Analytical and clinical validation of a novel CE-IVD kit for PIK3CA hotspot mutations in liquid biopsy samples

Abstract

PIK3CA mutations have been implicated in the prognosis and therapeutic response in HER2-positive early breast cancer, with variants located in exon 9 and exon 20 shown to modulate sensitivity to neoadjuvant chemotherapy. Several molecular diagnostic platforms based on PCR or NGS have been developed for the detection of PIK3CA mutations in both tumor tissues and plasma-derived cfDNA, and some commercially available assays have already received FDA approval. The aim of this study was to evaluate the analytical and clinical performance of a novel CE-IVD molecular assay (Oncolipsy, PIK3CA kit, Pharmassist, Greece) for hotspot PIK3CA mutations (p.E542K, p.E545K, p.E545Q, p.H1047R) in liquid biopsy (LB) specimens. Analytical validation included assessment of sensitivity, specificity and reproducibility using certified liquid biopsy reference standards. Subsequently, the clinical utility of the assay was evaluated by analyzing plasma-cfDNA, CTC-derived gDNA and primary tumor samples for PIK3CA hotspot mutations. A total of 55 peripheral blood samples from breast cancer (BrCa) patients and 30 primary tumors (FFPEs) were examined. The same samples were also tested with two commercially available assays, the cobas® PIK3CA Mutation Test (Roche Diagnostics) and the ddPCR PIK3CA mutation test (BioRad), and the results were directly compared. Our findings demonstrate that the Oncolipsy PIK3CA kit exhibits high analytical detectability and excellent specificity for detecting PIK3CA hotspot mutations, at low variant allele frequencies. Clinical evaluation confirmed its robustness for liquid biopsy applications, with p.H1047R identified as the most frequent PIK3CA mutation. Concordance with both commercially available assays was high, with minor discrepancies attributable to differences in mutation coverage or detection thresholds. In conclusion, the CE-IVD Oncolipsy PIK3CA kit represents a highly detectability, specific and cost-effective real-time PCR-based solution for high throughput detection of four clinically relevant PIK3CA hotspot mutations in liquid biopsy samples.

Abbreviations

ARMS-PCR

Amplification-Refractory Mutation System-PCR

BrCa

Breast Cancer

CDK

cyclin-dependent kinases

cfDNA

cell-free DNA

CTCs

circulating tumor cells

ctDNA

circulating tumor DNA

ddPCR

droplet digital PCR

ET

Endocrine Therapy

FFPE

Formalin-Fixed Paraffin-Embedded

IVD

in vitro diagnostic

LB

Liquid Biopsy

mBrCa

Metastatic Breast cancer

1. Introduction

Breast cancer (BrCa) remains one of the most significant causes of cancer-related mortality among women and continues to represent a major global health challenge [1]. Despite the substantial progress in its management over the past years, breast cancer heterogeneity continues to influence treatment response and ultimately impacts patient outcomes [1]. In the era of personalized medicine, predictive biomarkers allow to identify patients with BrCa who may benefit from a specific therapy and predict the effectiveness of a given treatment as well as tumor resistance [1]. Beyond the traditional markers, a variety of genomic and molecular predictive biomarkers have now been implemented for BrCa patients including PIK3CA and ESR1 mutations, HER-2 low-expression, PD-L1 expression, as well as the tumor mutational burden (TMB) [2]. The identification of relevant biomarkers—especially when working with limited or low-tumor-content material—depends on advanced technologies capable of accurately profiling both tissue specimens and liquid biopsy samples. This need is even more crucial in the context of liquid biopsy. The development and implementation of ultrasensitive molecular assays are therefore essential to support individualized treatment strategies, enable real-time disease monitoring, and ultimately improve patient outcomes.

PIK3CA (Phosphoinositide 3-kinase) mutations are common and represent important targets for personalized therapies particularly in breast cancer [3]. The most common PIK3CA hotspot mutations, p.E545K and p.E542K in exon 9, and p.H1047R in exon 20, are found in about 40% of hormone receptor-positive (HR+) breast cancer (BrCa), typically within the helical and kinase domains of the PIK3CA gene [4,5]. PIK3CA mutations have prognostic and predictive implications across BrCa subtypes. In HER2-positive early-stage BrCa, PIK3CA mutations influence clinical outcome with exon-specific variants reported to affect response to neoadjuvant chemotherapy [6,7]. In HR + metastatic breast cancer (mBrCa), the introduction of PIK3CA-targeted therapeutics, such as the alpha-selective PI3K-inhibitor, PIQRAY (alpelisib), has expanded treatment options, particularly when combined with endocrine therapy and cyclin-dependent kinases (CDK4/6) inhibitors [8,9]. Most recently, FDA approval of triplet therapy with inavolisib in combination with palbociclib and fulvestrant underscores the ongoing shift towards precision medicine for patients harboring PIK3CA-mutated tumors [10,11].

Liquid biopsy (LB) offers a minimally invasive alternative approach to tissue biopsy for monitoring disease progression, treatment efficacy, and resistance [12]. By analyzing circulating tumor cells (CTCs) and circulating tumor DNA (ctDNA), LB enables real-time assessment of tumor dynamics and detection mutations such as PIK3CA without repeated tissue sampling [[13], [14], [15]]. Numerous studies suggest that combined analysis of CTCs and plasma-cfDNA provides a more comprehensive mutation profile [[16], [17], [18]], supporting its clinical utility in guiding treatment decisions and disease monitoring [19].

A wide range of molecular diagnostic assays has been developed for the detection of PIK3CA mutations in primary tumors or plasma using PCR- and next-generation sequencing (NGS) platforms, some of which have received regulatory approval. The Foundation One CDx assay (Foundation Medicine, USA), is an FDA-approved NGS companion diagnostic for identifying PIK3CA mutations in advanced BrCa patients to guide alpelisib eligibility in both tissue and plasma [20,21]. The Therascreen PIK3CA RGQ PCR Kit (Qiagen, Germany), for the qualitative detection of 11 mutations in the PIK3CA gene by real-time PCR was clinically validated as an in vitro diagnostic (IVD) for alpelisib eligibility [20,22] but was recently recalled as a Class 2 Device by the FDA [23]. Additional PCR-based commercial assays include the cobas®PIK3CA Mutation Test (Roche Diagnostics) which detects 17 mutations in exons 2, 5, 8, 10, and 21 in FFPEs, and the QClamp®PIK3CA Mutation Detection Test (DiaCarta, US) which targets mutations within exons 9 and exon 20 in both FFPEs or plasma specimens [24]. Other available CE-IVD and/or RUO platforms encompass the AmoyDx® PIK3CA five mutations detection Kit (Amoy Diagnostics) [25], the PIK3CA Mutation Analysis Kit (EntroGen Inc.) [26], the FastPlex™ PIK3CA 11 Mutations Digital PCR Detection Kit (Precigenome LLC) [27] and the EasyPGX® ready PIK3CA (Diatech Pharmacogenetics) [28] which together cover a broad spectrum of clinically significant PIK3CA mutations relevant for therapeutic decision making.

Our group has previously developed and validated a highly sensitive methodology for the detection of PIK3CA mutations that combines allele-specific priming, asymmetric PCR (ARMS-PCR), and melting curve analysis. Using this approach, we demonstrated reliable detection of four clinically relevant PIK3CA hotspot mutations (p.E542K, p.E545K, p.E545Q, and p.H1047R) in formalin-fixed paraffin-embedded (FFPEs) primary tumors, CTCs and plasma-cfDNA [29].

The aim of the present study was to perform both analytical and clinical validation of a novel commercially available CE-IVD kit for PIK3CA hotspot mutations, the Oncolipsy PIK3CA kit (Pharmassist, Greece), which is based on the methodology described above [29]. Analytical validation included assessment of detectability, specificity and reproducibility using LB-certified reference standards. Clinical performance was subsequently evaluated through the analysis of plasma-cfDNA, CTC-derived gDNA and primary tumors from BrCa patients. In addition, our results were directly compared with two commercially available molecular assays for PIK3CA mutations, providing a benchmarking framework for clinical utility.

2. Materials and methods

The outline of the workflow of the study is shown in Fig. 1. All experimental procedures were performed according to the ISO-15189 requirements. Different rooms, dedicated labware and dedicated areas and dedicated labware were used to avoid contamination. All preparation steps for the ddPCR setup were performed in a dedicated pre-PCR room and a PCR hood dedicated for the preparation of ddPCR reactions.

Fig. 1.

Schematic flowchart of the study.

2.1. Analytical validation

Commercially cfDNA reference standards containing PIK3CA mutations (p.E545K, p.E542K and p.H1047R) at allele frequencies (AF) of 12.5%, 1%, 0.8%, and 0% (PIK3CA-11 mutations 12.5% AF cfDNA, Sens-ID & Multiplex cfDNA reference standard, Horizon Discovery) were used for the analytical validation of the assay. Additional reference materials for the p.E545K mutation at 1% and 0% AF in a synthetic matrix (Multiplex I cfDNA in Synthetic Matrix II, Horizon Discovery) were also employed to further verify assay performance. For all these LB reference standards, cfDNA extraction was performed using the same protocol applied to clinical plasma samples.

2.2. Clinical validation

The clinical performance of the Oncolipsy PIK3CA kit was evaluated through the analysis of PIK3CA hotspot mutations in primary tumors (FFPEs), plasma-derived cfDNA and CTCs-derived gDNA, followed, by direct comparison of the results with two commercially available molecular assays.

2.3. Clinical samples

Thirty primary tumor samples from ER+/HER2-mBrCa patients collected between 2020 and 2022 were analyzed. In parallel, peripheral blood (PB) in EDTA was collected from 20 ER+/HER2-patients with early BrCa and 35 ER+/HER2-mBrCa patients, after written informed consent. The study was approved by the Ethics Committees of all participating institutions.

2.4. Genomic DNA (gDNA) isolation from primary tumors (FFPEs)

FFPEs (10 mm tissue sections) containing >80% tumour cells were used for gDNA extraction, using the QIAamp DSP DNA FFPE Tissue Kit (CE-IVD, Qiagen, Hilden, Germany), according to the manufacturer's instructions. DNA quantification was performed using the NanoDrop™ 1000 Spectrophotometer (ThermoFisher Scientific). The quality of DNA extracted from FFPE samples was assessed by amplification of a 70 bp region of PIK3CA exon 3 using the assay included in the Oncolipsy PIK3CA kit. Only samples with acceptable DNA quality (Cq < 38) were further analyzed.

2.5. Isolation of plasma-cfDNA

Peripheral blood (20 mL for early BrCa and 10 mL for mBrCa) was collected in EDTA tubes as previously described [29]. Plasma was obtained by centrifugation at 530g for 10min at room temperature, and a second centrifugation at 2000g for 10min, transferred into clean 2 mL tubes, and stored at −70 °C. Plasma-cfDNA was extracted from 2 mL of plasma using the QIAamp DSP cNA Kit (CE-IVD, Qiagen, Hilden, Germany), and cfDNA was eluted in 30 μL elution buffer as previously described [30]. Plasma-cfDNA quality was assessed by amplification of a 70 bp region of PIK3CA exon 3 using the assay included in the Oncolipsy PIK3CA kit. Only samples with acceptable DNA quality (Cq < 38) were further analyzed.

2.6. CTC-derived gDNA isolation

In PB samples from 35 mBrCa patients, EpCAM⁽⁺⁾ CTCs were isolated using positive immunomagnetic enrichment as previously described [29]. gDNA was extracted from EpCAM(+) CTC fractions using Trizol LS reagent (Invitrogen™, Carlsbad, CA, USA) as previously described [[29], [30], [31]]. DNA quantification in all samples was performed using NanoDrop™ 1000 Spectrophotometer (Thermo Fisher Scientific). CTC-derived gDNA quality was assessed as above by amplification of a 70 bp region of PIK3CA exon 3 using the assay included in the Oncolipsy PIK3CA kit. Only samples with acceptable DNA quality (Cq < 38) were further analyzed.

2.7. PIK3CA mutation analysis using the Oncolipsy PIK3CA CE-IVD kit

The Oncolipsy PIK3CA kit was used for the detection of PIK3CA hotspot mutations (p.E542K, p.E545K, p.E545Q, p.H1047R) in FFPEs, plasma-derived cfDNA and CTC-derived gDNA samples. The Oncolipsy PIK3CA CE-IVD kit is a ready-to-use real-time qualitative PCR test that is based on the combination of allele-specific priming, asymmetric PCR, and melting curve analysis. The kit can be applied on DNA samples extracted from CTCs, plasma and FFPE tumor tissues samples of metastatic breast cancer patients. Each kit can detect four different mutations (p.E542K, p.E545K, p.E545Q, p.H1047R) in single-plex assays, and a 70 base pair region of PIK3CA exon 3 as a sample quality control. The Oncolipsy PIK3CA kit includes sufficient reagents to analyze 24 samples including PCR positive and negative controls. It contains PCR Buffer mix, fluorescent dye, master-mixes with primers and probes as well as positive controls specific to each individual mutation. Five different PCR mixes were prepared depending on the target; However, all PCR reaction mixes had identical composition. Each PCR reaction mix consisted of 10 μl Oncolipsy PIK3CA MMX1, 2 μl Oncolipsy PIK3CA MMX2 (previously diluted in nuclease-free water according to manufacturer's instructions with volumes adjusted to the number of reactions and the target), 2 μl Oncolipsy PIK3CA Dye, 4 μl Nuclease-free water and 2 μl of DNA sample (up to 50 ng/PCR DNA input) in a total volume of 20 μl. PCR thermal cycling was performed using different programs depending on the target. All PCR reactions started with an initial denaturation at 95 °C for 2min (Ramp Rate 4.40 °C/s). Subsequent cycling conditions varied as follows.

• •SQC: amplification step: 45 cycles of 95 °C for 5s (Ramp Rate 4.40 °C/s), 63 °C for 10s (Ramp Rate 2.20 °C/s), 72 °C for 10s (single acquisition mode, Ramp Rate 4.40 °C/s), followed by one cycle of melting step: 55 °C for 30s (Ramp Rate 2.20 °C/s), 95 °C for 0s (continuous acquisition mode, Ramp Rate 0.11 °C/s, 5 Acquisitions per ^oC) and 40 °C for 20s (Ramp Rate 2.20 °C/s).
• •E542K: amplification step: 70 cycles of 95 °C for 1s (Ramp Rate 4.40 °C/s), 63 °C for 4s (Ramp Rate 2.20 °C/s), 72 °C for 1s (single acquisition mode, Ramp Rate 4.40 °C/s), followed by one cycle of melting step: 50 °C for 1min (Ramp Rate 2.20 °C/s), 95 °C for 0s (continuous acquisition mode, Ramp Rate 0.11 °C/s, 5 Acquisitions per ^oC) and 40 °C for 30s (Ramp Rate 2.20 °C/s).
• •E545K: amplification step: 70 cycles of 95 °C for 1s (Ramp Rate 4.40 °C/s), 62 °C for 4s (Ramp Rate 2.20 °C/s), 72 °C for 1s (single acquisition mode, Ramp Rate 4.40 °C/s), followed by one cycle of melting step: 50 °C for 1min (Ramp Rate 2.20 °C/s), 95 °C for 0s (continuous acquisition mode, Ramp Rate 0.11 °C/s, 5 Acquisitions per ^oC) and 40 °C for 30s (Ramp Rate 2.20 °C/s).
• •E545Q: amplification step: 80 cycles of 95 °C for 1s (Ramp Rate 4.40 °C/s), 62 °C for 4s (Ramp Rate 2.20 °C/s), 72 °C for 1s (single acquisition mode, Ramp Rate 4.40 °C/s), followed by one cycle of melting step: 45 °C for 1min (Ramp Rate 2.20 °C/s), 95 °C for 0s (continuous acquisition mode, Ramp Rate 0.11 °C/s, 5 Acquisitions per ^oC) and 40 °C for 30s (Ramp Rate 2.20 °C/s).
• •H1047R: amplification step: 70 cycles of 95 °C for 1s (Ramp Rate 4.40 °C/s), 62 °C for 4s (Ramp Rate 2.20 °C/s), 72 °C for 1s (single acquisition mode, Ramp Rate 4.40 °C/s), followed by one cycle of melting step: 53 °C for 1min (Ramp Rate 2.20 °C/s), 95 °C for 0s (continuous acquisition mode, Ramp Rate 0.11 °C/s, 5 Acquisitions per ^oC) and 40 °C for 30s (Ramp Rate 2.20 °C/s).

The whole procedure was performed according to the detailed protocol provided by the manufacturer's instructions without any modifications, using the kit-compatible LightCycler®480 instrument (IVD instrument, Roche Diagnostics, Germany). Each run included one negative WT (wild-type) and one positive control provided in the kit for each mutation tested. Melting curve analysis was performed to determine the presence of PIK3CA mutations. Samples were classified as positive when two distinct melting peaks were observed under different temperature conditions. Mutation identification was based on the lower-temperature melting peak, which corresponds to the mutation-specific amplicon. The characteristic melting temperatures (Tm) used for interpretation were as follows: p.E542K: 59–62 °C, p.E545K: 58–60 °C, p.E545Q: 54–56 °C and p.H1047R: 60.5–62.5 °C.

2.8. PIK3CA mutation analysis using the cobas®PIK3CA mutation test

Thirty gDNA samples from FFPEs, analyzed for PIK3CA mutations using the Oncolipsy PIK3CA kit, were also analyzed by the cobas®PIK3CA mutation test (Roche Diagnostics). The real-time PCR cobas®PIK3CA Mutation test kit employs primers (85 to 155bp) targeting PIK3CA exons 1, 4, 7, 9, and 20, divided into three mixes for each sample and controls. According to the manufacturer, the test demonstrates >95% hit rate for variants with mutation levels of 0.7–3.5%, using 50 ng DNA input per PCR.

2.9. PIK3CA mutation analysis using droplet digital PCR (ddPCR)

A direct comparison study between the Oncolipsy PIK3CA kit and the PrimePCR™ ddPCR™ Mutation Detection Assay Kits targeting the PIK3CA p.E542K, p.E545K and p.H1047R mutations (#1863131, #1863132, #1863133, Bio-Rad Laboratories, Hercules, CA, USA) was performed using 55 plasma-derived cfDNA samples. For each reaction, the PCR mixture contained 10 μL 2 × ddPCR Supermix for Probes (No dUTP) (Bio-Rad Laboratories, Hercules, CA, USA), 1 μL 20x target primers/probe (FAM), 1 μL 20x reference primers/probe (ΗΕΧ), and 2 μL of cfDNA sample, adjusted to a final volume of 20 μL with sterile water, according to the manufacturer's instructions. Thermal cycling conditions of the three PIK3CA mutant detection assays were the same as follows: 95 °C for 10 min (1 cycle), 94 °C for 30 s, and 55 °C for 1min (40 cycles), and infinite hold at 12 °C. Each run included a wild-type (WT) negative control and a positive control consisting of cfDNA reference standards with defined allele frequencies. Finally, all samples were analyzed on a Bio-Rad QX200 droplet reader and the data were analyzed using QuantaSoft Software. A positivity threshold of 3 mutant droplets was applied for all three PIK3CA hotspot mutations.

2.10. Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics, version 28.0 (IBM Corp., Armonk, NY, USA). Cohen's kappa coefficient test was used to estimate the concordance between the Oncolipsy PIK3CA kit with cobas®PIK3CA mutation test and PrimePCR™ ddPCR™ Mutation Detection Assay Kits. A p value < 0.05 was considered statistically significant.

3. Results

3.1. Analytical validation of the Oncolipsy PIK3CA kit using Liquid Biopsy Certified Reference standards

The analytical validation of the Oncolipsy PIK3CA kit was based on the evaluation of analytical detectability, analytical specificity and precision, using Liquid Biopsy Certified Reference standards.

3.2. Analytical detectability

For the evaluation of analytical detectability, we firstly estimated the limit of detection (LOD) for each PIK3CA mutation using serial dilutions of synthetic mutant oligonucleotides (gene-blocks) in synthetic wild-type strands generating different ratios of 50%, 10%, 1%, 0.5% and 0.05% for each PIK3CA mutation. Μutant allele frequencies (MAF) of 50%, 10% and 1% were analyzed once, MAF of 0.5% was analyzed three times while the lower MAF of 0.05% was analyzed 6 times. In parallel, wild type (WT) and negative control were analyzed. Our results showed clearly that the Oncolipsy PIK3CA kit could reliably detect a MAF of 0.05% for all the mutations tested (Fig. 2A).

Fig. 2.

Α) Analytical detectability of the PIK3CA mutation assays of the kit using serial dilutions of synthetic oligonucleotides containing each of the four PIK3CA mutations mixed with synthetic WT sequences at allele ratios of 50%, 10%, 1%, 0.5%, 0.05%, and 0% (WT). (B, C) Representative melting curves for the p.E545K, p.E542K, and p.H1047R mutations generated using cfDNA reference material across multiple AFs. (D) Representative melting curves for the p.E545K mutation at 1% and 0% AF in a synthetic matrix.

Based on these results, we further investigated the analytical detectability of the assay using two types of cfDNA reference standards: (i) a Multiplex cfDNA reference standard containing all tested mutations, and (ii) a single-mutation reference standard harboring only the p.E545K variant at predefined AF values. The Oncolipsy PIK3CA kit successfully detected the p.E542K, p.E545K and p.H1047R mutations at 12.5% AF in the Multiplex cfDNA standard (Fig. 2B). The limit of detection (LOD) was determined at 1% AF for p.E542K, 1% AF for p.E545K, and 0.8% AF for p.H1047R (Fig. 2C). Using the single-mutation reference standard, the assay consistently detected the p.E545K mutation at 1% AF (Fig. 2D). No PIK3CA hotspot mutations were detected in the wild-type (WT) cfDNA reference standard.

3.3. Analytical specificity

The analytical specificity was evaluated for each single-plex assay of the Oncolipsy PIK3CA kit for cross reactions using LB-reference standards of 10% MAF for other PIK3CA mutation sites in triplicate. In each run, the WT synthetic oligonucleotide sequence corresponding to each exon region was also analyzed (n = 3). Non-specific PCR products were not detected in any case, demonstrating the high specificity of the assay for each mutation analyzed (Fig. 3). The analytical specificity was further confirmed through the analysis of a WT cfDNA certified reference standard, since no PIK3CA mutations were detected.

Fig. 3.

Analytical specificity of the PIK3CA mutation assays of the kit.

3.4. Analytical precision

Analytical precision was evaluated using three independent reagent lots of the Oncolipsy PIK3CA kit. Intermediate precision (within laboratory reproducibility) and repeatability were assessed based on the Tm values obtained from serial dilutions of synthetic PIK3CA mutant oligonucleotides (gene blocks) containing 50%, 10%, and 1% MAF in a WT background. Repeatability was evaluated by analyzing three replicates of each dilution within a single run, whereas intermediate precision was assessed by performing three independent analytical runs in three separate days using all three reagent lots. A WT reference standard was also analyzed in triplicate in each run. No PIK3CA hotspot mutations were detected in the WT-certified material across all runs. The assay demonstrated excellent precision across all mutations and lots tested. Tm values were highly consistent with CV% ranged from 0.04% to 0.41% for 50% MAF, 0.01% to 0.56% for 10% MAF, and 0.02% to 1.4% for 1% MAF (Suppl Table 1). Overall intermediate precision across all batches and experimental days remained robust, with CV% ranging from 0.36% to 0.96% for 50% MAF, 0.34% to 0.93% for 10% MAF and 0.21% to 1.10% for 1% MAF (Supp.Table 1) confirming high precision of the Oncolipsy PIK3CA kit for all tested hotspot mutations.

3.5. Clinical validation

3.5.1. Detection of PIK3CA hotspot mutations in FFPEs

gDNA from 30 primary tumors (FFPEs) were analyzed using the Oncolipsy PIK3CA kit. The H1047R mutation was the most prevalent, identified in 13/30(43.3%) of samples, followed by p.E542K in 9/29(31.0%) and p.E545K in 4/30(13.3%), whereas the p.E545Q was not detected in any case (Fig. 4A). The same FFPE samples were in parallel analyzed with the cobas®PIK3CA mutation kit (Roche Diagnostics). In this case, the p.H1047R mutation was detected in 9/30 (30%) samples, p.E545K in 6/30 (20%), while the p.E542K was not detected in any sample 0/29 (0%) and the p.E545Q was also not detected in any case.

Fig. 4.

Analysis of PIK3CA hotspot mutations in A) primary tumors (FFPEs gDNA) and B) plasma-cfDNA from early and metastatic breast cancer.

For the p.E542K mutation, the concordance between the two commercially available real-time PCR assays was 20/29(69%) (Table 1). Nine samples were classified as positive by the Oncolipsy PIK3CA kit but reported as negative by the cobas®PIK3CA assay (Table 1). To investigate these discordant results, the nine samples were further analyzed using ddPCR (BioRad). Eight of these nine samples were found positive by ddPCR, supporting the higher analytical detectability of the Oncolipsy PIK3CA kit compared to the cobas assay.

Table 1.

Direct comparison between the Oncolipsy PIK3CA CE-IVD kit and the cobas®PIK3CA Mutation Test kit for the detection of p.E542K, p.E545K and p.H1047R PIK3CA mutations in primary tumor samples (FFPEs).

		cobas®PIK3CA Mutation Test kit
Oncolipsy PIK3CA CE-IVD kit	p.E542K		+	-	Total
		+	0	9	9
		-	0	20	20
		Total	0	29	29
		Concordance: 20/29 (69%), k = 0.000
	p.E545K		+	-	Total
		+	4	0	4
		-	2	24	26
		Total	6	24	30
		Concordance:28/30 (93%), k = 0.762, p < 0.001
	p.H1047R		+	-	Total
		+	8	5	13
		-	1	16	17
		Total	9	21	30
		Concordance:24/30 (80%), k = 0.577, p < 0.001

For the p.E545K mutation, the concordance between the two commercially available real-time PCR kits was 28/30(93.3%) (Table 1). Two samples were reported as positive by the cobas®PIK3CA mutation test but negative using Oncolipsy PIK3CA kit. To elucidate this discrepancy, these two samples where further analyzed using ddPCR. ddPCR confirmed one of these two as negative. This discordance is likely attributable to the broader mutation coverage of the Cobas assay, which detects multiple p.E545X variants (p.E545A, p.E545D, p.E545G, or p.E545K), rather than exclusively the p.E545K variant.

For the p.H1047R mutation, the concordance between the two assays was 24/30 (80.0%) (Table 1). Five samples were detected as positive using the Oncolipsy PIK3CA kit but were reported as negative by the cobas®PIK3CA mutation test, whereas one sample, found negative by the Oncolipsy kit was identified as positive by the cobas assay (Table 1). This discrepancy could possibly be explained to the broader mutation coverage of the cobas assay, which targets p.H1047X variants (p.H1047L/R/Y) as opposed to the Oncolipsy kit, which specifically detects the p.H1047R variant. Further evaluation of the five discordant Oncolipsy-positive/cobas-negative cases using ddPCR confirmed four of these as true positives, while one was negative. These findings suggest that the Oncolipsy PIK3CA kit exhibits increased analytical sensitivity for the p.H1047R subtype compared to the cobas® assay.

3.5.2. Detection of PIK3CA hotspot mutations in plasma-cfDNA

We analyzed 55 plasma-cfDNA samples from ER+/HER2− breast cancer patients, including 20 patients with early and 35 patients with metastatic disease. Overall, PIK3CA hotspot mutations were detected in 14/55(26%) samples, comprising p.E545K, p.E542K and p.H1047R variants. Detection frequencies were lower in early-stage breast cancer (15%, 3/20) compared with metastatic disease (31%, 11/35), with the exception of p.E545K, which showed similar detection rates between the two groups; however, this finding should be interpreted cautiously due to the small number of early-stage samples. Across all cases, the p.H1047R was the most prevalent mutation (18.2%), followed by p.E545K (7.3%) and p.E542K (3.6%), while the p.E545Q was not detected in any sample (Fig. 4B).

Α direct comparison of the Oncolipsy PIK3CA kit and ddPCR (BioRad, US) was performed for all these 55 plasma-cfDNA samples (Table 2). The cut-off value was established based on the number of positive droplets detected in the healthy donor cohort for each mutation (n = 7). A threshold of 3 droplets was defined; accordingly, samples exhibiting ≥4 positive droplets were classified as mutation-positive. High concordance was observed for all mutations, with κ values indicating substantial agreement across assays (ranged from 0.411 to 0.848). Discrepancies between these two highly sensitive methodologies included isolated cases where mutations were selectively detected by each assay. These cases were confirmed upon re-analysis and are most likely attributable to stochastic sampling effects that are present at very low mutant allele frequencies (Poisson distribution), a recognized limitation in liquid biopsy testing at low-DNA input levels.

Table 2.

Direct comparison between the Oncolipsy PIK3CA CE-IVD kit and ddPCR for the detection of p.E542K, p.E545K and p.H1047R PIK3CA mutations in plasma-cfDNA.

		ddPCR PIK3CA Mutation Test kit
Oncolipsy PIK3CA CE-IVD kit	p.E542K		+	-	Total
		+	1	1	2
		-	1	41	41
		Total	2	42	44
		Concordance:42/44 (95%) (k = 0.478)
	p.E545K		+	-	Total
		+	3	0	3
		-	1	51	52
		Total	4	51	55
		Concordance: 54/55 (98%) (k = 0.848)
	p.H1047R		+	-	Total
		+	3	7	10
		-	0	44	44
		Total	3	51	54
		Concordance: 47/54 (87%) (k = 0.411)

3.5.3. Detection of PIK3CA hotspot mutations in CTC-derived gDNA samples

It was recently shown that information derived from CTC could be complementary to the cfDNA analysis, and is highly important such as in the case of another BrCa molecular biomarker, ESR1 [18]. We further investigated the applicability of the Oncolipsy PIK3CA kit in CTC-derived gDNA samples. Eighteen gDNA samples isolated from EpCAM(+) CTC fractions was available from a subset of eighteen BrCa patients and tested with the kit. In these samples p.E542K was detected in 1/13(7.7%), p.H1047R in 4/17(24%), while the p.E545K was not detected in any of the 18 samples.

4. Discussion

PIK3CA mutations are of high significance in personalized cancer treatment, especially since alpelicib started to be widely used in the treatment of ER+/HER2- MBC patients. In recent years an increasing number of treatment strategies targeting PIK3CA mutations is available, such as the triplet therapy with palbociclib, taselisib, and fulvestrant in PIK3CA-mutant BC [10]. Thus the detection of PIK3CA hotspot mutations becomes a mandatory tool to guide treatment's choice.

Beyond its role in breast cancer, PIK3CA mutation profiling is gaining importance in other types of cancer including colon cancer, gastric cancer, cervical cancer, prostate cancer, and lung cancer [32], reflecting its potential utility in precision oncology. For instance, Ntzifa et al. analyzed plasma-cfDNA and paired CTC-gDNA samples for PIK3CA mutations (p.H1047R, p.E542K, p.E545K) at two time points (baseline and disease progression) in NSCLC patients, detecting the specific mutations in 15.4% of cfDNA samples at baseline and 16.7% at disease progression, demonstrating the feasibility of liquid biopsy for monitoring of PIK3CA alterations in NSCLC patients [33]. Additionally, Markou et al. assessed the mutation status of ctDNA and paired CTCs in patients with early-stage NSCLC, revealing that PIK3CA was the most frequently mutated gene in both CTCs and plasma cfDNA, at baseline and at the time of relapse [34]. Recent ctDNA profiling of 1700 plasma samples from patients with advanced colorectal cancer revealed the presence of PIK3CA mutations in approximately 19.2% of cases, underscoring the clinical relevance of these mutations and supporting the development of targeted therapeutic strategies [35]. Overall, these findings among others are highlighting the detection of PIK3CA mutations through liquid biopsy and could thus provide a valuable, non-invasive tool for monitoring tumor molecular dynamics and guiding targeted therapy across multiple tumor types, not only in breast cancer.

PIK3CA mutations have already been studied in liquid biopsy samples and tumor tissues using NGS and ddPCR [[36], [37], [38], [39]]. Despite the potential of these technologies, there is still a crucial need for molecular diagnostic assays for the detection of PIK3CA mutations that are highly sensitive, rapid, inexpensive, and suitable for high throughput liquid biopsy analysis.

In the present study we validated the analytical and clinical performance of a novel commercially available CE-IVD kit for PIK3CA hotspot mutations. The Oncolipsy PIK3CA kit is designed for the detection of the four most common hotspot mutations of PIK3CA, in exons 9 and 20, that are all of clinical importance in various types of cancer [40]. Based on the analysis of certified liquid biopsy reference standards at specific MAF rates, we verified that the Oncolipsy PIK3CA kit successfully detects the p.E542K, p.E545K and p.H1047R mutations in all samples with very low MAF rates (for p.E542K: MAF 1%, for p.E545K: MAF 5% and 1% and for p.H1047R: MAF 0,8%).

In primary tumors (FFPEs), this kit has shown a higher PIK3CA positivity rate than the cobas®PIK3CA Mutation Test kit likely reflecting the latter's sensitivity threshold (5% mutant sequence in 50 ng of DNA) in BrCa patients. This observation is consistent with previously reported findings [41]. Moreover, the frequency of PIK3CA mutations in exon 20 (43.3%) was higher (36.7%) than in exon 9 in total (p.E545K and/or p.E542K mutant). In FFPEs, as expected, PIK3CA mutation frequency was higher when compared to plasma-cfDNA. Direct comparison of the results derived with the kit in FFPEs with those derived using the CE-IVD cobas®PIK3CA mutation test for the same FFPE samples revealed a better sensitivity for the Oncolipsy PIK3CA kit for all three PIK3CA hotspot mutations.

In plasma-cfDNA, the rates of PIK3CA positivity are relatively low both in early and metastatic breast cancer groups. In early breast cancer, using the Oncolipsy PIK3CA kit the p.E545K mutation was detected in 10% of plasma-cfDNA samples, the p.H1047R mutation in 15.0%, while both p.E542K and p.E545Q were not detected in any plasma-cfDNA sample tested. In metastatic breast cancer, the p.E545K mutation was detected in 5.7% of plasma-cfDNA samples, the p.H1047R mutation in 20.0% and p.E542K mutation in 5.7%. In our dataset for the early breast cancer group (n = 20), the p.E545K mutation was detected more frequently than the other hotspot mutations, including p.H1047R. Interestingly, the predominance of the E545K mutation in our cohort is consistent with findings reported by Tzanikou et al., who analyzed plasma-ctDNA from patients with early breast cancer (n = 77). In their study, the p.E545K hotspot mutation was detected in 30 out of 77 samples (39.0%), whereas the p.H1047R mutation was identified in only 7 out of 77 samples (9.1%) [30]. Nevertheless, given the small size of our cohort, this observed higher frequency of p.E545K should be interpreted cautiously, and further studies with larger sample sizes are warranted to confirm these findings. Direct comparison of the results derived with the Oncolipsy PIK3CA kit with those derived using ddPCR (BioRad) for the same plasma-cfDNA samples revealed that these two assays give comparable results for all these three mutations. More specifically, in the early breast cancer group, one sample that was found positive for p.E545K mutation using the ddPCR assay, was found negative with Oncolipsy PIK3CA kit. This sample was found to have positive droplets which is close to the determined cut-off of the ddPCR. Moreover, two plasma-cfDNA samples that were found positive for p.H1047R mutation by the Oncolipsy PIK3CA kit, were found negative by ddPCR as they have 0 and 1 droplet indicating the absence of positive events for this mutation. In the metastatic group, two samples were found positive for the p.E545K by both assays, one sample that was found positive for p.E542K mutation by the Oncolipsy PIK3CA kit, was found negative by ddPCR, while one sample that was found positive for the same mutation by ddPCR, was found negative by Oncolipsy PIK3CA kit. For the p.H1047R, the Oncolipsy PIK3CA kit detected five positive samples, that were found to have between 0 and 3 droplets in ddPCR, a result indicating lack of positive events, while for the p.E545K, two plasma-cfDNA samples were found positive by both assays. All the above results showed that the Oncolipsy PIK3CA kit has similar sensitivity to ddPCR and can provide accurate results in plasma-cfDNA samples in early and metastatic breast cancer. Our validation has shown that this new CE-IVD kit is highly sensitive, specific and reliable and can be used in the clinical routine setting. When compared to NGS and ddPCR this assay is high throughput and time-saving, while another major advantage is its significantly lower cost, simpler set-up, easier interpretation of results, and compatibility with widely used Real-Time PCR instrumentation.

The Oncolipsy PIK3CA CE-IVD kit targets the most clinically relevant and frequently actionable PIK3CA hotspot mutations in metastatic breast cancer. Although this focused approach does not capture the full molecular heterogeneity of the disease, it enables the identification of a substantial proportion of patients eligible for targeted therapies and retains significant clinical utility, particularly in settings requiring rapid, cost-effective, and high-throughput molecular screening. We fully acknowledge that the Oncolipsy PIK3CA CE-IVD kit is based on a predefined and limited mutation panel and therefore poses a risk to leave behind a group of BC patients eligible to target treatments. However, this group is expected to be small, since the mutations targeted by the Oncolipsy assay (p.E542K, p.E545K, p.E545Q, and p.H1047R) represent the most frequent and clinically actionable PIK3CA hotspot alterations in breast cancer, particularly in the hormone receptor–positive metastatic setting, where therapeutic decisions are currently driven by hotspot mutation status rather than rare variants. Consequently, while a small subset of patients harboring non-hotspot or less frequent alterations may not be identified by this approach, the assay is expected to detect a substantial proportion of patients eligible for PIK3CA-targeted therapies. The intended clinical use of this assay is to serve as a rapid, cost-effective, and highly sensitive screening tool that allows optimal patient stratification while balancing turnaround time, cost, and analytical sensitivity.

The concordance between the Oncolipsy PIK3CA CE-IVD kit, ddPCR and the cobas assay did not reach complete agreement, as expected. A potential explanation could be the predefined mutation panel of the Oncolipsy PIK3CA CE-IVD kit. More specifically, in primary tumor tissues, samples that tested negative with the Oncolipsy PIK3CA CE-IVD kit but were positive with the cobas® PIK3CA Mutation Test were further evaluated using digital PCR. This analysis revealed that these cases harbored alternative variants at the same genomic loci, which are not detected by the Oncolipsy assay, indicating differences in mutation coverage rather than true false-negative results. Notably, when restricted to the mutations specifically targeted by the assay, the Oncolipsy PIK3CA CE-IVD kit demonstrated higher analytical sensitivity than the cobas® PIK3CA Mutation Test in tissue samples. Most variants detected by the Oncolipsy kit were confirmed by digital PCR, supporting the analytical reliability of the detected alterations.

In plasma-derived cfDNA samples, concordance between the Oncolipsy PIK3CA CE-IVD kit and ddPCR was high across all mutation groups, further supporting the robustness of the assay for liquid biopsy applications. Nevertheless, mutant allele frequency (MAF) remains a critical determinant of detectability, particularly at very low allele fractions approaching the analytical limits of detection of each methodology. Under these conditions, stochastic sampling effects and methodological differences may contribute to residual discrepancies across platforms.

In conclusion, the Oncolipsy PIK3CA CE-IVD kit is highly sensitive and specific for the detection of four PIK3CA hotspot mutations in primary tumors, CTCs, and plasma-cfDNA samples. This kit could be used for a reliable PIK3CA mutation screening in breast cancer patients as a first step to clarify the treatment decision. The kit shows potential for broader application in detecting PIK3CA mutations across different cancer types, although further validation is required. The clinical performance of the Oncolipsy PIK3CA kit will be prospectively evaluated in a large number of well-characterized patient cohorts.

Author contributions

D.S., S.S. and E.T. carried out the experiments; A.M. and E.L. developed the PIK3CA assay; L.K., I.B., H.L., F.Z., E.R., S.K., A.P., V.G., and C.P. provided the clinical samples used in this study; A.M, D.S. and S.S. took the lead in writing the manuscript; E.L. designed the experiments, provided critical feedback and helped shape the research, analysis and manuscript. All authors have read and agreed to the published version of the manuscript.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: