Development and Validation of a Gene Signature Classifier for Consensus Molecular Subtyping of Colorectal Carcinoma in a CLIA-Certified Setting

PURPOSE

Consensus molecular subtyping (CMS) of colorectal cancer has potential to reshape the colorectal cancer landscape. We developed and validated an assay that is applicable on formalin-fixed, paraffin-embedded (FFPE) samples of colorectal cancer and implemented the assay in a Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory.

EXPERIMENTAL DESIGN

We performed an in silico experiment to build an optimal CMS classifier using a training set of 1,329 samples from 12 studies and validation set of 1,329 samples from 14 studies. We constructed an assay on the basis of NanoString CodeSets for the top 472 genes, and performed analyses on paired flash-frozen (FF)/FFPE samples from 175 colorectal cancers to adapt the classifier to FFPE samples using a subset of genes found to be concordant between FF and FFPE, tested the classifier's reproducibility and repeatability, and validated in a CLIA-certified laboratory. We assessed prognostic significance of CMS in 345 patients pooled across three clinical trials.

RESULTS

The best classifier was weighted support vector machine with high accuracy across platforms and gene lists (>0.95), and the 472-gene model outperforming existing classifiers. We constructed subsets of 99 and 200 genes with high FF/FFPE concordance, and adapted FFPE-based classifier that had strong classification accuracy (>80%) relative to "gold standard" CMS. The classifier was reproducible to sample type and RNA quality, and demonstrated poor prognosis for CMS1-3 and good prognosis for CMS2 in metastatic colorectal cancer (P < 0.001).

CONCLUSIONS

We developed and validated a colorectal cancer CMS assay that is ready for use in clinical trials, to assess prognosis in standard-of-care settings and explore as predictor of therapy response.

Detection of the MYD88 p.L265P Mutation in the CSF of a Patient With Secondary Central Nervous System Lymphoma

Primary Central Nervous System Lymphoma (PCNSL) and Metastatic (or Secondary) Central Nervous System Lymphoma (MCNSL) are rare central nervous system (CNS) malignancies that exhibit aggressive clinical behavior and have a poor prognosis. The majority of CNS lymphomas are histologically classified as diffuse large-B cell lymphoma (DLBCL). DLBCL harbors a high frequency of mutations in MYD88 and CD79b. The MYD88 p.L265P mutation occurs at high frequency in CNS lymphoma and is extremely rare in non-hematologic malignancies. Currently, brain biopsy is considered the gold standard for CNS lymphoma diagnosis. However, brain biopsy is invasive, carries a risk of complications, and can delay initiation of systemic therapy. Circulating tumor DNA (ctDNA) in the cerebrospinal fluid (CSF) can be utilized to detect tumor-derived mutations. Testing of CSF-ctDNA is a minimally-invasive methodology that can be used to assess the genomic alterations present in CNS malignancies. We present a case of an 82-year-old man with a history of testicular lymphoma who presented with speech difficulty and a multifocal enhancing left inferior frontal mass. Analysis for both CSF-cytology and flow cytometry did not show evidence of neoplastic cells. A brain biopsy was performed and microscopic examination showed DLBCL. We isolated CSF-ctDNA and used droplet digital PCR (ddPCR) to detect the most common lymphoma-associated mutations in MYD88, L265P, and V217F. In conjunction, we evaluated the patient-matched CNS lymphoma tissue for MYD88 mutations. We detected the MYD88 p.L265P mutation in formalin fixed paraffin embedded (FFPE) tissue from the brain biopsy and the CSF-ctDNA. In contrast, both the tumor tissue and the CSF ctDNA were negative for the MYD88 p.V217F mutation. This study shows that testing CSF ctDNA for MYD88 mutations is a potentially minimally-invasive approach to diagnosing patients with suspected CNS lymphomas.

A model to predict the histopathology of human stroke using diffusion and T2-weighted magnetic resonance imaging

BACKGROUND AND PURPOSE

We sought to identify MRI measures that have high probability in a short acquisition time to predict, at early time points after onset of ischemia, the eventual development of cerebral infarction in clinical patients who suffer occlusion of a cerebral artery.

METHODS

We developed an MR tissue signature model based on experimentally derived relationships of the apparent diffusion coefficient of water (ADCw) and T2 to ischemic brain tissue histopathology. In eight stroke patients we measured ADCw and T2 intensity using diffusion-weighted echo-planar imaging (DW-EPI). Tissue signature regions were defined, and theme maps of the ischemic focus at subacute time points after stroke onset were generated.

RESULTS

Five MR signatures were identified in human stroke foci: two that may predict either cell recovery or progression to necrosis, one that may mark the transition to cell necrosis, and two that may be markers of established cell necrosis.

CONCLUSIONS

An MR tissue signature model of ischemic histopathology using ADCw and T2 can now be tested for its potential to predict reversible and identify irreversible cellular damage in human ischemic brain regions.

Quantification of thallium-201 scintigrams in acute myocardial infarction

A method has been developed for measurement of myocardial infarct size from thallium-201 scintigrams that depends on computer measurement of levels of radioactivity in the myocardium. In 16 dogs, thallium-201 scintigrams were obtained in the left lateral and left anterior oblique projections 48 hours after ligation of the left anterior descending coronary artery. Scintigraphic results were obtained by two independent observers and were compared with tissue measurements of infarct volume calculated from thallium autoradiograms and nitro-blue tetrazolium (NBT)-stained tissue slices. Infarct volumes derived from tissue measurements were used to develop criteria for the computer scintigraphic technique. There was no significant difference in the scintigraphic measurements made by the two observers. Scintigraphic infarct size in the left lateral and left anterior oblique projections correlated with tissue infarct size with r values of 0.88 and 0.75, respectively, for thallium autoradiography and 0.71 and 0.70, respectively, for NBT tissue staining. The range of infarct volume was 3.3 to 14.8 percent of the left ventricular mass. Results of this study suggest that scintigraphic quantitation of infarct size is feasible in this dog model.