Patients’ Characteristics
74 patients had circulating tumor DNA molecular testing performed reviewed and analyzed (Table 2). There was a slight difference, but not statistically significant of female over male (55.4% vs. 44.6%). The median age of the patients presented was 56.5 years (range: 11–88 years). At time of tumor diagnosis the median age was 53 years (range: 10-78 years). Among the 74 patients investigated, 17 various cancers were analyzed. The majority had breast cancer (17/74), followed by lung cancer (13/74), colorectal cancer (10/74), pancreatic cancer (6/74), sarcoma (6/74), kidney cancer (2/74), gastric cancer (2/74). The other types represent one case tested (Table 2). We included also metastatic cancer patients with unknown tumour origin (8/74). At the time of tumor testing, all patients were in metastatic stages presenting at least one metastatic site and had escaped the standard first-line curative and metastatic treatment recommended by the clinical practice guidelines. Physicians ordered molecular genetic test to identify the aberrations leading to tumor resistance, to determine the alternative targeted treatment approved by the FDA for the indication and tentatively to understand the alterations which confer drug escape and resistance. Prior to molecular testing, 17 patients in our center had received all conventional therapies (chemotherapy, surgery and radiotherapy), 11 patients had undergone surgery and chemotherapy, 7 patients had chemotherapy and radiotherapy, 6 patients had received only chemotherapy, 1 patient had treated only with radiotherapy and 1 patient had undergone surgery and radiotherapy. Besides, no one was treated with targeted therapy before the genetic test.
Tissue biopsy & liquid biopsy technical concordance
From analysis of the results generated, we found that liquid biopsy data are similar to tissue biopsy, in term of library quality control, data throughput generated, data quality, bioinformatics analysis and genetic alteration detection. Libraries were prepared, sequenced and showed peak at expected size ~300 bp (Figure 1, A). The total quantity of the library ranged from 1.34 ng/µl to 35 ng/µl, as recommended by the company for library preparation. All libraries were sequenced and sufficient reads were generated per sample. Of the total generated reads, around 98 % were merged and around 2% were unmerged. Among the merged one, 98 % were mapped and 96 % were trimmed (Figure 1, B). According to genetic alteration detection, Of the 6 patients who had both ctDNA and tissue genetic testing, 3 patients had alterations in common between the tissue and matched ctDNA samples, however, the three other patients who had genetic alterations, were not detected in their tissue samples (Supplementary material, Table1).
Gene alterations analysis
NGS allowed us to detect the mutational profile of various genes. We detected gene mutations in 89% of the total patients tested. Only, 11% patients have either no ctDNA in their serum or, with detected ctDNA, but no detected mutations (Figure 2). In our cohort, 88% of our patients were eligible for treatment guidance using liquid biopsy: 64% with available FDA-approved drug, 9% could join clinical trials for experimental drugs and 15% have to avoid some drugs associated with tumor resistance. Only 1% of the investigated patients show detected non actionable mutation, then without any indication for available drug or open clinical trial.
NGS allowed us to detect various genetic alterations; in our cohort, 8% (7/85) of cases were detected with only one mutation, 81% (69/85) of cases were detected with at least two mutations, and 11% (9/85) of cases had no detected mutation. These data confirm the complexity of tumor development and dissemination. The reductionism analysis approach could be applied for small fraction of patients, but for most patients, the systemic analysis approach is required for successful patient follow up and treatment management.
According to identified alterations, from a total of 71 genes analysis we detected 353 distinct aberrations; which can be gathered on five genetic alteration classes: deletions (47%), transitions (30%), transversions (11%), duplications (11%) and deletion/insertion (1%) (Figure 3A). By analyzing the proportions of these mutations, we found that only six cancer types namely; breast cancer, lung cancer, colorectal cancer, pancreatic cancer, gastric cancer and cervix cancer present all the aberrations groups, six cancers namely; sarcoma, glioblastoma, kidney cancer, adenoid cystic cancer endometrial cancer and desmoid cancer had 4 mutations types, urothelial cancer and liver cancer had 3 mutations groups, only 2 mutations type were observed in Ewing cancer and brain cancer and no mutation was detected in adrenocortical cancer patient (Figure 3, B). Regarding the frequency of mutated genes detected in our cohort we observed that BRCA2, EGFR2, MSH6 and NF1 genes were the most mutated genes in the15 cancer types investigated (Figure 4).
Mutations Profile
In our cohort 10 genes had been found alterated in more than 10 cancer types including; BRCA2, EGFR, MSH6, NF1, PTEN, APC, ATM, PDGFRA, NOTCH1, FGFR1. Among these, the BRCA2 gene mutations found in the highest proportion of samples observed with 5 genetic alterations frequently detected namely c.68-7delT, c.9097delA, c.9257-10delT, c.1909+21_1909+22delTT, and c.5073delA. Regarding EGFR gene, two variants were frequently detected in our patients namely: c.*293_*294delAA and c.2361G>A. In MSH6 gene the following two variants; c.3261delC and c.3261dupC were the most identified. For NF1 gene, five variants were predominant in our cohort including: c.2251+25delT, c.4515-23delT, c.2033delC, c.2041C>T and c.1400C>T. For PTEN, four gene mutations were largely detected namely; c.254- 30delT, c.493- 34delT, c.493- 12delT, c.802- 4_802-3delTT. For APC gene the following two alterations; c.423-5_423-4delAA and c.423-17T>A were the most identified in our patients. In ATM gene only c.6007-35delT variant was usually detected. Regarding PDGFRA gene, two variants; c.629-10T>C and c.1701A>G were identified in most patients. For NOTCH1 gene the following two alterations c.7244_7246delCAC and c.4732_4734delGTG were frequently identified. Last but not least, only one variants c.396_398delTGA related to FGFR1 gene was often detected in our cohort. Other gene mutations had been found in less than 9 cancer types including mutations in the following genes: TP53, PIK3CA, RB1, PTCH1, ALK, FGFR3, NPM1, RET, MET, HRAS, IDH1, KDR, CSF1R, BRCA1, KRAS, STK11, FLT3, NF2, KIT, HIST1H3B, HNF1A, CDKN2A, DNMT3A, JAK3, SMARCB1, GNA11, ABL1, BRAF, DDR2, MSH2, NRAS, KMT2D, SMAD4, PMS2, MSH3, MLH3, FGFR2, MAP2K1, TERT, PMS1, AKT1, MTOR, POLE, FBXW7, RNF43, ARID1A, PTPN11, CDH1, MPL, ERBB3, ERBB4, CTNNB1, VHL, JAK2, PD-L1, ROS1, KMT2C, EZH2, GNAS, ZFHX3 and SMO (Figure 4).
Mutations associated with sensitivity
DNA sequencing has been used routinely to inform the choice of targeted therapy in specific cancer type. Genetic alterations associated with sensitivity to drugs across tumor types were observed in several genes. Of the 76 cases with genetic alterations, 73% of them were registered with potentially actionable alterations; 64% had FDA approved drug and 9% were eligible for clinical trials (Figure 2). Analysis of the results from these 76 tests revealed 84 distinct mutations associated with sensitivity to targeted therapies across the following investigated 19 genes: EGFR, MSH6, FGFR1, KRAS, NRAS, PDGFRA, BRCA2, ALK, JAK2, PD-L1, ROS1, TP53, IDH1, POLE, ERBB2, PIK3CA, ATM, PTCH1 and NF1.
ALK and EGFR was the most frequently altered gene associated with potential drug sensitivity (13 mutations), followed by BRCA2 gene (12 mutations), KRAS (9 mutations) and PIK3CA (6 mutations). Among the patients examined in our cohort, 10 cases harbored more than 7 potential actionable alterations; 5 cases had 7 distinct actionable alterations, 2 cases had 12 distinct actionable alterations and single cases have more than 8 alterations each (1 case had 13 actionable alterations, 1 case had 10 actionable alterations 1 case had 8 actionable alterations) (Supplementary material, Table2).
Interestingly for some genetic alterations, several treatment options were suggested. Like ALK gene mutations treatment, the following drugs Lorlatinib, Entrectinib, Crizotinib, Ceritinib, Brigatinib and Alectinib were proposed. Genetic aberrations found in EGFR gene could be targeted also by several drugs, to be used in combination or as single drug, such as; Panitumumab, Amivantamab, Bevacizumab, Cetuximab, Dacomitinib, Brigatinib, Necitumumab, Osimertinib, Gefitinib, Neratinib, Erlotinib, Afatinib, Mobocertinib and Lapatinib. For BRCA2 mutations, the following treatment options were suggested: Cisplatin, Olaparib, Platinum chemotherapy regimen, Tamoxifen, Anthracycline, Carboplatin, Talazoparib, Etoposide, Gemcitabine, Rucaparib, Niraparib, Bevacizumab, Pembrolizumab, Lurbinectedin, Topotecan and Durvalumab. Besides, all the mutations identified in PIK3CA, the anti-kinases treatments were suggested such as alpelisib. Part of the KRAS gene mutations detected in our cohort was sensitive to several drugs used in combination or alone including: Regorafenib, Oxaliplatin, Bevacizumab, and Erlotinib (Supplementary material, Table2).
Mutations associated with resistance
To further investigate the impact of tumor resistance on therapeutic strategies, the in-depth analysis of the ctDNA profiles from the 76 cases with mutations shows that 15% of patients had mutations which could be associated with drug-resistance (Figure 2). Table 3 revealed 24 distinct mutations associated with resistance to targeted therapies across 7 distinct genes. The results generated from this study discovered the presence of resistance remarkably frequent in KRAS (7 mutations), NF1 (5 mutations), NRAS (3 mutations), ALK (3 mutations), EGFR (2 mutations), TP53 (2 mutations) and BRAF (2 mutations).
KRAS alterations were the most common resistant mutations observed and present in 5 patients. The KRAS c.321A>T and c.35G>A mutations are the most common. Our analysis also revealed profound acquired resistance in 2 patients harbored 8 resistance alterations, including 6 KRAS mutations, and 2 distinct NRAS mutations, both patients were diagnosed with colorectal cancer. Another patient affected by gastric cancer, harbored seven different resistance alterations, including four NF1 mutations, two TP53 mutations as well as one ALK mutation. Other patients harbored one or two different resistance alterations.
KRAS and NRAS alterations were mostly resistant to Cetuximab and Panitumumab drugs with some differences related to each aberration. Mutations detected in NF1 gene were resistant to Imatinib drug. For ALK gene, aberrations quantified in our cohort were usually resistant to Crizotinib, Ceritinib and Alectinib drugs with minor differences related to two mutations: c.3593T>C, which are sensitive to Lorlatinib, and c.3605delG are sensitive to Brigatinib, Crizotinib, Lorlatinib and Erlotinib.
The c.2240_2254delTAAGAGAAGCAACAT mutation in EGFR gene was associated with resistance to Gefitinib and Erlotinib drugs while the c.2573T>G mutation was sensitive to Crizotinib drug. TP53 gene alterations were resistant to IFNA2, Rituximab, Fludarabine, Cyclophosphamide, Fludarabine phosphate chlorambucil, and Lenalidomide drugs and BRAF alterations were resistant to Cetuximab, Panitumumab and Trametinib drugs (Table 3).