Genomics has revolutionized biology, enabling the interrogation of whole transcriptomes, genome-wide binding sites for proteins, and many other molecular processes. However, individual genomic assays ...measure elements that interact in vivo as components of larger molecular machines. Understanding how these high-order interactions drive gene expression presents a substantial statistical challenge. Building on random forests (RFs) and random intersection trees (RITs) and through extensive, biologically inspired simulations, we developed the iterative random forest algorithm (iRF). iRF trains a feature-weighted ensemble of decision trees to detect stable, high-order interactions with the same order of computational cost as the RF. We demonstrate the utility of iRF for high-order interaction discovery in two prediction problems: enhancer activity in the early Drosophila embryo and alternative splicing of primary transcripts in human-derived cell lines. In Drosophila, among the 20 pairwise transcription factor interactions iRF identifies as stable (returned in more than half of bootstrap replicates), 80% have been previously reported as physical interactions. Moreover, third-order interactions, e.g., between Zelda (Zld), Giant (Gt), and Twist (Twi), suggest high-order relationships that are candidates for follow-up experiments. In human-derived cells, iRF rediscovered a central role of H3K36me3 in chromatin-mediated splicing regulation and identified interesting fifth- and sixth-order interactions, indicative of multivalent nucleosomes with specific roles in splicing regulation. By decoupling the order of interactions from the computational cost of identification, iRF opens additional avenues of inquiry into the molecular mechanisms underlying genome biology.
There is limited available information on patterns of utilization and efficacy of alternative medicine (AM) for patients with cancer. We identified 281 patients with nonmetastatic breast, prostate, ...lung, or colorectal cancer who chose AM, administered as sole anticancer treatment among patients who did not receive conventional cancer treatment (CCT), defined as chemotherapy, radiotherapy, surgery, and/or hormone therapy. Independent covariates on multivariable logistic regression associated with increased likelihood of AM use included breast or lung cancer, higher socioeconomic status, Intermountain West or Pacific location, stage II or III disease, and low comorbidity score. Following 2:1 matching (CCT = 560 patients and AM = 280 patients) on Cox proportional hazards regression, AM use was independently associated with greater risk of death compared with CCT overall (hazard ratio HR = 2.50, 95% confidence interval CI = 1.88 to 3.27) and in subgroups with breast (HR = 5.68, 95% CI = 3.22 to 10.04), lung (HR = 2.17, 95% CI = 1.42 to 3.32), and colorectal cancer (HR = 4.57, 95% CI = 1.66 to 12.61). Although rare, AM utilization for curable cancer without any CCT is associated with greater risk of death.
We did a phase 2 trial of pembrolizumab in patients with non-small-cell lung cancer (NSCLC) or melanoma with untreated brain metastases to determine the activity of PD-1 blockade in the CNS. Interim ...results were previously published, and we now report an updated analysis of the full NSCLC cohort.
This was an open-label, phase 2 study of patients from the Yale Cancer Center (CT, USA). Eligible patients were at least 18 years of age with stage IV NSCLC with at least one brain metastasis 5–20 mm in size, not previously treated or progressing after previous radiotherapy, no neurological symptoms or corticosteroid requirement, and Eastern Cooperative Oncology Group performance status less than two. Modified Response Evaluation Criteria in Solid Tumors (mRECIST) criteria was used to evaluate CNS disease; systemic disease was not required for participation. Patients were treated with pembrolizumab 10 mg/kg intravenously every 2 weeks. Patients were in two cohorts: cohort 1 was for those with PD-L1 expression of at least 1% and cohort 2 was patients with PD-L1 less than 1% or unevaluable. The primary endpoint was the proportion of patients achieving a brain metastasis response (partial response or complete response, according to mRECIST). All treated patients were analysed for response and safety endpoints. This study is closed to accrual and is registered with ClinicalTrials.gov, NCT02085070.
Between March 31, 2014, and May 21, 2018, 42 patients were treated. Median follow-up was 8·3 months (IQR 4·5–26·2). 11 (29·7% 95% CI 15·9–47·0) of 37 patients in cohort 1 had a brain metastasis response. There were no responses in cohort 2. Grade 3–4 adverse events related to treatment included two patients with pneumonitis, and one each with constitutional symptoms, colitis, adrenal insufficiency, hyperglycaemia, and hypokalaemia. Treatment-related serious adverse events occurred in six (14%) of 42 patients and were pneumonitis (n=2), acute kidney injury, colitis, hypokalaemia, and adrenal insufficiency (n=1 each). There were no treatment-related deaths.
Pembrolizumab has activity in brain metastases from NSCLC with PD-L1 expression at least 1% and is safe in selected patients with untreated brain metastases. Further investigation of immunotherapy in patients with CNS disease from NSCLC is warranted.
Merck and the Yale Cancer Center.
Abstract
Health economics research, defined as research that evaluates how patients, health-care providers, and governments make health-care decisions using economic theory, models, and empirical ...techniques, requires broad domains of knowledge that are not fully encompassed by a single discipline. Collaboration between disciplines provides different perspectives on problems, creates more comprehensive research questions, allows for more complex understanding of multifaceted determinants and processes, and thus, provides more realistic recommendations to address difficult questions of health economics. Realizing the importance of collaboration, the National Cancer Institute virtual conference on the Future of Cancer Health Economics included an interactive panel exploring how to foster effective collaborations in cancer health economics research. This manuscript summarizes the panel and participants’ discussion regarding the value, barriers, and potential facilitators to transdisciplinary collaboration within health economics research.
We have evaluated CZE‐ESI‐MS/MS for detection of trace amounts of host cell protein impurities in recombinant therapeutics. Compared to previously published procedures, we have optimized the buffer ...pH used in the formation of a pH junction to increase injection volume. We also prepared a 5‐point calibration curve by spiking 12 standard proteins into a solution of a human mAb. A custom CZE‐MS/MS system was used to analyze the tryptic digest of this mixture without depletion of the antibody. CZE generated a ∼70‐min separation window (∼90‐min total analysis duration) and ∼300‐peak capacity. We also analyzed the sample using ultra‐performance LC‐MS/MS. CZE‐MS/MS generated approximately five times higher base peak intensity and more peptide identifications for low‐level spiked proteins. Both methods detected all proteins spiked at ∼100 ppm level with respect to the antibody.
Extranodal extension (ENE) is a well-established poor prognosticator and an indication for adjuvant treatment escalation in patients with head and neck squamous cell carcinoma (HNSCC). Identification ...of ENE on pretreatment imaging represents a diagnostic challenge that limits its clinical utility. We previously developed a deep learning algorithm that identifies ENE on pretreatment computed tomography (CT) imaging in patients with HNSCC. We sought to validate our algorithm performance for patients from a diverse set of institutions and compare its diagnostic ability to that of expert diagnosticians.
We obtained preoperative, contrast-enhanced CT scans and corresponding pathology results from two external data sets of patients with HNSCC: an external institution and The Cancer Genome Atlas (TCGA) HNSCC imaging data. Lymph nodes were segmented and annotated as ENE-positive or ENE-negative on the basis of pathologic confirmation. Deep learning algorithm performance was evaluated and compared directly to two board-certified neuroradiologists.
A total of 200 lymph nodes were examined in the external validation data sets. For lymph nodes from the external institution, the algorithm achieved an area under the receiver operating characteristic curve (AUC) of 0.84 (83.1% accuracy), outperforming radiologists' AUCs of 0.70 and 0.71 (
= .02 and
= .01). Similarly, for lymph nodes from the TCGA, the algorithm achieved an AUC of 0.90 (88.6% accuracy), outperforming radiologist AUCs of 0.60 and 0.82 (
< .0001 and
= .16). Radiologist diagnostic accuracy improved when receiving deep learning assistance.
Deep learning successfully identified ENE on pretreatment imaging across multiple institutions, exceeding the diagnostic ability of radiologists with specialized head and neck experience. Our findings suggest that deep learning has utility in the identification of ENE in patients with HNSCC and has the potential to be integrated into clinical decision making.