Inactivation in vitro of the marine parasite Amyloodinium ocellatum.

The ectoparasite Amyloodinium ocellatum is a dinoflagellate that causes severe morbidity and mortality in both brackish and marine warmwater aquaculture fishes worldwide. A. ocellatum has a triphasic life cycle based on a free-living flagellate (the dinospore), a parasitic stage (the trophont) and a resting and reproductive cyst (the tomont). Current chemical treatments have shown some efficacy in eliminating dinospores but fail to inactivate the tomonts. Here we evaluated the efficacy of alternative treatments in vitro through sporulation tests and the quantification of dinospore production and motility. Hydrogen peroxide and peracetic acid efficiently decreased dinospore production at low concentrations, but total inactivation of tomonts was only achieved with higher dosages. Tomont inactivation was also observed with disinfectants such as sodium hypochlorite and Virkon S. This work provides insights into effective and environmentally friendly alternatives for the elimination of resistant forms of the marine parasite A. ocellatum.

Role of myocardial microRNAs in the long-term ventricular remodelling of patients with aortic stenosis.

Aims: We hypothesize that miRs are key players in the dynamics of the hypertrophy phenotype in aortic stenosis (AS) patients. In our study, we aimed to identify the transcriptional patterns (protein-coding transcripts and miRs) from myocardial sample biopsies that could be associated with the absence of left ventricular (LV) mass regression after aortic valve replacement (AVR) in patients with severe AS and LV hypertrophy. Methods and results: We prospectively included 40 patients with severe AS, LV hypertrophy, and preserved ejection fraction undergoing AVR. Myocardial biopsies obtained during surgery were analysed for transcriptomic analysis performed by next-generation sequencing. At a 1-year follow-up, no hypertrophy reversal was observed in about half of the patients in the absence of patient-prosthesis mismatch and prosthesis dysfunction of uncontrolled hypertension. Predictors of mass regression were assessed from clinical, echocardiographic, and biochemical variables as well as from 300 miRs obtained from myocardial specimens, allowing the identification 29 differentially expressed. miR-4709-3p was found as a positive independent predictor of hypertrophy regression together with high-sensitivity troponin T (cTNT-hs) as a negative predictor. Gene transcripts RFX1, SIX5, MAPK8IF3, and PKD1 were predicted as simultaneous targets of five upregulated miRs suggesting its importance in LV hypertrophy. Conclusion: In our cohort, tissue miR-4709-3p and cTNT-hs were independent predictors of hypertrophy regression. The hypertrophy reversal process will likely depend from a complex network where miRNAs may have an important role, allowing a potential opportunity for therapy.

The role of obstructive sleep apnea and nocturnal hypoxia as predictors of mortality in cancer patients.

BACKGROUND: Obstructive sleep apnea (OSA), due to its high prevalence, has been associated with a number of comorbidities, frequently impacting the overall course of these other diseases if left untreated. Recent studies highlight a potential association between OSA and cancer. This study investigates how OSA severity and hypoxia affect cancer prognosis, aiming to elucidate how they interplay. METHODS: Retrospective study including patients with a diagnosis of OSA after any cancer type followed up in a tertiary center during a 10-year period. OSA was mainly diagnosed after level III polysomnographic studies. RESULTS: Nocturnal hypoxia was significantly more prevalent in patients presenting lung cancer versus other malignancies and was associated with higher rates of oncologic disease progression. Overall survival was significantly lower in severe OSA patients and also in patients presenting nocturnal hypoxia. A composite hypoxia score considering both OSA severity and significant hypoxia was an independent predictor of mortality regardless of clinical cancer staging and treatment. Shorter time between cancer and OSA diagnosis was also associated with worse prognosis. CONCLUSION: This study suggests an association between OSA severity and nocturnal hypoxia and increased cancer mortality independently from possible confounding factors such as age, cancer clinical staging at diagnosis, treatment modality and also progression. Neoplastic patients with severe OSA and/or complex hypoxia seem to have lower overall survival rates than those with less severe OSA and nocturnal hypoxia.

Time since fire as a driver of taxonomic and phylogenetic patterns of grassland plant communities.

Fire plays a key role in grasslands, determining the distribution and evolution of species and boundaries with neighboring ecosystems. Evidence of community-wide responses to fire is largely based on taxonomic and functional descriptors, while the phylogenetic dimension is overlooked. Here we evaluated how the taxonomic and phylogenetic structure of grassland plant communities responded to a time since fire (TSF) gradient. We sampled 12 communities in Southern Brazil under varying TSF and calculated taxonomic species richness (S) and dominance (D), phylogenetic diversity (PD), and mean phylogenetic distances (MPD). We used Structural Equation Models to test the relationships between the environmental gradient and community descriptors. Communities with longer TSF presented higher PD and MPD but lower species richness and increased taxonomic dominance. These sites were dominated by monocots, specifically C4 grasses, but also presented exclusive clades, whereas recently-burned sites presented lower taxonomic dominance and more species distributed in a wider variety of clades. Our results indicate that these scenarios are interchangeable and dependent on fire management. Fire adaptation was not constrained by phylogenetic relatedness, contrasting with previous findings for tropical savannahs and indicating that temperate and tropical non-forest ecosystems from South America respond differently to fire, possibly due to different evolutionary histories.

Analysis of peri-implant bone tissue between hydrophilic and rough implant surfaces in spontaneously hypertensive rats treated with losartan.

OBJECTIVES: to evaluate the morphological and functional characteristics of the peri-implant bone tissue that was formed during the healing process by the placement implants using two different surface treatments: hydrophilic Acqua™ (ACQ) and rough NeoPoros™ (NEO), in spontaneously hypertensive (SHR) and normotensive rats (Wistar) whether or not treated with losartan. METHODOLOGY: In total, 96 male rats (48 Wistar and 48 SHR) were divided into eight subgroups: absolute control rough (COA NEO), absolute control hydrophilic (COA ACQ), losartan control rough (COL NEO), losartan control hydrophilic (COL ACQ), SHR absolute rough (SHR NEO), SHR absolute hydrophilic (SHR ACQ), SHR losartan rough (SHRL NEO), and SHR losartan hydrophilic (SHRL ACQ). The rats medicated with losartan received daily doses of the medication. NeoPoros™ and Acqua™ implants were installed in the tibiae of the rats. After 14 and 42 days of the surgery, the fluorochromes calcein and alizarin were injected in the rats. The animals were euthanized 67 days after treatment. The collected samples were analyzed by immunohistochemistry, biomechanics, microcomputerized tomography, and laser confocal scanning microscopy analysis. RESULTS: The osteocalcin (OC) and vascular endothelium growth factor (VEGF) proteins had moderate expression in the SHRL ACQ subgroup. The same subgroup also had the highest implant removal torque. Regarding microarchitectural characteristics, a greater number of trabeculae was noted in the control animals that were treated with losartan. In the bone mineralization activity, it was observed that the Acqua™ surface triggered higher values of MAR (mineral apposition rate) in the COA, COL, and SHRL groups (p<0.05). CONCLUSION: the two implant surface types showed similar responses regarding the characteristics of the peri-implant bone tissue, even though the ACQ surface seems to improve the early stages of osseointegration.

Impact of dydrogesterone use in cycles with low progesterone levels on the day of frozen embryo transfer.

PURPOSE: This study aims to evaluate whether the clinical outcomes of cycles with frozen embryo transfer (FET) in hormonal replacement treatment supplemented with dydrogesterone (DYD) following detection of low circulating levels of progesterone (P4) were comparable to the results of cycles with otherwise normal serum P4 values. METHODS: Extended analyses of a retrospective cohort that included FET cycles performed between July 2019 and March 2022 after a cycle of artificial endometrial preparation using valerate-estradiol and micronized vaginal P4 (400 mg twice daily). Whenever the serum P4 value was considered low on the morning of the planned transfer, 10 mg of DYD three times a day was added as a supplement. Only single-embryo transfers of a blastocyst were considered. The primary endpoint was live birth rate. RESULTS: Five-hundred thirty-five FET cycles were analyzed, of which 136 (25.4%) underwent treatment with DYD. There were 337 pregnancies (63%), 207 live births (38.6%), and 130 miscarriages (38.5%). The P4 values could be modeled by a gamma distribution, with a mean of 14.5 ng/ml and a standard deviation of 1.95 ng/ml. The variables female age on the day of FET, ethnicity, and weight were associated with a variation in the serum P4 values. There were no differences in the results between cycles with or without the indication for DYD supplementation. CONCLUSIONS: Live birth rate did not vary significantly in females with low and normal serum P4 levels on the day of FET when DYD was used as rescue therapy.

Health care expenses impact on the disability-adjusted life years in non-communicable diseases in the European Union.

Background: Non-communicable diseases are a global health problem. The metric Disability-Adjusted Life Years was developed to measure its impact on health systems. This metric makes it possible to understand a disease's burden, towards defining healthcare policies. This research analysed the effect of healthcare expenditures in the evolution of disability-adjusted life years for non-communicable diseases in the European Union between 2000 and 2019. Methods: Data were collected for all 27 European Union countries from Global Burden of Disease 2019, Global Health Expenditure, and EUROSTAT databases. Econometric panel data models were used to assess the impact of healthcare expenses on the disability-adjusted life years. Only models with a coefficient of determination equal to or higher than 10% were analysed. Results: There was a decrease in the non-communicable diseases with the highest disability-adjusted life years: cardiovascular diseases (-2,952 years/105 inhabitants) and neoplasms (-618 years/105 inhabitants). Health expenditure significantly decreased disability-adjusted life years for all analysed diseases (p < 0.01) unless for musculoskeletal disorders. Private health expenditure did not show a significant effect on neurological and musculoskeletal disorders (p > 0.05) whereas public health expenditure did not significantly influence skin and subcutaneous diseases (p > 0.05). Conclusion: Health expenditure have proved to be effective in the reduction of several diseases. However, some categories such as musculoskeletal and mental disorders must be a priority for health policies in the future since, despite their low mortality, they can present high morbidity and disability.

De novo detection of somatic variants in long-read single-cell RNA sequencing data.

In cancer, genetic and transcriptomic variations generate clonal heterogeneity, possibly leading to treatment resistance. Long-read single-cell RNA sequencing (LR scRNA-seq) has the potential to detect genetic and transcriptomic variations simultaneously. Here, we present LongSom, a computational workflow leveraging LR scRNA-seq data to call de novo somatic single-nucleotide variants (SNVs), copy-number alterations (CNAs), and gene fusions to reconstruct the tumor clonal heterogeneity. For SNV calling, LongSom distinguishes somatic SNVs from germline polymorphisms by reannotating marker gene expression-based cell types using called variants and applying strict filters. Applying LongSom to ovarian cancer samples, we detected clinically relevant somatic SNVs that were validated against single-cell and bulk panel DNA-seq data and could not be detected with short-read (SR) scRNA-seq. Leveraging somatic SNVs and fusions, LongSom found subclones with different predicted treatment outcomes. In summary, LongSom enables de novo SNVs, CNAs, and fusions detection, thus enabling the study of cancer evolution, clonal heterogeneity, and treatment resistance.

Evaluation of nanoscale versus hybrid micro/nano surface topographies for endosseous implants.

We examined the effect of a nanoscale titanium surface topography (D) versus two hybrid micro/nanoscale topographies (B and OS) on adherent mesenchymal stem cells (MSCs) and bone marrow derived macrophages (BMMs) function in cell culture and in vivo. In the in vitro study, compared to OS and B surfaces, D surface induced earlier and greater cell spreading, and earlier and profound mRNA expression of RUNX2, Osterix and BMP2 in MSCs. D surface induced earlier and higher expression of RUNX2 and BMP2 and lower expression of inflammatory genes in implant adherent cells in vivo. Measurement of osteogenesis at implant surfaces showed greater bone-to-implant contact at D versus OS surfaces after 21 days. We explored the cell population on the D and OS implant surfaces 24 h after placement using single-cell RNA sequencing and identified distinct cell clusters including macrophages, neutrophils and B cells. D surface induced lower expression and earlier reduction of inflammatory genes expression in BMMs in vitro. BMMs on D, B and OS surfaces demonstrated a marked increase of BMP2 expression after 1 and 3 days, and this increase was significantly higher on D surface at day 3. Our data implicates a dynamic process that may be influenced by nanotopography at multiple stages of osseointegration including initial immunomodulation, recruitment of MSCs and later osteoblastic differentiation leading to bone matrix production and mineralization. The results suggest that a nanoscale topography (D) favorably modulates adherent macrophage polarization toward anti-inflammatory and regenerative phenotypes and promotes the osteoinductive phenotype of adherent mesenchymal stem cells. STATEMENT OF SIGNIFICANCE: Our manuscript contains original data developed to define effects of a novel nanotopography on the process of osseointegration at the cell and tissue level.  Few studies have compared the effects of a nanoscale surface versus the more typical hybrid micro/nano-scale surfaces used today. We have utilized single-cell RNA sequencing for the first time to identify earliest cell populations on implant surfaces in vivo. We provide data indicating that the nanoscale surface acts upon both osteoprogenitor and immune cell (macrophages) to alter the process of bone formation in a surface-specific manner. This work represents new observations regarding osseointegration and immunomodulation.

Analysis of peri-implant bone tissue between hydrophilic and rough implant surfaces in spontaneously hypertensive rats treated with losartan

Abstract Hypertensive individuals present alterations in their calcium metabolism that consequently decrease the concentration of this mineral in the bone tissue. Objectives to evaluate the morphological and functional characteristics of the peri-implant bone tissue that was formed during the healing process by the placement implants using two different surface treatments: hydrophilic Acqua™ (ACQ) and rough NeoPoros™ (NEO), in spontaneously hypertensive (SHR) and normotensive rats (Wistar) whether or not treated with losartan. Methodology In total, 96 male rats (48 Wistar and 48 SHR) were divided into eight subgroups: absolute control rough (COA NEO), absolute control hydrophilic (COA ACQ), losartan control rough (COL NEO), losartan control hydrophilic (COL ACQ), SHR absolute rough (SHR NEO), SHR absolute hydrophilic (SHR ACQ), SHR losartan rough (SHRL NEO), and SHR losartan hydrophilic (SHRL ACQ). The rats medicated with losartan received daily doses of the medication. NeoPoros™ and Acqua™ implants were installed in the tibiae of the rats. After 14 and 42 days of the surgery, the fluorochromes calcein and alizarin were injected in the rats. The animals were euthanized 67 days after treatment. The collected samples were analyzed by immunohistochemistry, biomechanics, microcomputerized tomography, and laser confocal scanning microscopy analysis. Results The osteocalcin (OC) and vascular endothelium growth factor (VEGF) proteins had moderate expression in the SHRL ACQ subgroup. The same subgroup also had the highest implant removal torque. Regarding microarchitectural characteristics, a greater number of trabeculae was noted in the control animals that were treated with losartan. In the bone mineralization activity, it was observed that the Acqua™ surface triggered higher values of MAR (mineral apposition rate) in the COA, COL, and SHRL groups (p<0.05). Conclusion the two implant surface types showed similar responses regarding the characteristics of the peri-implant bone tissue, even though the ACQ surface seems to improve the early stages of osseointegration.

Stomatological profile of elderly patients who received dental care at a reference Hospital of Belo Horizonte, Minas Gerais - Brazil: a cross-sectional study

Aim: The aim of this study was to draw a stomatological profile of elderly patients treated at the Stomatology Service of a referral hospital in the city of Belo Horizonte, Minas Gerais ­ Brazil.Methods: Information was collected on the diagnostic hypotheses and definitive diagnoses of the elderly (aged 60 to 98 years) treated at the service, from September 2014 to December 2017. Results: Medical records of 1388 patients were evaluated, and 281 different diagnostic hypotheses were found. For males, there was a higher prevalence in both diagnostic and definitive diagnoses for benign and malignant epithelial neoplasms; In females, there was a higher prevalence, also in the hypotheses and diagnostic definitions, of inflammatory fibrous hyperplasia, related to the use of removable prostheses. Conclusion: Thus, this study evidenced the importance of knowledge about the main alterations that affect the oral cavity of elderly patients, which allows better attention to this population, besides being a pioneering stomatological study in hospital units.

A Minimally Invasive Model of Aortic Stenosis in Swine.

Large animal models of heart failure play an essential role in the development of new therapeutic interventions due to their size and physiological similarities to humans. Efforts have been dedicated to creating a model of pressure-overload induced heart failure, and ascending aortic banding while still supra-coronary and not a perfect mimic of aortic stenosis in humans, closely resembling the human condition. The purpose of this study is to demonstrate a minimally invasive approach to induce left ventricular pressure overload by placing an aortic band, precisely calibrated with percutaneously introduced high-fidelity pressure sensors. This method represents a refinement of the surgical procedure (3Rs), resulting in homogenous trans-stenotic gradients and reduced intragroup variability. Additionally, it enables swift and uneventful animal recovery, leading to minimal mortality rates. Throughout the study, animals were followed for up to 2 months after surgery, employing transthoracic echocardiography and pressure-volume loop analysis. However, longer follow-up periods can be achieved if desired. This large animal model proves valuable for testing new drugs, particularly those targeting hypertrophy and the structural and functional alterations associated with left ventricular pressure overload.

Judicialization and cancer: quality of life of patients and caregivers in the COVID-19 pandemic.

BACKGROUND: In Brazil, cancer patients and caregivers of cancer patients seek judicial intervention for free access to medications from the public health system. Indeed, the COVID-19 pandemic potentially affected the health-related quality of life of cancer patients and caregivers of cancer patients. This study aimed to describe the sociodemographic profile and assess the health-related quality of life of patients and caregivers in the state of Goias, Brazil, in 2020. METHODS: A cross-sectional study was conducted using the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) and a sociodemographic questionnaire. RESULTS: A total of 88 (67,7%) patients and 42 (32,3%) caregivers participated in the study, mostly women (55,5%); aged from 18 to 60 (66%) years old; with up to nine years of education (73,1%) and monthly family income lower than the minimum wage (69,2%); married or in a stable union (92,3%); living with multiple people in the same household (73,8%). The quality of life domains with the best scores were mental health for patients and pain for caregivers. The most affected quality of life domain was physical limitation for patients and caregivers. Factors associated with better quality of life were female gender and age between 18 and 60 years in patients, more than 9 years of education, living with multiple people in the same house, and having a monthly family income higher than US$200 for caregivers. CONCLUSION: The study found evidence of physical and emotional vulnerability during the pandemic, highlighting the need to strengthen public policies of assistance support to this population.

Transcatheter interventions for left-sided valvular heart disease complicated by cardiogenic shock: a consensus statement from the European Association of Percutaneous Cardiovascular Interventions (EAPCI) in collaboration with the Association for Acute Cardiovascular Care (ACVC) and the ESC Working Group on Cardiovascular Surgery.

Valvular heart disease (VHD) is one of the most frequent causes of heart failure (HF) and is associated with poor prognosis, particularly among patients with conservative management. The development and improvement of catheter-based VHD interventions have broadened the indications for transcatheter valve interventions from inoperable/high-risk patients to younger/lower-risk patients. Cardiogenic shock (CS) associated with severe VHD is a clinical condition with a very high risk of mortality for which surgical treatment is often deemed a prohibitive risk. Transcatheter valve interventions might be a promising alternative in this setting given that they are less invasive. However, supportive scientific evidence is scarce and often limited to small case series. Current guidelines on VHD do not contain specific recommendations on how to manage patients with both VHD and CS. The purpose of this clinical consensus statement, developed by a group of international experts invited by the European Association of Percutaneous Cardiovascular Interventions (EAPCI) Scientific Documents and Initiatives Committee, is to perform a review of the available scientific evidence on the management of CS associated with left-sided VHD and to provide a rationale and practical approach for the application of transcatheter valve interventions in this specific clinical setting.

Evaluation of oral mucositis, candidiasis, and quality of life in patients with head and neck cancer treated with a hypofractionated or conventional radiotherapy protocol: a longitudinal, prospective, observational study.

BACKGROUND: Due to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, recently, Radiotherapy (RT) protocols requiring fewer sessions (hypofractionated) have been used to shorten RT treatment and minimize patient exposure to medical centers, and decrease the risk of SARS-CoV-2 infection. METHODS: This longitudinal, prospective, observational study aimed to compare the quality of life (QoL) and the incidence of oral mucositis and candidiasis in 66 patients with head and neck cancer (HNC) who undergo a hypofractionated RT protocol (GHipo), total of 55 Gy for 4 weeks, or a conventional RT protocol (GConv), total of 66 - 70 Gy for 6 - 7 weeks. PURPOSE: To assess the incidence and severity of oral mucositis, the incidence of candidiasis, and QoL were evaluated using the World Health Organization scale, clinical evaluation, and the QLC-30 and H&N-35 questionnaires, respectively, at the beginning and the end of RT. RESULTS: The incidence of candidiasis did not show differences between the two groups. However, at the end of RT, mucositis had a higher incidence (p < 0.01) and severity (p < 0.05) in GHipo. QoL was not markedly different between the two groups. Although mucositis worsened in patients treated with hypofractionated RT, QoL did not worsen for patients on this regimen. CONCLUSIONS: Our results open perspectives for the potential use of RT protocols for HNC with fewer sessions in conditions that require faster, cheaper, and more practical treatments.

Brain metastases from hepatopancreatobiliary malignancies.

While colorectal and gastroesophageal cancer represent the two gastrointestinal (GI) tumor entities with the highest incidence of brain metastatic (BM) disease, data on the clinical course of BM patients from hepatopancreatobiliary malignancies are rare. Patients with cholangiocarcinoma (CCA), hepatocellular carcinoma (HCC), pancreatic ductal adenocarcinoma (PDAC) and gastroenteropancreatic neuroendocrine neoplasms (GEP NEN). Treated for BM between 1991 and 2017 at an academic care center were included. Brain metastases-free survival (BMFS) was defined as interval from first diagnosis until BM development. Overall survival (OS) was defined as interval from diagnosis of BM until death or last date of follow-up. Outcome was correlated with clinical and treatment factors. 29 patients from overall 6102 patients (0.6%) included in the Vienna Brain Metastasis Registry presented with BM from hepatopancreatobiliary primaries including 9 (31.0%) with CCA, 10 (34.5%) with HCC, 7 (24.1%) with PDAC and 3 (10.3%) with GEP NEN as primary tumor. Median BMFS was 21, 12, 14 and 7 months and median OS 4, 4, 6 and 4 months, respectively. Karnofsky Performance Status (KPS) below 80% (p = 0.08), age above 60 years (p = 0.10) and leptomeningeal carcinomatosis (LC) (p = 0.09) diagnosed concomitant to solid BM showed an inverse association with median OS (Cox proportional hazards model). In this cohort of patients with BM from hepatopancreatobiliary tumor entities, prognosis was shown to be very limited. Performance status, age and diagnosis of LC were identified as negative prognostic factors.

Revolutionizing Disease Modeling: The Emergence of Organoids in Cellular Systems.

Cellular models have created opportunities to explore the characteristics of human diseases through well-established protocols, while avoiding the ethical restrictions associated with post-mortem studies and the costs associated with researching animal models. The capability of cell reprogramming, such as induced pluripotent stem cells (iPSCs) technology, solved the complications associated with human embryonic stem cells (hESC) usage. Moreover, iPSCs made significant contributions for human medicine, such as in diagnosis, therapeutic and regenerative medicine. The two-dimensional (2D) models allowed for monolayer cellular culture in vitro; however, they were surpassed by the three-dimensional (3D) cell culture system. The 3D cell culture provides higher cell-cell contact and a multi-layered cell culture, which more closely respects cellular morphology and polarity. It is more tightly able to resemble conditions in vivo and a closer approach to the architecture of human tissues, such as human organoids. Organoids are 3D cellular structures that mimic the architecture and function of native tissues. They are generated in vitro from stem cells or differentiated cells, such as epithelial or neural cells, and are used to study organ development, disease modeling, and drug discovery. Organoids have become a powerful tool for understanding the cellular and molecular mechanisms underlying human physiology, providing new insights into the pathogenesis of cancer, metabolic diseases, and brain disorders. Although organoid technology is up-and-coming, it also has some limitations that require improvements.

A systematic evaluation of deep learning methods for the prediction of drug synergy in cancer.

One of the main obstacles to the successful treatment of cancer is the phenomenon of drug resistance. A common strategy to overcome resistance is the use of combination therapies. However, the space of possibilities is huge and efficient search strategies are required. Machine Learning (ML) can be a useful tool for the discovery of novel, clinically relevant anti-cancer drug combinations. In particular, deep learning (DL) has become a popular choice for modeling drug combination effects. Here, we set out to examine the impact of different methodological choices on the performance of multimodal DL-based drug synergy prediction methods, including the use of different input data types, preprocessing steps and model architectures. Focusing on the NCI ALMANAC dataset, we found that feature selection based on prior biological knowledge has a positive impact-limiting gene expression data to cancer or drug response-specific genes improved performance. Drug features appeared to be more predictive of drug response, with a 41% increase in coefficient of determination (R2) and 26% increase in Spearman correlation relative to a baseline model that used only cell line and drug identifiers. Molecular fingerprint-based drug representations performed slightly better than learned representations-ECFP4 fingerprints increased R2 by 5.3% and Spearman correlation by 2.8% w.r.t the best learned representations. In general, fully connected feature-encoding subnetworks outperformed other architectures. DL outperformed other ML methods by more than 35% (R2) and 14% (Spearman). Additionally, an ensemble combining the top DL and ML models improved performance by about 6.5% (R2) and 4% (Spearman). Using a state-of-the-art interpretability method, we showed that DL models can learn to associate drug and cell line features with drug response in a biologically meaningful way. The strategies explored in this study will help to improve the development of computational methods for the rational design of effective drug combinations for cancer therapy.