Evaluating the fairness and accuracy of machine learning-based predictions of clinical outcomes after anatomic and reverse total shoulder arthroplasty.

BACKGROUND: Machine learning (ML)-based clinical decision support tools (CDSTs) make personalized predictions for different treatments; by comparing predictions of multiple treatments, these tools can be used to optimize decision making for a particular patient. However, CDST prediction accuracy varies for different patients and also for different treatment options. If these differences are sufficiently large and consistent for a particular subcohort of patients, then that bias may result in those patients not receiving a particular treatment. Such level of bias would deem the CDST "unfair." The purpose of this study is to evaluate the "fairness" of ML CDST-based clinical outcomes predictions after anatomic (aTSA) and reverse total shoulder arthroplasty (rTSA) for patients of different demographic attributes. METHODS: Clinical data from 8280 shoulder arthroplasty patients with 19,249 postoperative visits was used to evaluate the prediction fairness and accuracy associated with the following patient demographic attributes: ethnicity, sex, and age at the time of surgery. Performance of clinical outcome and range of motion regression predictions were quantified by the mean absolute error (MAE) and performance of minimal clinically important difference (MCID) and substantial clinical benefit classification predictions were quantified by accuracy, sensitivity, and the F1 score. Fairness of classification predictions leveraged the "four-fifths" legal guideline from the US Equal Employment Opportunity Commission and fairness of regression predictions leveraged established MCID thresholds associated with each outcome measure. RESULTS: For both aTSA and rTSA clinical outcome predictions, only minor differences in MAE were observed between patients of different ethnicity, sex, and age. Evaluation of prediction fairness demonstrated that 0 of 486 MCID (0%) and only 3 of 486 substantial clinical benefit (0.6%) classification predictions were outside the 20% fairness boundary and only 14 of 972 (1.4%) regression predictions were outside of the MCID fairness boundary. Hispanic and Black patients were more likely to have ML predictions out of fairness tolerance for aTSA and rTSA. Additionally, patients <60 years old were more likely to have ML predictions out of fairness tolerance for rTSA. No disparate predictions were identified for sex and no disparate regression predictions were observed for forward elevation, internal rotation score, American Shoulder and Elbow Surgeons Standardized Shoulder Assessment Form score, or global shoulder function. CONCLUSION: The ML algorithms analyzed in this study accurately predict clinical outcomes after aTSA and rTSA for patients of different ethnicity, sex, and age, where only 1.4% of regression predictions and only 0.3% of classification predictions were out of fairness tolerance using the proposed fairness evaluation method and acceptance criteria. Future work is required to externally validate these ML algorithms to ensure they are equally accurate for all legally protected patient groups.

Data-driven development of an oral lipid-based nanoparticle formulation of a hydrophobic drug.

Due to its cost-effectiveness, convenience, and high patient adherence, oral drug administration normally remains the preferred approach. Yet, the effective delivery of hydrophobic drugs via the oral route is often hindered by their limited water solubility and first-pass metabolism. To mitigate these challenges, advanced delivery systems such as solid lipid nanoparticles (SLNs) and nanostructured lipid carriers (NLCs) have been developed to encapsulate hydrophobic drugs and enhance their bioavailability. However, traditional design methodologies for these complex formulations often present intricate challenges because they are restricted to a relatively narrow design space. Here, we present a data-driven approach for the accelerated design of SLNs/NLCs encapsulating a model hydrophobic drug, cannabidiol, that combines experimental automation and machine learning. A small subset of formulations, comprising 10% of all formulations in the design space, was prepared in-house, leveraging miniaturized experimental automation to improve throughput and decrease the quantity of drug and materials required. Machine learning models were then trained on the data generated from these formulations and used to predict properties of all SLNs/NLCs within this design space (i.e., 1215 formulations). Notably, formulations predicted to be high-performers via this approach were confirmed to significantly enhance the solubility of the drug by up to 3000-fold and prevented degradation of drug. Moreover, the high-performance formulations significantly enhanced the oral bioavailability of the drug compared to both its free form and an over-the-counter version. Furthermore, this bioavailability matched that of a formulation equivalent in composition to the FDA-approved product, Epidiolex®.

A dataset of formulation compositions for self-emulsifying drug delivery systems.

Self-emulsifying drug delivery systems (SEDDS) are a well-established formulation strategy for improving the oral bioavailability of poorly water-soluble drugs. Traditional development of these formulations relies heavily on empirical observation to assess drug and excipient compatibility, as well as to select and optimize the formulation compositions. The aim of this work was to leverage previously developed SEDDS in the literature to construct a comprehensive SEDDS dataset that can be used to gain insights and advance data-driven approaches to formulation development. A dataset comprised of 668 unique SEDDS formulations encompassing 20 poorly water-soluble drugs was curated. While there are still opportunities to enhance the quality and quantity of data on SEDDS, this research lays the groundwork to potentially simplify the SEDDS formulation development process.

Data demonstrating the in vivo anti-tumor efficacy of thermosensitive liposome formulations of a drug combination in pre-clinical models of breast cancer.

Thermosensitive liposomes in combination with localized mild hyperthermia can improve the delivery of drug to solid tumor sites. For this reason, thermosensitive liposome formulations of a range of chemotherapy drugs have been designed. Our group previously developed and characterized a thermosensitive liposome formulation of the heat shock protein 90 inhibitor alvespimycin as a companion therapeutic to a thermosensitive liposome formulation equivalent in composition to ThermoDox (i.e., ThermoDXR), with the goal of increasing the therapeutic index of doxorubicin as the combination was revealed to be highly synergistic in a panel of human breast cancer cell lines including MDA-MB-231 (Dunne et al., 2019). The data presented here further describes the effect of the doxorubicin (DXR) and alvespimycin (ALV) combination in vitro and in vivo. Specifically, the combination effect in mouse breast cancer 4T1 cells and the in vivo efficacy of this heat-activated chemotherapy combination in both immunocompromised (MDA-MB-231 tumor bearing female SCID mice) and immunocompetent (4T1 tumor bearing female BALB/c mice) models of breast cancer.

Revolutionizing drug formulation development: The increasing impact of machine learning.

Over the past few years, the adoption of machine learning (ML) techniques has rapidly expanded across many fields of research including formulation science. At the same time, the use of lipid nanoparticles to enable the successful delivery of mRNA vaccines in the recent COVID-19 pandemic demonstrated the impact of formulation science. Yet, the design of advanced pharmaceutical formulations is non-trivial and primarily relies on costly and time-consuming wet-lab experimentation. In 2021, our group published a review article focused on the use of ML as a means to accelerate drug formulation development. Since then, the field has witnessed significant growth and progress, reflected by an increasing number of studies published in this area. This updated review summarizes the current state of ML directed drug formulation development, introduces advanced ML techniques that have been implemented in formulation design and shares the progress on making self-driving laboratories a reality. Furthermore, this review highlights several future applications of ML yet to be fully exploited to advance drug formulation research and development.

Anisotropic wood-hydrogel composites: Extending mechanical properties of wood towards soft materials' applications.

Delignified wood (DW) offers a versatile platform for the manufacturing of composites, with material properties ranging from stiff to soft and flexible by preserving the preferential fiber directionality of natural wood through a structure-retaining production process. This study presents a facile method for fabricating anisotropic and mechanically tunable DW-hydrogel composites. These composites were produced by infiltrating delignified spruce wood with an aqueous gelatin solution followed by chemical crosslinking. The mechanical properties could be modulated across a broad strength and stiffness range (1.2-18.3 MPa and 170-1455 MPa, respectively) by varying the crosslinking time. The diffusion-led crosslinking further allowed to manufacture mechanically graded structures. The resulting uniaxial, tubular structure of the anisotropic DW-hydrogel composite enabled the alignment of murine fibroblasts in vitro, which could be utilized in future studies on potential applications in tissue engineering.

Self-driving laboratories: A paradigm shift in nanomedicine development.

Nanomedicines have transformed promising therapeutic agents into clinically approved medicines with optimal safety and efficacy profiles. This is exemplified by the mRNA vaccines against COVID-19, which were made possible by lipid nanoparticle technology. Despite the success of nanomedicines to date, their design remains far from trivial, in part due to the complexity associated with their preclinical development. Herein, we propose a nanomedicine materials acceleration platform (NanoMAP) to streamline the preclinical development of these formulations. NanoMAP combines high-throughput experimentation with state-of-the-art advances in artificial intelligence (including active learning and few-shot learning) as well as a web-based application for data sharing. The deployment of NanoMAP requires interdisciplinary collaboration between leading figures in drug delivery and artificial intelligence to enable this data-driven design approach. The proposed approach will not only expedite the development of next-generation nanomedicines but also encourage participation of the pharmaceutical science community in a large data curation initiative.

Data evaluating triamcinolone acetonide and triamcinolone hexacetonide loaded poly(δ-valerolactone-co-allyl-δ-valerolactone) microparticles.

Advanced drug delivery strategies can be used to enhance the therapeutic effectiveness of locally delivered corticosteroids. Poly(δ-valerolactone-co-allyl-δ-valerolactone) microparticles (PVL-co-PAVL MPs) were evaluated for delivery of two corticosteroids, triamcinolone acetonide and triamcinolone hexacetonide. PVL-co-PAVL MPs were prepared using a modified oil-in-water emulsification method, followed by a UV-initiated cross-linking process. The resulting PVL-co-PAVL MPs were purified with an excess amount of water and then acetone to remove residual surfactant, cross-linker, and catalyst before lyophilization. Triamcinolone acetonide and triamcinolone hexacetonide were independently loaded into the resulting PVL-co-PAVL MPs via a post-loading swelling-equilibrium method. The drug-loaded MPs were characterized in terms of drug loading (determined by high-performance liquid chromatography, HPLC), thermal properties (determined by differential scanning calorimetry, DSC), and in vitro drug release kinetics (with quantification of drug using HPLC) to better understand the suitability of PVL-co-PAVL MPs for delivery of corticosteroids. These data demonstrate the potential of PVL-co-PAVL MPs as a promising drug delivery platform for the sustained release of corticosteroids. Raw data have been made available on Mendeley Data. Additional details on PVL-co-PAVL MPs were previously reported [1].

2002-2017 anthropogenic emissions data for air quality modeling over the United States.

The United States Environmental Protection Agency (US EPA) has developed a set of annual North American emissions data for multiple air pollutants across 18 broad source categories for 2002 through 2017. The sixteen new annual emissions inventories were developed using consistent input data and methods across all years. When a consistent method or tool was not available for a source category, emissions were estimated by scaling data from the EPA's 2017 National Emissions Inventory with scaling factors based on activity data and/or emissions control information. The emissions datasets are designed to support regional air quality modeling for a wide variety of human health and ecological applications. The data were developed to support simulations of the EPA's Community Multiscale Air Quality model but can also be used by other regional scale air quality models. The emissions data are one component of EPA's Air Quality Time Series Project which also includes air quality modeling inputs (meteorology, initial conditions, boundary conditions) and outputs (e.g., ozone, PM2.5 and constituent species, wet and dry deposition) for the Conterminous US at a 12 km horizontal grid spacing.

Folate receptor targeted nanoparticles containing niraparib and doxorubicin as a potential candidate for the treatment of high grade serous ovarian cancer.

Combination chemotherapy is an established approach used to manage toxicities while eliciting an enhanced therapeutic response. Delivery of drug combinations at specific molar ratios has been considered a means to achieve synergistic effects resulting in improvements in efficacy while minimizing dose related adverse drug reactions. The benefits of this approach have been realized with the FDA approval of Vyxeos®, the first liposome formulation to deliver a synergistic drug combination leading to improved overall survival against standard of care. In the current study, we demonstrate the synergistic potential of the PARP inhibitor niraparib and doxorubicin for the treatment of ovarian cancer. Through in vitro screening in a panel of ovarian cancer cell lines, we find that niraparib and doxorubicin demonstrate consistent synergy/additivity at the majority of evaluated molar ratio combinations. Further to these findings, we report formulation of a nanoparticle encapsulating our identified synergistic combination. We describe a rational design process to achieve highly stable liposomes that are targeted with folate to folate-receptor-alpha, which is known to be overexpressed on the surface of ovarian cancer cells. With this approach, we aim to achieve targeted delivery of niraparib and doxorubicin at a pre-determined synergistic molar ratio via increased receptor-mediated endocytosis.

Re-envisioning the design of nanomedicines: harnessing automation and artificial intelligence.

INTRODUCTION: Interest in nanomedicines has surged in recent years due to the critical role they have played in the COVID-19 pandemic. Nanoformulations can turn promising therapeutic cargo into viable products through improvements in drug safety and efficacy profiles. However, the developmental pathway for such formulations is non-trivial and largely reliant on trial-and-error. Beyond the costly demands on time and resources, this traditional approach may stunt innovation. The emergence of automation, artificial intelligence (AI) and machine learning (ML) tools, which are currently underutilized in pharmaceutical formulation development, offers a promising direction for an improved path in the design of nanomedicines. AREAS COVERED: the potential of harnessing experimental automation and AI/ML to drive innovation in nanomedicine development. The discussion centers on the current challenges in drug formulation research and development, and the major advantages afforded through the application of data-driven methods. EXPERT OPINION: The development of integrated workflows based on automated experimentation and AI/ML may accelerate nanomedicine development. A crucial step in achieving this is the generation of high-quality, accessible datasets. Future efforts to make full use of these tools can ultimately contribute to the development of more innovative nanomedicines and improved clinical translation of formulations that rely on advanced drug delivery systems.

Machine learning models to accelerate the design of polymeric long-acting injectables.

Long-acting injectables are considered one of the most promising therapeutic strategies for the treatment of chronic diseases as they can afford improved therapeutic efficacy, safety, and patient compliance. The use of polymer materials in such a drug formulation strategy can offer unparalleled diversity owing to the ability to synthesize materials with a wide range of properties. However, the interplay between multiple parameters, including the physicochemical properties of the drug and polymer, make it very difficult to intuitively predict the performance of these systems. This necessitates the development and characterization of a wide array of formulation candidates through extensive and time-consuming in vitro experimentation. Machine learning is enabling leap-step advances in a number of fields including drug discovery and materials science. The current study takes a critical step towards data-driven drug formulation development with an emphasis on long-acting injectables. Here we show that machine learning algorithms can be used to predict experimental drug release from these advanced drug delivery systems. We also demonstrate that these trained models can be used to guide the design of new long acting injectables. The implementation of the described data-driven approach has the potential to reduce the time and cost associated with drug formulation development.

Triggered release from thermosensitive liposomes improves tumor targeting of vinorelbine.

Triggered drug delivery strategies have been shown to enhance drug accumulation at target diseased sites in comparison to administration of free drug. In particular, many studies have demonstrated improved targetability of chemotherapeutics when delivered via thermosensitive liposomes. However, most studies continue to focus on encapsulating doxorubicin while many other drugs would benefit from this targeted and localized delivery approach. The proposed study explores the therapeutic potential of a thermosensitive liposome formulation of the commonly used chemotherapy drug vinorelbine in combination with mild hyperthermia (39-43 °C) in a murine model of rhabdomyosarcoma. Rhabdomyosarcoma, the most common soft tissue sarcoma in children, is largely treated using conventional chemotherapy which is associated with significant adverse long-term sequelae. In this study, mild hyperthermia was pursued as a non-invasive, non-toxic means to improve the efficacy and safety profiles of vinorelbine. Thorough assessment of the pharmacokinetics, biodistribution, efficacy and toxicity of vinorelbine administered in the thermosensitive liposome formulation was compared to administration in a traditional, non-thermosensitive liposome formulation. This study shows the potential of an advanced formulation technology in combination with mild hyperthermia as a means to target an untargeted therapeutic agent and result in a significant improvement in its therapeutic index.

Harnessing immunotherapy to enhance the systemic anti-tumor effects of thermosensitive liposomes.

Chemotherapy plays an important role in debulking tumors in advance of surgery and/or radiotherapy, tackling residual disease, and treating metastatic disease. In recent years many promising advanced drug delivery strategies have emerged that offer more targeted delivery approaches to chemotherapy treatment. For example, thermosensitive liposome-mediated drug delivery in combination with localized mild hyperthermia can increase local drug concentrations resulting in a reduction in systemic toxicity and an improvement in local disease control. However, the majority of solid tumor-associated deaths are due to metastatic spread. A therapeutic approach focused on a localized target area harbors the risk of overlooking and undertreating potential metastatic spread. Previous studies reported systemic, albeit limited, anti-tumor effects following treatment with thermosensitive liposomal chemotherapy and localized mild hyperthermia. This work explores the systemic treatment capabilities of a thermosensitive liposome formulation of the vinca alkaloid vinorelbine in combination with mild hyperthermia in an immunocompetent murine model of rhabdomyosarcoma. This treatment approach was found to be highly effective at heated, primary tumor sites. However, it demonstrated limited anti-tumor effects in secondary, distant tumors. As a result, the addition of immune checkpoint inhibition therapy was pursued to further enhance the systemic anti-tumor effect of this treatment approach. Once combined with immune checkpoint inhibition therapy, a significant improvement in systemic treatment capability was achieved. We believe this is one of the first studies to demonstrate that a triple combination of thermosensitive liposomes, localized mild hyperthermia, and immune checkpoint inhibition therapy can enhance the systemic treatment capabilities of thermosensitive liposomes.

Reactive organic carbon air emissions from mobile sources in the United States.

Mobile sources are responsible for a substantial controllable portion of the reactive organic carbon (ROC) emitted to the atmosphere, especially in urban environments of the United States. We update existing methods for calculating mobile source organic particle and vapor emissions in the United States with over a decade of laboratory data that parameterize the volatility and organic aerosol (OA) potential of emissions from on-road vehicles, nonroad engines, aircraft, marine vessels, and locomotives. We find that existing emission factor information from Teflon filters combined with quartz filters collapses into simple relationships and can be used to reconstruct the complete volatility distribution of ROC emissions. This new approach consists of source-specific filter artifact corrections and state-of-the-science speciation including explicit intermediate-volatility organic compounds (IVOCs), yielding the first bottom-up volatility-resolved inventory of US mobile source emissions. Using the Community Multiscale Air Quality model, we estimate mobile sources account for 20 %-25 % of the IVOC concentrations and 4.4 %-21.4 % of ambient OA. The updated emissions and air quality model reduce biases in predicting fine-particle organic carbon in winter, spring, and autumn throughout the United States (4.3 %-11.3 % reduction in normalized bias). We identify key uncertain parameters that align with current state-of-the-art research measurement challenges.

The nutritional impact of a feeding protocol for infants on high flow nasal cannula therapy.

BACKGROUND: Clinicians may be reluctant to feed patients on high-flow nasal cannula (HFNC) therapy, despite studies suggesting it is beneficial and safe. We describe the implementation of a feeding protocol for patients with bronchiolitis on HFNC and determine its effect on nutrition goals. METHODS: Prospective bedside data on enteral volume, feed interruptions, and aspiration events were collected on patients with bronchiolitis who were <24 months of age, treated with HFNC, and fed per a developed protocol. Exclusion criteria included history of prematurity <32 weeks, congenital heart disease, or positive-pressure ventilation before feeding. Length of intensive care unit and hospital stay was compared with both a concurrent cohort (CC) of patients not fed per the protocol and a retrospective cohort (RC) admitted prior to protocol creation. RESULTS: Seventy-eight patients met the criteria for the prospective study arm: 24 patients were included in the CC, and 74 were included in the RC. Seventy-one percent of prospective patients received enteral nutrition (EN) on HFNC day 1 vs 42% of the CC. In the prospective cohort, feed interruption occurred in 23% of patients and was associated with higher flow rates; however, no aspiration events occurred. Patients fed per protocol were fed 8-10 h sooner and discharged 1 day earlier than those in the RC. CONCLUSION: The use of a feeding protocol for patients with bronchiolitis on HFNC was safe and associated with shorter time to initiate EN and shorter length of hospital stay.

Shifting the Paradigm on Cannabis Safety.

The global movement toward legalization of cannabis is resulting in an ever-increasing public perception that cannabis is safe. Cannabis is not the first drug to be available for nonmedical use, nor is it the first to have such an unfounded safety profile. The safety of long-term exposure to phytocannabinoids is misunderstood by, and under reported to, the general public. There is evidence to suggest that long-term use of recreational cannabis may be associated with an increased risk of undesirable side effects. This evidence warrants both appropriate caution from the general public and investment in further research by government and industry sectors that are profiting from the sale of these potent psychoactive agents. There is no doubt that these compounds have medical potential. However, in addition to the medical potential, we must also remain aware of the adverse health effects that are becoming synonymous with recreational cannabis use. This perspective highlights the privileged role that cannabis has as a perceived "safe drug" in society and summarizes some concerning side effects that are becoming associated with regular nonprescribed cannabis use.

Turning down the heat: The case for mild hyperthermia and thermosensitive liposomes.

"A single disappointing study does not mean an end to the future of ThermoDox®", writes Michael Tardugno (CEO of Celsion Corporation), after announcing the termination of Celsion's second Phase III clinical trial. The OPTIMA trial, as it was known, evaluated their thermosensitive liposome (TSL) formulation of doxorubicin (ThermoDox®) in combination with radiofrequency ablation for the treatment of hepatocellular carcinoma (HCC). The purpose of this perspective is to review the case of ThermoDox and to address questions related to its clinical translation. Specifically, what has prevented the clinical translation of this once highly regarded breakthrough technology? Is this the end of TSLs? What can we learn from the challenges faced in the clinical development of this multi-modal therapy? As formulation scientists working in the field, we continue to believe that heat-triggered drug delivery platforms have tremendous potential as chemotherapy. Herein, we highlight potential limitations in the design of many of the Thermodox clinical trials, and we propose that despite these setbacks, TSLs have the potential to become an effective component of cancer therapy.