Sarcomas are rare and heterogeneous malignancies of connective tissue, accounting for roughly 1% of adult cancers and approximately 15% of pediatric cancers. The rarity and biological complexity of the disease, spanning over 100 histological subtypes, has long made it a candidate for centralized, expert-driven care. Over recent decades, the field has transformed dramatically: limb salvage surgery now replaces amputation in more than 95% of extremity sarcoma cases, driven by advances in surgical technique combined with multimodal treatment strategies including chemotherapy, radiation, and targeted agents. This paradigm shift reflects a broader commitment to preserving quality of life while aggressively managing disease.
The centralization imperative: Specialized sarcoma centers have emerged as the cornerstone of modern management. These centers feature weekly multidisciplinary team (MDT) meetings and draw together pathologists, radiologists, sarcoma surgeons, radiation oncologists, and medical oncologists to collaboratively assess each case. Evidence from multiple national studies consistently shows that treatment at a specialized sarcoma center improves outcomes. Critically, centralization appears most important not just for surgery but for the entire diagnostic pathway, where biopsy errors and misdiagnosis at non-specialized settings can permanently compromise subsequent management.
The accessibility gap: Despite strong evidence for specialist care, approximately 50% of patients with sarcoma continue to receive treatment outside specialized centers. Geographic barriers requiring long-distance travel, specialist shortages from an aging workforce, financial pressures, and healthcare models that prioritize patient volume over quality all contribute to this gap. The result is unequal access to guideline-adherent care that directly affects patient outcomes. A purely centralized model, while ideal for clinical quality, has failed to scale to meet global demand.
This paper, published in ESMO Open in 2024 by Fuchs and Gronchi, proposes the Sarcoma Hub and Spoke Model (HASM) network as a structural solution: a framework that centralizes complex decision-making and expertise at hub centers while enabling peripheral spoke facilities to handle routine diagnostics and treatments, connected through digital infrastructure and shared data systems. The proposal integrates digital interoperable platforms, a Sarcoma Care Data Warehouse, and emerging predictive technologies including Sarcoma Digital Twins and causal machine learning (CML) to advance precision care at scale.
The authors describe a sarcoma center as more than a facility with high case volume and regular MDT meetings. Effective centers must demonstrate the capacity to manage complex cases, maintain specialized training programs, provide access to clinical trials, and sustain comprehensive prospective databases. The absence of formal specialty training pathways for sarcoma clinicians means that case volume alone cannot compensate for individual specialist inexperience. Moreover, the relationship between volume and outcome is not linear and depends on institutional resources, staffing, and support structures.
Diagnostic Coordination Hub: The first functional component of a sarcoma center is the Diagnostic Coordination Hub, where the radiologist, pathologist, and sarcoma surgeon jointly review referral consultations. This hub determines whether biopsy is indicated, what type of biopsy should be performed, and whether direct referral for treatment is appropriate. By centralizing this diagnostic expertise, the center ensures that all patients receive a unified, considered assessment rather than fragmented specialty opinions. This is particularly critical at the biopsy stage, where errors including inadequate sampling, inappropriate technique, or contamination of surgical planes can compromise limb salvage and overall treatment feasibility.
Therapeutic Integration Hub: The second component coordinates the actual execution of treatment plans. It manages and integrates surgery, chemotherapy, and radiotherapy, closely monitors treatment response, and adapts plans based on real-time evaluations. This hub ensures that each modality is delivered at the appropriate time and sequence for the individual patient's case, rather than in siloed departmental workflows. The hub also serves as a training ground for future sarcoma specialists and as a center for research and protocol development, disseminating advanced practices across the broader network.
The authors note that criteria for sarcoma center designation are inconsistent across countries and that minimum requirements such as MDT meetings and case volume thresholds do not guarantee quality. They argue that the true measure of a sarcoma center lies in its capacity to execute complex multidisciplinary pathways across both diagnostic and therapeutic functions, not in administrative metrics alone.
The HASM framework rests on a carefully structured division of care responsibilities between hub centers and peripheral spoke facilities. The authors emphasize that decentralization is not a binary choice but a spectrum that must be calibrated to the complexity of each clinical scenario, the capabilities of local facilities, and the educational mission of the broader network. Critically, even smaller sarcoma centers may benefit from performing lower-complexity cases as training opportunities for surgeons in development, while high-volume centers may prefer to refer these cases outward to focus resources on the most complex presentations.
Pathology: Straightforward cases where clinical, imaging, and pathological findings are concordant can be processed at local centers. A representative example is dedifferentiated liposarcoma of the retroperitoneum with clear concordance across all parameters. By contrast, diagnostically challenging entities such as extraskeletal myxoid chondrosarcoma require centralized pathological review to ensure accurate subtyping and appropriate management. The authors underscore that incorrect or imprecise histological diagnosis at the local level can cascade into inappropriate surgery, inadequate staging, or suboptimal systemic therapy selection.
Surgery: Procedures carrying lower technical complexity and risk, such as removal of atypical lipomatous tumors, superficial skin sarcomas including leiomyosarcoma and some dermatofibrosarcoma protuberans, and straightforward gastrointestinal or pulmonary metastasectomies, are appropriate for local execution. In contrast, large soft tissue sarcomas (greater than 5 cm, whether superficial or deep), all extremity and trunk wall sarcomas of significant size, and retroperitoneal or pelvic sarcomas must always be centralized given the specialized expertise required for margin-negative resection and the complexity of reconstructive planning.
Radiotherapy and chemotherapy: Routine postoperative radiotherapy for extremity sarcomas and palliative treatments can be administered locally. However, preoperative radiotherapy for large extremity, trunk wall, or retroperitoneal tumors requires centralized planning to coordinate with surgical timing and optimize coverage. For systemic therapy, standard anthracycline-based regimens in the advanced or adjuvant setting can be delivered locally, while neoadjuvant protocols for localized disease, concurrent chemoradiotherapy, and multidrug regimens for high-risk histologies such as Ewing sarcoma and rhabdomyosarcoma require hub-level expertise and infrastructure.
In the Sarcoma HASM network, strategic clinical decisions are centralized at hub facilities equipped with advanced technology and specialized multidisciplinary teams. Spokes are regional centers connected to the hub that manage standard diagnostics and, where appropriate, treatments, escalating complex cases to the hub. This configuration is designed not to replace patient-centered local care but to ensure that complexity-appropriate care reaches every patient regardless of geography. The hub provides not only advanced clinical management but also functions as the research and training engine for the network, disseminating protocols, generating evidence, and developing the next generation of sarcoma specialists.
Spoke roles and continuity of care: Spokes function as the first point of patient contact, providing accessible diagnostics and managing routine follow-up and lower-complexity treatments. This local presence alleviates demand on hub facilities and reduces patient travel burden, particularly for ongoing chemotherapy administration, surveillance imaging, and management of treatment-related complications. The authors emphasize that spoke facilities are not merely passive referral conduits but active participants in clinical pathways, with defined responsibilities and the capacity to escalate cases appropriately.
National and regional adaptation: The HASM network must be tailored to each country's healthcare structure, accounting for financial systems, specialist distribution, and geographic factors. The authors cite several European examples as evidence that such adaptation is feasible at scale. NETSARC in France is one of the most mature examples, demonstrating that a nationwide sarcoma reference network can produce measurable improvements in diagnosis accuracy, adherence to guidelines, and survival. Similar models have been implemented in the UK, Germany, Poland, and Italy, each adapted to national conditions. At the supranational level, the European Reference Network on Rare Adult Solid Cancers (EURACAN) extends this logic across EU member states, standardizing care and supporting cross-border collaboration for ultra-rare sarcoma subtypes that no single country encounters in sufficient numbers to develop independent expertise.
Resource matching: The number of hubs and spokes per country should be determined by population size, disease incidence, existing specialist capacity, and the geographic distribution of medical facilities. The authors argue that this model maximizes expert resource utilization: by centralizing only complex decision-making rather than all patient management, hubs can operate at clinical and intellectual capacity without being overwhelmed by routine care volume.
For the HASM network to function effectively, the authors argue that a dedicated digital data infrastructure is not optional but foundational. The platform must enable seamless, real-time exchange of patient records, imaging studies, pathology reports, and laboratory results across all hub and spoke facilities, ensuring that every clinician, regardless of location, has access to current and complete information at the point of care. This is a higher standard than most existing hospital information systems, which typically operate within institutional silos and require manual data transfer or delayed synchronization.
MDT meeting support: A critical function of the digital platform is supporting the logistics and clinical content of MDT meetings. These meetings are the quality cornerstone of sarcoma care, but they are logistically difficult to convene when specialists are distributed across hub and spoke sites. The platform enables telemedicine participation so that all team members, whether at the hub or any spoke, can log in to access real-time patient data and actively contribute to case discussions. This inclusivity ensures that spoke-based clinicians who know the patient's local context can participate, while hub-based specialists bring diagnostic and therapeutic expertise. The result is more representative and better-informed multidisciplinary decision-making.
Patient-reported outcome measures (PROMs): The platform also systematizes the collection of PROMs during follow-up visits. Patients regularly complete outcome questionnaires that update their health status in the platform in real time. Any documentation of local or systemic recurrence is automatically flagged for discussion at the next MDT meeting, ensuring that even patients managed at spoke facilities receive timely expert review when their disease status changes. Routine uneventful follow-up data is also captured, contributing to the longitudinal outcomes database that underpins benchmarking and research activities across the network.
Triage and decision support: Beyond data storage and sharing, the platform must support sophisticated data analysis to enable informed triage decisions between hub and spoke. The authors specify that the platform should analyze clinical parameters such as imaging characteristics, histological complexity, and patient performance status to recommend appropriate care pathways. This analytical capability is what distinguishes the digital infrastructure of the HASM network from simple electronic health record (EHR) systems, enabling it to function as an active participant in clinical decision-making rather than a passive repository.
The Sarcoma Care Data Warehouse, exemplified by systems such as Sarconnector, extends well beyond the function of a conventional clinical registry. It aggregates comprehensive patient data including detailed demographics, specific treatment protocols, immediate outcomes (early surgical success, initial response to systemic therapy), and long-term follow-up data encompassing local recurrence rates and metastasis-free survival. The integration of clinical-reported outcome measures (CROMs), PROMs, patient-reported experience measures (PREMs), economic measures (ECOMs), and patient-centric omics measures (PCOMs) enables a holistic, value-based healthcare assessment across the entire care cycle.
Sarcoma-specific quality indicators: The data warehouse tracks a minimal prospective dataset of quality indicators that the authors describe as essential for benchmarking sarcoma care. These include: time and accuracy to establish a diagnosis, thoroughness of pathological analysis, whether reference pathological review has occurred, margin status (R0, R1, or R2) after surgery, 30- and 90-day morbidity and mortality after surgical resection, the complexity of tumor resection (categorized by site, size, and procedure type), and adherence to established treatment guidelines. These metrics allow comparison and benchmarking not only within a single HASM network but across networks in different countries, identifying unexpected variations in quality and targeting interventions accordingly.
Automated analytics for real-time decision support: The authors describe automated analysis algorithms within the warehouse that support real-time triage decisions. For example, analysis of surgery complexity data can inform whether a case should be escalated to the hub or can be safely managed at a spoke. Automated dashboards updated on a regular cycle provide clinicians and policymakers with visibility into performance metrics across all hub and spoke facilities, supporting quality improvement initiatives and ensuring uniform care standards. The warehouse also uses federated data structures that enable analysis across multiple HASM networks spanning wide geographies without requiring the physical transfer or pooling of raw patient data, thus preserving institutional data sovereignty and patient privacy.
Research infrastructure: A major secondary function of the data warehouse is supporting research, particularly for rare sarcoma subtypes where no single institution generates sufficient cases for statistically meaningful analysis. By rapidly identifying eligible patients across the network, the warehouse facilitates enrollment in clinical trials. It also tracks long-term treatment efficacy and outcomes across a diverse patient population, providing the real-world evidence base needed to refine sarcoma treatment protocols over time.
The HASM network's robust, standardized, and longitudinally collected dataset provides the foundation for deploying advanced predictive tools. The authors describe two technologies that represent the next evolution in precision sarcoma care: Sarcoma Digital Twins and causal machine learning (CML). Both go substantially beyond what existing prognostic indices or conventional predictive models can offer, by moving from static risk stratification to dynamic, individualized treatment simulation.
Sarcoma Digital Twins: A Sarcoma Digital Twin is a continuously updated virtual model of an individual patient that integrates diverse real-time data streams including genetic information, clinical variables, imaging findings, treatment histories, and real-world health monitoring signals. The digital twin uses predictive algorithms to simulate how the patient might respond to different treatment options given their unique medical history, current health status, and evolving tumor biology. Clinicians can use the digital twin to visualize potential outcomes under various treatment scenarios before committing to a specific pathway, enabling risk-benefit analysis that was previously available only through clinical intuition and population-level evidence. This capability is particularly valuable for complex cases involving multiple viable treatment options, such as the sequencing of preoperative chemotherapy, surgery, and radiotherapy in locally advanced retroperitoneal sarcoma.
Causal machine learning: CML represents a fundamental advance over traditional predictive analytics. Where standard machine learning models identify statistical correlations in historical data to forecast outcomes, CML identifies the causal relationships between specific treatment actions and their consequences. Rather than simply predicting which patients are likely to survive, CML analyzes the counterfactual: what would happen to this specific patient's outcome if we chose treatment A versus treatment B? In sarcoma care, CML can quantify the expected balance between positive outcomes (such as gain in life years or metastasis-free survival) and potential side effects or toxicities, producing treatment recommendations that are optimized not just for efficacy but for the overall benefit-harm trade-off in the individual patient.
Integration with the HASM network: The authors frame Digital Twins and CML not as standalone research tools but as integral components of the HASM network's digital infrastructure, drawing on the continuously updated data warehouse to refine their predictive models with each new patient encounter. This creates a learning health system architecture where clinical evidence generated across the network continuously improves the accuracy of predictive analytics, which in turn informs increasingly precise individual treatment decisions throughout the network.
The authors close by addressing the practical challenges of implementing the HASM network across heterogeneous healthcare systems. Financial models, data governance frameworks, medicolegal liability for decentralized care, interoperability standards between digital platforms, and the training pipeline for spoke-based clinicians all represent genuine implementation barriers. The integration of AI-based tools into clinical workflows requires not only technical development but also regulatory approval pathways and robust validation against prospective outcomes data, domains where sarcoma-specific evidence remains limited compared to more common cancers.
Data standardization as a prerequisite: A harmonized minimal prospective dataset is identified as the non-negotiable foundation for HASM network function. This dataset must include patient demographics, treatment protocol details, immediate and long-term outcomes, and patient-reported measures. Standardization ensures that data generated at peripheral spokes can be meaningfully aggregated and compared with hub data, and that individual HASM networks can benchmark their performance against regional and international peers. The authors note that a recent benchmarking study confirmed that such comparison is not only technically feasible but also practically valuable, having identified unexpected inter-center variations in care quality that prompted targeted quality improvement interventions.
Broader applicability: One of the paper's closing arguments is that the HASM framework, while designed specifically for sarcoma, provides a template applicable to all rare and complex diseases. The combination of expertise centralization, structured decentralization of routine care, digital interconnection, real-world data aggregation, and AI-driven predictive analytics addresses challenges that are not unique to sarcoma but are universal across rare oncology and indeed rare disease management broadly. The authors position the Sarcoma HASM network as a proof-of-concept for a new model of global health network design where integrated care becomes the default rather than the exception.
Call to action: Healthcare providers, policymakers, and stakeholders are explicitly urged to support integration of the HASM model into national health systems. The authors acknowledge that this requires alignment across clinical, administrative, technological, and regulatory domains, and frame the task as an investment in infrastructure that will pay dividends not only in improved sarcoma outcomes but in the broader capability of health systems to manage low-incidence, high-complexity diseases that individual institutions cannot adequately address in isolation.