AstraZeneca (Finland)

Bacteria were first detected in human tumors more than 100 years ago, but the characterization of the tumor microbiome has remained challenging because of its low biomass. We undertook a comprehensive analysis of the tumor microbiome, studying 1526 tumors and their adjacent normal tissues across seven cancer types, including breast, lung, ovary, pancreas, melanoma, bone, and brain tumors. We found that each tumor type has a distinct microbiome composition and that breast cancer has a particularly rich and diverse microbiome. The intratumor bacteria are mostly intracellular and are present in both cancer and immune cells. We also noted correlations between intratumor bacteria or their predicted functions with tumor types and subtypes, patients' smoking status, and the response to immunotherapy.

The technological advances of the past century, marked by the computer revolution and the advent of high-throughput screening technologies in drug discovery, opened the path to the computational analysis and visualization of bioactive molecules. For this purpose, it became necessary to represent molecules in a syntax that would be readable by computers and understandable by scientists of various fields. A large number of chemical representations have been developed over the years, their numerosity being due to the fast development of computers and the complexity of producing a representation that encompasses all structural and chemical characteristics. We present here some of the most popular electronic molecular and macromolecular representations used in drug discovery, many of which are based on graph representations. Furthermore, we describe applications of these representations in AI-driven drug discovery. Our aim is to provide a brief guide on structural representations that are essential to the practice of AI in drug discovery. This review serves as a guide for researchers who have little experience with the handling of chemical representations and plan to work on applications at the interface of these fields.

BACKGROUND: In DECLARE-TIMI 58 (Dapagliflozin Effect on Cardiovascular Events-Thrombolysis in Myocardial Infarction 58), the sodium-glucose cotransporter 2 inhibitor dapagliflozin reduced the composite end point of cardiovascular death/hospitalization for heart failure (HHF) in a broad population of patients with type 2 diabetes mellitus. However, the impact of baseline left ventricular ejection fraction (EF) on the clinical benefit of sodium-glucose cotransporter 2 inhibition is unknown. METHODS: In the DECLARE-TIMI 58 trial, baseline heart failure (HF) status was collected from all patients, and EF was collected when available. HF with reduced EF (HFrEF) was defined as EF <45%. Outcomes of interest were the composite of cardiovascular death/HHF, its components, and all-cause mortality. RESULTS: Of 17 160 patients, 671 (3.9%) had HFrEF, 1316 (7.7%) had HF without known reduced EF, and 15 173 (88.4%) had no history of HF at baseline. Dapagliflozin reduced cardiovascular death/HHF more in patients with HFrEF (hazard ratio [HR], 0.62 [95% CI, 0.45-0.86]) than in those without HFrEF (HR, 0.88 [95% CI, 0.76-1.02]; P for interaction=0.046), in whom the treatment effect of dapagliflozin was similar in those with HF without known reduced EF (HR, 0.88 [95% CI, 0.66-1.17]) and those without HF (HR, 0.88 [95% CI, 0.74-1.03]). Whereas dapagliflozin reduced HHF both in those with (HR, 0.64 [95% CI, 0.43-0.95]) and in those without HFrEF (HR, 0.76 [95% CI, 0.62-0.92]), it reduced cardiovascular death only in patients with HFrEF (HR, 0.55 [95% CI, 0.34-0.90]) but not in those without HFrEF (HR, 1.08 [95% CI, 0.89-1.31]; P for interaction=0.012). Likewise, dapagliflozin reduced all-cause mortality in patients with HFrEF (HR, 0.59 [95% CI, 0.40-0.88;) but not in those without HFrEF (HR, 0.97 [95% CI, 0.86-1.10]; P for interaction=0.016). CONCLUSIONS: In the first sodium-glucose cotransporter 2 inhibitor cardiovascular outcome trial to evaluate patients with type 2 diabetes mellitus stratified by EF, we found that dapagliflozin reduced HHF in patients with and without HFrEF and reduced cardiovascular death and all-cause mortality in patients with HFrEF. CLINICAL TRIAL REGISTRATION: URL: https://www.clinicaltrials.gov . Unique identifier: NCT01730534.

The NCCN Guidelines for Prostate Cancer address staging and risk assessment after a prostate cancer diagnosis and include management options for localized, regional, and metastatic disease. Recommendations for disease monitoring and treatment of recurrent disease are also included. The NCCN Prostate Cancer Panel meets annually to reevaluate and update their recommendations based on new clinical data and input from within NCCN Member Institutions and from external entities. This article summarizes the panel's discussions for the 2021 update of the guidelines with regard to systemic therapy for metastatic castration-resistant prostate cancer.

BACKGROUND: Obesity shortens life expectancy. Bariatric surgery is known to reduce the long-term relative risk of death, but its effect on life expectancy is unclear. METHODS: We used the Gompertz proportional hazards regression model to compare mortality and life expectancy among patients treated with either bariatric surgery (surgery group) or usual obesity care (control group) in the prospective, controlled Swedish Obese Subjects (SOS) study and participants in the SOS reference study (reference cohort), a random sample from the general population. RESULTS: In total, 2007 and 2040 patients were included in the surgery group and the control group, respectively, and 1135 participants were included in the reference cohort. At the time of the analysis (December 31, 2018), the median duration of follow-up for mortality was 24 years (interquartile range, 22 to 27) in the surgery group and 22 years (interquartile range, 21 to 27) in the control group; data on mortality were available for 99.9% of patients in the study. In the SOS reference cohort, the median duration of follow-up was 20 years (interquartile range, 19 to 21), and data on mortality were available for 100% of participants. In total, 457 patients (22.8%) in the surgery group and 539 patients (26.4%) in the control group died (hazard ratio, 0.77; 95% confidence interval [CI], 0.68 to 0.87; P<0.001). The corresponding hazard ratio was 0.70 (95% CI, 0.57 to 0.85) for death from cardiovascular disease and 0.77 (95% CI, 0.61 to 0.96) for death from cancer. The adjusted median life expectancy in the surgery group was 3.0 years (95% CI, 1.8 to 4.2) longer than in the control group but 5.5 years shorter than in the general population. The 90-day postoperative mortality was 0.2%, and 2.9% of the patients in the surgery group underwent repeat surgery. CONCLUSIONS: Among patients with obesity, bariatric surgery was associated with longer life expectancy than usual obesity care. Mortality remained higher in both groups than in the general population. (Funded by the Swedish Research Council and others; SOS ClinicalTrials.gov number, NCT01479452.).

Nonclinical rodent and nonrodent toxicity models used to support clinical trials of candidate drugs may produce discordant results or fail to predict complications in humans, contributing to drug failures in the clinic. Here, we applied microengineered Organs-on-Chips technology to design a rat, dog, and human Liver-Chip containing species-specific primary hepatocytes interfaced with liver sinusoidal endothelial cells, with or without Kupffer cells and hepatic stellate cells, cultured under physiological fluid flow. The Liver-Chip detected diverse phenotypes of liver toxicity, including hepatocellular injury, steatosis, cholestasis, and fibrosis, and species-specific toxicities when treated with tool compounds. A multispecies Liver-Chip may provide a useful platform for prediction of liver toxicity and inform human relevance of liver toxicities detected in animal studies to better determine safety and human risk.

CRISPR-Cas genome editing induces targeted DNA damage but can also affect off-target sites. Current off-target discovery methods work using purified DNA or specific cellular models but are incapable of direct detection in vivo. We developed DISCOVER-Seq (discovery of in situ Cas off-targets and verification by sequencing), a universally applicable approach for unbiased off-target identification that leverages the recruitment of DNA repair factors in cells and organisms. Tracking the precise recruitment of MRE11 uncovers the molecular nature of Cas activity in cells with single-base resolution. DISCOVER-Seq works with multiple guide RNA formats and types of Cas enzymes, allowing characterization of new editing tools. Off-targets can be identified in cell lines and patient-derived induced pluripotent stem cells and during adenoviral editing of mice, paving the way for in situ off-target discovery within individual patient genotypes during therapeutic genome editing.

Deep learning methods applied to drug discovery have been used to generate novel structures. In this study, we propose a new deep learning architecture, LatentGAN, which combines an autoencoder and a generative adversarial neural network for de novo molecular design. We applied the method in two scenarios: one to generate random drug-like compounds and another to generate target-biased compounds. Our results show that the method works well in both cases. Sampled compounds from the trained model can largely occupy the same chemical space as the training set and also generate a substantial fraction of novel compounds. Moreover, the drug-likeness score of compounds sampled from LatentGAN is also similar to that of the training set. Lastly, generated compounds differ from those obtained with a Recurrent Neural Network-based generative model approach, indicating that both methods can be used complementarily.

This document was developed to enable greater consistency in the practice, application, and documentation of Model‐Informed Drug Discovery and Development (MID3) across the pharmaceutical industry. A collection of “good practice” recommendations are assembled here in order to minimize the heterogeneity in both the quality and content of MID3 implementation and documentation. The three major objectives of this white paper are to: i) inform company decision makers how the strategic integration of MID3 can benefit R&amp;D efficiency; ii) provide MID3 analysts with sufficient material to enhance the planning, rigor, and consistency of the application of MID3; and iii) provide regulatory authorities with substrate to develop MID3 related and/or MID3 enabled guidelines.

Background: Goals of management in patients with heart failure and reduced ejection fraction include reducing death and hospitalizations, and improving health status (symptoms, physical function, and quality of life). In the DAPA-HF trial (Dapagliflozin and Prevention of Adverse-Outcomes in Heart Failure), sodium–glucose cotransporter-2 inhibitor, dapagliflozin, reduced death and hospitalizations, and improved symptoms in patients with heart failure and reduced ejection fraction. In this analysis, we examine the effects of dapagliflozin on a broad range of health status outcomes, using the Kansas City Cardiomyopathy Questionnaire (KCCQ). Methods: KCCQ was evaluated at randomization, 4 and 8 months. Patients were divided by baseline KCCQ total symptom score (TSS); Cox proportional hazards models examined the effects of dapagliflozin on clinical events across these subgroups. We also evaluated the effects of dapagliflozin on KCCQ-TSS, clinical summary score, and overall summary score. Responder analyses were performed to compare proportions of dapagliflozin versus placebo-treated patients with clinically meaningful changes in KCCQ at 8 months. Results: A total of 4443 patients had available KCCQ at baseline (median KCCQ-TSS, 77.1 [interquartile range, 58.3–91.7]). The effects of dapagliflozin vs placebo on reducing cardiovascular death or worsening heart failure were consistent across the range of KCCQ-TSS (lowest to highest tertile: hazard ratio, 0.70 [95% CI, 0.57–0.86]; hazard ratio, 0.77 [95% CI, 0.61–0.98]; hazard ratio, 0.62 [95% CI, 0.46–0.83]; P for heterogeneity=0.52). Patients treated with dapagliflozin had greater improvement in mean KCCQ-TSS, clinical summary score, and overall summary score at 8 months (2.8, 2.5 and 2.3 points higher versus placebo; P &lt;0.0001 for all). Fewer patients treated with dapagliflozin had a deterioration in KCCQ-TSS (odds ratio, 0.84 [95% CI, 0.78–0.90]; P &lt;0.0001); and more patients had at least small, moderate, and large improvements (odds ratio, 1.15 [95% CI, 1.08–1.23]; odds ratio, 1.15 [95% CI, 1.08–1.22]; odds ratio, 1.14 [95% CI, 1.07–1.22]; number needed to treat=14, 15, and 18, respectively; P &lt;0.0001 for all; results consistent for KCCQ clinical summary score and overall summary score). Conclusions: Dapagliflozin reduced cardiovascular death and worsening heart failure across the range of baseline KCCQ, and improved symptoms, physical function, and quality of life in patients with heart failure and reduced ejection fraction. Furthermore, dapagliflozin increased the proportion of patients experiencing at least small, moderate, and large improvements in health status; these effects were clinically important. Clinical Trial Registration: URL: https://www.clinicaltrials.gov . Unique identifier: NCT03036124.

Epithelial ovarian cancer is the leading cause of death from gynecologic cancer in the United States, with less than half of patients living >5 years from diagnosis. A major challenge in treating ovarian cancer is that most patients have advanced disease at initial diagnosis. The best outcomes are observed in patients whose primary treatment includes complete resection of all visible disease plus combination platinum-based chemotherapy. Research efforts are focused on primary neoadjuvant treatments that may improve resectability, as well as systemic therapies providing improved long-term survival. These NCCN Guidelines Insights focus on recent updates to neoadjuvant chemotherapy recommendations, including the addition of hyperthermic intraperitoneal chemotherapy, and the role of PARP inhibitors and bevacizumab as maintenance therapy options in select patients who have completed primary chemotherapy.

Background: Many patients with heart failure and reduced ejection fraction (HFrEF) have chronic kidney disease that complicates pharmacological management and is associated with worse outcomes. We assessed the safety and efficacy of dapagliflozin in patients with HFrEF, according to baseline kidney function, in the DAPA-HF trial (Dapagliflozin and Prevention of Adverse-outcomes in Heart Failure). We also examined the effect of dapagliflozin on kidney function after randomization. Methods: Patients who have HFrEF with or without type 2 diabetes and an estimated glomerular filtration rate (eGFR) ≥30 mL·min –1 ·1.73 m –2 were enrolled in DAPA-HF. We calculated the incidence of the primary outcome (cardiovascular death or worsening heart failure) according to eGFR category at baseline (&lt;60 and ≥60 mL·min –1 ·1.73 m –2 ) and used eGFR at baseline as a continuous measure, as well. Secondary cardiovascular outcomes and a prespecified composite renal outcome (≥50% sustained decline eGFR, end-stage renal disease, or renal death) were also examined, along with a decline in eGFR over time. Results: Of 4742 patients with a baseline eGFR, 1926 (41%) had eGFR &lt;60 mL·min –1 ·1.73 m –2 . The effect of dapagliflozin on the primary and secondary outcomes did not differ by eGFR category or examining eGFR as a continuous measurement. The hazard ratio (95% CI) for the primary end point in patients with chronic kidney disease was 0.71 (0.59–0.86) versus 0.77 (0.64–0.93) in those with an eGFR ≥60 mL·min –1 ·1.73 m –2 (interaction P =0.54). The composite renal outcome was not reduced by dapagliflozin (hazard ratio=0.71 [95% CI, 0.44–1.16]; P =0.17) but the rate of decline in eGFR between day 14 and 720 was less with dapagliflozin, –1.09 (–1.40 to –0.77) versus placebo –2.85 (–3.17 to –2.53) mL·min –1 ·1.73 m –2 per year ( P &lt;0.001). This was observed in those with and without type 2 diabetes ( P for interaction=0.92). Conclusions: Baseline kidney function did not modify the benefits of dapagliflozin on morbidity and mortality in HFrEF, and dapagliflozin slowed the rate of decline in eGFR, including in patients without diabetes. Registration: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT03036124.

The NCCN Guidelines for Bladder Cancer provide recommendations for the diagnosis, evaluation, treatment, and follow-up of patients with bladder cancer and other urinary tract cancers (upper tract tumors, urothelial carcinoma of the prostate, primary carcinoma of the urethra). These NCCN Guidelines Insights summarize the panel discussion behind recent important updates to the guidelines regarding the treatment of non-muscle-invasive bladder cancer, including how to treat in the event of a bacillus Calmette-Guérin (BCG) shortage; new roles for immune checkpoint inhibitors in non-muscle invasive, muscle-invasive, and metastatic bladder cancer; and the addition of antibody-drug conjugates for metastatic bladder cancer.

PURPOSE Trastuzumab deruxtecan (T-DXd) 5.4 and 6.4 mg/kg showed robust antitumor activity in multiple cancer indications; however, T-DXd 5.4 mg/kg has not been evaluated in patients with previously treated human epidermal growth factor receptor 2–mutant ( HER2m; defined as single-nucleotide variants and exon 20 insertions) metastatic non–small-cell lung cancer (mNSCLC). METHODS DESTINY-Lung02, a blinded, multicenter, phase II study, investigated T-DXd 5.4 mg/kg once every 3 weeks for the first time in previously treated (platinum-containing therapy) patients with HER2m mNSCLC and further assessed T-DXd 6.4 mg/kg once every 3 weeks in this population. The primary end point was confirmed objective response rate (ORR) per RECIST v1.1 by blinded independent central review. RESULTS One hundred fifty-two patients were randomly assigned 2:1 to T-DXd 5.4 or 6.4 mg/kg once every 3 weeks. As of December 23, 2022, the median duration of follow-up was 11.5 months (range, 1.1-20.6) with 5.4 mg/kg and 11.8 months (range, 0.6-21.0) with 6.4 mg/kg. Confirmed ORR was 49.0% (95% CI, 39.0 to 59.1) and 56.0% (95% CI, 41.3 to 70.0) and median duration of response was 16.8 months (95% CI, 6.4 to not estimable [NE]) and NE (95% CI, 8.3 to NE) with 5.4 and 6.4 mg/kg, respectively. Median treatment duration was 7.7 months (range, 0.7-20.8) with 5.4 mg/kg and 8.3 months (range, 0.7-20.3) with 6.4 mg/kg. Grade ≥ 3 drug-related treatment-emergent adverse events occurred in 39 of 101 (38.6%) and 29 of 50 (58.0%) patients with 5.4 and 6.4 mg/kg, respectively. 13 of 101 (12.9%) and 14 of 50 (28.0%) patients had adjudicated drug-related interstitial lung disease (2.0% grade ≥ 3 in each arm) with 5.4 and 6.4 mg/kg, respectively. CONCLUSION T-DXd demonstrated clinically meaningful responses at both doses. Safety profile was acceptable and generally manageable, favoring T-DXd 5.4 mg/kg.

Purpose The purpose of this study was to examine outcomes associated with hormonal maintenance therapy (HMT) compared with routine observation (OBS) after primary cytoreductive surgery and platinum-based chemotherapy in women with stage II to IV low-grade serous carcinoma of the ovary or peritoneum. Patients and Methods Eligibility criteria for patients from our database were: treatment with primary surgery followed by platinum-based chemotherapy, stage II to IV disease, at least 2 years of follow-up for patients who had not experienced recurrence, and adequate clinical information. The two groups were compared for progression-free survival (PFS) and overall survival, and a multivariable Cox regression analysis was performed. Subset analyses for patients who were disease free or had persistent disease were also performed. Results Between 1981 and 2013, 203 eligible patients-133 who underwent OBS and 70 who received HMT-were seen at our institution. Median PFS for patients who underwent OBS was 26.4 months, compared with 64.9 months for those who received HMT ( P < .001). No statistically significant difference in overall survival was observed between the two groups (102.7 v 115.7 months, respectively). For subgroups of women who were disease free or had persistent disease, median PFS was superior for those who received HMT (81.1 v 30.0 months; P < .001 and 38.1 v 15.2 months; P < .001, respectively). Women who received HMT had a significantly lower risk of disease progression compared with those who underwent OBS (hazard ratio, 0.44; 95% CI, 0.31 to 0.64; P < .001). Conclusion Women with stage II to IV low-grade serous carcinoma who received HMT after primary treatment had significantly longer PFS compared with women who underwent OBS. These findings warrant further investigation using a prospective trial design.

Background: A systemic proinflammatory state has been hypothesized to mediate the association between comorbidities and abnormal cardiac structure/function in heart failure with preserved ejection fraction (HFpEF). We conducted a proteomic analysis to investigate this paradigm. Methods: In 228 patients with HFpEF from the multicenter PROMIS-HFpEF study (Prevalence of Microvascular Dysfunction in Heart Failure With Preserved Ejection Fraction), 248 unique circulating proteins were quantified by a multiplex immunoassay (Olink) and used to recapitulate systemic inflammation. In a deductive approach, we performed principal component analysis to summarize 47 proteins known a priori to be involved in inflammation. In an inductive approach, we performed unbiased weighted coexpression network analyses of all 248 proteins to identify clusters of proteins that overrepresented inflammatory pathways. We defined comorbidity burden as the sum of 8 common HFpEF comorbidities. We used multivariable linear regression and statistical mediation analyses to determine whether and to what extent inflammation mediates the association of comorbidity burden with abnormal cardiac structure/function in HFpEF. We also externally validated our findings in an independent cohort of 117 HFpEF cases and 30 comorbidity controls without heart failure. Results: Comorbidity burden was associated with abnormal cardiac structure/function and with principal components/clusters of inflammation proteins. Systemic inflammation was also associated with increased mitral E velocity, E/e′ ratio, and tricuspid regurgitation velocity; and worse right ventricular function (tricuspid annular plane systolic excursion and right ventricular free wall strain). Inflammation mediated the association between comorbidity burden and mitral E velocity (proportion mediated 19%–35%), E/e′ ratio (18%–29%), tricuspid regurgitation velocity (27%–41%), and tricuspid annular plane systolic excursion (13%) ( P &lt;0.05 for all), but not right ventricular free wall strain. TNFR1 (tumor necrosis factor receptor 1), UPAR (urokinase plasminogen activator receptor), IGFBP7 (insulin-like growth factor binding protein 7), and GDF-15 (growth differentiation factor-15) were the top individual proteins that mediated the relationship between comorbidity burden and echocardiographic parameters. In the validation cohort, inflammation was upregulated in HFpEF cases versus controls, and the most prominent inflammation protein cluster identified in PROMIS-HFpEF was also present in HFpEF cases (but not controls) in the validation cohort. Conclusions: Proteins involved in inflammation form a conserved network in HFpEF across 2 independent cohorts and may mediate the association between comorbidity burden and echocardiographic indicators of worse hemodynamics and right ventricular dysfunction. These findings support the comorbidity-inflammation paradigm in HFpEF.

Background: Hyperkalemia increases risk of cardiac arrhythmias and death and limits the use of renin-angiotensin-aldosterone system inhibitors and mineralocorticoid receptor antagonists, which improve clinical outcomes in people with chronic kidney disease or systolic heart failure. Sodium-glucose cotransporter 2 (SGLT2) inhibitors reduce the risk of cardiorenal events in people with type 2 diabetes at high cardiovascular risk or with chronic kidney disease. However, their effect on hyperkalemia has not been systematically evaluated. Methods: A meta-analysis was conducted using individual participant data from randomized, double-blind, placebo-controlled clinical outcome trials with SGLT2 inhibitors in people with type 2 diabetes at high cardiovascular risk or with chronic kidney disease in whom serum potassium levels were routinely measured. The primary outcome was time to serious hyperkalemia, defined as central laboratory–determined serum potassium ≥6.0 mmol/L, with other outcomes including investigator-reported hyperkalemia events and hypokalemia (serum potassium ≤3.5 mmol/L). Cox regression analyses were performed to estimate treatment effects from each trial with hazards ratios and corresponding 95% CIs pooled with random-effects models to obtain summary treatment effects, overall and across key subgroups. Results: Results from 6 trials were included comprising 49 875 participants assessing 4 SGLT2 inhibitors. Of these, 1754 participants developed serious hyperkalemia, and an additional 1119 investigator-reported hyperkalemia events were recorded. SGLT2 inhibitors reduced the risk of serious hyperkalemia (hazard ratio, 0.84 [95% CI, 0.76–0.93]), an effect consistent across studies ( P heterogeneity =0.71). The incidence of investigator-reported hyperkalemia was also lower with SGLT2 inhibitors (hazard ratio, 0.80 [95% CI, 0.68–0.93]; P heterogeneity =0.21). Reductions in serious hyperkalemia were observed across a range of subgroups, including baseline kidney function, history of heart failure, and use of renin-angiotensin-aldosterone system inhibitor, diuretic, and mineralocorticoid receptor antagonist. SGLT2 inhibitors did not increase the risk of hypokalemia (hazard ratio, 1.04 [95% CI, 0.94–1.15]; P heterogeneity =0.42). Conclusions: SGLT2 inhibitors reduce the risk of serious hyperkalemia in people with type 2 diabetes at high cardiovascular risk or with chronic kidney disease without increasing the risk of hypokalemia.

Zoning in on liver growth For organ homeostasis or regrowth after injury or disease, one or more stem cell populations is needed to rebuild lost tissue. There is considerable debate about the source of new cells in the liver. Two groups now identify the source of new hepatocytes (see the Perspective by Andersson). Although the liver may seem to lack major variation across its structure, its lobule is organized into concentric zones where hepatocytes express different metabolic enzymes. Wei et al. sought to systematically define the source of new liver cells by comparing 14 fate-mapping mice that label different liver cell types. They found that different regions of the liver lobule exhibit differences in hepatocyte turnover, with zone 2 representing a primary source of new hepatocytes during homeostasis and regeneration. Similarly, He et al. designed a genetic approach to record cell proliferation in vivo with high spatial and temporal resolution to enable continuous recording of proliferative events of any specific cell type at the whole-cell population level. Using this method, they identified zone 2 as having the highest proliferative activity and contributing the most to liver regrowth. These findings have implications for the cellular basis of chronic disease pathogenesis, cancer development, and regenerative medicine strategies. Science , this issue p. eabb1625 , p. eabc4346 ; see also p. 887

OBJECTIVE: This study examined the risk of antidepressant-induced manic switch in patients with bipolar disorder treated either with antidepressant monotherapy or with an antidepressant in conjunction with a mood stabilizer. METHOD: Using Swedish national registries, the authors identified 3,240 patients with bipolar disorder who started treatment with an antidepressant and had no antidepressant treatment during the previous year. Patients were categorized into those receiving antidepressant monotherapy and those receiving an antidepressant plus a mood stabilizer. A within-individual design was used to control for confounding by disorder severity, genetic makeup, and early environmental factors. Cox regression analyses conditioned on individual were used to compare the rate of mania 0-3 months and 3-9 months after the start of antidepressant treatment with a preceding non-treatment period. RESULTS: Nearly 35% of the patients were treated with antidepressant monotherapy. The increased risk of treatment-emergent mania was confined to patients on antidepressant monotherapy (hazard ratio=2.83, 95% CI=1.12, 7.19). Among patients treated with a concurrent mood stabilizer, no acute change in risk of mania was observed during the 3 months after the start of antidepressant treatment (hazard ratio=0.79, 95% CI=0.54, 1.15), and a decreased risk was observed during the period 3-9 months after treatment initiation (hazard ratio=0.63, 95% CI=0.42, 0.93). CONCLUSIONS: In this national registry study, antidepressant monotherapy was associated with an increased risk of mania. However, no risk of mania was seen in patients receiving an antidepressant while treated with a mood stabilizer. The results highlight the importance of avoiding antidepressant monotherapy in the treatment of bipolar disorder.

REINVENT 4 is a modern open-source generative AI framework for the design of small molecules. The software utilizes recurrent neural networks and transformer architectures to drive molecule generation. These generators are seamlessly embedded within the general machine learning optimization algorithms, transfer learning, reinforcement learning and curriculum learning. REINVENT 4 enables and facilitates de novo design, R-group replacement, library design, linker design, scaffold hopping and molecule optimization. This contribution gives an overview of the software and describes its design. Algorithms and their applications are discussed in detail. REINVENT 4 is a command line tool which reads a user configuration in either TOML or JSON format. The aim of this release is to provide reference implementations for some of the most common algorithms in AI based molecule generation. An additional goal with the release is to create a framework for education and future innovation in AI based molecular design. The software is available from https://github.com/MolecularAI/REINVENT4 and released under the permissive Apache 2.0 license. Scientific contribution. The software provides an open-source reference implementation for generative molecular design where the software is also being used in production to support in-house drug discovery projects. The publication of the most common machine learning algorithms in one code and full documentation thereof will increase transparency of AI and foster innovation, collaboration and education.