Corrigendum to "3D-printed tri-element-doped hydroxyapatite/ polycaprolactone composite scaffolds with antibacterial potential for osteosarcoma therapy and bone regeneration" [Bioact. Mater. 31 (January 2024) 18-37].

[This corrects the article DOI: 10.1016/j.bioactmat.2023.07.004.].

3D-printed tri-element-doped hydroxyapatite/ polycaprolactone composite scaffolds with antibacterial potential for osteosarcoma therapy and bone regeneration.

The resection of malignant osteosarcoma often results in large segmental bone defects, and the residual cells can facilitate recurrence. Consequently, the treatment of osteosarcoma is a major challenge in clinical practice. The ideal goal of treatment for osteosarcoma is to eliminate it thoroughly, and repair the resultant bone defects as well as avoid bacterial infections. Herein, we fabricated a selenium/strontium/zinc-doped hydroxyapatite (Se/Sr/Zn-HA) powder by hydrothermal method, and then employed it with polycaprolactone (PCL) as ink to construct composite scaffolds through 3D printing, and finally introduced them in bone defect repair induced by malignant osteosarcoma. The resultant composite scaffolds integrated multiple functions involving anti-tumor, osteogenic, and antibacterial potentials, mainly attributed to the anti-tumor effects of SeO32-, osteogenic effects of Sr2+ and Zn2+, and antibacterial effects of SeO32- and Zn2+. In vitro studies confirmed that Se/Sr/Zn-HA leaching solution could induce apoptosis of osteosarcoma cells, differentiation of MSCs, and proliferation of MC3T3-E1 while showing excellent antibacterial properties. In vivo tests demonstrated that Se/Sr/Zn-HA could significantly suppress tumors after 8 days of injection, and the Se/Sr/Zn-HA-PCLs scaffold repaired femoral defects effectively after 3 months of implantation. Summarily, the Se/Sr/Zn-HA-PCLs composite scaffolds developed in this study were effective for tumor treatment, bone defect repair, and post-operative anti-infection, which provided a great potential to be a facile therapeutic material for osteosarcoma resection.

Tissue engineering in growth plate cartilage regeneration: Mechanisms to therapeutic strategies.

The repair of growth plate injuries is a highly complex process that involves precise spatiotemporal regulation of multiple cell types. While significant progress has been made in understanding the pathological mechanisms underlying growth plate injuries, effectively regulating this process to regenerate the injured growth plate cartilage remains a challenge. Tissue engineering technology has emerged as a promising therapeutic approach for achieving tissue regeneration through the use of functional biological materials, seed cells and biological factors, and it is now widely applied to the regeneration of bone and cartilage. However, due to the unique structure and function of growth plate cartilage, distinct strategies are required for effective regeneration. Thus, this review provides an overview of current research on the application of tissue engineering to promote growth plate regeneration. It aims to elucidates the underlying mechanisms by which tissue engineering promotes growth plate regeneration and to provide novel insights and therapeutic strategies for future research on the regeneration of growth plate.

Risk factor analysis for tibial tubercle avulsion fractures in children.

PURPOSE: Tibial tubercle avulsion fractures (TTAFs) are rare in children, particularly bilateral TTAFs. This study aimed to elucidate the associating factors of TTAF, and compare the risk factor profiles of unilateral and bilateral injuries, thus provide clinical theoretical basis for reducing the occurrence of TTAFs. METHODS: TTAF paediatric patients who were hospitalized between April 2017 and November 2022 were retrospectively analysed. Children who presented for physical examination during the same period were randomly selected, and were age- and sex-matched as controls. A subgroup analysis based on endocrine function was also performed. A risk factor analysis for bilateral TTAF was performed as well. Data were collected via medical records and a questionnaire. All variables were assessed for association with TTAF using univariate and multiple logistic regression analyses. RESULTS: A total of 64 TTAF patients and controls were respectively included. Multivariate analysis demonstrated BMI (P = 0.000，OR = 3.172), glucose (P = 0.016，OR = 20.878), and calcium (P = 0.034，OR = 0.000) as independent associating factors of TTAF. Subgroup analysis showed significant differences in oestradiol (P = 0.014), progesterone (P = 0.006) and insulin levels (P = 0.005) between the TTAF and control groups. Bilateral TTAF was found to significantly associate with a history of knee joint pain (P = 0.026). CONCLUSION: High BMI, hyperglycaemia, and low calcium levels were found as independent risk factors for TTAF in children. In addition, decreased oestradiol, elevated progesterone, and insulin resistance were identified as potential risk factors for TTAF. A history of knee pain may be suggestive of bilateral TTAF.

A Novel Method for Treating Distal Radius Diaphyseal Metaphyseal Junction Fracture in Children.

BACKGROUND The treatment of distal radius diaphyseal metaphyseal junction (DMJ) fracture in children is a clinical problem; several treatments are available, but none are very effective. Therefore, this study aimed to report a novel method for treating this fracture using limited open reduction and transepiphyseal intramedullary fixation with Kirschner wire. MATERIAL AND METHODS From January 2018 to December 2019, a total of 15 children (13 boys and 2 girls) with distal radius DMJ fractures with a mean age of 10 years (range: 6-14 years) were included in the study. The operation time, incision length, and X-ray radiation exposure were precisely recorded. All children were followed up regularly. At the final follow-up, clinical outcomes were evaluated according to Price criteria, and complications were recorded. RESULTS The mean operation time of the 15 children was 21.4 min, and the mean incision length was 1.9 cm. The intraoperative X-ray was performed 3.7 times on average. The mean radiographic union of fracture was 4.7 weeks, and the mean time to remove the Kirschner wire was 4.8 weeks for radial instrumentation and 4.7 months for ulnar instrumentation. According to the Price grading evaluation system, clinical outcome was excellent in 14 cases and good in 1 case. Moreover, there were no major complications related to loss of reduction, malunion, nonunion, and physeal arrest of the distal radius. CONCLUSIONS Limited open reduction and transepiphyseal intramedullary fixation with Kirschner wire are effective for treating distal radius DMJ fracture in children, which has the advantages of simple surgical procedures, short operation time, small incision, and less radiation exposure, making it an excellent choice for treating this fracture.

3D bioprinted hydrogel/polymer scaffold with factor delivery and mechanical support for growth plate injury repair.

Introduction: Growth plate injury is a significant challenge in clinical practice, as it could severely affect the limb development of children, leading to limb deformity. Tissue engineering and 3D bioprinting technology have great potential in the repair and regeneration of injured growth plate, but there are still challenges associated with achieving successful repair outcomes. Methods: In this study, GelMA hydrogel containing PLGA microspheres loaded with chondrogenic factor PTH(1-34) was combined with BMSCs and Polycaprolactone (PCL) to develop the PTH(1-34)@PLGA/BMSCs/GelMA-PCL scaffold using bio-3D printing technology. Results: The scaffold exhibited a three-dimensional interconnected porous network structure, good mechanical properties, biocompatibility, and was suitable for cellchondrogenic differentiation. And a rabbit model of growth plate injury was appliedto validate the effect of scaffold on the repair of injured growth plate. The resultsshowed that the scaffold was more effective than injectable hydrogel in promotingcartilage regeneration and reducing bone bridge formation. Moreover, the addition ofPCL to the scaffold provided good mechanical support, significantly reducing limbdeformities after growth plate injury compared with directly injected hydrogel. Discussion: Accordingly, our study demonstrates the feasibility of using 3D printed scaffolds for treating growth plate injuries and could offer a new strategy for the development of growth plate tissue engineering therapy.

Injectable hydrogel loaded with bilayer microspheres to inhibit angiogenesis and promote cartilage regeneration for repairing growth plate injury.

Introduction: The repair and regeneration of growth plate injuries using tissue engineering techniques remains a challenge due to large bone bridge formation and low chondrogenic efficiency. Methods: In this study, a bilayer drug-loaded microspheres was developed that contains the vascular endothelial growth factor (VEGF) inhibitor, Bevacizumab, on the outer layer and insulin-like growth factor-1 (IGF-1), a cartilage repair factor, on the inner layer. The microspheres were then combined with bone marrow mesenchymal stem cells (BMSCs) in the gelatin methacryloyl (GelMA) hydrogel to create a composite hydrogel with good injectability and biocompatibility. Results: The in vitro drug-release profile of bilayer microspheres showed a sequential release, with Bevacizumab released first followed by IGF-1. And this hydrogel simultaneously inhibited angiogenesis and promoted cartilage regeneration. Finally, in vivo studies indicated that the composite hydrogel reduced bone bridge formation and improved cartilage regeneration in the rabbit model of proximal tibial growth plate injury. Conclusion: This bilayer microsphere-based composite hydrogel with sequential controlled release of Bevacizumab and IGF-1 has promising potential for growth plate injury repair.

Multifunctional 3D-printed bioceramic scaffolds: Recent strategies for osteosarcoma treatment.

Osteosarcoma is the most prevalent bone malignant tumor in children and teenagers. The bone defect, recurrence, and metastasis after surgery severely affect the life quality of patients. Clinically, bone grafts are implanted. Primary bioceramic scaffolds show a monomodal osteogenesis function. With the advances in three-dimensional printing technology and materials science, while maintaining the osteogenesis ability, scaffolds become more patient-specific and obtain additional anti-tumor ability with functional agents being loaded. Anti-tumor therapies include photothermal, magnetothermal, old and novel chemo-, gas, and photodynamic therapy. These strategies kill tumors through novel mechanisms to treat refractory osteosarcoma due to drug resistance, and some have shown the potential to reverse drug resistance and inhibit metastasis. Therefore, multifunctional three-dimensional printed bioceramic scaffolds hold excellent promise for osteosarcoma treatments. To better understand, we review the background of osteosarcoma, primary 3D-printed bioceramic scaffolds, and different therapies and have a prospect for the future.

Biomaterial-assisted tumor therapy: A brief review of hydroxyapatite nanoparticles and its composites used in bone tumors therapy.

Malignant bone tumors can inflict significant damage to affected bones, leaving patients to contend with issues like residual tumor cells, bone defects, and bacterial infections post-surgery. However, hydroxyapatite nanoparticles (nHAp), the principal inorganic constituent of natural bone, possess numerous advantages such as high biocompatibility, bone conduction ability, and a large surface area. Moreover, nHAp's nanoscale particle size enables it to impede the growth of various tumor cells via diverse pathways. This article presents a comprehensive review of relevant literature spanning the past 2 decades concerning nHAp and bone tumors. The primary goal is to explore the mechanisms responsible for nHAp's ability to hinder tumor initiation and progression, as well as to investigate the potential of integrating other drugs and components for bone tumor diagnosis and treatment. Lastly, the article discusses future prospects for the development of hydroxyapatite materials as a promising modality for tumor therapy.

Application of three-dimensional printed navigation templates to correct lower limb deformities in children by the guided growth technique.

Objective: Currently, individualized navigation templates are rarely applied in pediatric orthopedic surgery. This study aimed to explore the potential of navigation templates obtained using computer-aided design and three-dimensional (3D) printing to correct lower limb deformities in children by the guided growth technique. Methods: We prospectively studied 45 children with leg length discrepancy (LLD) or lower limb angular deformities, who underwent guided growth surgery involving 8-plate. In total, 21 and 24 children were included in the navigation template (group A) group and in the traditional surgery (group B) group, respectively. Mimics software was used for designing and printing navigation templates. The operation time, X-ray radiation exposure, damage to cartilage, and postoperative complications were recorded. Results: The mean operation time in groups A and B were 20.78 and 28.39 min, respectively, and the difference was statistically significant. Compared with group B, the intraoperative exposure of X-rays in group A was reduced by 25% on average. After 9-24 months of follow-up, the deformities were corrected in both groups. No significant differences in the treatment effect were noted between the groups, and no complications occurred. Conclusions: Using the individualized navigation template in the guided growth technique made the surgical procedure convenient and simple to perform. In addition, the operation time and intraoperative exposure to X-rays were reduced. We consider that 3D printed navigation templates can facilitate the accurate completion of corrective surgeries for lower limb deformities in children, which is worthy of promotion and application.

A New Method for Modeling Rabbit Growth Plate Injury for the Study of Tissue Engineering Scaffolds.

Establishing a suitable animal model of growth plate injury is necessary to evaluate the effect of tissue engineering scaffolds on repairing the injured growth plate. However, the currently used animal models have limitations. Therefore in this study, we reported and evaluated a new modeling method termed the longitudinal disruption method, which is to make a longitudinal defect in the region of growth plate. To compare this new method with the traditional transverse disruption method, we constructed the models by both methods, respectively. To observe whether bone bridges were formed, histological sections were analyzed by hematoxylin-eosin (HE) and Masson staining at 3 weeks after modeling. The HE and Masson staining results showed the formation of bone bridges in both groups, implying that the two methods successfully injured the growth plate. However, it was unclear whether the exact injury to growth plate caused by both methods was consistent. Therefore, to evaluate the accuracy and precision of modeling method, the X-ray and micro-computed tomography (CT) were performed immediately after modeling. The percentages of accurate defect position in the longitudinal and transverse modeling groups were 88.89% and 55.56%, respectively. The micro-CT results revealed irregularly shaped defect cross sections in the transverse modeling group, whereas the defects in the longitudinal modeling group had regular shapes. The mean defect areas were 10.06 ± 0.86 and 12.30 ± 2.13 mm2 in the longitudinal and transverse modeling groups, respectively, while the difference between the actual area and the expected area were -1.94 ± 0.86 and -7.70 ± 2.13 mm2, respectively, showing the high precision of this new method. Altogether, we successfully demonstrated a new method for establishing a rabbit model of growth plate injury, which provides a simple and rapid modeling process, good modeling effect, high modeling accuracy, and convenient scaffold implantation. The new method provides an effective animal model for tissue engineering research on the repair and regeneration of injured growth plate. Impact Statement In recent years, an increasing number of studies have used tissue engineering scaffolds in the repair and regeneration of growth plate. However, the currently used animal models have certain limitations in the study of tissue engineering scaffold for growth plate. In this study, a new method is presented to establish a rabbit model of growth plate injury. This method is characterized by simple and rapid modeling process, good modeling effect, high modeling accuracy, and convenient scaffold implantation, which is suitable for the study of the repair effects of tissue engineering scaffolds. Altogether, this method provides an effective animal model for tissue engineering research on growth plate and facilitates the development of tissue engineering research on the repair and regeneration of injured growth plate.

Anteroposterior pelvic radiograph is not sufficient to confirm hip reduction after conservative treatment of developmental dysplasia of the hip.

The purpose of this study was to investigate whether an anteroposterior pelvic radiograph alone is sufficient to confirm hip reduction after conservative treatment or whether MRI could be alternatively performed. A total of 133 children (145 hips) were enrolled. All children were examined by anteroposterior pelvic radiographs and MRI. Three experts interpreted anteroposterior pelvic radiographs and then verified these results on MRI. For patients with inconsistent results between anteroposterior pelvic radiographs and MRI, the continuity of Shenton's line and Calve's line was recorded, and the medial clear space of bilateral hips was measured for unilateral cases. There was complete agreement between the three experts in the interpretation of anteroposterior pelvic radiographs of 111 (76.55%) hips; there was disagreement in the remaining 34 hips, with two experts diagnosing satisfactory reduction in 13 hips and dislocation in 21 hips. Assuming that the judgment of two or more doctors on anteroposterior pelvic radiographs was taken as the final result, 17 hips (11.72%) were misjudged. There was no statistically significant difference between the actual in-position group and the actual dislocation group in terms of the continuity of Shenton's line ( P = 0.62) and Calve's line ( P = 0.10) and the medial clear space of bilateral hips ( P = 0.08). In children less than 1 year of age with developmental dysplasia of the hip treated conservatively, the use of anteroposterior pelvic radiographs alone to judge hip reduction might result in misdiagnosis and missed diagnosis. MRI could be alternatively used to detect hip reduction after conservative treatment, especially when the doctor was not familiar with ultrasound in the presence of plaster.

Limited Open Reduction and Transepiphyseal Intramedullary Kirschner Wire Fixation for Treatment of Irreducible Distal Radius Diaphyseal Metaphyseal Junction Fracture in Older Children.

Objective: This study aimed to compare limited open reduction and transepiphyseal intramedullary fixation with Kirschner wire (LOR-TIKW) versus open reduction and internal fixation with plate and screw (ORIF-PS) for treatment of irreducible distal radius diaphyseal-metaphyseal junction (DMJ) fracture in older children. Methods: Data of children (aged 10-14 years) treated in our hospital for distal radius DMJ fractures with LOR-TIKW or ORIF-PS from January 2018 to December 2019 were retrospectively analyzed. Follow-up was until radiographic union. Demographic, clinical, and radiographic data; treatment cost; healing time; functional outcome (by Price criteria); complications; and postoperative angulation and displacement were compared between children treated by the two methods. Statistical analysis was performed with alpha set at P < 0.05. Results: A total of 26 children were included: 14 treated with LOR-TIKW and 12 with ORIF-PS. Operation time was less (22.1 min vs. 46.7 min, P < 0.0001), surgical incision smaller (2.43 cm vs. 5.00 cm, P < 0.0001), cost of internal fixation lower (US$, 40.6 vs. 2020, P < 0.0001), and healing time shorter (4.79 weeks vs. 5.64 weeks, P = 0.03) with LOR-TIKW; however, postoperative fracture angulation was slightly larger (1.07° vs. 0.83°, P = 0.85) and displacement slightly more (0.86 mm vs. 0.58 mm, P = 0.44) in the LOR-TIKW group. Rate of union, functional outcome, and complications were not significantly different between the groups. Conclusion: For irreducible DMJ fracture of distal radius in older children, LOR-TIKW appears to be a promising method with several advantages over ORIF-PS.

Closed Reduction Percutaneous Intramedullary Fixation with Kirschner Wires in 4 Children with Displaced Fractures of the Distal Humerus.

BACKGROUND This study describes the use of closed reduction percutaneous intramedullary fixation with Kirschner wires in 4 children with displaced metaphyseal-diaphyseal junction (MDJ) fractures of the distal humerus. MATERIAL AND METHODS Between August 2016 and August 2019, 4 patients (3 boys and 1 girl), whose mean age was 4 years 5 months (range: 3 years 6 months to 5 years 4 months), with displaced MDJ fractures of the distal humerus were treated using closed reduction percutaneous intramedullary fixation with Kirschner wires. Three of the fractures were oblique and 1 was transverse. The operation time and the frequency of intraoperative fluoroscopy were recorded. All children were followed up for greater than 18 months, taking anteroposterior and lateral radiographs of the elbow joint to evaluate the outcomes. At the last follow-up, the Flynn elbow joint function score was used to evaluate clinical outcomes, and complications were recorded. RESULTS The mean operation time was 37.5 min (range: 35-40 min) and the frequency of intraoperative fluoroscopy was 11.7 times (range: 8-15 times). All of the fractures were confirmed to be healed based on radiographic results at 4 weeks after surgery. At the last follow-up, 4 children had normal elbow joint motion without elbow deformity. The Flynn score showed their outcomes were excellent. CONCLUSIONS Closed reduction percutaneous intramedullary fixation using Kirschner wires was an effective treatment for displaced MDJ fractures of the distal humerus in the 4 children described and was shown to be easy to perform with a short operation time.

Integrating deep learning and unbiased automated high-content screening to identify complex disease signatures in human fibroblasts.

Drug discovery for diseases such as Parkinson's disease are impeded by the lack of screenable cellular phenotypes. We present an unbiased phenotypic profiling platform that combines automated cell culture, high-content imaging, Cell Painting, and deep learning. We applied this platform to primary fibroblasts from 91 Parkinson's disease patients and matched healthy controls, creating the largest publicly available Cell Painting image dataset to date at 48 terabytes. We use fixed weights from a convolutional deep neural network trained on ImageNet to generate deep embeddings from each image and train machine learning models to detect morphological disease phenotypes. Our platform's robustness and sensitivity allow the detection of individual-specific variation with high fidelity across batches and plate layouts. Lastly, our models confidently separate LRRK2 and sporadic Parkinson's disease lines from healthy controls (receiver operating characteristic area under curve 0.79 (0.08 standard deviation)), supporting the capacity of this platform for complex disease modeling and drug screening applications.

The comparison of teaching efficiency between virtual reality and traditional education in medical education: a systematic review and meta-analysis.

BACKGROUND: Virtual reality (VR) technology has developed rapidly in recent years and has been applied in many fields, including medical education. A meta-analysis was performed to compare the examination pass rate of medical students educated using VR and those receiving traditional education to evaluate the teaching effect of VR in medical education. METHODS: The PubMed, Springer Link, Science Direct, and Wiley Online Library were searched from inception to May 2020. Articles meeting the inclusion criteria were then evaluated, relevant information extracted and a meta-analysis conducted. Students were allocated to a VR group, those trained using VR technology, and a traditional education group, those who received a traditional medical education. RESULTS: Six studies were included in the meta-analysis. The results indicate a significant difference between the pass rate of students educated using VR and those receiving traditional medical education. The odds ratios and confidence intervals of individual studies and our meta-analysis are illustrated with a forest plot. CONCLUSIONS: Students in the VR group performed better than those in the traditional education group. Teaching with VR may enhance student learning in medical education. Medical schools should consider making greater use of VR when educating students.

Multiresolution Modeling of High-Latitude Ionospheric Electric Field Variability and Impact on Joule Heating Using SuperDARN Data.

The most dynamic electromagnetic coupling between the magnetosphere and ionosphere occurs in the polar upper atmosphere. It is critical to quantify the electromagnetic energy and momentum input associated with this coupling as its impacts on the ionosphere and thermosphere system are global and major, often leading to considerable disturbances in near-Earth space environments. The current general circulation models of the upper atmosphere exhibit systematic biases that can be attributed to an inadequate representation of the Joule heating rate resulting from unaccounted stochastic fluctuations of electric fields associated with the magnetosphere-ionosphere coupling. These biases exist regardless of geomagnetic activity levels. To overcome this limitation, a new multiresolution random field modeling approach is developed, and the efficacy of the approach is demonstrated using Super Dual Auroral Radar Network (SuperDARN) data carefully curated for the study during a largely quiet 4-hour period on February 29, 2012. Regional small-scale electrostatic fields sampled at different resolutions from a probabilistic distribution of electric field variability conditioned on actual SuperDARN LOS observations exhibit considerably more localized fine-scale features in comparison to global large-scale fields modeled using the SuperDARN Assimilative Mapping procedure. The overall hemispherically integrated Joule heating rate is increased by a factor of about 1.5 due to the effect of random regional small-scale electric fields, which is close to the lower end of arbitrarily adjusted Joule heating multiplicative factor of 1.5 and 2.5 typically used in upper atmosphere general circulation models. The study represents an important step toward a data-driven ensemble modeling of magnetosphere-ionosphere-atmosphere coupling processes.

A machine learning approach to define antimalarial drug action from heterogeneous cell-based screens.

Drug resistance threatens the effective prevention and treatment of an ever-increasing range of human infections. This highlights an urgent need for new and improved drugs with novel mechanisms of action to avoid cross-resistance. Current cell-based drug screens are, however, restricted to binary live/dead readouts with no provision for mechanism of action prediction. Machine learning methods are increasingly being used to improve information extraction from imaging data. These methods, however, work poorly with heterogeneous cellular phenotypes and generally require time-consuming human-led training. We have developed a semi-supervised machine learning approach, combining human- and machine-labeled training data from mixed human malaria parasite cultures. Designed for high-throughput and high-resolution screening, our semi-supervised approach is robust to natural parasite morphological heterogeneity and correctly orders parasite developmental stages. Our approach also reproducibly detects and clusters drug-induced morphological outliers by mechanism of action, demonstrating the potential power of machine learning for accelerating cell-based drug discovery.

Learning causal networks using inducible transcription factors and transcriptome-wide time series.

We present IDEA (the Induction Dynamics gene Expression Atlas), a dataset constructed by independently inducing hundreds of transcription factors (TFs) and measuring timecourses of the resulting gene expression responses in budding yeast. Each experiment captures a regulatory cascade connecting a single induced regulator to the genes it causally regulates. We discuss the regulatory cascade of a single TF, Aft1, in detail; however, IDEA contains > 200 TF induction experiments with 20 million individual observations and 100,000 signal-containing dynamic responses. As an application of IDEA, we integrate all timecourses into a whole-cell transcriptional model, which is used to predict and validate multiple new and underappreciated transcriptional regulators. We also find that the magnitudes of coefficients in this model are predictive of genetic interaction profile similarities. In addition to being a resource for exploring regulatory connectivity between TFs and their target genes, our modeling approach shows that combining rapid perturbations of individual genes with genome-scale time-series measurements is an effective strategy for elucidating gene regulatory networks.

Correcting nuisance variation using Wasserstein distance.

Profiling cellular phenotypes from microscopic imaging can provide meaningful biological information resulting from various factors affecting the cells. One motivating application is drug development: morphological cell features can be captured from images, from which similarities between different drug compounds applied at different doses can be quantified. The general approach is to find a function mapping the images to an embedding space of manageable dimensionality whose geometry captures relevant features of the input images. An important known issue for such methods is separating relevant biological signal from nuisance variation. For example, the embedding vectors tend to be more correlated for cells that were cultured and imaged during the same week than for those from different weeks, despite having identical drug compounds applied in both cases. In this case, the particular batch in which a set of experiments were conducted constitutes the domain of the data; an ideal set of image embeddings should contain only the relevant biological information (e.g., drug effects). We develop a general framework for adjusting the image embeddings in order to "forget" domain-specific information while preserving relevant biological information. To achieve this, we minimize a loss function based on distances between marginal distributions (such as the Wasserstein distance) of embeddings across domains for each replicated treatment. For the dataset we present results with, the only replicated treatment happens to be the negative control treatment, for which we do not expect any treatment-induced cell morphology changes. We find that for our transformed embeddings (i) the underlying geometric structure is not only preserved but the embeddings also carry improved biological signal; and (ii) less domain-specific information is present.