Opportunistic screening for broad range of clinically relevant secondary findings: Outcomes of exome analysis in the Incidental Genomics RCT
Laboratory Genetics and Genomics
-
Primary Categories:
- Laboratory Genetics
-
Secondary Categories:
- Laboratory Genetics
Introduction:
Practice is shifting towards use of exome and genome sequencing, offering the opportunity to expand analyses beyond variants related to the primary diagnostic indication to identify secondary findings (SFs). While medically actionable SFs are prioritized by guidelines, a much broader spectrum of results could be analyzed, aligned with patients' preferences. We characterized the outcomes of analyzing a broad range of clinically relevant SFs from exome sequencing in the Incidental Genomics RCT (NCT03597165), and the laboratory resources required for sequence analysis.
Methods:
Participants were adults with suspicion of a hereditary cancer syndrome and uninformative results from standard-of-care genetic testing. Participants had exome sequencing and were randomized to receive only results related to cancer (control arm), or cancer results in addition to a choice of SFs (intervention arm). SFs included: medically actionable disease risks (121 genes, including 42 from the ACMG list [v3.2, non-cancer genes only]), Mendelian disease risks (3,837 genes), early-onset neurodegenerative disease risks (60 genes), carrier status for recessive and X-linked conditions (684 genes), pharmacogenomic variants (24 Class A variants from the Clinical Pharmacogenetics Implementation Consortium [CPIC]), and risk variants for common complex diseases (26 variants). Following filtration, single nucleotide variants (SNVs) and small insertions/deletions (indels) were classified according to the American College of Medical Genetics and Genomics (ACMG)/Association of Molecular Pathology (AMP) criteria. Pathogenic (P) and likely pathogenic (LP) variants in genes associated with monogenic disorders were reported as SFs, as well as pharmacogenomic variants and risk variants for common complex disease. P/LP variants and variants of uncertain significance (VUS) were reported as primary cancer findings. We characterized the genetic variants that were classified and compared resource requirements for analysis between trial arms.
Results:
Exome analysis was performed for 279 participants (140 intervention, 139 control). On average, there were 35,855.5 (SD 11,279.2) raw variants and 39.5 (SD 15.6) filtered variants per case in the intervention arm, and 1,265.4 (SD 128.1) raw variants and 4.0 (SD 2.2) filtered variants in the control arm. In total, across all participants, there were 4441 unique variants in the SF genes, 5.0% (221) of which were classified as P/LP and were reportable, and 81.4% (3615) were classified as VUS and not reportable. In the intervention arm, there were on average 2.6 (SD 1.66, range 0-9) P/LP variants per case and 29.5 VUS (SD 13.2, range 2-74). The Mendelian category generated the most VUS, which is expected given that it had the largest gene list; our expanded medically actionable gene list and the ACMG subset generated fewer VUS. Given the broader scope of analysis, variant filtration, variant classification and report preparation were substantially more time consuming in the interention arm compared the control arm. Variant filtration took 7.7 times longer per case (95% CI 5.3 to 11.3, p<0.0001; 65.7 vs. 8.5 minutes on average per case). Variant classification took 13.3 times longer (95% CI 10.6 to 16.5, p<0.0001; 497.8 minutes vs. 37.4 minutes on average per case), and report preparation took 3.3 times longer (95% CI 2.6 to 4.1, p<0.0001; 157.2 minutes vs. 47.9 minutes on average per case). SFs related to monogenic disease risk were reported in 35.3% (49/139) of participants, ACMG-recommended SFs in 1.4% (2/139), carrier status in 89.3% (117/131), pharmacogenomic variants in 97.8% (135/138), and risk variants in 89.4% (118/132).
Conclusion:
While the yield of reportable SFs was high, this was accompanied by many non-reportable VUS, which substantially increased the time required for exome analysis and reporting. Analytic approaches that streamline variant prioritization and classification, such as through greater use of automation, may help promote the feasibility of offering a broad range of clinically relevant SFs.
Practice is shifting towards use of exome and genome sequencing, offering the opportunity to expand analyses beyond variants related to the primary diagnostic indication to identify secondary findings (SFs). While medically actionable SFs are prioritized by guidelines, a much broader spectrum of results could be analyzed, aligned with patients' preferences. We characterized the outcomes of analyzing a broad range of clinically relevant SFs from exome sequencing in the Incidental Genomics RCT (NCT03597165), and the laboratory resources required for sequence analysis.
Methods:
Participants were adults with suspicion of a hereditary cancer syndrome and uninformative results from standard-of-care genetic testing. Participants had exome sequencing and were randomized to receive only results related to cancer (control arm), or cancer results in addition to a choice of SFs (intervention arm). SFs included: medically actionable disease risks (121 genes, including 42 from the ACMG list [v3.2, non-cancer genes only]), Mendelian disease risks (3,837 genes), early-onset neurodegenerative disease risks (60 genes), carrier status for recessive and X-linked conditions (684 genes), pharmacogenomic variants (24 Class A variants from the Clinical Pharmacogenetics Implementation Consortium [CPIC]), and risk variants for common complex diseases (26 variants). Following filtration, single nucleotide variants (SNVs) and small insertions/deletions (indels) were classified according to the American College of Medical Genetics and Genomics (ACMG)/Association of Molecular Pathology (AMP) criteria. Pathogenic (P) and likely pathogenic (LP) variants in genes associated with monogenic disorders were reported as SFs, as well as pharmacogenomic variants and risk variants for common complex disease. P/LP variants and variants of uncertain significance (VUS) were reported as primary cancer findings. We characterized the genetic variants that were classified and compared resource requirements for analysis between trial arms.
Results:
Exome analysis was performed for 279 participants (140 intervention, 139 control). On average, there were 35,855.5 (SD 11,279.2) raw variants and 39.5 (SD 15.6) filtered variants per case in the intervention arm, and 1,265.4 (SD 128.1) raw variants and 4.0 (SD 2.2) filtered variants in the control arm. In total, across all participants, there were 4441 unique variants in the SF genes, 5.0% (221) of which were classified as P/LP and were reportable, and 81.4% (3615) were classified as VUS and not reportable. In the intervention arm, there were on average 2.6 (SD 1.66, range 0-9) P/LP variants per case and 29.5 VUS (SD 13.2, range 2-74). The Mendelian category generated the most VUS, which is expected given that it had the largest gene list; our expanded medically actionable gene list and the ACMG subset generated fewer VUS. Given the broader scope of analysis, variant filtration, variant classification and report preparation were substantially more time consuming in the interention arm compared the control arm. Variant filtration took 7.7 times longer per case (95% CI 5.3 to 11.3, p<0.0001; 65.7 vs. 8.5 minutes on average per case). Variant classification took 13.3 times longer (95% CI 10.6 to 16.5, p<0.0001; 497.8 minutes vs. 37.4 minutes on average per case), and report preparation took 3.3 times longer (95% CI 2.6 to 4.1, p<0.0001; 157.2 minutes vs. 47.9 minutes on average per case). SFs related to monogenic disease risk were reported in 35.3% (49/139) of participants, ACMG-recommended SFs in 1.4% (2/139), carrier status in 89.3% (117/131), pharmacogenomic variants in 97.8% (135/138), and risk variants in 89.4% (118/132).
Conclusion:
While the yield of reportable SFs was high, this was accompanied by many non-reportable VUS, which substantially increased the time required for exome analysis and reporting. Analytic approaches that streamline variant prioritization and classification, such as through greater use of automation, may help promote the feasibility of offering a broad range of clinically relevant SFs.