General Applicability of Existing College of American Pathologists Accreditation Requirements to Clinical Implementation of Machine Learning-Based Methods in Molecular Oncology Testing.

CONTEXT.­: The College of American Pathologists (CAP) accreditation requirements for clinical laboratory testing help ensure laboratories implement and maintain systems and processes that are associated with quality. Machine learning (ML)-based models share some features of conventional laboratory testing methods. Accreditation requirements that specifically address clinical laboratories' use of ML remain in the early stages of development. OBJECTIVE.­: To identify relevant CAP accreditation requirements that may be applied to the clinical adoption of ML-based molecular oncology assays, and to provide examples of current and emerging ML applications in molecular oncology testing. DESIGN.­: CAP accreditation checklists related to molecular pathology and general laboratory practices (Molecular Pathology, All Common and Laboratory General) were reviewed. Examples of checklist requirements that are generally applicable to validation, revalidation, quality management, infrastructure, and analytical procedures of ML-based molecular oncology assays were summarized. Instances of ML use in molecular oncology testing were assessed from literature review. RESULTS.­: Components of the general CAP accreditation framework that exist for traditional molecular oncology assay validation and maintenance are also relevant for implementing ML-based tests in a clinical laboratory. Current and emerging applications of ML in molecular oncology testing include DNA methylation profiling for central nervous system tumor classification, variant calling, microsatellite instability testing, mutational signature analysis, and variant prediction from histopathology images. CONCLUSIONS.­: Currently, much of the ML activity in molecular oncology is within early clinical implementation. Despite specific considerations that apply to the adoption of ML-based methods, existing CAP requirements can serve as general guidelines for the clinical implementation of ML-based assays in molecular oncology testing.

CLONAL DYNAMICS AND RELAPSE RISK REVEALED BY HIGH SENSITIVITY FLT3-ITD DETECTION IN ACUTE MYELOID LEUKEMIA.

The ability to detect low level disease is key to our understanding of clonal heterogeneity in acute myeloid leukemia (AML) and residual disease that elude conventional assays and seed relapse. We have developed a high sensitivity next generation sequencing (HS-NGS) clinical assay, able to reliably detect low levels (1x10-5) of FLT3-ITD, a frequent, therapeutically targetable and prognostically relevant mutation in AML. By applying this assay to 289 longitudinal samples from 62 patients at initial diagnosis and/or clinical follow up (mean follow-up of 22 months) we reveal the frequent occurrence of FLT3-ITD subclones at diagnosis and demonstrate a significantly decreased relapse risk when FLT3-ITD is cleared after induction or thereafter. We perform pairwise sequencing of diagnosis and relapse samples from 23 patients to uncover more detailed patterns of FLT3-ITD clonal evolution at relapse than is detectable by less sensitive assays. Lastly, we show that rising ITD level during consecutive biopsies is a harbinger of impending relapse. Our findings corroborate the emerging clinical utility of high sensitivity FLT3-ITD testing and expands our understanding of clonal dynamics in FLT3-ITD positive AML.

No Remdesivir Resistance Observed in the Phase 3 Severe and Moderate COVID-19 SIMPLE Trials.

Remdesivir (RDV) is a broad-spectrum nucleotide analog prodrug approved for the treatment of COVID-19 in hospitalized and non-hospitalized patients with clinical benefit demonstrated in multiple Phase 3 trials. Here we present SARS-CoV-2 resistance analyses from the Phase 3 SIMPLE clinical studies evaluating RDV in hospitalized participants with severe or moderate COVID-19 disease. The severe and moderate studies enrolled participants with radiologic evidence of pneumonia and a room-air oxygen saturation of ≤94% or >94%, respectively. Virology sample collection was optional in the study protocols. Sequencing and related viral load data were obtained retrospectively from participants at a subset of study sites with local sequencing capabilities (10 of 183 sites) at timepoints with detectable viral load. Among participants with both baseline and post-baseline sequencing data treated with RDV, emergent Nsp12 substitutions were observed in 4 of 19 (21%) participants in the severe study and none of the 2 participants in the moderate study. The following 5 substitutions emerged: T76I, A526V, A554V, E665K, and C697F. The substitutions T76I, A526V, A554V, and C697F had an EC50 fold change of ≤1.5 relative to the wildtype reference using a SARS-CoV-2 subgenomic replicon system, indicating no significant change in the susceptibility to RDV. The phenotyping of E665K could not be determined due to a lack of replication. These data reveal no evidence of relevant resistance emergence and further confirm the established efficacy profile of RDV with a high resistance barrier in COVID-19 patients.

The cytological features of effusions with mesothelioma in situ: A report of 9 cases.

INTRODUCTION: The diagnosis of mesothelioma in situ (MIS) is now accepted by the WHO as a pre-invasive neoplastic mesothelial proliferation and considered a diagnosis based on histologic evaluation only. Although the definition of MIS includes recurrent effusions, little is known about the cytologic features of such effusions. Since mesothelioma is usually diagnosed at an advanced stage and has a poor prognosis, early detection of a neoplastic mesothelial population in such effusions can potentially have a positive impact on the management of such a dire disease. MATERIALS AND METHODS: We reviewed a total of 18 pleural effusions from nine patients with recurrent effusions. Of these, five patients had follow-up biopsies diagnosed as MIS and the remaining four cases had negative radiology and malignant cytology proven by molecular markers (BAP1, MTAP or CDKN2A deletion) and at least 1 year follow-up with no overt mass identified by radiology. RESULTS: Initial effusions may mimic reactive mesothelial hyperplasia or exhibit atypia. As effusions recur, the cellularity and atypia increase and the mesothelial proliferation becomes morphologically indistinguishable from mesothelioma. Molecular alterations diagnostic of mesothelioma can be detected in these effusions, even in the initial-benign/reactive appearing ones. The cellularity and atypia detected in such effusions surpassed those noted on the biopsies, raising questions regarding the cause of such discrepancy. CONCLUSION: The diagnosis of MIS can be suspected based on malignant effusion cytology supported by molecular alterations. We propose that the proliferation of neoplastic mesothelial clones represent a clinically silent "liquid phase MIS stage" corresponding to in situ stage in other organs.

Tailored Treatment Based on Helicobacter pylori Genetic Markers of Resistance Is Associated With Higher Eradication Success.

INTRODUCTION: Increasing antimicrobial resistance with Helicobacter pylori infection has focused efforts to tailor eradication therapy based on identifying genetic markers of resistance to predict antimicrobial susceptibility. METHODS: In this retrospective study, we report the effect of routine inclusion of antimicrobial susceptibility testing and recommendations for eradication therapy with gastric specimens with H. pylori . RESULTS: The use of a recommended treatment regimen based on genetic markers of resistance was associated with an 84% rate of eradication success and 4.4 greater odds of eradication relative to unrecommended treatment. DISCUSSION: This is the first study describing the use of H. pylori genetic resistance testing as standard of care.

Late Recurrence of Colorectal Carcinoma in Patients with Malignant Polyp and Risk Factors.

Malignant polyps are polypoid lesions that appear benign endoscopically but harbor invasive adenocarcinoma microscopically. Patient with diagnosis of malignant polyp can be managed by surgical resection or endoscopic surveillance. Current literature on long term recurrence is sparse. A total of 76 patients with malignant polyp and follow-up period of over one year are included. Of these, 28 patients underwent endoscopic polypectomy followed by surveillance (group 1). Forty-eight patients underwent segmental colectomy (group 2). In group 1, three patients developed local recurrent pT3 adenocarcinoma (5.9 to 9.7 years) and one patient developed liver metastasis (7.3 years). One patient presented with malignant polyp in another segment of colon (4.0 years). Two of the malignant polyps with local recurrence do not have commonly reported high-risk features, including tumor ≤ 1 mm from resection margin, presence of lymphovascular invasion and high grade tumor, they had invasion depth of >4 mm and harbored a TP53 missense mutation. In group 2, during the follow-up period (1.0-21.8 years, median 9.3 years), none of the patients developed local recurrence. In this study, surveillance group had a local late recurrence rate of 10.7% versus no local recurrence in surgical resection group (0%). Our study shows that depth of invasion of over 4 mm in malignant polyp is a risk factor for late local recurrence if managed by endoscopic surveillance. Further study is needed to explore whether certain molecular alterations, such as TP53 mutation, is a risk factor for late recurrence.

Identification of a Cancer-Predisposing Germline POT1 p.Ile49Metfs*7 Variant by Targeted Sequencing of a Splenic Marginal Zone Lymphoma.

Germline disruptive variants in Protection of Telomeres 1 (POT1) predispose to a wide variety of cancers, including melanoma, chronic lymphocytic leukemia (CLL), Hodgkin lymphoma, myeloproliferative neoplasms, and glioma. We report the first case of splenic marginal zone lymphoma (SMZL) arising in a patient with a germline POT1 variant: a 65-year-old male with an extensive history of cancer, including melanoma and papillary thyroid carcinoma, who presented with circulating atypical lymphocytosis. Bone marrow biopsy revealed 20% involvement by a CD5-CD10- B-cell lymphoma that was difficult to classify. During the clinical workup of his low-grade lymphoma, targeted next-generation sequencing (NGS) identified POT1 p.I49Mfs*7 (NM_015450:c. 147delT) at a variant allele frequency (VAF) of 51%. NGS of skin fibroblasts confirmed the POT1 variant was germline. This likely pathogenic POT1 loss-of-function variant has only been reported once before as a germline variant in a patient with glioma and likely represents one of the most deleterious germline POT1 variants ever linked to familial cancer. The spectrum of cancers associated with germline pathogenic POT1 variants (i.e., autosomal dominant POT1 tumor predisposition syndrome) should potentially be expanded to include SMZL, a disease often associated with the loss of chromosome 7q: the location of the POT1 genetic locus (7q31.33).

Deep Convolutional Neural Networks Implementation for the Analysis of Urine Culture.

BACKGROUND: Urine culture images collected using bacteriology automation are currently interpreted by technologists during routine standard-of-care workflows. Machine learning may be able to improve the harmonization of and assist with these interpretations. METHODS: A deep learning model, BacterioSight, was developed, trained, and tested on standard BD-Kiestra images of routine blood agar urine cultures from 2 different medical centers. RESULTS: BacterioSight displayed performance on par with standard-of-care-trained technologist interpretations. BacterioSight accuracy ranged from 97% when compared to standard-of-care (single technologist) and reached 100% when compared to a consensus reached by a group of technologists (gold standard in this study). Variability in image interpretation by trained technologists was identified and annotation "fuzziness" was quantified and found to correlate with reduced confidence in BacterioSight interpretation. Intra-testing (training and testing performed within the same institution) performed well giving Area Under the Curve (AUC) ≥0.98 for negative and positive plates, whereas, cross-testing on images (trained on one institution's images and tested on images from another institution) showed decreased performance with AUC ≥0.90 for negative and positive plates. CONCLUSIONS: Our study provides a roadmap on how BacterioSight or similar deep learning prototypes may be implemented to screen for microbial growth, flag difficult cases for multi-personnel review, or auto-verify a subset of cultures with high confidence. In addition, our results highlight image interpretation variability by trained technologist within an institution and globally across institutions. We propose a model in which deep learning can enhance patient care by identifying inherent sample annotation variability and improving personnel training.

High-Throughput Adaptable SARS-CoV-2 Screening for Rapid Identification of Dominant and Emerging Regional Variants.

OBJECTIVES: Emerging severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variant strains can be associated with increased transmissibility, more severe disease, and reduced effectiveness of treatments. To improve the availability of regional variant surveillance, we describe a variant genotyping system that is rapid, accurate, adaptable, and able to detect new low-level variants built with existing hospital infrastructure. METHODS: We used a tiered high-throughput SARS-CoV-2 screening program to characterize variants in a supraregional health system over 76 days. Combining targeted reverse transcription-polymerase chain reaction (RT-PCR) and selective sequencing, we screened SARS-CoV-2 reactive samples from all hospitals within our health care system for genotyping dominant and emerging variants. RESULTS: The median turnaround for genotyping was 2 days using the high-throughput RT-PCR-based screen, allowing us to rapidly characterize the emerging Delta variant. In our population, the Delta variant is associated with a lower cycle threshold value, lower age at infection, and increased vaccine-breakthrough cases. Detection of low-level and potentially emerging variants highlights the utility of a tiered approach. CONCLUSIONS: These findings underscore the need for fast, low-cost, high-throughput monitoring of regional viral sequences as the pandemic unfolds and the emergence of SARS-CoV-2 variants increases. Combining RT-PCR-based screening with selective sequencing allows for rapid genotyping of variants and dynamic system improvement.

Flow Cytometric Findings in Clonal Cytopenia of Undetermined Significance.

OBJECTIVES: To examine flow cytometric (FCM) findings in clonal cytopenia of undetermined significance (CCUS) in relation to variant allele fraction (VAF) and mutation risk. METHODS: Nine FCM parameters, including 5 FCM metrics (Meyerson-Alayed scoring scheme [MASS] parameters) we previously used to identify myelodysplastic syndromes (MDS), were compared among 96 CCUS samples, 100 low-grade MDS samples and 100 samples from patients without somatic alterations (controls). RESULTS: FCM findings did not differ between CCUS samples with less than 20% VAF and controls. CCUS samples with more than 20% VAF (CCUS >20% VAF) demonstrated more than 1 abnormal FCM parameter at a frequency between MDS and controls. Abnormalities in CCUS with high-risk alterations (CCUS(hi)) were similar to MDS, with no statistical difference in the percentage of cases with more than 1 FCM abnormality or a positive MASS score. The positive predictive value (PPV) for clinically significant myeloid processes; MDS, CCUS(hi), and CCUS >20% VAF compared with other CCUS samples and controls was 94.8%, with 96.5% specificity and 61% sensitivity using a modified MASS score. A subset of MDS (43%) was distinguished from CCUS(hi) and CCUS >20% VAF using 3 parameters, with a 93.5% PPV and 83.3% specificity. CONCLUSIONS: FCM abnormalities can distinguish high-risk CCUS based on VAF or alteration type from low-risk CCUS and MDS in many cases. The findings are of potential utility in the evaluation of patients with cytopenias.

Clinical Utility of Targeted Next-Generation Sequencing in the Evaluation of Low-Grade Lymphoproliferative Disorders.

OBJECTIVES: We investigated the usefulness of a custom-designed 31-gene next-generation sequencing (NGS) panel implemented on a routine basis for the evaluation of low-grade lymphoproliferative disorders (LPDs). METHODS: In total, 147 blood, bone marrow, and tissue specimens were sequenced, including 81% B-cell, 15% T-cell, and 3% natural killer (NK)-cell neoplasms. RESULTS: Of the cases, 92 (63%) of 147 displayed at least one pathogenic variant while 41 (28%) of 147 had two or more. Low mutation rates were noted in monoclonal B-cell lymphocytoses and samples with small T- and NK-cell clones of uncertain significance. Pathogenic molecular variants were described in specific disorders and classified according to their diagnostic, prognostic, and potential therapeutic value. Diagnostically, in addition to confirming the diagnosis of 15 of 15 lymphoplasmacytic lymphomas, 10 of 12 T large granular lymphocytic leukemias, and 2 of 2 hairy cell leukemias (HCLs), the panel helped resolve the diagnosis of 10 (62.5%) of 16 challenging cases lacking a specified diagnosis based on standard morphology, phenotype, and genetic analysis. CONCLUSIONS: Overall, implementation of this targeted lymphoid NGS panel as part of regular hematopathology practice was found to be a beneficial adjunct in the evaluation of low-grade LPDs.

Specificity of SARS-CoV-2 Real-Time PCR Improved by Deep Learning Analysis.

Real-time PCR (RT-PCR) is widely used to diagnose human pathogens. RT-PCR data are traditionally analyzed by estimating the threshold cycle (CT ) at which the fluorescence signal produced by emission of a probe crosses a baseline level. Current models used to estimate the CT value are based on approximations that do not adequately account for the stochastic variations of the fluorescence signal that is detected during RT-PCR. Less common deviations become more apparent as the sample size increases, as is the case in the current SARS-CoV-2 pandemic. In this work, we employ a method independent of CT value to interpret RT-PCR data. In this novel approach, we built and trained a deep learning model, qPCRdeepNet, to analyze the fluorescent readings obtained during RT-PCR. We describe how this model can be deployed as a quality assurance tool to monitor result interpretation in real time. The model's performance with the TaqPath COVID19 Combo Kit assay, widely used for SARS-CoV-2 detection, is described. This model can be applied broadly for the primary interpretation of RT-PCR assays and potentially replace the CT interpretive paradigm.

Clonal cytopenia of undetermined significance (CCUS) with dysplasia is enriched for MDS-type molecular findings compared to CCUS without dysplasia.

OBJECTIVES: Although morphologic dysplasia is not typically considered a feature of CCUS, we have consistently observed low-level bone marrow (BM) dysplasia among CCUS patients. We sought to determine whether sub-diagnostic BM dysplasia in CCUS patients is associated with other clinico-pathologic findings of myelodysplastic syndrome (MDS). METHODS: We identified 49 CCUS patients, 25 with sub-diagnostic dysplasia (CCUS-D), and 24 having no dysplasia (CCUS-ND). We compared the clinical, histologic, and laboratory findings of CCUS-D and CCUS-ND patients to 49 MDS patients, including blood cell counts, BM morphology, flow cytometry, cytogenetics, and results of next-generation sequencing. RESULTS: No statistically significant differences were observed between CCUS-D and CCUS-ND patients in the degree of cytopenias, BM cellularity, myeloid-to-erythroid ratio, or the presence of flow cytometric abnormalities. However, compared to CCUS-ND, CCUS-D patients exhibited increased mutations in myeloid malignancy-associated genes, including non-TET2/DNMT3A/ASXL1 variants, spliceosome (SF3B1, SRSF2, ZRSR2, or U2AF1) variants, and IDH2/RUNX1/CBL variants. CCUS-D patients were also enriched for higher variant allele frequencies and co-mutation of TET2/DNMT3A/ASXL1 with other genes. CONCLUSIONS: CCUS-D patients exhibit a molecular (but not clinical) profile more similar to MDS patients than CCUS-ND, suggesting CCUS-D may represent a more immediate precursor to MDS and may warrant closer clinical follow-up.

Molecular Profiling of Pediatric and Adult Glioblastoma.

OBJECTIVES: Although glioblastoma (GBM) is rare in the pediatric population, it is the most common cause of death among children with central nervous system neoplasms. Recent molecular profiling of these neoplasms has demonstrated distinct differences in comparison to their adult counterparts. Moreover, many pediatric GBMs occur within the context of cancer predisposition syndromes, such as constitutional mismatch repair deficiency syndrome (CMMRD). Children with CMMRD who develop GBM exhibit a high tumor mutational burden and may benefit from treatment with immune checkpoint inhibitors. METHODS: We performed next-generation sequencing and immunohistochemistry for mismatch repair proteins in our cohort of pediatric and adult GBMs to further characterize the molecular profiles of these groups. RESULTS: We examined a total of 11 pediatric and 11 adult GBMs. Pediatric patients had a higher number of alterations compared to their adult counterparts. They also had a higher frequency of alterations in the mismatch repair genes, which can be detected by immunohistochemistry (IHC). We also identified one pediatric patient with CMMRD syndrome. CONCLUSIONS: Our study highlighted the distinct molecular differences between pediatric and adult GBM. We also demonstrated that pediatric patients have a higher frequency of alterations in the mismatch repair genes, which may render them susceptible to treatment with immune checkpoint inhibitors. These alterations can be detected using routine IHC and should be performed on all pediatric GBM.

Use of cytology centrifuged supernatants improves cost and turnaround time for targeted next generation sequencing.

BACKGROUND: Molecular testing is an essential step in providing patients with advanced non-small-cell lung cancer (NSCLC), the most appropriate front-line targeted therapies. We recently implemented targeted NGS on previously discarded cytology centrifuged supernatant (CCS). METHODS: In this study, we reviewed our implementation process to evaluate its performance. Performance and turnaround time (TAT) of molecular testing on all cytology NSCLC cases submitted for targeted NGS from June 2018 to September 2019 were evaluated, which included 46 and 62 cytology cases before and after implementation of CCS, respectively. Associated cost savings using CCS was also analyzed. RESULTS: The mean TAT defined as the time of collection to time of reporting was 8.5 ± 1.8 days in CCS cohort (range 5-13) as compared with 12.2 ± 5.3 days in the (FFPE) cell block (CB) cohort (range: 6-27). The success rate of sequencing was similar for both cohorts (100% in CCS and 96% in FFPE CB). CONCLUSION: Our results demonstrate that NGS using CCS improves TAT, preserves FFPE CB for other testing, and results in cost savings of $50 per case.

Molecular Detection of SARS-CoV-2 Infection in FFPE Samples and Histopathologic Findings in Fatal SARS-CoV-2 Cases.

OBJECTIVES: To report methods and findings of 2 autopsies with molecular evaluation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) positive individuals. METHODS: Postmortem examination was completed following Centers for Disease Control and Prevention public guidelines. Numerous formalin-fixed paraffin-embedded (FFPE) tissue types from each case were surveyed for SARS-CoV-2 RNA by quantitative reverse transcription polymerase chain reaction (qRT-PCR). SARS-CoV-2 viral genome was sequenced by next-generation sequencing (NGS) from FFPE lung tissue blocks. RESULTS: Postmortem examinations revealed diffuse alveolar damage, while no viral-associated hepatic, cardiac, or renal damage was observed. Viral RNA was detected in lungs, bronchi, lymph nodes, and spleen in both cases using qRT-PCR method. RNA sequencing using NGS in case 1 revealed mutations most consistent with Western European Clade A2a with ORF1a L3606F mutation. CONCLUSIONS: SARS-CoV-2 testing and viral sequencing can be performed from FFPE tissue. Detection and sequencing of SARS-CoV-2 in combination with morphological findings from postmortem tissue examination can aid in gaining a better understanding of the virus's pathophysiologic effects on human health.

Tiny but mighty: use of next generation sequencing on discarded cytocentrifuged bile duct brushing specimens to increase sensitivity of cytological diagnosis.

Bile duct brushing (BDB) is used to evaluate pancreatobiliary lesions as it widely samples lesions with a low complication rate. Cytological evaluation of BDB is a specific but insensitive test. There is limited literature on the use of post-cytocentrifuged (PCC) samples, which are usually discarded, for next-generation sequencing (NGS) as an adjunct to cytological diagnosis of BDB. In this study we investigate whether molecular analysis by NGS of PCC specimens improves the sensitivity of diagnosis. PCC samples from 100 consecutive BDB specimens spanning 93 unique patients were retained. DNA was extracted and mutational analysis was performed agnostic of morphologic or clinical findings. Each BDB specimen was characterized as negative, atypical or positive based on morphological analysis by trained cytopathologists. Performance characteristics for mutational profiling and morphological analysis were calculated on the basis of clinicopathologic follow-up. There was sufficient clinicopathologic follow-up to classify 94 of 100 cases as either malignant (n = 43) or benign (n = 51). Based on morphologic analysis of cytology, these 94 cases were classified as either benign (n = 55), atypical (n = 18), or as at least suspicious or positive for malignancy (n = 21). Morphologic analysis of cytology showed a sensitivity of 49% and a specificity of 100% if atypical cases were considered negative. NGS revealed oncogenic alterations in 40/43 (93%) of malignant cases based on clinicopathologic follow-up. The most common alterations were in KRAS and TP53, observed in 77% and 49% of malignant cases respectively. No alterations were observed in the 51 benign cases classified based on clinicopathologic follow-up. Supplementing cytomorphologic analysis with molecular profiling of PCC by targeted NGS analysis increased the sensitivity to 93% and maintained specificity at 100%. This study provides evidence for the utility of NGS molecular profiling of PCC specimens to increase the sensitivity of BDB cytology samples, although studies with larger cohorts are needed to verify these findings.