1% mutant allele frequency (AF), and at 50ng and 10ng input. This material contains SNVs, InDels, CNVs and gene fusions.

- An in-house characterized reference pool of patient plasma, containing SNVs and InDels at various AF (0.07% - 5.51%), was analyzed in four replicates at 50ng and 10ng input.

- Plasma and serum samples obtained from patients with tissue confirmed presence of CNV were used to validate the robustness of CNV detection. Twelve samples were selected to represent a range of background DNA (serum had 2-4 times higher wildtype DNA levels compared to plasma), amplification levels (CNV levels 1.4-fold to 4.2-fold by ddPCR) and total input (10-50ng cfDNA input).

- Plasma obtained from five patients with EML4-ALK gene fusion confirmed on tissue was used to further validate fusion detection.

Sequencing was performed on an Illumina NextSeq on High Output mode with sixteen samples per run (20-33M read pairs per sample).

Results

Sensitivity for detection of SNVs and InDels was 100% for all variants down to 0.18% AF when using 50ng input, and 50% for variants at 0.10% AF. For 10ng input the variants were detected down to 0.60% AF with sensitivity of 92%. For variants with AF between 0.07% and 0.53%, sensitivity was 67%.

Results from twelve plasma and serum samples obtained from patients with EGFR CNV were 100% concordant with ddPCR results. These results were confirmed in SeraCare samples at 10ng and 50ng input.

Sensitivity of fusion detection in SeraCare material was 100% for AF of 0.5% and 50ng input, and 38% for input of 10ng. No fusions were detected at 50ng or 10ng input for AF of 0.1%. We furthermore confirmed 3 out of the 5 fusions found in tissue.

Conclusions

We have determined the lower limits of detection for SNVs, CNVs and gene fusions. Sensitivity was dependent on DNA input with performance optimized at 50ng.

Our results confirm the feasibility of in-house use of the AVENIO ctDNA Targeted Kit for broad molecular profiling in liquid biopsies.

Citation Format: Daan C. Vessies, Theodora C. Linders, Kalpana L. Ramkisoensing, Petra M. Nederlof, Gerrit A. Meijer, Kim Monkhorst, Daan van den Broek. Technical validation of Roche AVENIO sequencing platform for liquid biopsies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1381.

Cell-free DNA in the supernatant of pleural effusion can be used to detect driver and resistance mutations, and can guide tyrosine kinase inhibitor treatment decisions

Objectives

Molecular profiling of tumours has become the mainstay of diagnostics for metastasised solid malignancies and guides personalised treatment, especially in nonsmall cell lung cancer (NSCLC). In current practice, it is often challenging to obtain sufficient tumour material for reliable molecular analysis. Cell-free DNA (cfDNA) in blood or other bio-sources could present an alternative approach to obtain genetic information from the tumour. In a retrospective cohort we analysed the added value of cfDNA analysis in pleural effusions for molecular profiling.

Methods

We retrospectively analysed both the supernatant and the cell pellet of 44 pleural effusions sampled from 39 stage IV patients with KRAS (n=23) or EGFR (n=16) mutated tumours to detect the original driver mutation as well as for EGFR T790M resistance mutations. Patients were diagnosed with either NSCLC (n=32), colon carcinoma (n=4), appendiceal carcinoma (n=2) or adenocarcinoma of unknown primary (n=1). Samples collected in the context of routine clinical care were stored at the Netherlands Cancer Institute biobank. We used droplet digital PCR for analysis.

Results

The driver mutation could be detected in 36 of the 44 pleural effusions by analysis of both the supernatant (35 out of 44 positive) and the cell pellet (31 out of 44 positive). In seven out of 20 pleural effusions from patients with EGFR mutation-positive tumours, a T790M mutation was detected. All seven supernatants and cell pellets were positive.

Conclusions

cfDNA in pleural effusion can be used to detect driver mutations as well as resistance mechanisms like EGFR T790M in pleural effusion with high accuracy and is therefore a valuable bio-source.

Choice of tumour markers in patients with neuroendocrine tumours is dependent on the histological grade. A marker study of Chromogranin A, Neuron specific enolase, Progastrin-releasing peptide and cytokeratin fragments

BACKGROUND

Chromogranin A (CgA) is the most important tumour marker for well-differentiated neuroendocrine tumours (NET) and neuron specific enolase (NSE) for poorly differentiated neuroendocrine carcinoma (NEC). This study investigated whether the markers progastrin-releasing peptide (proGRP) and cytokeratin fragments (CKfr) CK8, CK18 and CK19 (MonoTotal) can be of additional value to the histological classification and help predict survival in these patients.

METHODS

CgA, NSE, proGRP and CKfr were measured in 242 patients with grade 1 NET (G1NET), 38 with grade 2 NET (G2NET), 42 with large cell NEC (LCNEC), 251 with small cell NEC (SCNEC) and in 282 healthy persons. Results were compared with tumour characteristics and survival by means of Receiver Operating Characteristics (ROC) curves and Cox regression analyses.

RESULTS

The largest area under the ROC curve was for CgA (0.86, 0.91 and 0.90, respectively) when comparing patients with G1NET, G2NET and LCNEC with healthy persons. ProGRP showed the highest sensitivity (73%) at 95% specificity in patients with SCNEC. In a multivariate survival analysis, only CKfr was associated with survival (P<0.0001) for patients with well-differentiated NET (G1NET and G2NET). For patients with poorly differentiated NEC, both CKfr and NSE were associated with survival (P<0.0001 and P=0.003, respectively).

CONCLUSION

Within all histological groups a combination of tumour markers proved to be more informative as diagnostic and prognostic marker than each marker alone. In patients with well-differentiated NET and LCNEC we recommend the use of CgA and CKfr, whilst in patients with SCNEC, proGRP and CKfr are preferred.

S-100B protein and melanoma inhibitory activity protein in uveal melanoma screening. A comparison with liver function tests

AIM

Our purpose was to determine whether S-100B or melanoma inhibitory activity (MIA) concentrations in the serum of patients with large uveal melanomas were better markers for the presentation of metastases than liver function tests. We also investigated whether increased marker levels were related to known clinical and histopathological prognostic parameters.

METHODS

Total S-100B (A1B + BB) and MIA concentrations were measured in the sera from 104 patients with uveal melanoma prior to enucleation and in the sera from 50 healthy controls. Concentrations were also determined in the sera from 30 patients with known uveal melanoma metastases. Liaison Sangtec 100, an automated immunoluminometric assay measuring the total S-100B, and Roche MIA ELISA were used to quantify these proteins in serum. Results were compared with liver function tests [alkaline phosphatase, lactate dehydrogenase (LD), aspartate aminotransferase, alanine aminotransferase and gamma-glutamyl transpeptidase].

RESULTS

The mean S-100B and MIA concentrations were significantly higher in patients with metastases compared to melanoma patients without metastases. At the time of enucleation, S-100B and MIA were not prognostic for metastases in uveal melanoma, but S-100B and LD were the best tests to predict the occurrence of metastatic disease during the follow-up period.

CONCLUSIONS

In our study, the S-100B and MIA serum concentrations were not correlated with any tested established prognostic parameter. S-100B and LD showed better performance in identifying melanoma metastases than gamma-glutamyl transpeptidase and MIA. A prospective follow-up study is needed to evaluate S-100B and MIA in identifying early micrometastasis in uveal melanoma.

Effect of paclitaxel (Taxol) on CA 125 expression and release by ovarian cancer cell lines

Two ovarian cancer cell lines, OVC432 and the newly established CVU4I, were used to study the effect of Taxal on cell growth and simultaneous CA 125 antigen expression. Growth of both cell lines was effectively inhibited by drug concentrations of 0.1 microM and higher. Complete inhibition of cell growth may result from high concentrations of Cremophor EL present in the Taxol formulation. Immunohistochemical analysis demonstrated that both cell lines retained the CA 125 expression on the cell surface during exposure to paclitaxel. This was reflected in a constant statistically significant correlation between cell numbers and CA 125 concentrations found in cell lysates. CA 125 levels in the culture medium showed a significant relation to cell numbers and, consequently, to the response of the cell line to the administered anticancer drug. It may be concluded from this study that CA 125 seems to be a reliable tumor marker in monitoring tumor response during paclitaxel treatment.

Insulin-like growth-factor-binding protein 3 is decreased in early-stage operable pre-menopausal breast cancer

Insulin-like growth factor I (IGF-I) is a potent mitogen for human breast-cancer cells in vitro. In circulation, most of IGF-I is bound to IGF-binding protein 3 (IGFBP-3). This high-affinity binding is thought to have an important limiting effect on the availability of IGF-I for biological activity. To assess the availability of IGF-I for receptor binding, we determined serum levels of IGF-I and IGFBP-3 and IGF-I/IGFBP-3 ratios. In a case-control study, 150 women aged 38 to 75 years presenting with stage-I or -II breast cancer were investigated just prior to surgery (n = 76), or to irradiation one month after surgery (n = 74). The population-based control group consisted of 441 women of the same age having no breast cancer. Women reporting diabetes mellitus or other hormonal abnormalities were excluded. Premenopausal cases showed elevated IGF-I serum concentrations, decreased IGFBP-3 levels and increased IGF-I/IGFBP-3 ratios. The IGF-I/IGFBP-3 ratio was a significant breast-cancer risk factor, also after adjustment for age, family history, height, body-mass index, body-fat distribution, and serum levels of C-peptide. The relative risk was 7.34 for the highest compared with the lowest quintile of IGF-I/IGFBP-3. The presence or absence of tumor had no influence on these results. Increased levels of available IGF-I in the circulation of pre-menopausal women may contribute to the development of breast cancer.