1. Introduction

2. Materials and Methods

2.1. Sample Collection and Targeted Exon Sequencing

A total of 63 patient samples including industry reference standards (cfDNA and gDNA), clinical samples (blood and formalin-fixed paraffin embedded tissue (FFPE) samples), and cross-laboratory samples were utilized in the study. The ctDNA or FFPE DNA were extracted using commercially available DNA extraction kits (QIAamp minELute ccfDNA DNA kit, Qiagen, Germany). The extracted DNA was subjected to quality check (≥80% ccfDNA content) on TapeStation instrument (Agilent, USA) and the concentration was determined on Qubit 4.0 fluorometer (ThermoFisher Scientific, USA). Quality assessment was followed by preparation of Illumina-compatible libraries using target hybridization method. DNA libraries were then sequenced on the NextSeq 2000 platform in a paired-end fashion. Variant calling, and annotation was then performed using indigenously developed iCare platform. An illustration of the NGS sequencing workflow is illustrated in Figure 1.

Figure 1. Illustration of OncoIndx^® workflow. Figure presents the detailed workflow of OncoIndx NGS assay from sample collection, DNA extraction, DNA sequencing, variant calling, variant annotation, and data analysis.

Figure 1. Illustration of OncoIndx^® workflow. Figure presents the detailed workflow of OncoIndx NGS assay from sample collection, DNA extraction, DNA sequencing, variant calling, variant annotation, and data analysis.

3. Results

3.1. OncoIndx^® Detected Genomic Alterations from Ngs Standard Reference Samples with High Concordance and Analytical Precision

OncoIndx^® was tested to detect SNVs, INDELs, CNAs, and translocations (fusion) from industry NGS reference samples which were pre-synthesized with variants. The outcomes are discussed in detail in the following sections.

3.1.1. Single Nucleotide Variants, and INDELs

Clinical implementation of genomic tests requires high analytical precision. Several parameters including accuracy, sensitivity and specificity stand vital for such investigations. Thus, OncoIndx^® variant calling pipeline (iCare’s VCP)^© was validated using 43 NGS reference samples with variant distributions at 5 %, 1% and 0.1% frequency. From the test outcomes, variant detection was most sensitive and accurate at 5% variant allele frequency (VAF) was reported in Table 1. CNAs were detected with 100% accuracy, sensitivity, and specificity, with no false positive and/or false negative occurrences. From a total of 264 SNVs detected by OncoIndx^®, 156 were true positives and 108 rendered true negative outcomes with excellent concordance against reference standards. Highest positive and negative predictive values (NPV and PPV) obtained from OncoIndx^® assay were 100% for SNVs. Maximum accuracy of 97.40% was obtained for small INDELs with 100% specificity and 95.60% sensitivity. In addition, translocations (fusions) were detected with a high accuracy of 98.48%. It also yielded 100% specificity and a sensitivity of 97.44%. OncoIndx^® assay pipeline yielded no false positive hits even at 1% VAF. This highlights the reliability and sensitivity of the assay to exhibit acceptable performance for variant calling even at 1% VAF.

Table 1. Outcomes of statistical analysis obtained OncoIndx^® comprehensive genomic panel at 5% variant allele frequency.

Table 1. Outcomes of statistical analysis obtained OncoIndx^® comprehensive genomic panel at 5% variant allele frequency.

Alteration type	Total number of alterations	True positives	False positives	True negatives	False negatives	*PPV	*NPV	Accuracy	Specificity	Sensitivity
SNVs	264	156	0	108	0	100	100	100	100	100
Small INDELs	154	87	0	63	4	100	94.03	97.40	100	95.60
CNA	66	39	0	27	0	100	100	100	100	100
Translocations (Fusions)	66	38	0	27	1	100	96.43	98.48	100	97.44

* NPV: Negative Predictive Value. PPV: Positive Predictive Value.

3.1.2. Copy Number Alterations and Translocations (Fusions)

Detection of CNAs and fusions can be challenging due to inconsistencies of rearranged chromosomal regions, breakpoints, read lengths, etc. SV such as translocations, especially fusions may be important in solid tumors as it is in hematological cancers (Taniue and Akimitsu 2021). In OncoIndx^®, the detection of CNAs and fusion variants were validated using customized reference materials at VAF from 5%, 1% to 0.1%. The validation set was mainly focused on 3 genes for CNAs namely, ERBB2, MET, MYC, and 3 translocations namely, EML4-ALK, NCOA4-RET, and CD74-ROS1. CNAs were detected with highest PPV, NPV, specificity and accuracy of 100%. In addition, no false positives and false negatives were detected from OncoIndx^® assay for CNAs at 5% VAF, whose distributions are shown in Figure 2A. 100% of patients with ERBB2, and MYC amplification were detected at a VAF as low as 0.1% (Figure 2A). Next, for fusions, a detection accuracy of 98.48%, PPV and specificity of 100% were obtained with no false positive detection. At all three dilutions of VAF (5%, 1%, and 0.1%), fusions were detected at a specificity of 100%. NCOA4-RET and CD74-ROS1 fusions were predominantly detected at 0.1%VAF indicating the sensitivity of OncoIndx^® assay at a low VAF of 0.1% (Figure 2B).

Figure 2. Distribution of copy number alterations and translocations (fusions) detected by OncoIndx. (A) From the total number of samples in the cohort of 5% VAF, figure indicates 39 true positive copy number alterations with 0 false positives detected across the reference sample validation cohort. (B) Figure indicates the detection of fusions with 38 true positives and 1 false positive. In both A and B, no CNA/Fusions were detected in Wild-type controls.

Figure 2. Distribution of copy number alterations and translocations (fusions) detected by OncoIndx. (A) From the total number of samples in the cohort of 5% VAF, figure indicates 39 true positive copy number alterations with 0 false positives detected across the reference sample validation cohort. (B) Figure indicates the detection of fusions with 38 true positives and 1 false positive. In both A and B, no CNA/Fusions were detected in Wild-type controls.

3.1.3. Limit of Detection of OncoIndx^®

For identifying the limit of detection (LOD) of OncoIndx^® assays, serial dilutions of standard reference Seraseq™ samples were performed from 0.2% to 1% VAF. Gene variants including AKT1:p.E17K, EGFR:p.L858R, EGFR:p.E746_A750del and ERBB2:p.Y772_A775dup were detected in each of the diluted samples. Expected vs. observed VAF% were visualized for all variants to further determine LOD. A total of 83.33% (n=5) of SNVs could be detected at VAF less than 1% namely, EGFR (p.T790M), NRAS (p.Q61R), PIK3CA (p.H1047R), KRAS (p.G12D), ALK (p.G1202R) respectively (Figure 3A). Similarly for small InDels, 60% (n=3) of the variants could be detected at VAF less than 1% namely, BRCA1 (p.K654fs*47), BRCA2 (p.R2645fs*3), and PIK3CA (p.*1069Mfs*4) (Figure 3B). The observed % VAF thus denotes the LOD of OncoIndx^® assays which were determined for both SNVs and INDELs. Thus, the overall sensitivity of OncoIndx^® appears at 0.1% from LOD values as indicated by the arrows in Figure 3A,B. List of SNVs and INDELs validated from the industrial samples are presented in Table 2.

Figure 3. Limit of detection of OncoIndx^® test. (A) and (B) Detection limit observed at 0.1% VAF for majority of SNVs and INDELs pre-established by standard reference samples tested by OncoIndx^® assay (shown by arrows).

Figure 3. Limit of detection of OncoIndx^® test. (A) and (B) Detection limit observed at 0.1% VAF for majority of SNVs and INDELs pre-established by standard reference samples tested by OncoIndx^® assay (shown by arrows).

Table 2. Table presents the list of SNVs and INDELs detected and validated from the industrial samples.

Table 2. Table presents the list of SNVs and INDELs detected and validated from the industrial samples.

List of SNVs and INDELs validated from the industrial samples
AKT1:p.E17K	EGFR:p.T790M
ALK:p.F1174L	ERBB2:p.Y772_A775dup
ALK:p.G1202R	KIT:p.D816V
BRAF:p.V640E	KRAS:p.G12C
BRCA1:p.K654fs*47	KRAS:p.G12D
BRCA2:p.R2645fs*3	KRAS:p.Q61H
EGFR:p.E746_A750del	KRAS:p.Q61R
EGFR:p.L747_P753delinsS	NRAS:p.Q61R
EGFR:p.L858R	PIK3CA:p.*1069Mfs*4
EGFR:p.S752_I759del	PIK3CA:p.H1047R

3.2. High Concordance of Genomic Alterations Obtained from Clinical Samples

Clinical sample cohort with 14 samples were tested across genomic alterations for 5 hotspot genes such as EGFR, ALK, KRAS, PIK3CA, and BRCA2. OncoIndx^® assay showed 100% concordance for alterations detected in EGFR, ALK, KRAS, and BRCA2 genes, and 50% concordance for PIK3CA alterations (Table 3). In ALK positive samples, the OncoIndx^® assay efficiently determined ALK fusion partners beyond EML4 including NPM1. In addition, we predict that the 50% concordance obtained for PIK3CA alteration in one clinical sample may be most likely due to tumor heterogeneity presented by testing different sample types, i.e., former clinical finding being in FFPE sample and OncoIndx^® assay in ctDNA from blood sample.

Table 3. Table shows the concordance levels of genomic alterations in clinical samples as detected by OncoIndx^® assay.

Table 3. Table shows the concordance levels of genomic alterations in clinical samples as detected by OncoIndx^® assay.

S.No.	Genes	Concordant Genomic findings from OncoIndx® assay	Concordance levels obtained from OncoIndx® assay
1	EGFR	L858R E746_A750del L747_S752del	100%
2	ALK	NPM1-ALK ALK-EML4 Fusion G1202R G1269A	100%
3	KRAS	A146T	100%
4	PIK3CA	H1047R	50%
5	BRCA2	S636*	100%

Tumor heterogeneity is a major roadblock in effective treatment decision making which is also one of the main challenges of FFPE DNA testing. Thus, a ctDNA based genomic detection may offer benefits in these cases by analyzing the totality of tumor rather than individual sections like the former testing method. Thus, OncoIndx^® may be a beneficial complementary tool in such scenarios. In addition to the concordance among genomic alterations, the OncoIndx^® assay was also validated against important biomarkers including MSI, TMB, LOH, LST, and TAI whose outcomes are detailed in the next section.

3.3. Validation of Biomarker Signatures against Reference Laboratories: Microsatellite Instability and Tumor Mutation Burden

In addition to detecting genomic variants, the OncoIndx^® test identifies the status of biomarkers to predict immunotherapy response in patients. From the blood samples collected, MSI, and TMB were detected. MSI in tumors has been studied to promote “immune-hot” conditions (Wang et al. 2023; Bai et al. 2021). Thus, presence of high MSI favors tumor immune responses. To validate the reproducibility of MSI detected from OncoIndx^®, 2 MSI detection tools were utilized against the outcomes of clinically stratified reference laboratory samples. From our pipeline, MSI were detected with an accuracy of 95%, sensitivity of 90%, and a specificity of 100%. Overall, PPV was identified to be 100% and NPV was 90.91% (Table 4).

Table 4. Table presents the analytical parameters of MSI detection by OncoIndx^® test.

Table 4. Table presents the analytical parameters of MSI detection by OncoIndx^® test.

Statistics of MSI detection in OncoIndx® (Percentage %)
Positive Predictive Value (PPV)	100
Negative Predictive Value (PPV)	90.91
Sensitivity	90
Specificity	100
Accuracy	95

All MSI high and MSI low samples were appropriately predicted with 1 misclassified sample whose value was in the intermediate range (Figure 4). Intermediate MSI values are challenging as they are neither high nor low and are missed in many classifier tools. Thus, more optimizations are necessary and are constantly underway in OncoIndx^® for better accuracy.

Figure 4. Validation of MSI status from cross-laboratory documentations across MSI prediction in OncoIndx^®.

Figure 4. Validation of MSI status from cross-laboratory documentations across MSI prediction in OncoIndx^®.

The summarized outcomes from reference laboratory validation of MSI against the food and drug administration (FDA) approved companion diagnostic tests are presented in Table 5. These outcomes suggest high sensitivity and specificity of OncoIndx^® assay in detecting MSI from NGS.

With rapidly emerging immune-oncological biomarkers, MSI stands important in predicting immunotherapy and a highly sensitive assay like OncoIndx^® can improve responsive standards of immunotherapy in cancer. MSI threshold criteria utilized for our CGP test is presented in Table 6.

Table 5. OncoIndx^® test prediction of MSI status validated against FDA approved test outcomes.

Table 5. OncoIndx^® test prediction of MSI status validated against FDA approved test outcomes.

Sample Type	FDA approved test prediction	OncoIndx® test prediction
Blood	MSS	3.2 (MSI-low)
Blood	MSS	1.55 (MSI-low)
Blood	MSS	0.79 (MSI-low)
Blood	MSS	3.07 (MSI-low)
Blood	MSS	3.61 (MSI-low)

Table 6. MSI thresholds of high, Intermediate, Low, and stable scores utilized for various sample types in OncoIndx^® test.

Table 6. MSI thresholds of high, Intermediate, Low, and stable scores utilized for various sample types in OncoIndx^® test.

Biomarker/s	Outcome	Blood/Pleural Effusion	FFPE/RNALater
MSI	MSI-H	≥ 20	≥ 20
	MSI-I	≥ 10	≥ 10
	MSI-L	< 10	< 10
	MSI-S	0	0

Likewise, TMB predictions from OncoIndx^® assay were cross validated with reference standards like Seraseq^™ and the College of American pathologists (CAP) accredited genomic DNA lab reference samples. Reference laboratory validations against the FDA-approved tests were also performed using OncoIndx^® CGP (Table 7, Table 8).

Table 7. TMB validation performed against NGS reference standards.

Table 7. TMB validation performed against NGS reference standards.

Sample	Sample type	True prediction	OncoIndx® test prediction
Control	Healthy control	Negative control	1.5
	Healthy control	Negative control	1.5
	Healthy control	Negative control	1.5
	Healthy control	Negative control	2.5
	Healthy control	Negative control	1.67
	Healthy control	Negative control	0
	Healthy control	Negative control	0
	Healthy control	Negative control	0.5
SeraSeqTM reference samples	TMB Mix Score 7 (0%)	5.8-9.2	8.67
	TMB Mix Score 7 (0.5%)	10.5-15.7 (d=3.5-7.7)	10.83
	TMB Mix Score 7 (2%)	16.6-19.2 (d=3.5-7.7)	6.67
	TMB Mix Score 20 (0%)	6.1-8.9	6.5
	TMB Mix Score 20 (0.5%)	23.7-28.3 (d=15.8-21.2)	8.17
	TMB Mix Score 20 (2%)	34.6-36.6 (d=26.4-29.8)	5.83
CAP gDNA samples	gDNA	9	11.5
	gDNA	26	19.5
	gDNA	9	9.83
	gDNA	26	5.67

Table 8. Cross-laboratory validation of TMB with OncoIndx^® test predictions.

Table 8. Cross-laboratory validation of TMB with OncoIndx^® test predictions.

Sample Type	FDA approved test prediction	OncoIndx® test prediction
Blood	1	3.2
Blood	7.26	1.55
Blood	6.7	0.79
Blood	3	3.07
Blood	4	3.61
Blood	4.77	3.167

TMB thresholds utilized in the CGP assay is presented in Table 9. The outcomes have been categorized as low and High TMB scores which are then predictive of immunotherapy benefits.

Table 9. TMB thresholds of OncoIndx^® test.

Table 9. TMB thresholds of OncoIndx^® test.

Biomarker/s	Outcomes	Blood/Pleural effusion	FFPE/RNALater
TMB	TMB-H	≥ 10	≥ 10
	TMB-L	< 10	< 10

In addition, the predicted TMB values from targeted OncoIndx^® panel was also validated by comparing with whole exome sequencing (WES) data from the cancer genome atlas program (TCGA) database. Pearson correlation coefficient of r=0.9885 was obtained (Figure 5). Overall, these outcomes indicate concordance between the OncoIndx^® CGP against established FDA-approved companion diagnostic tests.

Figure 5. Pearson correlation outcomes for Tumor mutation burden from Whole exome sequencing and OncoIndx gene panel. Figure shows a high correlation between the predicted TMB scores of whole exome sequencing (WES) vs. the targeted comprehensive gene panel from OncoIndx.

Figure 5. Pearson correlation outcomes for Tumor mutation burden from Whole exome sequencing and OncoIndx gene panel. Figure shows a high correlation between the predicted TMB scores of whole exome sequencing (WES) vs. the targeted comprehensive gene panel from OncoIndx.

4. Conclusions

References

1. Karol, D., McKinnon, M., Mukhtar, L., Awan, A., Lo, B. and Wheatley-Price, P., 2021. The impact of foundation medicine testing on cancer patients: a single academic centre experience. Frontiers in oncology, 11, p.687730.
2. Bai, J., Chen, H. and Bai, X., 2021. Relationship between microsatellite status and immune microenvironment of colorectal cancer and its application to diagnosis and treatment. Journal of Clinical Laboratory Analysis, 35(6), p.e23810.
3. Bharde, A., Bhonde, M., Bose, C., Raut, N.V., Kothavade, H., D’Souza, A., Kad, T., Jadhav, B., Prajapati, S., Jadhav, V.B.L. and Joshi, S.K., Hariramani, K., Uttarwar, M., Khandare, J., Shafi, G., 2023. Comprehensive ctDNA analysis to identify genome-instability and actionable mutation landscape in gynecologic cancers. Journal of Clinical Oncology, 41, p.e17503.
4. Bharde, A., Noronha, V., Ratnaparkhi, M., Patil, V., Jadhav, B., Prajapati, S., Choughule, A., Chandrani, P., Haldar, S., Moubeen, F. and Menon, N., Hariramani K., Basavalingegowda, M., Ramesh, A., Khandare, J., Chaturvedi, P., Shafi, G., Prabhash, K., 2024. Comprehensive genomic profiling of ctDNA reveals distinct genomic signatures and therapeutic implications for immunotherapy response in advanced head and neck cancer. Cancer Research, 84(6_Supplement), pp.2408-2408.
5. Fabrizio, D., Cristescu, R., Albacker, L., Snyder, A., Ward, A., Lunceford, J., Aurora-Garg, D., Jin, F., Hopkins, J., Rubin, E. and Hegde, P., 2021. Real-world prevalence across 159 872 patients with cancer supports the clinical utility of TMB-H to define metastatic solid tumors for treatment with pembrolizumab. Annals of Oncology, 32(9), pp.1193-1194.
6. Hiemenz, M.C., Graf, R.P., Schiavone, K., Harries, L., Oxnard, G.R., Ross, J.S. and Huang, R.S., 2022. Real-world comprehensive genomic profiling success rates in tissue and liquid prostate carcinoma specimens. The Oncologist, 27(12), pp.e970-e972.
7. Jennings, L.J., Arcila, M.E., Corless, C., Kamel-Reid, S., Lubin, I.M., Pfeifer, J., Temple-Smolkin, R.L., Voelkerding, K.V. and Nikiforova, M.N., 2017. Guidelines for validation of next-generation sequencing–based oncology panels: a joint consensus recommendation of the Association for Molecular Pathology and College of American Pathologists. The Journal of molecular diagnostics, 19(3), pp.341-365.
8. Kadri, S., Long, B.C., Mujacic, I., Zhen, C.J., Wurst, M.N., Sharma, S., McDonald, N., Niu, N., Benhamed, S., Tuteja, J.H. and Seiwert, T.Y., 2017. Clinical validation of a next-generation sequencing genomic oncology panel via cross-platform benchmarking against established amplicon sequencing assays. The Journal of Molecular Diagnostics, 19(1), pp.43-56.
9. Krzyszczyk, P., Acevedo, A., Davidoff, E.J., Timmins, L.M., Marrero-Berrios, I., Patel, M., White, C., Lowe, C., Sherba, J.J., Hartmanshenn, C. and O’Neill, K.M., 2018. The growing role of precision and personalized medicine for cancer treatment. Technology, 6(03n04), pp.79-100.
10. Li, M.M., Datto, M., Duncavage, E.J., Kulkarni, S., Lindeman, N.I., Roy, S., Tsimberidou, A.M., Vnencak-Jones, C.L., Wolff, D.J., Younes, A. and Nikiforova, M.N., 2017. Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. The Journal of molecular diagnostics, 19(1), pp.4-23.
11. Mathew, A., Joseph, S., Boby, J., Benny, S., Veedu, J., Rajappa, S., Rohatgi, N., Sirohi, B., Jain, R., Agarwala, V. and Shukla, D.K., 2022. Clinical benefit of comprehensive genomic profiling for advanced cancers in India. JCO Global Oncology, 8, p.e2100421.
12. Shafi, G., Dongare, M., Bharde, A., Fauzul, M., Hariramani, K., D’Souza, A., Jadhav, B., Kad, T., Prajapati, S., Jadhav, V. and Kumaran, M., 2023. Abstract PR007: Comprehensive ctDNA profiling reveals potential metastatic genomic signatures in treatment-naive early-stage breast cancer patients. Cancer Research, 83(2_Supplement_2), pp.PR007-PR007.
13. Takeda, M., Takahama, T., Sakai, K., Shimizu, S., Watanabe, S., Kawakami, H., Tanaka, K., Sato, C., Hayashi, H., Nonagase, Y. and Yonesaka, K., 2021. Clinical application of the FoundationOne CDx assay to therapeutic decision-making for patients with advanced solid tumors. The oncologist, 26(4), pp.e588-e596.
14. Taniue, K. and Akimitsu, N., 2021. Fusion genes and RNAs in cancer development. Non-coding RNA, 7(1), p.10.
15. Waarts, M.R., Stonestrom, A.J., Park, Y.C. and Levine, R.L., 2022. Targeting mutations in cancer. The Journal of clinical investigation, 132(8).
16. Wang, L., Geng, H., Liu, Y., Liu, L., Chen, Y., Wu, F., Liu, Z., Ling, S., Wang, Y. and Zhou, L., 2023. Hot and cold tumors: Immunological features and the therapeutic strategies. MedComm, 4(5), p.e343.