Dataset Information

Performance of a Genomic Sequencing Classifier for the Preoperative Diagnosis of Cytologically Indeterminate Thyroid Nodules.

ABSTRACT:

Importance

Use of next-generation sequencing of RNA and machine learning algorithms can classify the risk of malignancy in cytologically indeterminate thyroid nodules to limit unnecessary diagnostic surgery.

Objective

To measure the performance of a genomic sequencing classifier for cytologically indeterminate thyroid nodules.

Design, setting, and participants

A blinded validation study was conducted on a set of cytologically indeterminate thyroid nodules collected by fine-needle aspiration biopsy between June 2009 and December 2010 from 49 academic and community centers in the United States. All patients underwent surgery without genomic information and were assigned a histopathology diagnosis by an expert panel blinded to all genomic information. There were 210 potentially eligible thyroid biopsy samples with Bethesda III or IV indeterminate cytopathology that constituted a cohort previously used to validate the gene expression classifier. Of these, 191 samples (91.0%) had adequate residual RNA for validation of the genomic sequencing classifier. Algorithm development and independent validation occurred between August 2016 and May 2017.

Exposures

Thyroid nodule surgical histopathology diagnosis by an expert panel blinded to all genomic data.

Main outcomes and measures

The primary end point was measurement of genomic sequencing classifier sensitivity, specificity, and negative and positive predictive values in biopsies from Bethesda III and IV nodules. The secondary end point was measurement of classifier performance in biopsies from Bethesda II, V, and VI nodules.

Results

Of the 183 included patients, 142 (77.6%) were women, and the mean (range) age was 51.7 (22.0-85.0) years. The genomic sequencing classifier had a sensitivity of 91% (95% CI, 79-98) and a specificity of 68% (95% CI, 60-76). At 24% cancer prevalence, the negative predictive value was 96% (95% CI, 90-99) and the positive predictive value was 47% (95% CI, 36-58).

Conclusions and relevance

The genomic sequencing classifier demonstrates high sensitivity and accuracy for identifying benign nodules. Its 36% increase in specificity compared with the gene expression classifier potentially increases the number of patients with benign nodules who can safely avoid unnecessary diagnostic surgery.

SUBMITTER: Patel KN

PROVIDER: S-EPMC6583881 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:ImportanceApproximately 20% of fine-needle aspirations (FNA) of thyroid nodules have indeterminate cytology, most frequently Bethesda category III or IV. Diagnostic surgeries can be avoided for these patients if the nodules are reliably diagnosed as benign without surgery.ObjectiveTo determine the diagnostic accuracy of a multigene classifier (GC) test (ThyroSeq v3) for cytologically indeterminate thyroid nodules.Design, setting, and participantsProspective, blinded cohort study conducted at 10 medical centers, with 782 patients with 1013 nodules enrolled. Eligibility criteria were met in 256 patients with 286 nodules; central pathology review was performed on 274 nodules.InterventionsA total of 286 FNA samples from thyroid nodules underwent molecular analysis using the multigene GC (ThyroSeq v3).Main outcomes and measuresThe primary outcome was diagnostic accuracy of the test for thyroid nodules with Bethesda III and IV cytology. The secondary outcome was prediction of cancer by specific genetic alterations in Bethesda III to V nodules.ResultsOf the 286 cytologically indeterminate nodules, 206 (72%) were benign, 69 (24%) malignant, and 11 (4%) noninvasive follicular thyroid neoplasms with papillary-like nuclei (NIFTP). A total of 257 (90%) nodules (154 Bethesda III, 93 Bethesda IV, and 10 Bethesda V) had informative GC analysis, with 61% classified as negative and 39% as positive. In Bethesda III and IV nodules combined, the test demonstrated a 94% (95% CI, 86%-98%) sensitivity and 82% (95% CI, 75%-87%) specificity. With a cancer/NIFTP prevalence of 28%, the negative predictive value (NPV) was 97% (95% CI, 93%-99%) and the positive predictive value (PPV) was 66% (95% CI, 56%-75%). The observed 3% false-negative rate was similar to that of benign cytology, and the missed cancers were all low-risk tumors. Among nodules testing positive, specific groups of genetic alterations had cancer probabilities varying from 59% to 100%.Conclusions and relevanceIn this prospective, blinded, multicenter study, the multigene GC test demonstrated a high sensitivity/NPV and reasonably high specificity/PPV, which may obviate diagnostic surgery in up to 61% of patients with Bethesda III to IV indeterminate nodules, and up to 82% of all benign nodules with indeterminate cytology. Information on specific genetic alterations obtained from FNA may help inform individualized treatment of patients with a positive test result.

Project description:Background: Gene panels are routinely used to assess predisposition to hereditary cancers by simultaneously testing multiple susceptibility genes and/or variants. More recently, genetic panels have been implemented as part of solid tumor malignancy testing assessing somatic alterations. One example is targeted variant panels for thyroid nodules that are not conclusively malignant or benign upon fine-needle aspiration (FNA). We systematically reviewed published studies from 2009 to 2018 that contained genetic data from preoperative FNA specimens on cytologically indeterminate thyroid nodules (ITNs) that subsequently underwent surgical resection. Pooled prevalence estimates per gene and variant, along with their respective positive predictive values (PPVs) for malignancy, were calculated. Summary: Our systematic search identified 540 studies that were supplemented by 18 studies from bibliographies or personal files. Sixty-one studies met all inclusion criteria and included >4600 ITNs. Overall, 26% of nodules contained at least 1 variant or fusion. However, half of them did not include details on the specific gene, variant, and/or complete fusion pair reported for inclusion toward PPV calculations. The PPVs of genomic alterations reported at least 10 times were limited to BRAFV600E (98%, 95% confidence interval [CI 96-99%]), PAX8/PPARG (55% [CI 34-78%]), HRASQ61R (45% [CI 22-72%]), BRAFK601E (42% [CI 19-68%]), and NRASQ61R (38% [CI 23-55%]). Excluding BRAFV600E, the pooled PPV for all other specified variants and fusions was 47%. Multiple variants within the same nodule were identified in ∼1% of ITN and carried a cumulative PPV of 77%. Conclusions: The chance that a genomic alteration predicts malignancy depends on the individual variant or fusion detected. Only five alterations were reported at least 10 times; BRAFV600E had a PPV of 98%, while the remaining four had individual PPVs ranging from 38% to 55%. The small sample size of most variants and fusion pairs found among ITNs, however, limits confidence in their individual PPV point estimates. Better specific reporting of genomic alterations with cytological category, histological subtype, and cancer staging would facilitate better understanding of cancer prediction, and the independent contribution of the genomic profile to prognosis.