AI Comes of Age in Hematology-Oncology at ASH 2025
Creado por Jorge Durand, PhD, M Eng; Dane Lund
What you’ll learn:
ASH 2025 confirmed that AI and Machine Learning in hematology-oncology are transitioning from proof-of-concept to practice-ready, guideline-adjacent tools that can meaningfully improve diagnostic precision, treatment selection, and operational efficiency within the next two years.
The primary barriers to adoption are no longer model performance, but trust, explainability, bias, infrastructure readiness, payer validation requirements, and an increasingly stringent and complex regulatory landscape.
For pharma, early engagement through companion AI, synthetic real-world evidence, and AI-enabled clinical workflows offers a near-term opportunity for durable differentiation, but carries growing commercial and regulatory risks if not approached proactively and strategically.
American Society of Hematology (ASH) 2025 marked a decisive shift in hematology-oncology (hem-onc): AI and Machine Learning (ML) are no longer a showcase of promising prototypes but a maturing ecosystem of practice-ready, clinically relevant tools. Across multiple sessions, investigators unveiled new applications in diagnostics, Minimal/Measurable Residual Disease (MRD) assessment, CAR-T eligibility prediction, pathology automation, and transplant monitoring, illustrating how AI and ML are being embedded into day-to-day hem-onc workflows, with potential to reduce diagnostic and operational workload by more than 50%.
At the same time, the sessions underscored a set of non-negotiable watch-outs—including data quality variability, model explainability, demographic bias, operational constraints, and a rapidly tightening regulatory environment. These factors will significantly shape adoption speed, payer acceptance, and guideline updates. It was interesting to see minimal coverage on the role, needs, and experience of Advanced Practice Providers (APPs) in relation to some of these advanced AI-powered tools.
For pharma, the window is clearly open: clinicians will increasingly expect companion AI that enhances patient selection, safety monitoring, and evidence generation. Organizations that seed these capabilities and solutions now, by co-developing or aligning with emerging AI workflows, can gain durable brand differentiation. At the same time, regulatory frameworks (FDA PCCP guidance for AI-enabled device software functions, EU AI Act) continue to evolve in 2026.
“We are seeing an increase in AI adoption in healthcare. But we're also seeing that the levels of adoption [and] institutional readiness is very heterogeneous. There's tremendous confusion, there's a lot of well-known taken-year gifts that really have the most fear-worthy AI that we've been seeing today." —Dr. Shannon McWeeney OHSU, Portland; ASH 2025 Joint Session: Advancing Hematology Through Artificial Intelligence
"While the integration of these techniques into routine hematologic practice remains a developing frontier, the capacity for AI to refine disease diagnosis, prognosis, and therapeutic strategy is becoming increasingly evident." —Andrew Srisuwananukorn, MD, The Ohio State University Comprehensive Cancer Center
1. Practice Changing Work on the Horizon
Several ASH 2025 presentations illustrate how AI is moving from proof-of-concept to clinic-ready capabilities, such as:
One-click MRD flow cytometry delivering ~1-minute gating
Ultra-fast DNA methylation-based leukemia classifiers running in under four hours
Explainable CAR-T outcome predictors using radiomics + labs
AI-standardized PD-L1 scoring for small-biospy T/NK lymphomas
Daily smartphone-based GVHD severity tracking with Long Short-Term Memory models
Collectively, these innovations promise earlier diagnoses, more tailored therapy selection, lower costs, and 40–98% workflow efficiency gains—outcomes realistically deployable within two years.
2. AI in Guideline Integration and Operational Efficiency
Many AI/ML presentations explicitly map their algorithms to NCCN, ELN, or ESMO decision nodes, including diagnosis, risk stratification, treatment choice, and monitoring. Multi-center validation and FDA companion diagnostic filings are accelerating adoption, but payers and pathologists still require external validation data before guidelines shift.
Some of the ASH presentations also showed that AI/ML implementation brings value as an operational accelerator, slashing lab or chart-review time by 40–98%, enlarging rare disease cohorts 10-fold with synthetic data, and speeding virtual tumor boards. The impact of this integration ranges from incremental workflow relief to disruptive automation that can rebalance staffing and shorten trial timelines. This integration could create an uneven adoption curve across U.S. and global centers.
3. Commercial and Marketing Implications
Opportunities for brand teams’ partnerships center on bundling companion-AI tests, tapping synthetic real-world evidence for rapid label expansion, and arming the field and omnichannel strategies with live guideline engines, all differentiators that resonate with payers and HCPs. Risks loom if third-party triage algorithms steer patients to rivals or if open-source diagnostics erode premium pricing, making proactive data-sharing and IP refresh critical.
Companion-AI differentiation: Integrating AI-based patient eligibility or response tools into the brand ecosystem can speed decisions, make switching less appealing, and strengthen economic value stories
RWE Generation at Scale: Tap privacy-safe synthetic data from many hospitals to run real-world studies in weeks, not years
HCP Engagement 2.0: Tools like AI guideline engines, ASH-G bots, and virtual tumor boards allow MSLs to elevate conversations from slide reviews to real-time clinical problem-solving
Watchouts
Even the strongest AI tools at ASH came with consistent risks that will shape adoption speed and market impact:
Trust & Fairness: Explainability is mandatory for high-stakes decisions; variable data quality and untested bias can erode accuracy, especially in underrepresented populations, so human-in-the-loop oversight remains essential.
Guidelines & Regulation: Payers and regulators expect multi-center, real-world validation and clear plans to manage model drift and updates; pathologist/APP comfort will lag without transparent, interpretable tools.
Operations & Infrastructure: Synthetic data can quietly reproduce existing biases, foundation models may overfit to a few centers, and uneven IT maturity (e.g., EHRs) will slow scale up.
Commercial Risk: Third-party AI triage can steer patients to competitors, AI diagnostics may become commoditized, and clinical-grade AI software bundled with therapeutics will face growing regulatory burden and the need for specialized AI/ML regulatory capabilities.
Closing Remarks
The ASH 2025 focus on AI/ML tools marks a tipping point from “cool demos” to clinically actionable, guideline‑aware tools that may reshape hematology-oncology within the next few years. But adoption will hinge on addressing the session’s key watch-outs—explainability, data quality, fairness, operational readiness, and regulatory alignment.
“The Society generally supports the use of AI in the health care system and to improve the safety and efficacy of drugs and biologics approved by the FDA. With the appropriate foundation, and the additional guardrails around AI models, we believe AI has the potential to improve and transform clinical research, particularly in the field of hematology.” —ASH, comment letter to FDA on draft AI guidance, April 9, 2025
Pharma marketing teams that begin integrating, co-developing, or aligning with these tools now will be best positioned to effectively support clinical decision-making, influence care pathways, strengthen competitive differentiation, and create durable value before standards and competitors solidify.
Klick Health is the world’s largest independent commercialization partner for life sciences and a leading full-service pharma marketing partner, serving as agency of record for leading pharma, biotech, and healthcare brands. Klick’s specialized offerings are rooted in deep medical and scientific understanding, including market insights, award-winning creative, and proprietary AI and data models to craft impactful brand narratives and seamless customer journeys. Backed by nearly 250 medical experts and advanced healthcare analytics, Klick delivers integrated marketing strategy and communications, from new product launch strategy to MLR review with real-world evidence, helping brands thrive in today’s complex healthcare landscape. Learn more at Klick.com.
Etiquetas
Autores
Jorge Durand, PhD, M Eng
Executive Director, Medical Strategy
With over 15 years of medical strategy expertise, Jorge has helped build healthcare brands across diverse categories in the US and globally, including oncology, immunology, neuroscience, gastroenterology, pediatric vaccines, and rare diseases, among others.
Jorge has supported marketing and medical affairs efforts, from the development of new messaging platforms and campaigns to elevating and optimizing omnichannel strategy as well as building collaboration with thought leaders and patient-advocate organizations. Since joining Klick in 2017, Jorge has played a crucial role in growing our brand portfolio as well as expanding the medical strategy department tenfold. He is particularly passionate about bringing forward practice-changing ideas rooted in a deep analysis of the science and market dynamics, especially in brands that have a significant impact on the lives of patients.
Jorge graduated with a BS and an MS in nuclear engineering from Balseiro Institute (Argentina). He completed his PhD/MS in biomedical sciences at Albert Einstein College of Medicine, where he was also a postdoctoral research fellow.
Dane Lund
Medical Strategist
Dane brings deep experience in oncology and rare diseases, with a focus on lung, breast, prostate, gastric, and hematologic cancers, as well as autoimmune and genetic conditions. He is passionate about translating complex science into clear, compelling stories that resonate with both healthcare providers and patients. Prior to entering medical strategy, Dane completed a PhD in molecular biology and a postdoctoral fellowship at Memorial Sloan Kettering Cancer Center, where he conducted research on cancer-related muscle wasting and AI-driven drug discovery to perturb cancer cell growth and metastasis.
Contenido relacionado
Ready to Drive Life Sciences Forward?
Experience the transformative power of Klick Health, where deep industry expertise meets cutting-edge AI-driven wisdom.
As your trusted partners in life sciences commercialization, we combine a storied history in healthcare with the latest technologies to elevate every facet of your omnichannel strategy. From crafting engaging narratives to enabling data-driven decision-making, our integrated capabilities ensure you lead the way in transforming patient outcomes through digital health innovation.
Let's create something transformative together.
