This paper investigated the effects of hospital, cancer type, specimen acquisition method, formalin-fixed, paraffin-embedded (FFPE) block storage duration, percentage of tumor nuclei based on H&E staining of slides, percentage of tumor nuclei estimated by Foundation Med, tumor purity estimated by sequencing, success of microsatellite status testing, and DNA integrity on successful comprehensive genomic profiling (CGP) analysis using the F1CDx assay by retrospectively analyzing data from 1271 cancer patients.
Real-world data analysis for factors influencing the quality check status in FoundationOne CDx cancer genomic profiling tests.
Niitsu H, Nakahara H, Ishida K, Hinoi T, Hayes CN, Arihiro K, Kaneko Y
Purpose
Conclusion
Of the 1204 remaining cases that underwent F1Cdx testing, 1024 passed and 180 were “qualified”. F1Cdx testing failures were attributed to low DNA yields in 41 specimens (33 biopsies and 8 surgical), to failing quality control in 16 specimens (12 biopsies and 4 surgical) and to technical problems such as contamination in 10 cases. Specimens that passed had higher microsatellite instability (>95% versus approximately 65%, P<0.001), a higher percentage of tumor nuclei (by H&E at hospital and Foundation Med evaluation) or purity (computationally determined based on sequencing) (P<0.001, all), and shorter FFPE block storage duration (P<0.001), but no effect of DNA integrity, collection institution, patient age, or patient gender was found. The percentage of tumor nuclei determined by H&E staining, percentage tumor nuclei as determined by Foundation Med and computational tumor purity were all modestly correlated (all pairwise, R=0.445-0.507, ρ=0.444-0.517 and P<0.001). Based on area under the curve analysis, computational tumor purity >27.3% was a better predictor of passing (0.783) than percentage tumor nuclei (0.722 and 0.712 for H&E and Foundation Med, respectively), or block storage for <560 days (0.57). Pancreatic and biliary duct tumor specimens had a higher rate of qualified results than other tumor types. When broken down by tumor type, the percentage tumor nuclei determined at the hospital in seven of the twenty-nine tumor types was lower in cases that yielded a qualified result compared to those that passed (P<0.05), but block storage duration only significantly affected the pass versus qualified rate in pancreatic and ovarian cancers (P<0.05); further, the pass versus qualified rate only differed between biopsy and surgical specimens in pancreatic tumors (P<0.05). In multivariate analysis, low percentage tumor nuclei (<35%), long FFPE block storage duration (≥560 days), and pancreas or biliary duct cancers rather than other cancer types significantly increased the risk of a qualified rather than a pass result; of these, the largest effect was attributed to a low percentage of tumor nuclei. DNA integrity number (DIN) and the storage duration of FFPE blocks were not correlated, but DIN did vary among the cancer types investigated.
Studies (1)
Purpose
This study investigated the effects of hospital, cancer type, specimen acquisition method, formalin-fixed, paraffin-embedded (FFPE) block storage duration, percentage of tumor nuclei based on H&E staining of slides, percentage of tumor nuclei estimated by Foundation Med, tumor purity estimated based on sequencing, success of microsatellite status testing, and DNA integrity on successful comprehensive genomic profiling (CGP) analysis using the F1CDx assay by retrospectively analyzing data from 1271 cancer patients. FFPE surgical or biopsy specimens of 29 cancer types collected at six different Japanese hospitals between September 2019 and April 2024 were analyzed in the study. Unstained slides of FFPE blocks were sent to Foundation Med for testing (further details not provided). The FD1Cx assay required >20% tumor nuclei (determined by Foundation Med analysis) and 50-1000 ng of DNA. F1CDx results are classified as pass if all of the quality standards were met and qualified if any failed. For a specimen to be classified as “passed” it must have ≥ 545 ng after library construction, ≥ 140 ng after hybridization capture, ≥ 250x median exon coverage, > 100x coverage of ≥95% of the target, and < 1% sequencing error rate. DIN was determined using 10 µm FFPE sections and the DNeasy Blood & Tissue Kit and a TapeStation.
Summary of Findings
Of the 1204 remaining cases that underwent F1Cdx testing, 1024 passed and 180 were “qualified”. F1Cdx testing failures were attributed to low DNA yields in 41 specimens (33 biopsies and 8 surgical), to failing quality control in 16 specimens (12 biopsies and 4 surgical) and to technical problems such as contamination in 10 cases. Specimens that passed had higher microsatellite instability (>95% versus approximately 65%, P<0.001), a higher percentage of tumor nuclei (by H&E at hospital and Foundation Med evaluation) or purity (computationally determined based on sequencing) (P<0.001, all), and shorter FFPE block storage duration (P<0.001), but no effect of DNA integrity, collection institution, patient age, or patient gender was found. The percentage of tumor nuclei determined by H&E staining, percentage tumor nuclei as determined by Foundation Med and computational tumor purity were all modestly correlated (all pairwise, R=0.445-0.507, ρ=0.444-0.517 and P<0.001). Based on area under the curve analysis, computational tumor purity >27.3% was a better predictor of passing (0.783) than percentage tumor nuclei (0.722 and 0.712 for H&E and Foundation Med, respectively), or block storage for <560 days (0.57). Pancreatic and biliary duct tumor specimens had a higher rate of qualified results than other tumor types. When broken down by tumor type, the percentage tumor nuclei determined at the hospital in seven of the twenty-nine tumor types was lower in cases that yielded a qualified result compared to those that passed (P<0.05), but block storage duration only significantly affected the pass versus qualified rate in pancreatic and ovarian cancers (P<0.05); further, the pass versus qualified rate only differed between biopsy and surgical specimens in pancreatic tumors (P<0.05). In multivariate analysis, low percentage tumor nuclei (<35%), long FFPE block storage duration (≥560 days), and pancreas or biliary duct cancers rather than other cancer types significantly increased the risk of a qualified rather than a pass result; of these, the largest effect was attributed to a low percentage of tumor nuclei. DNA integrity number (DIN) and the storage duration of FFPE blocks were not correlated, but DIN did vary among the cancer types investigated. among the cancer types.
Biospecimens
TissueBladderTissueBoneTissueBreastTissueProstateTissuePancreasTissueSmall BowelTissueColorectalTissueLungTissueEsophagusTissueNot specifiedTissueOtherTissueStomachTissueOvaryTissueBrainTissueKidneyTissueLiverTissueUterusTissueCervixTissueSkinTissueThyroid GlandTissueThymus GlandTissueTestisTissueAdrenal GlandTissueEyePreservation Types
FormalinDiagnoses
NeoplasticCarcinomaPlatforms
DNANext generation sequencingDNAAutomated electrophoresis/BioanalyzerPre-analytical Factors
Classification Pre-Analytical Factor Value(s) Preaquisition Diagnosis/ patient condition Biliary duct cancerOtherPancreatic cancerBiospecimen Acquisition Biospecimen location Bile DuctOtherPancreasBiospecimen Acquisition Method of tissue acquisition BiopsySurgical resectionBiospecimen Aliquots and Components Biospecimen heterogeneity Differing levels of cellularityStorage Storage duration <560 days≥560 days