General considerations on study design

The correct collection and storage of samples are of primary importance and should be planned before commencement of the trial, with a goal of running a few highly pointed assays or analyses first—prior to storing remaining samples for later analysis informed by follow-on questions, patient outcomes, and/or clinical data.

The disease, specimen type, and number of specimens that can be obtained—as well as the immunotherapies or other drugs onboard with FUS—will guide the analyses and their timing. The analyses and their timing will have to be optimized for each study as a drug may act only on one cell subset, for example, and its delivery and/or pharmacokinetics will influence the necessary type and collection time point(s) for samples of interest.

There are still outstanding biological unknowns underscoring FUS-mediated enhancement of the antitumor immune response. These unknowns limit our ability to design analyses based purely on efficacy wherein survival (e.g.progression-free survival, overall survival) is the primary endpoint. We therefore strongly suggest placing an emphasis on the inclusion of biological primary endpoints and designing trials to collect biological specimens that can retrospectively generate insights into these unknowns. Through this approach, study design can then be guided by fundamental biological questions that need to be addressed to further develop the promise of FUS technology and its combinatorial alliance. Among the most pressing issues are assessment of whether FUS treatment is associated with changes in the composition of the tumorous immune infiltrate (i.e. phenotype and function of innate and adaptive immune cells), the spatial distribution of immune cells, the spectrum of T-cell activation and exhaustion, and T-cell clonal expansion.

We would like to stress the importance of doing window-of-opportunity trials, with an emphasis on the role of neoadjuvant immuno-oncology approaches. Such trial designs capitalize on the intervening time leading up to a planned resection, offering the opportunity to investigate novel interventions in a manner that does not impede normal resection time for the patient. The neoadjuvant approach provides opportunities to glean meaningful immunological insights in the absence of further disruption to the in situ microenvironment, while post-treatment tumor resection offers a prime opportunity to directly interrogate tumor biology as well as understand concordance between blood and tissue signatures where liquid biopsy is involved.

Finally, it is also imperative to document FUS treatment parameters. For details on appropriate reporting of acoustic exposures, we refer the reader to the Guidelines on Treatment Reporting.28

Tissue samples

Tumor biopsies (incisional or core needle) should be collected with as many samples as possible at each time point—with at least three to five biopsies being highly desired. The biopsy samples should be processed into formalin-fixed, paraffin-embedded (FFPE) blocks to allow stable and durable conservation of the specimens for later analysis as needed. Options for downstream analysis include but are not limited to H&E, multispectral (multiplexed) immunohistochemistry (IHC), and transcriptomics. Additional tissues should be stored to best preserve quality for RNA sequencing—that is, immersion in TRIzol or flash-freezing. Otherwise, if phenotypic analyses or functional assays on immune cells are envisioned (using flow or mass cytometry), biopsies should be appropriately disaggregated and preserved in liquid nitrogen (otherwise, −80°C) for future interrogation.

Blood samples

For blood samples collected in EDTA, plasma and peripheral blood mononuclear cells (PBMCs; separated with Ficoll gradient) isolated from whole blood should be preserved at −80°C for future interrogation. Because polymorphonuclear cells have reduced recovery after freezing, it is recommended that a fraction of any such samples be analyzed fresh when this approach is accessible.

Timing of collection

Baseline sampling is required, and when possible, on-treatment samples should be collected as well. Particularly in the context of ablative procedures with spatially complex bioeffects, defining the degree of intact versus treated tissue is not a simple task and requires generous sampling of multiple regions. As such, for biopsies, efforts should be taken to obtain samples in the directly treated region, in the region(s) reflecting the peri-treatment zone, and from a region that has not been directly targeted with FUS.

The ability to get the on-trial tissue biopsy will be dependent on (1) safety for the patient, (2) disease location (e.g. central nervous system vs extracranial), (3) patient consent, and (4) clinical feasibility.

Because clinical feasibility will often be the limiting factor to the number of biopsies, we suggest a minimum of two biopsies: baseline and post-treatment. If progression occurs, having an additional biopsy will help assess changes in disease that led to its advancement.

For optimized timing of sample collection, a few points can be considered. For tissue samples, a baseline or pretreatment sample is necessary. However, archival tissue may be used if obtained proximal to time of treatment, typically within 6 weeks of treatment initiation, and without the confounding influence of intervening therapies. Post-treatment, acute time points are expected to bear utility for gaging the innate inflammatory response elicited by FUS. Based on previous FUS studies, sampling 1–7 days post-treatment should allow assessment of the acute inflammatory response. Such assessments would be well-served by physical measurements (e.g. temperature or cavitation recordings) and correlative histology reflecting the degree of tissue destruction or coagulation. When FUS is applied in monotherapy format, a biopsy 2–3 weeks after treatment is recommended to assess changes in the local innate and adaptative immune sequelae, with more chronic time points expected to better capture elaboration of the adaptive immune response. For longitudinal trials or those that seek to compare across locally ablative modalities, serial sampling needs may differ. For example, the majority of clinical studies evaluating radiotherapy have assessed the on-treatment effects of treatment after two cycles of study intervention (Cycle 2) and have generally aimed to biopsy between day 2 and 10. Subsequent monthly sampling would then be ideal to surveil the evolution of the response. Where acute repeat biopsy is not possible, sampling within a few months (3–4 months) after the treatment will be informative—especially to assess changes associated with disease progression.

Similar recommendations apply to blood samples, wherein sampling should include a pretreatment time point for baseline standardization. It is advisable to obtain a sample at screening and before treatment on the first day of drug administration (Cycle 1, day 1 pre-infusion for combination trials) or FUS treatment. Acute sampling (e.g. 24 hours post-treatment) where possible will allow for assessment of the short-term inflammatory response. Where there is a desire to capture elements of the response reflected within cell-free/circulating tumor DNA, it will likely be necessary to sample even more acutely (i.e. immediately post-FUS). Sampling at 2–3 weeks post-treatment (when an expanding immune response might be expected) and subsequently performing weekly or monthly sampling will enable ample temporal resolution to facilitate monitoring of immune response evolution via PBMCs, cytokines, etc. Of note, where disease progression is observed, a blood sample at this time would also be of interest for contextualizing whether loss of disease control is associated with significant alterations in the circulating immune cell population.

Blood draws and biopsy timing will also be informed by FUS and/or drug dosing, treatment schedule, and pharmacokinetics. If the treatment consists of a combination of FUS and immunotherapies, it may be sensible to coincide the timing of blood sampling with immunotherapy infusion sessions—the frequency of which will depend on the study protocol.

Finally, selection of time points will be influenced by drug choice and associated pharmacokinetics (PK), as the time to objective response to immunotherapy—or the peak concentration of the drug in the blood or the target tumor—depends on the drug and tumor type. After the incitement of evident response to immunotherapy, or after max PK has been reached, it should be possible to schedule longer intervals between blood draws.

Summary of recommendations for sample collection and storage

The timing for collections and analyses are summarized in the flowchart below (figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Recommended timing of analyses (clinical) for tissue and blood samples. cfDNA: cell-free DNA; CTCs: circulating tumor cells; DNA: cell-free DNA; CTCs: circulating tumor cells; FUS, focused ultrasound; IHC: immunohistochemistry; PBMCs: peripheral blood mononuclear cells; RNAseq: RNA sequencing; WES: whole exome sequencing.

• Tissue samples should be collected pre-FUS (baseline) and within 2–3 weeks post-treatment to assess immunomodulation. Continued monthly sampling would then be ideal to monitor the evolution of response. If monthly sampling is not possible, sample acquisition within a few months (3–4 months) after treatment is desired.

• When possible, efforts should be taken to obtain baseline and on-treatment biopsies from the same lesion as well as from a separate observational lesion that has not been treated with FUS in order to interrogate abscopal effects.

• Blood samples should be collected at baseline (at screening) and before treatment (on the first day of treatment). An early time point (day 1, if possible) will offer insight into the acute inflammatory response to FUS. Meanwhile, collection closer to 2–3 weeks is expected to inform whether expansion or activation of a cellular immune response is taking place.

• Where clinically feasible to obtain, tissue and/or blood specimens acquired at the time of disease progression will be of interest for assessing immunological changes underscoring disease advancement.

• FUS and/or drug dosing, treatment scheduling, and resulting PK may necessitate adjustment of the aforementioned collection time points.

Staining of tissue specimens

One of the leading hypotheses to explain the localized immunogenic effects of FUS is that the damage induced by thermal or mechanical ablation with FUS leads to a remodeling of the spatial relationship between immune cells within the treated TME, promoting their colocalization in the peri-ablative zone. Thus, there is a critical need to define and locate immune cell subsets within and surrounding ablated regions, and this underpins the importance of staining techniques that preserve spatial information. While traditional and multispectral IHC have been mainstays for spatial profiling of tissues, the advent of novel spatial multiomics technologies is expected to provide an unprecedented wealth of high-dimensional information that will allow us to address FUS-related hypotheses with an even higher degree of specificity.

Recommendations for tissue staining

• Tissue samples should be stained to assess frequency and localization of T cell and myeloid cell subsets—including CD4+T cells, CD8+T cells, FOXP3+regulatory T cells, monocytes/macrophages [and their expression of major histocompatibility complex (MHC)-I, programmed cell death protein 1 (PD-1) and programmed death-ligand 1 (PD-L1)]. Standard IHC can provide preliminary information about many of these markers and be further supplemented by multicolor IHC. If available, multispectral immunofluorescence (e.g. Vectra) is recommended due to its ability to simultaneously interrogate multiple cell surface markers as well as those within the cytoplasm and nucleus of the same cell (in contrast to multicolor IHC, which can be used when markers are on different cells). In the absence of FFPE specimens that lend to these approaches, IHC on frozen sections can serve as an alternative.

• Where possible, samples should be tested for T-cell activation (granzyme B, interferon-gamma, Ki67), B-cell prevalence (CD20), and for cell devitalization such as necrosis or apoptosis (cleaved caspase-3 or TUNEL staining).

• Because immune modulation following FUS is thought to involve expression of danger signals, samples should be stained for damage associated molecular patterns as initial evidence for the elicitation of immunogenic cell death. Where thermal FUS is used, heat shock proteins are a relevant target to include in this evaluation as well.

• Finally, when known or suspected, samples should be stained for tumor or subtype-specific markers (e.g. panCK, HER2).

Phenotypic and functional analyses of immune cells

There are still open questions regarding the phenotypic and functional characteristics of the cellular immune response to FUS. The markers and panels used for phenotypic characterization should be guided by the biological primary endpoints in the study design as well as the expected effects of FUS. For example, does FUS alter dendritic cell (DC) presence, location and activation status? T-cell differentiation or proliferation? T-cell function, activation, or exhaustion? Immunosuppressive myeloid cell prevalence (e.g. granulocytes)? The recommendations below have been collated to serve as guidelines for general approaches toward the more specific methods of characterization necessary to answer the aforementioned questions.

It is expected that both acute and durable changes in cytokine production will be triggered in response to tumor disruption or destruction by FUS. Cytokine assessment should be tailored for these two cases as it is likely that these short-term and long-term responses will be different.

It will be important to identify relevant time points to perform these analyses. Some acute cytokine production might be expected in response to FUS, in addition to some durable changes due to tumor ablation, and these two categories would likely involve different cytokines.

Summary of recommendations for phenotypic and functional analyses of immune cells in tissue and circulation

• Effectors including CD4+ and CD8+ T cells, natural killer cells.

• Antigen presenting cells (APCs) including DCs and macrophages.

• Suppressive immune cells: monocytes and granulocytes (and myeloid-derived suppressor cell, or MDSC, subsets), M2 macrophages; regulatory T cells.

• B cells.

• Immune activation and memory phenotype within T-cell subsets: naive, central memory and effector memory (e.g. CD44; CD62L; CD38, T-bet, Tox, TCF7).

• Gamma-delta T cells.

• Expression of immune checkpoint receptors and ligands (e.g. PD-1, PD-L1; Tim3, Gal9, Lag3, MHC-II, CD39).

We recommend focusing functional assays for PBMC on T-cell cytokines and cytolytic proteins, as well as on DCs, after stimulation. To guide the design and deployment of these analyses, we refer readers to two comprehensive authority publications offering detailed guidelines on design of flow cytometry panel design—and considerations therein—for phenotypic and functional subsetting of human and murine cells.29 30

Gene expression analysis

Due to our expanding capacity to interrogate gene expression within both fresh and archival tissues, transcriptional and whole exome sequencing are expected to provide invaluable insights into the nature of the global tissue response to FUS. In one respect, whole genome analysis may provide guidance as to the presence of mutations that lead to responsiveness or resistance to the effects of FUS—whether in monotherapy or combination format. This is in alignment with clinical practice as common mutations are generally screened as part of patient care. In another respect, transcriptional analysis of relative changes between baseline and post-treatment biopsies can provide insights as to how elements of the TME—including the cancer cells and stromal constituents such as immune cells and fibroblasts—respond to and adapt following treatment.

Analysis of the transcriptome, when paired with Gene Set Enrichment Analysis or cell type deconvolution algorithms such as CIBERSORT(x), can provide insight on how best to deploy more specific protein-level assays at the level of flow cytometry or IHC. It is worth emphasizing that RNA sequencing analyses require intensive quality control procedures, especially as ablation approaches may be anticipated to impact RNA quality; as such, samples should be aggregated in a sequencing run. RNA sequencing is recommended for exploratory analyses or hypothesis generation, given that it enables broad, unbiased evaluation of tissues. While it is known that transcript-level and protein-level data do not always correlate directly, broad gene expression profiling readily lends to informative analysis of pathways enriched for inflammatory genes, tumor immune signatures and immune checkpoint genes that can then be further interrogated by complementary methods. If tumor tissue samples are limited in RNA amount and/or of poor RNA quality, other approaches (e.g. NanoString or other hybridization approaches) for high-dimensional assessment of immune-related gene signatures are available and considered to be robust to such limitations.

T cell receptor (TCR) clonality and diversity will be expanded following FUS treatment, especially when administered in combination with immunotherapies known to drive such effects. RNA-sequencing of tumor tissues in parallel with PBMCs are invaluable for testing this hypothesis. As such, we suggest comparing TCR repertoires across tumor and blood samples across pretreatment and post-treatment time points where possible.

Recent published data suggest that FUS potentiates improved release of cell-free/circulating tumor DNA into the blood following disruption of the blood brain barrier in glioblastomas.31 Analysis of tumor-derived DNA present in plasma samples is an emerging field of study underscored by the burgeoning role of liquid biopsy in cancer diagnosis and surveillance. For tumor types with high tumor mutational burden or well-characterized tumor mutations, isolating and banking the plasma for DNA analysis (e.g. in Streck blood collection tubes) should be considered. Because field-wide consensus for standardization approaches remains in development for liquid biopsy, we suggest adhering to the recommendations offered by the leaders of the BloodPAC Consortium; their Minimum Technical Data Element provides published recommendations on critical pre-analytic attributes and a detailed workflow for blood specimen isolation, processing, and analysis2. In the context of CNS tumors specifically, additional considerations can be made on the use of cerebrospinal fluid (CSF) for liquid biopsy. CSF is accessible through lumbar punctures—which are commonly deployed in gliomas, much as they are in Alzheimer disease or multiple sclerosis. While we postulate that CSF-based liquid biopsy may offer utility in monitoring of the brain tumor-immune landscape following FUS, the relationship between tumor, blood, and CSF analyte profiles is yet unknown in this context. Thus, we suggest incorporating and aligning these readouts into future clinical studies where possible.

Summary of recommendations for genetic analyses

• When possible, leverage RNA-sequencing and DNA-sequencing to compare tumor mutational burden and TCR clonality/diversity before and after FUS.

• When FUS blood brain barrier (BBB) opening is performed, collect plasma and/or CSF for cell free/circulating tumor DNA analyses.

Circulating tumor cells

There are no reports to date linking FUS treatment to variations in circulating tumor cells. We therefore suggest that analyses of circulating tumor cells be performed as they would be on a standard clinical basis.

Summary of recommended assays

Recommended routes of analysis are summarized in the flowcharts below (figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Recommended routes of analyses (clinical) for tissues and blood samples. ctDNA: circulating tumor DNA; FFPE: formalin-fixed, paraffin-embedded; IHC: immunohistochemistry; PBMC: peripheral blood mononuclear cell; RNAseq: RNA sequencing; TCR: T cell receptor.

The recommended assays for tumor biopsy or resection are:

1. Tumor histology: H&E+multispectral IHC.

2. RNA sequencing or NanoString analysis of tumor biopsies or resections before (if applicable) and after treatment.

3. Flow cytometry analysis (or Cytometry by Time of Flight (CyTOF, e.g. mass cytometry), when possible) of immune cell phenotype and function in tissues (where sufficient material is available).

4. TCR sequencing and analysis of TCR clonality.

5. DNA (whole exome sequencing or other) analysis of tumor tissues.

The recommended assays for blood samples are:

1. Flow cytometry analysis (or CyTOF, when possible) of PBMC phenotype and function.

2. Protein assessment of serum sample (e.g. cytokines/chemokines).

3. TCR sequencing and analysis of TCR clonality in PBMCs.

4. If tumor antigens are known, PBMCs can be assessed for reactivity by ELISpot.

5. Quantification and sequencing of cell free/circulating tumor DNA.

6. Quantification and profiling of circulating tumor cells.

General considerations on study design

The recommendations for preclinical studies follow the recommendations for clinical trials—with the general rationale similarly being to provide information that will support the development of optimized FUS treatment parameters while further increasing the likelihood of therapeutic success through allied immunostimulatory drug combinations.

However, the preclinical nature of animal studies brings about two major differences. First, the possibility exists for more detailed, mechanistic analyses. Flow or mass cytometry are critical tools for deep profiling of the cells comprising the tumor-immune landscape. With modern advancements in these multiparametric techniques, many interesting biological questions can be addressed regarding the impact of FUS treatment on changes in myeloid population (and their corresponding expression of cytokines), DC recruitment and activation, the nature of cell death in the TME, and the influence of FUS on the lymphatics and vasculature. Additionally, while multimodality in vivo imaging (e.g. ultrasound, fluorescence, bioluminescence, MRI, CT, positron emission tomography; PET) is outside the scope of these recommendations, we note briefly that it can serve as a powerful complementary tool for spatiotemporally monitoring tumor progression, regression, and recurrence.

Model selection is of paramount importance for preclinical studies and should be carefully considered. First, it is important to understand the degree to which the tumor model is infiltrated by immune cells prior to FUS; understanding the basal immunogenicity of the tumor and its response to immunotherapy is a critically important step. In complementarity to genetically engineered mouse models, implantable tumor models provide an opportunity for consistency between preclinical studies. It is critical that tumor cells used in these models are authenticated and confirmed to be free of Mycoplasma. In particular, many tumor lines have been sequenced for mutations, and MHC class I and class II neoantigen epitopes have been defined. The availability of this information enables a greater degree of acuity in assessing T-cell responses and can limit the need for tumors designed to express model antigens. Where transplantable models are limited in how well they recapitulate clinical pathology, genetically engineered mouse models yielding spontaneous tumors are also a powerful asset in enabling clinically relevant preclinical FUS immunotherapy investigations.

In order to enhance the rigor and reproducibility of preclinical results and conclusions, attempts should be made to understand the difference in responses through the lens of three important factors: (1) different types of tumors (i.e. how consistent the observed response is according to tumor genetics and tissue of origin); (2) different mouse genetic backgrounds; and (3) different sex and age of animals.

The second point concerns study design considerations. As noted above, questions remain about the exact biological mechanisms responsible for FUS-mediated enhancement of an antitumor immune response. When designing studies, we therefore suggest first performing a study with overall survival (OS) as the primary endpoint. If OS is improved, then validate the involvement of the immune system in the response through assessment of the systemic immune response by evaluating effects on untreated contralateral tumors (e.g. through bilateral tumor implantation models) or distal lesion involvement in spontaneously metastatic models. This approach can be coupled with targeted immune cell depletion studies (e.g. CD8+ or CD4+ T-cell depletion) or use of transgenic mouse models (e.g. RAG-1^–/– mice). Selected assays targeting a role for adaptive immunity effects can assess CD8+T cells, CD4+T cells, regulatory T cells (Tregs), and (where deemed appropriate) TCR clonality. In many mouse models, tumor-specific neoantigens have been characterized and can be used to trace the tumor-specific immune response using MHC-multimers, ELISpot, and intracellular cytokine staining assays. We also recommend storing excised tissues, either as single cell suspensions or in FFPE blocks, to allow for post hoc hypothesis-driven analyses.

An alternative approach is to perform systematic analyses of the response to FUS treatment—alone or in combination with immunotherapies—to identify changes induced by FUS treatment and subsequently tune FUS exposure conditions to increase (or decrease) these changes. While this approach has merit and can be informative, its implementation will be considerably limited by the large parameter space associated with FUS treatment paradigms. Nonetheless, there is considerable utility in understanding whether alterations of a single FUS parameter have cascading effects on innate and adaptive immune responses. FUS treatment spans a wide range of mechanisms of action that can rely on thermal mechanisms (e.g. thermal ablation or hyperthermia) or on mechanical mechanisms (e.g. blood brain barrier opening or histotripsy), to name the most used. Within each modality, many treatment parameters can be adjusted (e.g. ultrasound frequency, power, treatment duration, number and spatial distribution of treatment points, and the pulse sequence at each treatment point). This large parameter space can be modulated to produce ablative or subablative treatments over a specified fraction of the tumor, either in a single treatment or in multiple treatments. When combined with immunotherapies, additional specifications regarding dosage and sequence must be considered as well. Due to this broad treatment parameter space, we suggest applying these approaches on first intent for treatment modalities where the parameters have already been considerably reduced (e.g. for BBB opening or thermal ablation) or otherwise following the approach described earlier. This is summarized in the flowchart below (figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Experimental designs for preclinical studies. IHC, immunohistochemistry; OS, overall survival.

Most preclinical animal studies testing immunostimulatory effects of FUS for cancer treatment have focused on rodent models. These models have been critical for moving the field forward and will continue to be essential for providing mechanistic insight. There are, however, limitations in the translational value and in the scale and anatomical relevance of these models. To address these limitations, comparative oncology trials leveraging veterinary models have emerged as a powerful complement for more accurate studies of cancer, reflecting the disease in a large animal model bearing naturally-occurring spontaneous tumors. As an example, a review of histotripsy ablation in spontaneous tumors in veterinary patients was recently published.32 Combining rodent models and veterinary patients should provide realistic avenues for developing and testing combined treatments of FUS and immunotherapies more comprehensively, which can help improve the development of these treatments for both animals and humans. We therefore recommend initiating these veterinary studies where complementary to laboratory animal studies.

Finally, while outside of the scope of these guidelines to review in detail, it is also imperative to document FUS treatment parameters in a systematic and standardized manner. See FUSF Guidelines on Treatment Reporting3 or their published version.28

Data sharing

With recent policies from federal funding agencies mandating that supporting data from publications be made publicly accessible, planning and budgeting for data management and sharing must be incorporated more thoughtfully into study designs going forward. This will involve:

• Identifying what data will be shared (type, quantity, need for security compliance for repositories storing human data).

• Identifying appropriate methods/approaches and repositories for managing and sharing the data.

• Developing a Data Sharing and Management (DMS) plan for managing and sharing scientific data.

• Budgeting the funds needed for data management and sharing activities.

Many data repositories are available publicly and the most appropriate choice of repository will depend on data type and discipline. The National Institutes of Health (NIH), for example, offers a listing of NIH-supported repositories. Other available tools, such as re3data.org, provide a registry of research data repositories that can be used to identify the most appropriate repository.

Bioinformatics and biostatistics

To support the robust and rigorous interrogation of treatments combining FUS and immunotherapies, several specific bioinformatics resources are needed. These include:

• Multiomics data integration tools for effective integration of genomic, transcriptomic, and proteomic data. This will be essential for understanding the complex interactions unveiled by multiomics approaches.

• Data visualization tools to visualize high-dimensional data (e.g. network maps, pathway analyses, and interactive plots).

• Predictive modeling and artificial intelligence-driven approaches, to fuse multimodal data, predict patient responses to combination therapies, and stratify groups that are most likely to benefit from specific treatment combinations—thereby facilitating individualized treatment decisions.

• Data sharing and collaborative platforms: Platforms that allow for easy data sharing, such as cloud-based storage, and can support bioinformatics analyses.

In addition to bioinformatics resources, specific biostatistical support is also needed across preclinical and clinical studies to develop statistical methodology that can integrate data from multiple sources (e.g. clinical, genomic, and imaging data) to identify subgroups that are most likely to respond to combination therapies and that can account for the vast heterogeneity in treatment responses across animal models and human subjects. A critical and underappreciated component of this support pillar is the development of approaches to accommodate data dropout, which is a larger liability in the relatively small Phase 0/1 trials in which FUS is presently being investigated. Data dropout refers to instances where expected data are not obtained or cannot be used for analysis, owing to a variety of reasons falling broadly into two categories: technical issues and issues related to participant compliance or trial conformity. Technical issues involve problems that might occur during data collection, preservation, or analysis, such as degradation of a biological specimen or equipment failure during data collection that force sample exclusion from analysis. Participant compliance issues involve situations where a participant does not comply with the study protocol, such as missed clinic visits, treatments, or tests leading to missing data points; the participant choosing to leave the trial; or the need for a participant to be removed due to an adverse event or unexpected health issue.

In animal studies, similar principles apply. For example, a mouse might need to be removed from a study due to an adverse event such as unexpected reaction to an intervention or an unforeseen health issue, resulting in missing data points. Data dropout can significantly impact the analysis and interpretation of preclinical and clinical trials, as it could introduce bias or reduce the power of the study. As such, it is a critical aspect to proactively consider in trial design and data analysis efforts. We expect that future FUS clinical trials interrogating combinatorial approaches will make use of adaptive trial designs that can adjust treatment protocols based on interim analysis of treatment outcomes.