The utilization of low-depth shotgun metagenomics provides a highly cost-efficient mechanism for the high-frequency, longitudinal monitoring of patient colonization dynamics (31, 32). Unlike resource-intensive deep sequencing, this scalable approach allows for affordable and replicable profiling, e.g., of the gut microbiome throughout treatment of cancer patients (33). However, the risk that low-depth metagenomics may limit the ability to perform transmission analyses should be taken into account. Thus, a balance is required at each context, between the need to lower costs and the introducing a tradeoff in which AMR detection may not outperform existing diagnostic or surveillance methods. The current experiences with shallow shotgun sequencing (in clinical and mock microbiome samples) demonstrate that a depth of 0.5 million sequences per sample was sufficient to produce a quality similar level of species and functional profiles to that of deep WGS sequencing such as understanding which species and functions are present and identifying biomarkers related to outcomes (31, 32). This surveillance is sufficiently sensitive to capture the subtle expansion of AMR genes. By identifying the proliferation of resistant clones or specific mutations prior to clinical manifestation, or at specific timepoints of the patient journey, such as near the time of diagnosis, this strategy allows clinicians to adjust prophylactic regimens or initiate preemptive stewardship measures before life-threatening infections occur (34). Costs can be reduced further by a shift to more conventional sequencing approaches when NGS is not required. Additionally, in the case of pathogens with high viral load samples where sensitivity rates are low, sample pooling strategies with validated deconvolution can also enhance efficiency.

Effective routine surveillance also requires the establishment of a local genomic database by clinical microbiology laboratories, complemented by environmental samples from cancer patient wards. Such a database enables high-resolution phylogenetic analysis, which is essential for distinguishing persistent institutional clones from imported strains (35). Currently, environmental sampling, e.g. for high-touch surfaces, is typically recommended only in specific outbreak contexts, where results directly inform targeted interventions. However, shallow-depth shotgun metagenomics can also be leveraged in highly targeted environmental sampling guided interventions, in the latter case able to unveil ecological niches of microbes and antibiotic resistance genes (36). Consequently, such an approach can guide evidence-based outbreak responses, allowing infection control teams to implement targeted containment strategies such as specific environmental decontamination, or staff screening (37).

The maximization of pathogen genomics utility is contingent upon the existence of a functional laboratory infrastructure and coordinated surveillance networks. In settings with severe resource limitations, the immediate adoption of genomic technologies is often unfeasible. For existing laboratory infrastructures, experts from the WHO and other institutions recently released a tool to assist with commodity forecasting and costing for pathogen genomics. Such a tool can be used for national budgeting purposes (38). As hospitals operate within fixed budgets and often see no direct financial or operational return on such investments, the cost of WGS/NGS can be embedded into national health budgets as part of broader policy-driven disease control and preparedness programs. However, the affordable cost threshold depends on context and the ability of countries' Ministries of Health to advocate for domestic funding and enhanced routine surveillance. A list of priority pathogens has been developed by WHO to enhance outbreak preparedness by advocating for investments in research, development, and innovation in medical countermeasures. To this end, the prioritized pathogens included under routine NGS/WGS surveillance to yield measurable benefit for cancer patients could be a subset from this priority list, providing a global comparable reference point.

The ability to shift from lower to increased diagnostic resolution in a stepwise fashion and within a clinical environment, predisposes that all of these options would be running in parallel. Thus, the flexibility provided by a tiered diagnostic testing can be more readily achieved in larger, tertiary clinical units, where the available infrastructure, consumables and expertise are most likely to be present (39). In terms of cancer care, several cancers are treatable only within such units, and therefore the additional tiered approach would be complementary to existing laboratory functions. An example of such a tiered approach would be to first use low-cost rapid tests (e.g., antigen tests, CRP/PCT for triage), followed by targeted PCR for likely pathogens, and finally NGS sequencing for complex or unresolved cases. The NGS data can be compared to the sequenced environmental samples derived from the same hospital to provide a more comprehensive interpretation of results. Comparing NGS data from clinical samples with environmental surveillance data enables differentiation between endogenous infection and exogenous acquisition and confirmation of indirect transmission pathways. This has been shown in the case of identifying sources of environmental contamination (40), as well as in the reconstruction of the potential outbreak transmission route of drug-resistant pathogens in a hospital environment (37) and within acceptable operational costs (41).

However, to achieve data interoperability, a centralized, standardized genomic analysis pipeline should be implemented across all participating sampling locations within the same hospital. This necessitates either the use of existing, or the establishment of harmonized standard operating procedures (SOPs) covering the entire workflow from DNA extraction, library preparation to minimize technical batch effects and experimental bias (42). Crucially, on the computational side, the network should also adopt containerized bioinformatic workflows to support the use of identical software versions and parameter settings across the same functions of different hospital departments (43). In parallel, adopting FAIR (Findable, Accessible, Interoperable, Reusable) data principles would further strengthen interoperability across departments by ensuring that genomic datasets are consistently organized, documented, and easily integrated into shared analytical frameworks (44). By aligning sampling, laboratory and bioinformatics standards, such an approach will guarantee rigorous quality control. This unified laboratory framework ensures that resistance profiles and transmission clusters identified in one patient and/or ward are directly comparable to those in another, facilitating accurate local AMR surveillance.

The genomic data streams from cancer patients should be consolidated into a centralized, secure data platform that integrates sequencing outputs with clinical and epidemiological metadata, such as antibiotic exposure history, immune status, and treatment outcomes (45). This integration can power the deployment of near real-time decision-support dashboards, which visualize transmission networks and, when sufficient data is available, leverage machine-learning methods to generate automated early-warning alerts for high-risk clones or anomalous resistance profiles (46), ultimately transforming raw genomic data into actionable intelligence that guides infection control interventions and therapeutic adjustments. However, such flexible bioinformatics infrastructures capable of integrating genomic data with electronic clinical records are difficult to implement in general, and even more so in middle-income settings. Successful models of twinning between high-income and resource-restricted facilities, have shown that technology transfer, local adaptation and implementation of surveillance and bioinformatics capacities are feasible (47). While these models are indeed successful, they are slow and limited by funding support, as well as the numbers and availabilities of willing institutions and experts.

Data privacy and biosecurity are increasingly rigorously regulated globally, which means that any suggested data flow and integration need to be aligned with existing regulatory requirements (48, 49). While such structures exist for infectious disease control units, it may need to be extended for cancer units too—or for the latter to update their data handling to include a clear mention for the use of infectious disease data from cancer patients. Access to this high-resolution data is typically governed by institutional review boards, granting full access to authorized clinical and public health personnel while providing anonymized, aggregated data for broader research (50). However, in the case of cancer patients, a dedicated interdisciplinary oversight would be required, bridging the specialties of oncology, epidemiology and infectious diseases—leveraging the experiences in which antimicrobial stewardship has been integrated with surveillance in several clinical settings (51, 52). This consideration for continuous data flow and integration is warranted both as cancer patients represent a susceptible population, as well as often serve as long-term reservoirs for multidrug resistant organisms due to prolonged hospitalization and heavy antibiotic selection pressure (53).