Skip to main content

Identification of Factors Affecting the Success of Next-Generation Sequencing Testing in Solid Tumors.

Goswami RS, Luthra R, Singh RR, Patel KP, Routbort MJ, Aldape KD, Yao H, Dang HD, Barkoh BA, Manekia J, Medeiros LJ, Roy-Chowdhuri S, Stewart J, Broaddus RR, Chen H

Publication: Am J Clin Pathol (2016) • Vol 145 • Page 222-37PMID: 27124905View on PubMed

Purpose

This paper investigated how acquisition method, tumor and tissue type, size of the microdissected area, formalin-fixed paraffin-embedded (FFPE) block storage duration, and tumor cellularity affects the percentage of specimens for which next-generation sequencing (NGS) was successful, as well as the impact of reducing  DNA input requirements. 

Conclusion

NGS was successful in 87% of FFPE specimens with 91% of failures stemming from insufficient DNA yield; however, further study revealed that NGS was successful for 73% of specimens that were initially excluded due to low DNA yield.  NGS failure rates (including those due to insufficient DNA) differed among tumor and tissue types and different sizes of the microdissected area. NGS failure rates were higher for fine-needle aspiration (FNA) or biopsy specimens than for resection specimens. The area chosen for microdissection was associated with specimen acquisition method, tumor cellularity, and DNA yield. Importantly, changing DNA input requirements such that specimens with more than 10 ng/µL DNA were included allowed NGS in 95% of specimens.Block age did not affect NGS success rates.

Studies (1)

Purpose

This study investigated how acquisition method, tumor and tissue type, size of the microdissected area, FFPE block storage duration, and tumor cellularity affects the percentage of specimens for which NGS was successful, as well the impact of reducing DNA input requirements. Areas of solid tumors containing >20% tumor cells were microdissected using an H&E stained slide as a guide, from 5 µm unstained sections of FFPE biopsy (237), surgical (342), and fine-needle aspiration (FNA) cell block (34) specimens. Bone specimens were decalcified in formic acid or remained non-decalcified. DNA was extracted from up to 20 sections of tiny tumors (1-10 mm2), 6-10 sections of small tumors (10-20 mm2), 3-5 sections of medium tumors (20-60 mm2), and 1-2 sections of large tumors (>60 mm2) using the PicoPure DNA extraction kit and purified using the Agencourt AMPure XP kit.  DNA was quantified using Qubit. If initial extraction resulted in insufficient DNA for NGS, extraction was repeated using additional sections before a specimen was considered an NGS failure.  NGS libraries were constructed using the Ion AmpliSeq Cancer Hotspot Panel v2.

Summary of Findings

Of the 614 specimens from which DNA was extracted, 67 (11%) failed NGS because less than 10 ng DNA was available, 3 (0.5%) failed library construction due to lack of amplification and 4 (0.7%) failed due to variably-sized amplicons.  NGS success rates were significantly higher for melanomas than hepatocellular carcinomas (95% versus 55%, P=0.001), gynecologic organs (100%) than for bone (63%, P<0.001) or liver (77%, P=0.002), and for resection specimens (333 of 342, 97%) than biopsy specimens (189 of 237, 80%, P<0.001) or FNA specimens (17 of 34, 50%, P<0.001), while NGS success rates were not significantly different among other tumor types nor were they different between metastasis and primary tumors. Success rates were also lower for specimens when the microdissected area was tiny (<10mm2, 69%) than when small (10-20mm2, 97%; P < 0.001), medium (20- 60mm2, 98%, P<0.01), or large (60mm2, 99.6%, P<0.01); and among small compared to  large specimens (97% versus 99.6%, P<0.01); but was not affected by block storage duration (<1 year, 1-2 years, 3-6 years, or 7-12 years). Decalcification of specimens resulted in decreased NGS success compared to non-decalcified specimens (62% versus 89%, P=0.013), but the effect may be dependent on solution used as the success rate was higher for formic acid decalcified specimens than OSH decalcified specimens (73% versus 0%). Further, NGS success rates were comparable for non-decalcified (3 of 5, 60%) decalcified (8 of 13, 62%) bone specimens.

Importantly, the size of the microdissected area may be a confounding factor to NGS success, as DNA yield was significantly correlated with the size of the microdissected area (P<0.01). The percentage of specimens that had a microdissected area classified as tiny differed based on procurement method: 6% of resection specimens, 68% (160 of 237) of biopsy specimens, and 59% of FNA specimens (10 of 17). Invasive tumor cellularity was also associated with the size of the microdissected area, as less than 30% cellularity was observed in 60% of specimens with a tiny microdissected area and 33% of specimens with a medium or large microdissected area.

Of the 22 specimens that were originally excluded from NGS analysis due to insufficient DNA (<10 ng), NGS was successful for 16 while 6 failed library preparation. NGS library construction was possible for 15 of the 17 specimens with 1.2 to 10 ng DNA but was successfully generated for only 1 of the 5 (20%) specimens with <1.2 ng DNA. Sanger sequencing was used to confirm variants in 10 cases with eight allowing for complete verification and two only partial verification. Using another cohort of 408 specimens, NGS was successful in 88% of specimens including 28 of 33 specimens (85%) with 1.2-10 ng DNA. Further, exclusion of specimens with <10 ng increased the NGS success rate to 95%. The effect of excluding specimens with <10 ng had the biggest effect on cytology blocks where success increased from 58% to 87% (P<0.01) and biopsy specimens where success increased from 84% to 95% (P<0.01). Even with the increased success observed after exclusion of specimens with <10 ng, a small difference remained between FNA and resection specimens (87% versus 96%, P=0.03). Block age did not affect NGS success rates.

  • Biospecimens

    TissueBone
    TissueBrain
    TissueBreast
    TissueColorectal
    TissueOvary
    TissueVagina
    TissueUterus
    TissueKidney
    TissueHead and Neck
    TissueLiver
    TissueLung
    TissueLymph Node
    TissueEsophagus
    TissueStomach
    TissueSkin
  • Preservation Types

    Formalin
  • Diagnoses

    NeoplasticCarcinoma
    NeoplasticSarcoma
    NeoplasticMelanoma
    NeoplasticOther
    NeoplasticBenign
  • Platforms

    DNANext generation sequencing
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    PreaquisitionDiagnosis/ patient condition
    Adenocarcinoma
    Astrocytic tumor
    Hepatocellular carcinoma
    Melanoma
    Neuroendocrine carcinoma
    Sarcoma
    Serous carcinoma
    Squamous cell carcinoma
    Unspecified Carcinoma
    Unspecified brain tumor
    Biospecimen AcquisitionBiospecimen location
    Bone
    Brain
    Breast
    Colorectal
    Genitourinary organs
    Gynecologic organs
    Head/Neck
    Liver
    Lung
    Lymph
    Metastasis
    Other deep-seated organs
    Primary tumor
    Skin
    Soft tissue
    Upper gastrointestinal organs
    Biospecimen Aliquots and ComponentsAliquot size/volume
    10-20 mm2
    20-60 mm2
    <10mm2
    >60 mm2
    StorageStorage duration
    1-2 years
    3-6 years
    7-12 years
    <1 year
    Analyte Extraction and PurificationDecalcification solution/ duration
    Decalcified
    Not decalcified
    Next generation sequencing SpecificTechnology platform
    Sanger Sequencing
    Next generation sequencing SpecificTemplate/input amount
    1.2-10 ng DNA
    >10 ng DNA