Skip to main content

Feasibility of Comprehensive Genomic Profiling (CGP) in Real-Life Clinical Practice.

Nibid L, Sabarese G, Righi D, Rossi SM, Perrone G, Crucitti P, Vincenzi B, Tonini G, Merlini G

Publication: Diagnostics (Basel) (2023) • Vol 13PMID: 36832270View on PubMed

Purpose

This paper investigated the effects of specimen type (primary versus metastasis tumor), collection method (surgical resection versus biopsy or cell block), formalin-fixed paraffin-embedded (FFPE) specimen storage, surface area and percentage tumor nuclei on successful Comprehensive Genomic Profiling (CGP) analysis by retrospectively comparing success rates for 184 unmatched FFPE tumor specimens. CGP profiling results were confirmed by in-house next-generation sequencing (NGS), real-time PCR, immunohistochemistry (IHC) and/or fluorescent in situ hybridization (FISH) in 110 specimens.

Conclusion

Of the 184 CGP reports produced for the same number of specimens, 138 were classified as completely satisfying (complete data including microsatellite instability (MSI) and tumor mutational burden (TMB) data was generated) and 12 were partially satisfying (complete data without MSI or TMB data was generated), with the remaining 34 reports classified as inconclusive (incomplete genetic data was obtained). There was no difference in the success rate (defined as the percentage of completely or partially satisfying CGP reports) between primary tumors and metastases (66 of 83 versus 84 of 101), but the success rate was significantly and negatively correlated with the time that elapsed between specimen collection and analysis (r=-0.302, P<0.0001). The success rate was 89.3% for specimens stored for ≤ 6 months before analysis, but was 76.2% for specimens stored for 7–12 months, 64.0% for specimens stored for 13–24 months, and 56.2% for specimens stored for >24 months (P<0.001).  CGP success rates were also higher for surgical specimens than biopsies or cell blocks (96.7% versus 74.3% and 71.4%, respectively). The 34 inconclusive reports included 13 and 14 instances where the specimen was considered acceptable or suboptimal based on surface area and/or the percentage of tumor nuclei and 7 instances where the specimen was considered to be optimal, which indicates that a lower tumor cell content and/or surface area increases the likelihood of an inconclusive report. TMB data was obtained from 138 specimens, of which 30 had a TMB ≥10 mutations per Mb.

Of the 110 FFPE specimens selected for in-house NGS analysis, 109 were successfully analyzed. The single specimen that failed in-house testing was a transbronchial needle aspirate classified as suboptimal (with a surface area of 9 mm2 and 20% tumor nuclei), which was classified as inadequate for CGP. Of the 109 specimens successfully analyzed, 24 had a targetable mutation detected, and 22 were confirmed by CGP analysis. The two mutations detected by the in-house assays but not CGP included an ALK rearrangement detected with immunohistochemistry (IHC) and fluorescent in situ hybridization (FISH) in a non-small cell lung cancer specimen that was considered suboptimal (had a surface area of 9 mm2 and 30% tumor nuclei) and a KRAS mutation detected by real-time PCR in a colon cancer metastasis that was acceptable (surface area 25 mm2, 50% tumor nuclei). Among lung cancer specimens, CGP reports matched in-house assays for SNVs in all 72 specimens tested by both methods, but CGP missed one of two ALK rearrangements that were detected by the in-house method. While 7 of 36 breast cancer specimens were HER2 positive by the in-house assays, 2 of these (both metastases) were missed by CGP. MSI status matched for all three gastric cancer specimens analyzed by both CGP and the in-house assay, but HER-2 amplification was detected by the in-house method and not CGP in one of three specimens.

Studies (1)

Purpose

This study investigated the effects of specimen type (primary versus metastasis tumor), collection method (surgical resection versus biopsy or cell block), FFPE block storage, surface area and the percentage of tumor nuclei on successful Comprehensive Genomic Profiling (CGP) analysis by NGS by retrospectively comparing success rates for 184 unmatched FFPE tumor specimens. CGP profiling results were confirmed by in-house NGS, real-time PCR, immunohistochemistry (IHC) and/or FISH in 110 specimens.   CPG data for 184 FFPE tumor specimens analyzed using FoundationOneCDx assays were obtained. The specimens were collected during clinical trial enrollment based on clinical evaluation by an oncologist. Specimens included 89 lung (50 primary tumors and 39 metastases), 36 breast (10 primary and 26 metastases), 12 colon (1 primary and 11 metastases), 10 sarcoma (all primary), 5 pancreatic  (3 primary and 2 metastases), 4 stomach  (all primary), 4 thyroid (1 primary and 3 metastases), 4 melanoma (all metastases), 3 germ cell/testicular (all metastases), 2 hepatocellular carcinoma (1 primary and 1 metastasis), 2 cholangiocarcinoma (bile duct, both primary), 2 salivary gland (1 primary and 1 metastasis), 1 thymus (primary), 1 nasopharynx (metastasis), 1 kidney (metastasis), 1 esophagus (primary), 1 mesothelium (metastasis) and 1 penis (primary) cancer tumors as well as 5 metastasis of unknown origin. In total, 61 of the specimens were collected during surgical resection, 109 were biopsies and 14 were cell block specimens. Each tumor was formalin-fixed and paraffin-embedded (FFPE; additional details not provided). CGP analysis was conducted within 6 months of collection (122 specimens), between 6 and 12 months after collection (21 specimens), 12 and 24 months (25 specimens), and more than 24 months after collection (16 specimens). CGP reports were classified as: 1.) completely satisfying when they contained complete genomic data including MSI and TMB, 2.) partially satisfying if they included genomic data but not MSI or TMB and 3.) inconclusive when they did not have complete genomic data. CGP success was defined as the percentage of reports that contained completely satisfying or partially satisfying data. The surface area and percentage of tumor nuclei were determined from an H&E-stained section by two pathologists, and specimens were classified as optimal (surface area ≥ 25 mm2 and ≥35% tumor nuclei), acceptable (surface area ≥ 15 mm2 and ≥20% tumor nuclei, but one parameter not optimal), or suboptimal (either < 15 mm2 or  <20% tumor nuclei).  For 110 specimens, in-house real-time PCR or NGS assays were used to evaluate EGFR, KRAS, NRAS, BRAF, BRCA1, BRCA2, HRAS, c-KIT, PDGFRA, HER2 (SNVs), and ALK (SNVs) as well as MSI. ALK (rearrangements), ROS-1, MET, and RET were assayed by RT-PCR on RNA-Sequencing. ALK (rearrangements), ROS-1, and HER2 (amplification) status were evaluated by FISH. ALK, ROS1, MMR, and PD-L1 expression were visualized by immunohistochemistry.  

Summary of Findings

Of the 184 CGP reports produced for the same number of specimens, 138 were classified as completely satisfying (complete data including microsatellite instability (MSI) and tumor mutational burden (TMB) data was generated) and 12 were partially satisfying (complete data without MSI or TMB data was generated), with the remaining 34 reports classified as inconclusive (incomplete genetic data was obtained). There was no difference in the success rate (defined as the percentage of completely or partially satisfying CGP reports) between primary tumors and metastases (66 of 83 versus 84 of 101), but the success rate was significantly and negatively correlated with the time that elapsed between specimen collection and analysis (r=-0.302, P<0.0001). The success rate was 89.3% for specimens stored for ≤ 6 months before analysis, but was 76.2% for specimens stored for 7–12 months, 64.0% for specimens stored for 13–24 months, and 56.2% for specimens stored for >24 months (P<0.001).  CGP success rates were also higher for surgical specimens than biopsies or cell blocks (96.7% versus 74.3% and 71.4%, respectively). The 34 inconclusive reports included 13 and 14 instances where the specimen was considered acceptable or suboptimal based on surface area and/or the percentage of tumor nuclei and 7 instances where the specimen was considered to be optimal, which indicates that a lower tumor cell content and/or surface area increases the likelihood of an inconclusive report. TMB data was obtained from 138 specimens, of which 30 had a TMB ≥10 mutations per Mb.

Of the 110 FFPE specimens selected for in-house NGS analysis, 109 were successfully analyzed. The single specimen that failed in-house testing was a transbronchial needle aspirate classified as suboptimal (with a surface area of 9 mm2 and 20% tumor nuclei), which was classified as inadequate for CGP. Of the 109 specimens successfully analyzed, 24 had a targetable mutation detected, and 22 were confirmed by CGP analysis. The two mutations detected by the in-house assays but not CGP included an ALK rearrangement detected with immunohistochemistry (IHC) and fluorescent in situ hybridization (FISH) in a non-small cell lung cancer specimen that was considered suboptimal (had a surface area of 9 mm2 and 30% tumor nuclei) and a KRAS mutation detected by real-time PCR in a colon cancer metastasis that was acceptable (surface area 25 mm2, 50% tumor nuclei). Among lung cancer specimens, CGP reports matched in-house assays for SNVs in all 72 specimens tested by both methods, but CGP missed one of two ALK rearrangements that were detected by the in-house method. While 7 of 36 breast cancer specimens were HER2 positive by the in-house assays, 2 of these (both metastases) were missed by CGP. MSI status matched for all three gastric cancer specimens analyzed by both CGP and the in-house assay, but HER-2 amplification was detected by the in-house method and not CGP in one of three specimens.method and not CGP.

  • Biospecimens

    TissueEsophagus
    TissueLung
    TissueBreast
    TissueColorectal
    TissueSkin
    TissueBone
    TissuePancreas
    TissueTestis
    TissueStomach
    TissueThyroid Gland
    TissueLiver
    TissueSalivary Gland
    TissueThymus Gland
    TissueNasopharynx
    TissueKidney
    TissueOther
  • Preservation Types

    Formalin
  • Diagnoses

    NeoplasticCarcinoma
    NeoplasticGerm Cell
    NeoplasticSarcoma
    NeoplasticMelanoma
    NeoplasticNot specified
  • Platforms

    DNANext generation sequencing
    DNAReal-time qPCR
    DNAFISH
    ProteinImmunohistochemistry
    RNAReal-time qRT-PCR
    RNANext generation sequencing
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    Biospecimen AcquisitionBiospecimen location
    Breast
    Lung
    Metastasis
    Primary tumor
    Stomach
    Biospecimen AcquisitionMethod of tissue acquisition
    Aspirate
    Biopsy
    Surgical resection
    Biospecimen Aliquots and ComponentsAliquot size/volume
    15-25 mm2
    < 15 mm2
    ≥ 25 mm2
    Biospecimen Aliquots and ComponentsBiospecimen components
    20-35% tumor nuclei
    <20% tumor nuclei
    ≥35% tumor nuclei
    StorageStorage duration
    13–24 months
    7–12 months
    > 24 months
    ≤ 6 months
    Next generation sequencing SpecificTechnology platform
    FISH
    IHC
    In house NGS
    Real-time PCR