Developmental Deconvolution for Classification of Cancer Origin.

Cancer is partly a developmental disease, with malignancies named based on cell or tissue of origin. However, a systematic atlas of tumor origins is lacking. Here we map the single-cell organogenesis of 56 developmental trajectories to the transcriptomes of over 10,000 tumors across 33 cancer types. We deconvolute tumor transcriptomes into signals for individual developmental trajectories. Using these signals as inputs, we construct a developmental multilayer perceptron (D-MLP) classifier that outputs cancer origin. D-MLP (ROC-AUC: 0.974 for top prediction) outperforms benchmark classifiers. We analyze tumors from patients with cancer of unknown primary (CUP), selecting the most difficult cases in which extensive multimodal workup yielded no definitive tumor type. Interestingly, CUPs form groups distinguished by developmental trajectories, and classification reveals diagnosis for patient tumors. Our results provide an atlas of tumor developmental origins, provide a tool for diagnostic pathology, and suggest developmental classification may be a useful approach for patient tumors. SIGNIFICANCE: Here we map the developmental trajectories of tumors. We deconvolute tumor transcriptomes into signals for mammalian developmental programs and use this information to construct a deep learning classifier that outputs tumor type. We apply the classifier to CUP and reveal the developmental origins of patient tumors. See related commentary by Wang, p. 2498. This article is highlighted in the In This Issue feature, p. 2483.

Diagnostic Value of MAML2 Rearrangements in Mucoepidermoid Carcinoma.

Mucoepidermoid carcinoma (MEC) is often seen in salivary glands and can harbor MAML2 translocations (MAML2+). The translocation status has diagnostic utility as an objective confirmation of the MEC diagnosis, for example, when distinction from the more aggressive adenosquamous carcinoma (ASC) is not straightforward. To assess the diagnostic relevance of MAML2, we examined our 5-year experience in prospective testing of 8106 solid tumors using RNA-seq panel testing in combinations with a two-round Delphi-based scenario survey. The prevalence of MAML2+ across all tumors was 0.28% (n = 23/8106) and the majority of MAML2+ cases were found in head and neck tumors (78.3%), where the overall prevalence was 5.9% (n = 18/307). The sensitivity of MAML2 for MEC was 60% and most cases (80%) were submitted for diagnostic confirmation; in 24% of cases, the MAML2 results changed the working diagnosis. An independent survey of 15 experts showed relative importance indexes of 0.8 and 0.65 for "confirmatory MAML2 testing" in suspected MEC and ASC, respectively. Real-world evidence confirmed that the added value of MAML2 is a composite of an imperfect confirmation test for MEC and a highly specific exclusion tool for the diagnosis of ASC. Real-world evidence can help move a rare molecular-genetic biomarker from an emerging tool to the clinic.

t(4;12)(q12;p13) ETV6-rearranged AML without eosinophilia does not involve PDGFRA: relevance for imatinib insensitivity.

Acute myeloid leukemia (AML) with t(4;12)(q12;p13) translocation is rare and often associated with an aggressive clinical course and poor prognosis. Previous reports based on fluorescence in situ hybridization (FISH) analysis have suggested that ETV6::PDGFRA fusions are present in these patients, despite the absence of eosinophilia, which is typically found in other hematopoietic malignancies with PDGFRA-containing fusions. We first detected an ETV6-SCFD2 fusion by targeted RNA sequencing in a patient with t(4;12)(q12;p13) who had been diagnosed with an ETV6-PDGFRA fusion by FISH analysis but failed to respond to imatinib. We then retrospectively identified 4 additional patients with AML and t(4;12)(q12;p13) with apparent ETV6-PDGFRA fusions using chromosome and FISH analysis and applied targeted RNA sequencing to archival material. We again detected rearrangements between ETV6 and non-PDGFRA 4q12 genes, including SCFD2, CHIC2, and GSX2. None of the 3 patients who received imatinib based on the incorrect assumption of an ETV6-PDGFRA fusion responded. Our findings highlight the importance of using a sequencing-based assay to confirm the presence of targetable gene fusions, particularly in genomic regions, such as 4q12, with many clinically relevant genes that are too close to resolve by chromosome or FISH analysis. Finally, combining our data and review of the literature, we show that sequence-confirmed ETV6-PDGFRA fusions are typically found in eosinophilic disorders (3/3 cases), and patients with t(4;12)(q12;p13) without eosinophilia are found to have other 4q12 partners on sequencing (17/17 cases).

A Pathologist-Annotated Dataset for Validating Artificial Intelligence: A Project Description and Pilot Study.

PURPOSE: Validating artificial intelligence algorithms for clinical use in medical images is a challenging endeavor due to a lack of standard reference data (ground truth). This topic typically occupies a small portion of the discussion in research papers since most of the efforts are focused on developing novel algorithms. In this work, we present a collaboration to create a validation dataset of pathologist annotations for algorithms that process whole slide images. We focus on data collection and evaluation of algorithm performance in the context of estimating the density of stromal tumor-infiltrating lymphocytes (sTILs) in breast cancer. METHODS: We digitized 64 glass slides of hematoxylin- and eosin-stained invasive ductal carcinoma core biopsies prepared at a single clinical site. A collaborating pathologist selected 10 regions of interest (ROIs) per slide for evaluation. We created training materials and workflows to crowdsource pathologist image annotations on two modes: an optical microscope and two digital platforms. The microscope platform allows the same ROIs to be evaluated in both modes. The workflows collect the ROI type, a decision on whether the ROI is appropriate for estimating the density of sTILs, and if appropriate, the sTIL density value for that ROI. RESULTS: In total, 19 pathologists made 1645 ROI evaluations during a data collection event and the following 2 weeks. The pilot study yielded an abundant number of cases with nominal sTIL infiltration. Furthermore, we found that the sTIL densities are correlated within a case, and there is notable pathologist variability. Consequently, we outline plans to improve our ROI and case sampling methods. We also outline statistical methods to account for ROI correlations within a case and pathologist variability when validating an algorithm. CONCLUSION: We have built workflows for efficient data collection and tested them in a pilot study. As we prepare for pivotal studies, we will investigate methods to use the dataset as an external validation tool for algorithms. We will also consider what it will take for the dataset to be fit for a regulatory purpose: study size, patient population, and pathologist training and qualifications. To this end, we will elicit feedback from the Food and Drug Administration via the Medical Device Development Tool program and from the broader digital pathology and AI community. Ultimately, we intend to share the dataset, statistical methods, and lessons learned.

Temporary Regulatory Deviations and the Coronavirus Disease 2019 (COVID-19) PCR Labeling Update Study Indicate What Laboratory-Developed Test Regulation by the US Food and Drug Administration (FDA) Could Look Like.

The coronavirus disease 2019 (COVID-19) response necessitated innovations and a series of regulatory deviations that also affected laboratory-developed tests (LDTs). To examine real-world consequences and specify regulatory paradigm shifts, legislative proposals were aligned on a common timeline with Emergency Use Authorization (EUA) of LDTs and the US Food and Drug Administration (FDA)-orchestrated severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) labeling update study. The initial EUA adoption by LDT developers shows that the FDA can have oversight over LDTs. We used efficiency-corrected microcosting of our EUA PCR assay to estimate the national cost of the labeling update study to $0.3 to $1.4 million US dollars. Labeling update study performance data showed lower average detection limits in commercial in vitro diagnostic (IVD) assays versus LDTs (32,000 ± 75,000 versus 71,000 ± 147,000 nucleic acid amplification tests/mL; P = 0.04); however, comparison also shows that FDA review of IVD assays and LDTs did not prevent differences between initial and labeling update performance (IVD assay, P < 0.0001; LDT, P = 0.003). The regulatory shifts re-emphasized that both commercial tests and LDTs rely heavily on laboratory competence and procedures; however, lack of performance data on authorized tests, when clinically implemented, precludes assessment of the benefit related to regulatory review. Temporary regulatory deviations during the pandemic and regulatory science tools (ie, reference material) have generated valuable real-world evidence to inform pending legislation regarding LDT regulation.

PD-1 inhibition in advanced myeloproliferative neoplasms.

Myelofibrosis (MF) is a clonal stem cell neoplasm characterized by abnormal JAK-STAT signaling, chronic inflammation, cytopenias, and risk of transformation to acute leukemia. Despite improvements in the therapeutic options for patients with MF, allogeneic hematopoietic stem cell transplantation remains the only curative treatment. We previously demonstrated multiple immunosuppressive mechanisms in patients with MF, including increased expression of programmed cell death protein 1 (PD-1) on T cells compared with healthy controls. Therefore, we conducted a multicenter, open-label, phase 2, single-arm study of pembrolizumab in patients with Dynamic International Prognostic Scoring System category of intermediate-2 or greater primary, post-essential thrombocythemia or post-polycythemia vera myelofibrosis that were ineligible for or were previously treated with ruxolitinib. The study followed a Simon 2-stage design and enrolled a total of 10 patients, 5 of whom had JAK2V617mutation, 2 had CALR mutation, and 6 had additional mutations. Most patients were previously treated with ruxolitinib. Pembrolizumab treatment was well tolerated, but there were no objective clinical responses, so the study closed after the first stage was completed. However, immune profiling by flow cytometry, T-cell receptor sequencing, and plasma proteomics demonstrated changes in the immune milieu of patients, which suggested improved T-cell responses that can potentially favor antitumor immunity. The fact that these changes were not reflected in a clinical response strongly suggests that combination immunotherapeutic approaches rather than monotherapy may be necessary to reverse the multifactorial mechanisms of immune suppression in myeloproliferative neoplasms. This trial was registered at www.clinicaltrials.gov as #NCT03065400.

Development of a qualitative real-time RT-PCR assay for the detection of SARS-CoV-2: a guide and case study in setting up an emergency-use, laboratory-developed molecular microbiological assay.

Developing and deploying new diagnostic tests are difficult, but the need to do so in response to a rapidly emerging pandemic such as COVID-19 is crucially important. During a pandemic, laboratories play a key role in helping healthcare providers and public health authorities detect active infection, a task most commonly achieved using nucleic acid-based assays. While the landscape of diagnostics is rapidly evolving, PCR remains the gold-standard of nucleic acid-based diagnostic assays, in part due to its reliability, flexibility and wide deployment. To address a critical local shortage of testing capacity persisting during the COVID-19 outbreak, our hospital set up a molecular-based laboratory developed test (LDT) to accurately and safely diagnose SARS-CoV-2. We describe here the process of developing an emergency-use LDT, in the hope that our experience will be useful to other laboratories in future outbreaks and will help to lower barriers to establishing fast and accurate diagnostic testing in crisis conditions.

A rapid genotyping panel for detection of primary central nervous system lymphoma.

Diagnosing primary central nervous system lymphoma (PCNSL) frequently requires neurosurgical biopsy due to nonspecific radiologic features and the low yield of cerebrospinal fluid (CSF) studies. We characterized the clinical evaluation of suspected PCNSL (N = 1007 patients) and designed a rapid multiplexed genotyping assay for MYD88, TERT promoter, IDH1/2, H3F3A, and BRAF mutations to facilitate the diagnosis of PCNSL from CSF and detect other neoplasms in the differential diagnosis. Among 159 patients with confirmed PCNSL, the median time to secure a diagnosis of PCNSL was 10 days, with a range of 0 to 617 days. Permanent histopathology confirmed PCNSL in 142 of 152 biopsies (93.4%), whereas CSF analyses were diagnostic in only 15/113 samplings (13.3%). Among 86 archived clinical specimens, our targeted genotyping assay accurately detected hematologic malignancies with 57.6% sensitivity and 100% specificity (95% confidence interval [CI]: 44.1% to 70.4% and 87.2% to 100%, respectively). MYD88 and TERT promoter mutations were prospectively identified in DNA extracts of CSF obtained from patients with PCNSL and glioblastoma, respectively, within 80 minutes. Across 132 specimens, hallmark mutations indicating the presence of malignancy were detected with 65.8% sensitivity and 100% specificity (95% CI: 56.2%-74.5% and 83.9%-100%, respectively). This targeted genotyping approach offers a rapid, scalable adjunct to reduce diagnostic and treatment delays in PCNSL.

A Regulatory Science Initiative to Harmonize and Standardize Digital Pathology and Machine Learning Processes to Speed up Clinical Innovation to Patients.

Unlocking the full potential of pathology data by gaining computational access to histological pixel data and metadata (digital pathology) is one of the key promises of computational pathology. Despite scientific progress and several regulatory approvals for primary diagnosis using whole-slide imaging, true clinical adoption at scale is slower than anticipated. In the U.S., advances in digital pathology are often siloed pursuits by individual stakeholders, and to our knowledge, there has not been a systematic approach to advance the field through a regulatory science initiative. The Alliance for Digital Pathology (the Alliance) is a recently established, volunteer, collaborative, regulatory science initiative to standardize digital pathology processes to speed up innovation to patients. The purpose is: (1) to account for the patient perspective by including patient advocacy; (2) to investigate and develop methods and tools for the evaluation of effectiveness, safety, and quality to specify risks and benefits in the precompetitive phase; (3) to help strategize the sequence of clinically meaningful deliverables; (4) to encourage and streamline the development of ground-truth data sets for machine learning model development and validation; and (5) to clarify regulatory pathways by investigating relevant regulatory science questions. The Alliance accepts participation from all stakeholders, and we solicit clinically relevant proposals that will benefit the field at large. The initiative will dissolve once a clinical, interoperable, modularized, integrated solution (from tissue acquisition to diagnostic algorithm) has been implemented. In times of rapidly evolving discoveries, scientific input from subject-matter experts is one essential element to inform regulatory guidance and decision-making. The Alliance aims to establish and promote synergistic regulatory science efforts that will leverage diverse inputs to move digital pathology forward and ultimately improve patient care.

Development of a qualitative real-time RT-PCR assay for the detection of SARS-CoV-2: A guide and case study in setting up an emergency-use, laboratory-developed molecular assay.

Developing and deploying new diagnostic tests is difficult, but the need to do so in response to a rapidly emerging pandemic such as COVID-19 is crucially important for an effective response. In the early stages of a pandemic, laboratories play a key role in helping health care providers and public health authorities detect active infection, a task most commonly achieved using nucleic acid-based assays. While the landscape of diagnostics is rapidly evolving, polymerase chain reaction (PCR) remains the gold-standard of nucleic acid-based diagnostic assays, in part due to its reliability, flexibility, and wide deployment. To address a critical local shortage of testing capacity persisting during the COVID-19 outbreak, our hospital set up a molecular based laboratory developed test (LDT) to accurately and safely diagnose SARS-CoV-2. We describe here the process of developing an emergency-use LDT, in the hope that our experience will be useful to other laboratories in future outbreaks and will help to lower barriers to fast and accurate diagnostic testing in crisis conditions.

Clinical sensitivity and interpretation of PCR and serological COVID-19 diagnostics for patients presenting to the hospital.

The diagnosis of COVID-19 requires integration of clinical and laboratory data. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) diagnostic assays play a central role in diagnosis and have fixed technical performance metrics. Interpretation becomes challenging because the clinical sensitivity changes as the virus clears and the immune response emerges. Our goal was to examine the clinical sensitivity of two most common SARS-CoV-2 diagnostic test modalities, polymerase chain reaction (PCR) and serology, over the disease course to provide insight into their clinical interpretation in patients presenting to the hospital. We conducted a single-center, retrospective study. To derive clinical sensitivity of PCR, we identified 209 PCR-positive SARS-CoV-2 patients with multiple PCR test results (624 total PCR tests) and calculated daily sensitivity from date of symptom onset or first positive test. Clinical sensitivity of PCR decreased with days post symptom onset with >90% clinical sensitivity during the first 5 days after symptom onset, 70%-71% from Days 9 to 11, and 30% at Day 21. To calculate daily clinical sensitivity by serology, we utilized 157 PCR-positive patients with a total of 197 specimens tested by enzyme-linked immunosorbent assay for IgM, IgG, and IgA anti-SARS-CoV-2 antibodies. In contrast to PCR, serological sensitivity increased with days post symptom onset with >50% of patients seropositive by at least one antibody isotype after Day 7, >80% after Day 12, and 100% by Day 21. Taken together, PCR and serology are complimentary modalities that require time-dependent interpretation. Superimposition of sensitivities over time indicate that serology can function as a reliable diagnostic aid indicating recent or prior infection.

Lymph node metastases develop through a wider evolutionary bottleneck than distant metastases.

Genetic diversity among metastases is poorly understood but contains important information about disease evolution at secondary sites. Here we investigate inter- and intra-lesion heterogeneity for two types of metastases that associate with different clinical outcomes: lymph node and distant organ metastases in human colorectal cancer. We develop a rigorous mathematical framework for quantifying metastatic phylogenetic diversity. Distant metastases are typically monophyletic and genetically similar to each other. Lymph node metastases, in contrast, display high levels of inter-lesion diversity. We validate these findings by analyzing 317 multi-region biopsies from an independent cohort of 20 patients. We further demonstrate higher levels of intra-lesion heterogeneity in lymph node than in distant metastases. Our results show that fewer primary tumor lineages seed distant metastases than lymph node metastases, indicating that the two sites are subject to different levels of selection. Thus, lymph node and distant metastases develop through fundamentally different evolutionary mechanisms.

MET Alterations Are a Recurring and Actionable Resistance Mechanism in ALK-Positive Lung Cancer.

PURPOSE: Most ALK-positive lung cancers will develop ALK-independent resistance after treatment with next-generation ALK inhibitors. MET amplification has been described in patients progressing on ALK inhibitors, but frequency of this event has not been comprehensively assessed. EXPERIMENTAL DESIGN: We performed FISH and/or next-generation sequencing on 207 posttreatment tissue (n = 101) or plasma (n = 106) specimens from patients with ALK-positive lung cancer to detect MET genetic alterations. We evaluated ALK inhibitor sensitivity in cell lines with MET alterations and assessed antitumor activity of ALK/MET blockade in ALK-positive cell lines and 2 patients with MET-driven resistance. RESULTS: MET amplification was detected in 15% of tumor biopsies from patients relapsing on next-generation ALK inhibitors, including 12% and 22% of biopsies from patients progressing on second-generation inhibitors or lorlatinib, respectively. Patients treated with a second-generation ALK inhibitor in the first-line setting were more likely to develop MET amplification than those who had received next-generation ALK inhibitors after crizotinib (P = 0.019). Two tumor specimens harbored an identical ST7-MET rearrangement, one of which had concurrent MET amplification. Expressing ST7-MET in the sensitive H3122 ALK-positive cell line induced resistance to ALK inhibitors that was reversed with dual ALK/MET inhibition. MET inhibition resensitized a patient-derived cell line harboring both ST7-MET and MET amplification to ALK inhibitors. Two patients with ALK-positive lung cancer and acquired MET alterations achieved rapid responses to ALK/MET combination therapy. CONCLUSIONS: Treatment with next-generation ALK inhibitors, particularly in the first-line setting, may lead to MET-driven resistance. Patients with acquired MET alterations may derive clinical benefit from therapies that target both ALK and MET.

Synthesis and Characterization of a Magnetically Active 19F Molecular Beacon.

Gene expression is used extensively to describe cellular characteristics and behaviors; however, most methods of assessing gene expression are unsuitable for living samples, requiring destructive processes such as fixation or lysis. Recently, molecular beacons have become a viable tool for live-cell imaging of mRNA molecules in situ. Historically, beacon-mediated imaging has been limited to fluorescence-based approaches. We propose the design and synthesis of a novel molecular beacon for magnetic resonance detection of any desired target nucleotide sequence. The biologically compatible synthesis incorporates commonly used bioconjugation reactions in aqueous conditions and is accessible for laboratories without extensive synthesis capabilities. The resulting beacon uses fluorine (19F) as a reporter, which is broadened, or turned "off", via paramagnetic relaxation enhancement from a stabilized nitroxide radical spin label when the beacon is not bound to its nucleic acid target. Therefore, the 19F NMR signal of the beacon is quenched in its hairpin conformation when the spin label and the 19F substituent are held in proximity, but the signal is recovered upon beacon hybridization to its specific complementary nucleotide sequence by physical separation of the radical from the 19F reporter. This study establishes a path for magnetic resonance-based assessment of specific mRNA expression, providing new possibilities for applying molecular beacon technology in living systems.

Gene expression-based enrichment of live cells from adipose tissue produces subpopulations with improved osteogenic potential.

INTRODUCTION: Mesenchymal stem cells have been increasingly used for cell-based therapies. Adipose-derived stem/stromal cells (ASCs) from the stromal vascular fraction (SVF) of fat tissue are a particularly attractive option for cell based therapy given their accessibility and relative abundance. However, their application in both clinical and basic science investigations is complicated by the isolation of differentiable cells within the SVF. Current enrichment strategies, such as monolayer passaging and surface marker-based sorting, can be time-consuming or overly stringent. Ideally, a population of cells with great regenerative capacity could be isolated with high yields so that extensive in vitro manipulation is not necessary. The objective of this study was to determine whether SVF cells sorted based on expression of alkaline phosphatase liver/bone/kidney (ALPL) resulted in populations with increased osteogenic differentiation potential. METHODS: SVF samples were obtained from four, human donors and processed to isolate initial, heterogeneous cell populations. These SVF cells underwent a four day osteogenic priming period, after which they were treated with a fluorescent, oligodeoxynucleotide molecular beacon probe specific for ALPL mRNA. Cells were separated into positive and negative groups using fluorescence-activated cell sorting (FACS) then differentiated down the osteogenic lineage. Differentiation was assessed by measuring calcified matrix production in each sample. RESULTS: Cells positive for ALPL expression (ALPL+) represented approximately 34% of the gated population, while cells negative for ALPL expression (ALPL-) represented approximately 18%. ALPL+ cells produced 3.7-fold and 2.1-fold more calcified matrix than ALPL- and unsorted SVF cells, respectively, indicating a significant improvement in osteogenic differentiation. Further, ALPL+ cells showed increases in metabolite production for both adipogenesis and chondrogenesis, suggesting that the enrichment process yields an enhanced multipotent phenotype. Osteogenic differentiation response and cell yields for ALPL+ cells were markedly improved over surface marker-sorted samples. CONCLUSION: This study demonstrates a novel method to enrich heterogeneous SVF cells for increased osteogenic potential. The procedure requires less time and results in higher yields of therapeutically useful cells than other existing approaches. Gene expression-based sorting of MSCs is a potentially paradigm-shifting approach that could benefit applications spanning from basic science to clinical therapy.