The latest advances in diagnostics and monitoring: Insights from Heart Failure 2026

• It is time to move away from a reactive approach, toward a proactive, community-based model; the goal must be to intercept HF before it progresses to symptomatic disease
• The referral process and high demand for echocardiography creates significant waiting times; adding risk-prediction models to diagnostic algorithms could help to ensure high-risk patients are prioritized
• NT-proBNP testing retains a pivotal role, particularly at point of care and when combined with the latest digital technologies; however, cut-off thresholds for certain patient groups should be adjusted 

Academics, physicians, and industry representatives from around the world came together to discover the latest developments in screening, diagnostics, monitoring, and treatment for heart failure (HF). Each update represents an important step towards the ESC’s goal of ensuring patients worldwide receive the best possible care.

An overarching message at ESC Heart Failure Congress 2026 was the need to transition from reactive, hospital-based care to a proactive, decentralized community model for diagnosis and management.

Screening helps to facilitate early diagnosis before disease progression occurs. Once diagnosis is confirmed, guideline-directed medical therapy (GDMT) can be initiated safely and quickly. Ongoing monitoring is then critical for tracking patient status, ensuring disease management is effective and identifying the need to adjust treatment.

In this article, we will focus on advances and trends from Heart Failure 2026 relating to screening, diagnosis, and monitoring.

PORTHOS is a cross-sectional, population-based, observational trial conducted in Portuguese individuals aged ≥50 years.¹ It demonstrated how obesity can act as “diagnostic camouflage”, with high BMI associated with lower N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels in some patients, which led to HF being “ruled out”.²

Based on 6189 participants, the estimated HF prevalence was 16.54%.¹ However, out of 285 patients who were “ruled out” due to normal NT-proBNP levels, 26.7% were obese (BMI ≥30 kg/m2).² And after adjustment for age, sex, and estimated glomerular filtration rate, obese patients with normal NT-proBNP had higher odds of symptomatic HF than lean participants (BMI <30 kg/m2) with NT-proBNP ≥125 pg/ml.²

These findings demonstrate that if NT-proBNP is normal (<125 pg/mL) in an obese patient, HF cannot be ruled out without an echocardiogram.² Additionally, experts have suggested lowering the NT-proBNP “rule-out” threshold for this population, for example to <50 pg/mL, to avoid missing clinically significant disease.

Atrial fibrillation and chronic kidney disease also impact NT-proBNP. Both conditions are associated with elevated NT-proBNP levels, and higher thresholds may help to more accurately diagnose HF. Any changes to NT-proBNP diagnostic thresholds should be validated in randomized controlled trials.

Heart Failure 2026 showcased how artificial intelligence (AI) is being integrated into the HF diagnostic workflow. Experts emphasized that the "optimal approach" should combine biomarkers with digital technologies.

Highlights from Heart Failure 2026:

• A physician and AI assistant work well together: In the ASSIST-HF study, the combination of a physician with an AI assistant was associated with superior GDMT optimization and fewer hospitalizations, compared with a physician alone.³
• AI can review patient medical history: ChatGPT-4 outperformed physicians in some HF diagnostic scenarios, suggesting that the integration of electronic health records (EHR) with AI is a clear next step.⁴
• Physical assessments could be digitized: An AI-enhanced stethoscope (TRICORDER) has been tested for point-of-care detection of HF.⁵

Ongoing heart failure monitoring is essential for optimal patient care. However, it places a significant burden on patients, is associated with high costs and takes up precious healthcare resources. Research presented at Heart Failure 2026 showed that progress is being made in this area too.

Highlights from Heart Failure 2026:

• Telemonitoring can be close to cost neutral: While telemonitoring has historically faced budget barriers, real-world data suggest it leads to cost savings in high-severity cohorts due to lower acute care use.⁶
• Ongoing heart failure monitoring could also be digitized: Smartphone-recorded respiratory acoustics have been investigated for remote monitoring of hemodynamic decompensation; results suggest high accuracy when referenced to NT-proBNP.⁷
• Devices could support self-management: An implantable V-LAP system measures left atrial pressure and notifies patients when to take their next dose.⁸

These technologies can support the STRONG-HF protocol, which encourages safe and rapid up-titration of GDMT. Telemonitoring, or home-based nursing care, could help to resolve the "up-titration gap", allowing for careful heart failure monitoring and optimization of treatment without overwhelming healthcare capacity.

Thank you to everyone who joined the Roche Diagnostics symposium at Heart Failure 2026.

The next section of this article is based on discussions from the symposium, which centered around how the pathway to HF diagnosis can be supported and streamlined.

At Heart Failure 2026, Roche Diagnostics brought together leading experts in the field, to discuss the need to identify and treat HF patients sooner to improve outcomes. Three speakers explored how we can close this gap by making use of rapid biomarker testing, leveraging the role of primary care, and harnessing the power of AI.

Professor Antoni Bayes-Genis started the symposium with a presentation about strategies for early identification and management of HF. He opened his presentation by contrasting the urgency of diagnosing HF with other critical diseases.

Patients suspected of having cancer have a biopsy within days. In contrast, HF patients may have to wait several months before receiving a confirmed diagnosis. This is largely due to the reliance on echocardiography, which creates a bottleneck due to limited capacity.

Professor Bayes-Genis highlighted the importance of community-based de novo diagnosis of HF using the FIND-HF tool. It is a mnemonic that stands for “Fatigue, Increased congestion, Natriuretic peptide testing, Dyspnea”, which reminds clinicians that even vague symptoms should be treated with suspicion.

A strategy combining education of primary care physicians and point-of-care NT-proBNP testing using Roche’s LumiraDx system is being investigated in the FIND-HF clinical trial.⁹ The goal of the study is to show that this strategy can accelerate the pathway to HF diagnosis, with NT-proBNP testing central to patient triage.

Next, Dr Clare Taylor explored the role of primary care in accelerating HF diagnosis. She started by highlighting that 8 out of 10 patients are diagnosed at the time of hospitalization.¹⁰ Put simply, this represents a failure of the outpatient system.

As a GP, Dr Taylor wants to identify ways to improve HF diagnosis in primary care for the benefit of patients, their families, and healthcare systems. For this reason, she developed “BEAT”, which stands for “Breathless, Exhausted, Ankle swelling, Time to test”, and has been advocating for its use in awareness campaigns.

However, the challenge remains: How can we ensure at-risk patients are not being missed while also protecting valuable healthcare resources? ESC and National Institute for Health and Care Excellence guidelines recommend NT-proBNP testing for patient triage,^11,12 but Dr Taylor recognizes the need for age- and BMI-adjusted cut-offs to increase specificity.

In the final presentation of the symposium, Dr Ambarish Pandey argued that biomarkers remain an important tool for HF risk prediction and detection despite the arrival of digital, multimodal diagnostic technologies.

In fact, the performance of AI-assisted tools is enhanced when paired with biomarkers. This has been demonstrated with three of the latest and most advanced models:

• ML-based risk prediction algorithms: Clinical risk scores, such as WATCH-DM, followed by biomarker testing demonstrated improved identification of patients at high risk, who require intervention to prevent HF progression.¹³
• Electrocardiogram (ECG)-AI: Various studies have investigated the clinical utility of combining ECG with AI algorithms to screen for and diagnose HF^.14–16
• Echocardiography in risk-prediction: Echocardiography in combination with biomarkers can improve predictive accuracy versus echocardiography alone.¹⁷

Diagnosing HF is the first step. Next, the “implementation gap” must be overcome, to drive timely initiation of lifesaving GDMT.

The STRONG-DM trial is investigating the utility of electronic health record-embedded alerts in prompting primary care physicians to implement NT-proBNP testing in at-risk patients. This is combined with virtual consultations with HF specialists who provide treatment recommendations, to support the timely intervention.¹⁸

From tools that raise awareness of symptoms to the latest digital technologies, the Roche Diagnostics symposium demonstrated the potential to accelerate HF diagnosis. And there remains an important role for NT-proBNP testing, with its use in primary care and its integration within multi-modal approaches.

The insights and latest research shared at Heart Failure 2026 underscore the importance of transitioning toward a personalized, proactive model for diagnosis and management.

A major point of discussion was the need for an additional layer to support more specific and cost-effective patient screening. Integrating risk-prediction models could help to identify the highest risk patients who should be prioritized for biomarker testing and echocardiography.

The need to adjust NT-proBNP thresholds was explored. Physicians should be aware of “diagnostic camouflage” when interpreting the results of obese patients. Eventually, this could lead to guideline-recommended thresholds being revisited, to avoid patients with clinically relevant disease from being missed.

As we move away from a "one size fits all" approach and embrace AI-driven diagnostics, the potential to intercept HF before symptoms appear has never been greater. At the Roche Symposium, we saw that there are tools and technologies that have the potential to speed up HF diagnosis. The challenge now lies in bringing these advances to the clinic and implementing them in daily practice.

At Roche, we were inspired to see the HF field come together in pursuit of the shared goal to bring optimal care to patients with HF. The congress provided valuable insights from clinical practice, which can guide us as we continue to support clinicians with high quality NT-proBNP assays and develop point-of-care and digital solutions that can help overcome major challenges facing healthcare systems.