Citations

Introduction: Lineage switch (LS) refers to the transformation of acute leukemia from one cell lineage to another (e.g., lymphoid (ALL) to myeloid (AML)). This rare phenomenon, distinct from treatment induced malignancy (i.e., t-AML) and mixed phenotype acute leukemia (MPAL), is increasingly being observed following cell surface antigen targeting. Given challenges in the diagnosis and treatment of LS, Project EVOLVE was developed to globally identify cases of LS following immunotherapy. Methods: A protocol and data collection form for this IRB exempt study was widely disseminated to capture deidentified LS case information across multiple centers and cancer consortia. This included collection of data from published reports of LS, with authors contacted to obtain additional information. For this analysis, LS was defined by the emergence of cell-surface markers warranting reclassification of the original leukemia as a different lineage derivation based on the immunophenotype, including cases of AML emerging alongside the original ALL. “Confirmed” LS cases included those where retention of original cytogenetic or molecular aberrations and/or the clonal immunoglobulin rearrangement patterns were verified to confirm the clonal relationship. “Suspected” LS cases included those where samples to confirm clonality marker retention could not be obtained, but the clinical presentation was consistent with LS. The term “transition” was applied when immunophenotypic shifts were transient. Cases with new cytogenetic abnormalities unrelated to the original diagnosis and compatible with t-AML were excluded. Statistical analysis was primarily descriptive. Data cut-off was July 26, 2023. Results: A total of 58 cases were identified, including 19 cases referenced in prior publications. Of the 58, 11 cases were excluded from this initial analysis, including: 4 cases of t-AML, 2 cases of T-ALL to AML, and 5 “transition” cases. Collectively, 47 cases of LS (43 “confirmed”; 4 “suspected”), spanning across 3 continents and 8 countries, were analyzed. (Fig. 1A) Cases included transition from B-ALL to AML in 36 (76.6%) and to MPAL/biphenotypic/ambiguous lineage in 11 (23.4%). Two likely had MPAL at diagnosis and a third with chronic myeloid leukemia. The median age at initial diagnosis was 8.4 years (range, 1 day-76.5 years); median age at LS presentation was 11.0 years (range 0.4-77.3 years). Twenty-three (48.9%) were male and 19 (40.4%) had prior stem cell transplant (SCT). Baseline cytogenetics showed KMT2A rearrangement in 27 (57.4%). Baseline myeloid antigen co-expression was seen in 22 of 23 patients with submitted data, but variation in degree of expression across patients and antigens was substantial. The most proximal immunotherapy prior to development of LS was CAR T-cells in 23 (48.9%) and blinatumomab in 24 (51.1%). The median time from the most proximal immunotherapy to development of LS was 1.6 months (range, 7 days-36.5 months). In 7 (29.2%) patients receiving blinatumomab, LS developed during the infusion. LS presented as isolated marrow relapse in 27 (57.4%), isolated CNS in 2 (4.3%), isolated non-CNS extramedullary disease in 3 (6.4%), combined relapse in 14 (29.8%) and unreported in 1. First line treatment for LS was chemotherapy induction in 31 (66.0%), palliative care/no therapy in 8 (17.0%), and alternative therapies in the remaining 8 (17.0%). In 35 (74.5%) patients, remission induction was the goal of first line therapy. Fifteen (31.9%) achieved a complete remission (CR) of whom 2 were treated with palliative intent. Across all patients, 8 (17.0%) are alive in remission, the longest being 4.7 years from LS diagnosis; 2 (4.3%) are alive with active LS and 37 (78.7%) died from LS, their original disease, or complications of treatment of LS (Fig 1B). For those alive in CR (n=8), 7 (87.5%) proceeded to a consolidative SCT for treatment of LS and 1 patient, an octogenarian remains in remission after AML directed therapy. Conclusions: LS is emerging as an important mechanism of immune escape following immunotherapy. Project EVOLVE provides the largest dataset of patients with LS to date. This systematic, international effort reveals substantial complexity and variability in diagnosing and managing LS. Unfortunately, outcomes following LS are grim, underscoring the critical need to identify risk factors of and optimal therapies for this leukemia transformation.

Birt-Hogg-Dubé (BHD) syndrome is associated with an increased risk of multifocal renal tumors, including hybrid oncocytic tumor (HOT) and chromophobe renal cell carcinoma (chRCC). HOT exhibits heterogenous histologic features overlapping with chRCC and benign renal oncocytoma, posing challenges in diagnosis of HOT and renal tumor entities resembling HOT. In this study, we performed integrative analysis of bulk and single-cell RNA sequencing data from renal tumors and normal kidney tissues, and nominated candidate biomarkers of HOT, L1CAM, and LINC01187, which are also lineage-specific markers labeling the principal cell and intercalated cell lineages of the distal nephron, respectively. Our findings indicate the principal cell lineage marker L1CAM and intercalated cell lineage marker LINC01187 to be expressed mutually exclusively in a unique checkered pattern in BHD-associated HOTs, and these 2 lineage markers collectively capture the 2 distinct tumor epithelial populations seen to co-exist morphologically in HOTs. We further confirmed that the unique checkered expression pattern of L1CAM and LINC01187 distinguished HOT from chRCC, renal oncocytoma, and other major and rare renal cell carcinoma subtypes. We also characterized the histopathologic features and immunophenotypic features of oncocytosis in the background kidney of patients with BHD, as well as the intertumor and intratumor heterogeneity seen within HOT. We suggest that L1CAM and LINC01187 can serve as stand-alone diagnostic markers or as a panel for the diagnosis of HOT. These lineage markers will inform future studies on the evolution and interaction between the 2 transcriptionally distinct tumor epithelial populations in such tumors.

Biodegradable and all-natural gelatin-based nanoemulsions encapsulate and stabilize antifungal essential oils to provide efficient penetration and eradication of drug-resistant fungal biofilms without resistance generation.

Abstract Intraductal carcinoma of the prostate (IDC-P) is a lethal prostate cancer subtype that generally coexists with invasive high-grade prostate acinar adenocarcinoma (PAC) but exhibits distinct biological features compared with concomitant adenocarcinoma. In this study, we performed whole-exome, RNA, and DNA-methylation sequencing of IDC-P, concurrent invasive high-grade PAC lesions, and adjacent normal prostate tissues isolated from 22 radical prostatectomy specimens. Three evolutionary patterns of concurrent IDC-P and PAC were identified: early divergent, late divergent, and clonally distant. In contrast to those with a late divergent evolutionary pattern, tumors with clonally distant and early divergent evolutionary patterns showed higher genomic, epigenomic, transcriptional, and pathologic heterogeneity between IDC-P and PAC. Compared with coexisting PAC, IDC-P displayed increased expression of adverse prognosis–associated genes. Survival analysis based on an independent cohort of 505 patients with metastatic prostate cancer revealed that IDC-P carriers with lower risk International Society of Urological Pathology (ISUP) grade 1–4 adenocarcinoma displayed a castration-resistant free survival as poor as those with the highest risk ISUP grade 5 tumors that lacked concurrent IDC-P. Furthermore, IDC-P exhibited robust cell-cycle progression and androgen receptor activities, characterized by an enrichment of cellular proliferation–associated master regulators and genes involved in intratumoral androgen biosynthesis. Overall, this study provides a molecular groundwork for the aggressive behavior of IDC-P and could help identify potential strategies to improve treatment of IDC-P. Significance: The genomic, transcriptomic, and epigenomic characterization of concurrent intraductal carcinoma and adenocarcinoma of the prostate deepens the biological understanding of this lethal disease and provides a genetic basis for developing targeted therapies.

INTRODUCTION: The American Gastroenterological Association (AGA) has compiled risk factors that may be predictive of disease complications in Crohn's disease (CD) and ulcerative colitis (UC). The aim of this study was to evaluate the performance of the AGA risk factors for risk stratification in UC and CD. METHODS: We included participants of 2 cohorts: the Ocean State Crohn's and Colitis Area Registry cohort and the Mayo Clinic cohort. Baseline clinical risk factors were extracted according to the AGA pathway. Our primary end point was defined as follows: (i) any inflammatory bowel disease related–hospitalization, (ii) any inflammatory bowel disease–related bowel surgery, or (iii) any progression of disease. We analyzed the association of the number of AGA risk factors with our end point. Statistical multivariable modeling was performed with Cox proportional hazards model. RESULTS: A total of 412 patients with CD were included. Comparing ≥3 risk factors with 0–1 risk factor, we found a significantly increased risk of complications in both the Ocean State Crohn's and Colitis Area Registry cohort (hazard ratio [HR] 2.75, 95% confidence interval 1.71–4.41) and Mayo Clinic cohort (HR 2.07, 95% confidence interval 1.11–3.84). Diagnosis at younger age (HR 2.07), perianal disease (HR 1.99), and B2/B3 behavior (HR 1.92) were significantly associated with disease complications. We did not observe a consistent association between number of risk factors nor any specific individual risk factors and risk of disease complications in the 265 patients with UC included. DISCUSSION: We found a significant association between the number of AGA risk factors and the risk of disease complication in CD; this association was not significant in UC. The presence of ≥ 3 risk factors in CD leads to the highest risk of complications. The AGA care pathway is a useful tool to stratify patients who are at higher risk of disease complications in patients with CD.

We assessed the efficacy and safety of combining bevacizumab with temsirolimus in patients with advanced extra-pancreatic neuroendocrine tumors. This NCI-sponsored multicenter, open-label, phase II study (NCT01010126) enrolled patients with advanced, recurrent, or metastatic extra-pancreatic neuroendocrine tumors. All patients were treated with temsirolimus and bevacizumab until disease progression or unacceptable toxicity. Temsirolimus 25 mg was administered i.v. on days 1, 8, 15, and 22 and bevacizumab 10 mg/kg i.v. on days 1 and 15 of a 4-week cycle. Discontinuation of temsirolimus or bevacizumab did not require discontinuation of the other agent. The primary endpoints were objective response rate and 6-month progression-free survival rate. Fifty-nine patients were enrolled in this study, and 54 were evaluated for efficacy and adverse events. While median progression-free survival was 7.1 months, the median duration of treatment with temsirolimus was 3.9 months and that with bevacizumab was 3.5 months. The objective response rate of combination therapy was 2%, and 6-month progression-free survival was 48%. The most frequently reported grade 3–4 adverse events included fatigue (13%), hypertension (13%), and bleeding (13%). Close to 54% of the patients discontinued treatment due to adverse events, refusal of further treatment, or treatment delays. Three deaths occurred in the study, of which two were due to treatment-related bowel perforations. Given the minimal efficacy and increased toxicity seen with the combination of bevacizumab and temsirolimus, we do not recommend the use of this regimen in patients with advanced extra-pancreatic neuroendocrine tumors.

Abstract Objective Transformer-based language models are prevailing in the clinical domain due to their excellent performance on clinical NLP tasks. The generalizability of those models is usually ignored during the model development process. This study evaluated the generalizability of CancerBERT, a Transformer-based clinical NLP model, along with classic machine learning models, i.e., conditional random field (CRF), bi-directional long short-term memory CRF (BiLSTM-CRF), across different clinical institutes through a breast cancer phenotype extraction task. Materials and methods Two clinical corpora of breast cancer patients were collected from the electronic health records from the University of Minnesota (UMN) and Mayo Clinic (MC), and annotated following the same guideline. We developed three types of NLP models (i.e., CRF, BiLSTM-CRF and CancerBERT) to extract cancer phenotypes from clinical texts. We evaluated the generalizability of models on different test sets with different learning strategies (model transfer vs locally trained). The entity coverage score was assessed with their association with the model performances. Results We manually annotated 200 and 161 clinical documents at UMN and MC. The corpora of the two institutes were found to have higher similarity between the target entities than the overall corpora. The CancerBERT models obtained the best performances among the independent test sets from two clinical institutes and the permutation test set. The CancerBERT model developed in one institute and further fine-tuned in another institute achieved reasonable performance compared to the model developed on local data (micro-F1: 0.925 vs 0.932). Conclusions The results indicate the CancerBERT model has superior learning ability and generalizability among the three types of clinical NLP models for our named entity recognition task. It has the advantage to recognize complex entities, e.g., entities with different labels.

Abstract Despite recent methodology advancements in clinical natural language processing (NLP), the adoption of clinical NLP models within the translational research community remains hindered by process heterogeneity and human factor variations. Concurrently, these factors also dramatically increase the difficulty in developing NLP models in multi-site settings, which is necessary for algorithm robustness and generalizability. Here, we reported on our experience developing an NLP solution for Coronavirus Disease 2019 (COVID-19) signs and symptom extraction in an open NLP framework from a subset of sites participating in the National COVID Cohort (N3C). We then empirically highlight the benefits of multi-site data for both symbolic and statistical methods, as well as highlight the need for federated annotation and evaluation to resolve several pitfalls encountered in the course of these efforts.

Abstract Androgen receptor (AR) inhibition by androgen deprivation and/or antiandrogen administration is the mainstay therapy for advanced prostate cancer. However, most prostate cancers ultimately become resistant to these therapies, indicating the importance of identifying mechanisms driving resistance to improve patient outcomes. Here we demonstrated that acute treatment with the antiandrogen enzalutamide (ENZ) decreased glutathione (GSH) production, increased lipid peroxidation, and induced ferroptosis in prostate cancer cells. Consistently, meta-analysis of transcriptomic data linked the androgen-AR axis to metabolism-related biological processes, including lipid metabolism. The cystine transporter gene SLC7A11 was a key AR target, and full-length AR (AR-FL) transactivated SLC7A11 transcription by directly occupying the SLC7A11 promoter and putative enhancer regions. AR variants (AR-V) preferentially bound the SLC7A11 enhancer and upregulated SLC7A11 expression, thereby conferring resistance to ferroptosis induced by ENZ treatment. However, this effect was abolished following downregulation of AR-Vs using the dual CBP/p300 and BET inhibitor NEO2734. These findings reveal ferroptosis induction as an anticancer mechanism of antiandrogens and SLC7A11 as a direct target gene of AR-FL and AR-Vs. AR-V-mediated SLC7A11 expression represents a mechanism coupling ferroptosis resistance to prostate cancer progression. Significance: Upregulation of SLC7A11 can be induced by androgen receptor variants to inhibit antiandrogen-induced prostate cancer cell ferroptosis and to drive castration resistance in prostate cancer.