Introduction

Lung cancer remains the leading cause of cancer-related deaths worldwide. In Brazil, it is the second most common cancer among men and the fourth among women1. According to the American Cancer Society, non-small cell lung cancer (NSCLC) accounts for 80-85% of lung cancer cases, while small cell lung cancer represents the remaining 15-20%. Beyond histological and clinical diagnosis, molecular profiling has become an essential component to identify specific NSCLC subtypes. The detection of biomarkers in tissue and/or blood is now a cornerstone of precision oncology2. Identifying actionable driver mutations and fusions enables the use of targeted therapies, which have demonstrated superior response rates, improved survival outcomes, and distinct toxicity profiles compared to platinum-based chemotherapy3. To accurately detect the expanding list of targetable alterations, Next-Generation Sequencing (NGS) assays, particularly comprehensive genomic profiling (CGP), have emerged as the gold standard4.

Tissue biopsy remains the gold standard for molecular testing in NSCLC5, as it is needed for histopathological diagnosis and Programmed Death-Ligand 1 (PD-L1) expression analysis. First-line therapies are defined according to histology, pertinent positive or negative results for driver alterations plus PD-L1 status. However, tissue-based molecular profiling has significant limitations6. Adequate genomic data from tissue samples is not always feasible. Preanalytical challenges, such as insufficient nucleic acid quantity (common in small biopsies or upfront histopathological diagnosis) or poor deoxyribonucleic acid and ribonucleic acid (DNA/RNA) quality7,8 may impair test performance. Furthermore, the turnaround time for tissue-based testing can delay treatment decisions, impacting patient outcomes9.

Liquid biopsy has emerged as a less invasive alternative for detecting actionable alterations through circulating tumor DNA (ctDNA), which is released into body fluids via necrosis or apoptosis of cancer cells10. This approach offers several advantages, including the ability to assess spatial and temporal tumor heterogeneity and reduce turnaround times, enabling faster molecular diagnoses and earlier treatment initiation compared to tissue biopsies11. Additionally, it can function as a complementary assay and may serve as a rescue strategy when tissue-based NGS fails. Despite that, liquid biopsy also faces limitations. Sensitivity for complex alterations and fusions is lower than tissue NGS, and limited availability or accessibility related to high cost restricts its widespread use, particularly in low- and middle-income countries12. Additionally, non-informative results are common in scenarios with low tumor burden, or suboptimal timing of sample collection relative to therapy initiation, all of which can significantly impact test sensitivity12.

In resource-limited settings, molecular testing faces even greater challenges, where not only tissue inadequacy but also operational barriers and implementation gaps often limit access to comprehensive biomarker testing. In Brazil, up to 35% of tumors may be eligible to single-gene non-NGS assays only due to NGS failures13, potentially excluding patients from targeted therapies. These limitations underscore the unmet need for a molecular testing alternative that is both accurate and feasible for use with small or low tumor cell content samples.

PlexBio uses tiny silicon discs called πCODE MicroDiscs, each just 40 micrometers in diameter. Every disc has a unique pattern imprinted on its surface, acting like a microscopic barcode. This allows multiple tests to be run simultaneously in a single sample (multiplexing). To perform a multiplex test, discs coated with specific nucleic acid probes (designed to capture different targets) are mixed. This mixture is exposed to pre amplified samples for target regions of interest. During this step, targets in the sample bind to their matching probes on the discs (hybridization), and fluorescent labels attach simultaneously to mark the bound targets. After binding and labeling, the disc mixture is analyzed by the PlexBio 100 Analyzer. This instrument uses a Charge-Coupled Device (CCD) camera to perform two key tasks: first, it reads each disc’s unique pattern under bright-field illumination to identify which target is being detected. Second, it measures the intensity of the fluorescent signal on each disc under dark-field illumination to quantify how much of the target is present. The DeXipher software then processes this data. It cross-references the disc’s identity (from its pattern) with its fluorescence signal intensity. This turns the raw imaging data into an automated, user-friendly report with interpreted results. The system supports tests for single genes or entire gene panels. For example, their KRAS test, clinically validated for use in colorectal cancer, is available using this technology14.

The IntelliPlex Lung Cancer Panel DNA and RNA utilizes π-CODE technology to perform multiplex reactions, enabling the evaluation of 74 single-nucleotides variations (SNVs) or insertions/deletions (indels) across 8 genes: KRAS, NRAS, PIK3CA, BRAF, EGRF, ERBB2, MEK1, and AKT1 and 28 fusion variants in 5 genes: ALK, ROS1, RET, NTRK1, and MET. This approach offers a faster laboratory workflow compared to NGS7,8, providing comprehensive detection of actionable biomarkers in NSCLC. With its requirement for minimal DNA and RNA input, the IntelliPlex panel has the potential to address key limitations of current workflow, particularly in cases of small or limited tissue samples. In this setting, the objective of this study is to evaluate the performance of the IntelliPlex Lung Cancer Panel DNA and RNA kits against results obtained from a comprehensive NGS-based panel.

Results

Concordance assessment

Out of 60 samples, 25 were analyzed for concordance in IntelliPlex Lung Cancer Panel DNA and 34 in IntelliPlex Lung Cancer Panel RNA. Considering the DNA panel, we observed 98% agreement between the results obtained from the CGP NGS and IntelliPlex Lung Cancer Panel DNA (Table 1).

Table 1 Concordance analysis of variants identified by both NGS and IntelliPlex Lung Cancer Panel DNA tests.
Full size table

It is important to mention the two inconsistencies in the concordance analysis for the EGFR exon 20 p.Thr790Met variant presented a limitation regarding the allelic frequency (samples 2022891, 2118817). In the NGS assay, VAF greater than 1% is reported while the manufacturer’s limit of detection for this variant with IntelliPlex Lung Cancer Panel DNA is 3.17%. Therefore, these discordances were not considered as false negatives, and both were excluded from the final analysis. On the other hand, there was disagreement in the nomenclature regarding the detection of an insertion in exon 20 of EGFR (sample 1076215.) While in NGS the alteration was detected as p. A767_V769dup, the IntelliPlex Lung Cancer Panel DNA alteration was reported as p.D770_N771insSVD or p.V769_D770insASV.

A real discordance occurred with the reference control HD827 when IntelliPlex Lung Cancer Panel DNA failed to detect two co-occuring KRAS mutations in exon 2 (KRAS p.G12D and KRAS p.G13D) (Table 1). It is possible that the probes compete for this region that carry two different mutations 3 base pairs apart from each other. However, in a second experiment (LOD experiment), we were able to detect both mutations, suggesting a punctual discordance. The IntelliPlex Lung Cancer Panel–RNA, presented 100% agreement between the results obtained by NGS, as described in Table 2.

Table 2 Comparison of variants detected by NGS and IntelliPlex Lung Cancer panel RNA.
Full size table

Considering a high pre analytical failure rate for RNA in NGS, we tested 13 samples that did not have quality/quantity metrics for our NGS panel. Basically, cases with insufficient RNA input (< 200ng) and samples that presented a cycle threshold (Ct) higher than 28 in a qPCR quality check test were considered as QC failure. Of these, 61.5% (8/13) presented valid results for the IntelliPlex Lung Cancer Panel RNA, one of which was positive for ROS1 fusion, which was orthogonally confirmed by FISH. The remaining 38.5% (5/13) resulted in failure in one or more controls, confirming the low quality of the extracted material (Table 3).

Table 3 Comparison of low-quality samples for NGS and IntelliPlex Lung Cancer panel RNA.
Full size table

In summary, the test presented a satisfactory concordance with 97,73% sensitivity, 100% specificity and 98,15% accuracy (Table 4).

Table 4 Sensitivity, specificity and accuracy evaluation for the IntelliPlex Lung Cancer panel DNA and RNA.
Full size table

Limit of detection (LOD)

In order to determine the limit of detection (LOD) for DNA mutations, serial dilutions were performed using the reference control OncoSpan gDNA Reference Standard HD827 (Horizon Dx) in a wild type sample (NA12878, Coriell Institute) at 50%, 25%, 12.5%, 6.25% and 3.12%, as described in Table 5. Based on these results, the LOD of the assay was determined at 5% VAF. It is important to mention it was observed to have difficulty in detecting KRAS G13D using reference control.

Table 5 Limit of detection analysis according to variants present in OncoSpan gDNA reference standard HD827.
Full size table

Discussion

In this study, we evaluated the performance of IntelliPlex Lung Cancer Panel DNA and RNA against the gold standard comprehensive NGS-based panel. The panel demonstrated high concordance rates of 98% for DNA and 100% for RNA, with an overall accuracy of 98.15%. The discordance found in the reference control containing two co-occurring KRAS mutations in exon 2 (KRAS p.G12D and KRAS p.G13D) can be partially explained by the possibility that the probes compete for this region that carry two different mutations 3 base pairs apart from each other. However, this does not occur frequently in the routine, since KRAS mutations are mutually exclusive. Notably, among samples that did not meet the quantity or quality metrics required for RNA sequencing, 61.5% still yielded valid results using the IntelliPlex panel. These findings highlight the panel’s robustness in generating reliable molecular results even in challenging preanalytical conditions, such as those involving limited nucleic acids. When compared to our broad NGS panels, which require 50 ng of DNA and 200 ng of RNA, the input of IntelliPlex panels is significantly lower (10 ng of DNA and 20 ng of RNA). The IntelliPlex panel’s capacity to deliver valid results with minimal nucleic acid input or suboptimal quality addresses a critical gap in molecular diagnostics, particularly in cases of preanalytical failures often associated with small biopsy samples or upfront histopathological diagnosis. Furthermore, the panel can be completed in about 5 h, and most laboratories are able to provide results within a 7-day turnaround when accounting for the entire workflow. In contrast, traditional methods like Next-Generation Sequencing (NGS) may require up to 14 days for results. This faster process can help accelerate treatment decisions, especially in urgent clinical situations. Other commercially available platforms such as Oncomine DxTT, an NGS-based method, require lower DNA/RNA input but has a longer total workflow turnaround time. AmoyDx most closely resembles the π-CODE system in hands-on time and workflow duration, though it demands higher minimum DNA/RNA input. (60 ng DNA; 120 ng RNA. https://www.amoydiagnostics.com/products/amoydx-pan-lung-cancer-pcr-panel).

These advantages not only enhance workflow efficiency but also reduce the rate of test failures, thereby increasing the detection of actionable biomarkers. In resource-limited settings, where access to advanced molecular testing is often limited, the IntelliPlex panel offers an alternative to improve diagnostic efficiency and patient care. To fully support the benefits of the IntelliPlex panel, it is essential to integrate it effectively into existing laboratory workflows and refine sample selection strategies. This requires close collaboration between pathology and genomics teams to optimize tissue allocation. Laboratories should prioritize preserving the minimum tissue necessary for accurate histopathological diagnosis while reserving samples with higher tumor cell content for comprehensive molecular testing via NGS. For cases with limited tissue quantity or quality, the π-CODE technology used in the IntelliPlex panel should be prioritized, as it is better suited to such challenging samples. This strategic approach increases the likelihood of successfully assessing the genomic landscape of patient samples and reduces the failure rate, ensuring more patients have access to molecular testing results. Even with optimized sample selection to minimize biospecimen-related NGS failure, other causes of NGS failure, such as test performance, remain a challenge15. Therefore, in cases where the traditional NGS fails, not only due to sample quality or quantity but also due to other technical limitations, the IntelliPlex panel can potentially serve as a viable rescue strategy. Given the existing failure rates of broad NGS panels unrelated to biospecimen performance, this approach provides an alternative pathway to molecular information.

Our study has limitations that may impact the routine use of the IntelliPlex panel in laboratory workflows. First, while the panel adequately covers driver mutations in NSCLC, it does not include co-occurring mutations such as TP53, KEAP1, STK11, and CDKN2A/B, which are increasingly recognized for their prognostic value16 and role in guiding treatment intensification strategies, as supported by recent data17,18,19. The absence of broader genomic coverage may complicate treatment decision-making, as it limits the ability to fully characterize the tumor’s molecular profile. Additionally, the panel demonstrated lower sensitivity in detecting EGFR p.Thr790Met, a well-established resistance biomarker to first- and second-generation EGFR tyrosine kinase inhibitors (TKIs) that is often subclonal20,21. This limitation is likely attributable to the panel’s higher limit of detection for this variant (3.17%) compared to NGS, which can detect variant allele frequencies as low as 1%. However, this reduced sensitivity is unlikely to significantly impact clinical practice, as current guidelines recommend osimertinib, a third-generation EGFR TKI, as a first-line treatment option for EGFR-mutated NSCLC, thereby diminishing the need for EGFR T790M detection in most cases. Third, while NSCLC samples wild-type for all drivers mutations represents only 5% of cases when assessed via broad NGS technology13, negative results with the IntelliPlex panel will require clear communication between the laboratory and treating physicians. Clinicians often expect to identify at least some type of mutation, even if non-driver mutations, and the absence of any detected alterations might need additional explanation, particularly during integration of the panel into daily clinical workflows. Therefore, the true value of the IntelliPlex panel will only be fully understood through its adoption in routine clinical practice, where its role in real-world practice and its acceptance by physicians, laboratory users, and patients can be comprehensively evaluated.

Continuous analytical improvements for NGS technology are ongoing to enhance test performance and reduce pre-sequencing failure rates. However, the IntelliPlex Lung Cancer Panel demonstrates clear value in various scenarios within the laboratory workflow, particularly in addressing challenges related to sample quality, quantity, and turnaround time. Finally, molecular tumor boards (MTBs) dedicated to discussing genomic alterations, as well as the panel’s capabilities and limitations, may serve as a strategic tool to enhance test acceptance and facilitate its definitive incorporation into the molecular diagnostic journey.

The IntelliPlex Lung Cancer Panel DNA and RNA demonstrated high concordance rates of 98% and 100%, respectively, with NGS for NSCLC-relevant variants. The panel’s faster turnaround time when compared to broad NGS, ability to analyze samples with limited tissue quantity, and utility as a rescue strategy following NGS failure highlight its potential to positively impact the molecular diagnostic journey. The lack of coverage of non-driver mutations and lower sensitivity in detecting EGFR: pThr790Met does not preclude its use in daily laboratory workflow. Molecular tumor boards and educational initiatives will be critical to facilitate the panel’s adoption into clinical practice, ultimately ensuring broader access to molecular information.

Methods

Sample acquisition and nucleic acid extraction

The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice (ICH E6R2) and in accordance with the laws and regulations applicable to Brazilian legislation for clinical research. Participants’ names and other HIPAA identifiers were not included in the analysis or report.

Sixty (60) FFPE samples (26 and 34 samples for DNA and RNA, respectively) derived from 53 patients diagnosed with advanced lung adenocarcinoma and 2 reference controls (DNA only) were selected according to the presence of actionable mutations or fusions in the genes described in the IntelliPlex Lung Cancer Panel DNA and RNA kits (PlexBio, Taiwan), previously detected by a validated NGS panel carried out in the genomics laboratory of Oncoclínicas Medicina de Precisão (OCPM) using SureSelect CGP (Agilent) in Illumina NovaSeq and MGI G400 sequencing platforms. TNA (total nucleic acid) were extracted using the ReliaPrep FFPE Total RNA Miniprep System – Promega kit, without the use of DNAse. Nucleic acid quantification was carried out by fluorometric method using Qubit dsDNA Assay Kit for DNA and Qubit RNA HS Assay Kit for RNA. In this study, DNA extractions were used with quantification greater than 1 ng/µL (20 ng) and greater than 10 ng/µL (10 ng) for RNA.

The determination of the limit of detection (LOD) was performed with the OncoSpan gDNA Reference Standard - HD827 control. From the initial concentration (100%), serial dilutions were carried out at 50%, 25%, 12.5%, 6.25% and 3.12%, diluted in a wild-type sample (NA12878, Coriell Institute).

Mutation analysis by IntelliPlex Lung Cancer panel

The IntelliPlex Lung Cancer Panel DNA qualitatively evaluates 74 mutations in the KRAS, NRAS, PIK3CA, BRAF, EGRF, ERBB2, MEK1 and AKT1 genes and the IntelliPlex Lung Cancer Panel RNA evaluates 28 fusion variants in ALK, ROS1, RET, NTRK1 and MET (exon 14 skipping). The π-CODE micro disc technology can access up to 85,000 distinct circular images multiplexed in a single reaction. Each micro disc has a distinct circular image that can be detected when present in the sample. Amplification is carried out by PCR reaction, following hybridization with the micro discs and fluorescence detection of the positive mutation identification. A positive control (POS) and a negative control (NEG) are used for each run. Each sample is monitored by internal control, which evaluates the quality of nucleic acids. Additional controls for non-specific fluorescence (background), fluorescence signal intensity and micro disc checking are included in the reaction and will validate the assay performance.