ESMO 2026 Insight | Part 3

Refining and re-defining standard of care

This is the third of four installments from ESMO 2026, released throughout the week. Check out our prior insights:

• ESMO 2026 Insights Part 1
• ESMO 2026 Insights Part 2

This October, Lumanity’s oncology experts joined the crowds in Madrid for Europe’s largest cancer conference. Across tumor types, we heard the central question shift from “does it work?” to “where does it belong, for whom, and can patients get it?” Potent new agents are moving into earlier lines of therapy. That makes patient selection, sequencing, evidence quality, pricing and access as important as efficacy.

Executive summary: Major leaps forward grab the headlines. But in clinical care and clinical development, much work remains to troubleshoot and optimize new therapies before they become established standards of care.

Below, our experts cover three themes:

• Bladder cancer: ADC plus IO combinations in muscle-invasive disease, and how new biomarker approaches may inform when to dial therapy up or down
• Gioblastoma: Continued struggles to make progress, with lessons from recent failures and promising signals
• Lymphoma and myeloma: How first-line therapy continues to evolve

What’s coming next:

• Part 4 / Friday: Looking forward to new paradigms. New immunotherapy modalities, radiopharmaceuticals and delivery, and new models of evidence generation.

---

When Is Enough, Enough in MIBC?

Kirstie Norgate, PhD, Director, Insight

ESMO 2026 shifted the conversation to perioperative muscle-invasive bladder cancer (MIBC). The question is no longer whether to add immunotherapy, but how much is needed, in which combination, and for which patients. The aim is to improve survival while limiting long-term toxicity.

Treatment has changed quickly in recent years. Perioperative pembrolizumab entered practice first. EV plus pembrolizumab (EV+P) is now replacing cisplatin-based chemotherapy in MIBC, following the same path as in locally advanced or metastatic urothelial cancer. The FDA approved EV+P for perioperative treatment of all MIBC on July 10^th 2026. EV is highly effective but hard to tolerate, particularly because of skin toxicity and peripheral neuropathy.

VOLGA tested a reduced-EV approach. Patients received three cycles of EV plus the anti-PD-L1 agent durvalumab (EV+D) before surgery, then nine cycles of adjuvant durvalumab. The regimen cut the risk of an EFS event by 49% versus upfront cystectomy (HR 0.51). Median EFS was not reached versus 12.1 months, and 24-month EFS was 64% versus 45%. Median OS was not reached in either arm, but it favored EV+D (HR 0.58), with 24-month rates of 78% versus 69%. The pCR rate was 45% versus 10% by local review and 40% versus 8% by central review. That is a clear benefit from three neoadjuvant EV cycles. Adverse events were those expected with durvalumab or EV, including skin reactions. Immune-related events of any type occurred in 11% of patients and peripheral neuropathy in only 0.4%. This suggests that limiting EV exposure limited its long-term toxicity.

VOLGA also tested a triplet, EV+D plus the anti-CTLA-4 agent tremelimumab. It improved EFS (median not reached versus 12.1 months; 24-month rate 60% versus 45%) but was less compelling overall. Median OS was not reached, and unlike the doublet, the curves did not separate well (HR 0.77; p=0.1230). The triplet also added toxicity and more treatment discontinuations. In this setting, less may lead to more. As discussant Michiel van der Heijden noted, a lighter regimen could reduce the treatment burden for patients recovering from major cystectomy surgery.

KEYNOTE-866 explored the same idea from the other direction by adding perioperative pembrolizumab to cisplatin/gemcitabine. The trial was negative. EFS did not improve significantly (HR 0.88; p=0.137), and five-year OS was 70.8% versus 70.9%. pCR was higher with pembrolizumab (44.4% versus 35.8%), but this did not carry through to EFS. Here, less was not more. A stronger immunotherapeutic effect appears to be needed, and the result is a useful caution about pCR as a surrogate. The investigators also noted that control-arm outcomes were better than historical expectations for neoadjuvant cisplatin/gemcitabine with radical cystectomy (RC) and pelvic lymph node dissection (PLND).

The KEYNOTE-905 ad hoc analysis reinforces the value of EV+P in the perioperative setting. The trial included patients who were ineligible for cisplatin as well as those who refused it, so its message applies beyond a narrowly defined group. The analysis compared all three arms in concurrently randomized patients (n=255). Pembrolizumab alone was active, with a pCR rate of 25.3% versus 10.3% with surgery alone. That is well below the 58.8% pCR reported with EV+P. The question is therefore not whether immunotherapy alone works, but whether it is enough. Discussant Min Yuen Teo argued that these data may mark the end of the neoadjuvant checkpoint inhibitor monotherapy era. Taken with the negative KEYNOTE-866 result, they strengthen the case that adding EV to immunotherapy, not adding immunotherapy to chemotherapy, is what moves outcomes.

KEYNOTE-B15 provides some additional insight and potential future treatment selection guidance through its ctDNA analysis. ctDNA is a strong prognostic marker, and it also shows how well treatment is working. With EV+P, ctDNA positivity fell from 52.4% at baseline to 13.0% presurgery and 4.8% postsurgery. With cisplatin/gemcitabine, it fell from 56.8% to 19.7% and 13.5%. Among ctDNA-positive patients, pCR was only 3.7% with EV+P and 2.2% with chemotherapy. Among ctDNA-negative patients it was 72.4% and 43.7%.

Will data mark the end of the neoadjuvant checkpoint inhibitor monotherapy era?

EFS followed the same pattern. At 24 months, EFS in patients who stayed ctDNA-negative was 94.6% with EV+P and 87.8% with chemotherapy. Patients who converted from positive to negative reached 81.1% and 63.2%. Those who stayed positive reached only 33.3% and 35.3%. Two points stand out. Clearing ctDNA is associated with good outcomes, and EV+P appears to clear it more often and with a better EFS result than chemotherapy. But persistent positivity carries a poor prognosis in both arms, and EV+P showed no clear advantage in that small group. That is a signal for escalation strategies and new agents.

The EFS HR favored EV+P in baseline ctDNA-positive patients (0.54). It also favored EV+P in ctDNA-negative patients (0.68), but that 95% confidence interval crosses 1 (0.37–1.26). While extremely promising, the data do not yet show that ctDNA-negative patients can safely avoid future treatment.  As the discussant Teo noted, using ctDNA to select patients for bladder preservation or to omit adjuvant therapy needs further exploration and prospective validation.

Takeaways

So, when is enough, enough? ESMO 2026 suggests the answer is not one regimen but a threshold that differs by patient. For some, three cycles of EV with durvalumab appear to be enough. For others, enough has not yet been achieved.

ctDNA may eventually help define that threshold patient by patient. Patients who clear ctDNA do well. Patients who stay ctDNA-positive do poorly on either regimen, indicating new strategies are needed. Neither idea is proven. The ctDNA-negative subgroup data do not yet show that treatment can be safely omitted, and ctDNA’s link to micrometastatic disease still needs validation. Together, though, these data give hope that bladder-sparing treatment may be achievable in the near future.

Answering the question will take trials designed for it. They need ctDNA-guided escalation and de-escalation arms, and endpoints such as bladder-intact EFS and patient-reported outcomes. The field’s next task is to find out how much treatment each patient needs, not to keep adding more.

---

Glioblastoma at ESMO 2026: Turning Early Signals into Evidence

Michael Freeman, Senior Associate, Medical Affairs Consulting

More than two decades after the Stupp protocol became standard of care, glioblastoma (GBM) remains one of oncology’s most stubborn challenges. Maximal safe resection followed by temozolomide (TMZ)-based chemoradiotherapy is still the backbone of treatment. Tumor Treating Fields is the only addition since then to show even a modest survival benefit. ESMO 2026 brought no practice-changing breakthrough. What it did offer was encouraging early-stage signals, a clearer view of where the field is heading, and data that reinforce the need to select the right patients for trial enrollment. 

Activity with Targeted Therapies

The randomized phase 2 trial adding the BET inhibitor trotabresib to chemoradiotherapy was another reminder that the Stupp protocol is hard to beat. Outcomes were not improved and toxicity was increased. It joins a long list of agents that have struggled to improve on the Stupp protocol. Like many GBM trials with targeted agents, it enrolled an unselected patient population.

The IMPRESS-Norway trial offers an alternative approach, molecularly profiling patients with advanced tumors who have exhausted standard-of-care treatment, and matching them with targeted agents approved in other indications. Among patients with CNS malignancies, 20 with GBM were evaluable for response and 30% experienced durable disease control ≥16 weeks, all using drugs with no GBM indication. While this activity is modest, the point stands: matching the drug to the patient got activity out of agents that have no established record in GBM.

Immunotherapy Advances

GBM is a notoriously “cold” tumor, with low T cell infiltration and a highly immunosuppressive tumor microenvironment. However, encouraging data from novel immunotherapy targets and administration techniques were presented at ESMO. Preliminary results from a phase 1b/2 trial in China with ivonescimab and TMZ in recurrent GBM demonstrated:

• DCR of 95% (19/20)
• 6-month PFS of 77.8% (14/18)
• 12-month OS of 44.4% (8/18, mOS not reached)

Long term follow-up from the GLITIPNI trial with intracerebrally injected ipilimumab, nivolumab, and autologous myeloid dendritic cells following maximal safe resection highlighted:

• 2-year PFS: 26% (5/19)
• 2-year OS: 42% (8/19)

While the ivonescimab and GLITIPNI trials are limited by small populations, both demonstrated manageable safety profiles and durable responses (18+ mo.) in certain patients. Investigation of both regimens is ongoing in larger, more diverse patient populations. If these signals are going to hold up, developers will likely need to identify which patient subgroups are receiving durable disease control.

During his keynote lecture on CNS malignancies, Prof. Wolfgang Wick highlighted the NeoVax trial at DFCI that combines a personalized neoantigen vaccine with pembrolizumab for newly diagnosed GBM (originally presented at ASCO 2026).  The regimen induced neoantigen-specific intratumoral T cell responses and improved PFS and OS compared to propensity score-matched institutional controls. Patients with ex vivo neoantigen-specific T cell responses had significantly improved OS compared to non-responders (HR 0.26, p=0.015).

While validation of NeoVax is needed in a larger randomized trial, these data are promising, particularly in light of Moderna/Merck’s phase 3 success with their personalized vaccine in melanoma. Notably, transcriptional analyses in the NeoVax study revealed that the mesenchymal subtype was associated with greater T cell infiltration and survival. While the transcriptional state of GBM is heterogeneous, deeper understanding of factors that improve T cell killing will guide future trials and research.  

Developing Beyond Early Signals

ESMO 2026 points to a more selective development strategy in GBM: identify the right patient subgroup, demonstrate meaningful tumor exposure, and deliver benefit without compromising existing treatment. Across treatment modalities, a minority of patients achieve durable responses and long-term survival, emphasizing the need for further research to elucidate which treatments are effective in which patients.

The strongest opportunities remain developmental, and encouraging early signals must be distinguished from evidence that can support registration and adoption. GBM’s unmet need is clear, but the investable opportunity lies in a precisely defined treatment proposition supported by comparative evidence and a deliverable care pathway.

Takeaways

GBM remains one of oncology’s toughest challenges. There was promising data presented at ESMO from phase 1/2 trials testing targeted agents and immunotherapies, but encouraging early signals must be distinguished from evidence that can support registration and adoption.

ESMO 2026 points to a more selective development strategy in GBM: identify the right patient subgroup, demonstrate meaningful tumor exposure, and deliver benefit without compromising existing treatment. GBM’s unmet need is clear, but the investable opportunity lies in a precisely defined treatment proposition supported by comparative evidence and a deliverable care pathway.

---

The Changing Paradigm in Diffuse Large B-cell Lymphoma (DLBCL) and Multiple Myeloma (MM): Beyond R-CHOP and Upfront Transplant

Bianca Lee, PharmD, MSL-BC, Senior Consultant, Medical Affairs Consulting

Lymphoma and myeloma discussions at ESMO 2026 highlighted a common shift towards earlier immune-based therapies and more personalized treatment selection, guided by tumor biology and MRD.

DLBCL: Moving beyond R-CHOP

After approximately 50 years as the frontline backbone, R-CHOP is being challenged by targeted combinations. The following trials were highlighted during ESMO 2026 discussions to illustrate the evolving DLBCL treatment landscape. In POLARIX, polatuzumab vedotin plus R-CHP maintained a PFS benefit at five years (HR 0.77), while frontMIND reported improved PFS with tafasitamab plus lenalidomide and R-CHOP (HR 0.75) with added toxicity. Another strategy introduces bispecific antibodies into frontline treatment, with three phase 3 trials: OLYMPIA-3 (odronextamab), EPCORE DLBCL-2 (epcoritamab) and SKYGLO (glofitamab). EPCORE DLBCL-2 met its primary PFS endpoint in a topline announcement (HR 0.49; full data pending). Despite these advances, molecular heterogeneity remains important. Molecular profiling may guide treatment selection. ViPOR has shown preferential activity in MCD and EZB-MYC-positive subtypes, while tumor microenvironment and T-cell exhaustion signatures may help explain response variability to bispecific antibodies and CAR-T, respectively.

MM: Rethinking transplantation

The question is shifting from whether to transplant to which patients benefit most. Anti-CD38 quadruplets are increasingly established in frontline treatment, while trials are exploring MRD-guided transplant strategies and earlier use of bispecific antibodies and CAR-T. Two ESMO 2026 abstracts explored belantamab-based combinations in transplant-ineligible patients. Early-phase belantamab combinations have shown encouraging activity: Bela-DRd achieved 81.3% MRD negativity among evaluable patients, while Bela-Rd reported a 97.6% overall response rate. Ocular toxicity remains an important management consideration. These early findings support further investigation of ADC-based combinations in patients unable to tolerate standard quadruplets.

Takeaways

In DLBCL, targeted agents and bispecific antibodies are moving into frontline treatment. In MM, MRD-guided strategies are challenging the role and timing of upfront transplantation. Both diseases are moving towards earlier use of targeted immunotherapy and more individualized treatment. Molecular profiling in DLBCL and MRD assessment alongside cytogenetic risk in MM may help refine treatment selection and intensity; however, prospective validation and longer-term outcomes remain necessary.