Research News for the Third Two Months of 2026

This is the thirty-seventh article of the project whose aim is to periodically collect (every two months) the research news on possible treatments for glioblastoma multiforme. Below I list the news items we considered most significant that emerged over the last two months. As with the previous articles in the series, each news item will be preceded by the original title with a link to the source and followed by a brief comment. The criterion by which the news is selected is always to include, in general, only news relating to research in the clinical phase, unless the research’s potential for the treatment of glioblastoma is truly remarkable.

Focused Ultrasound Substantially Improves Survival in Glioblastoma Drug-Delivery Trial 
One of the historical obstacles to glioblastoma therapy is the blood–brain barrier, which prevents most drugs from reaching the tumor: of temozolomide, the standard chemotherapy, it is estimated that less than 20% actually reaches the brain. An international multicenter phase 1/2 trial (the BT008NA study, sponsored by Insightec, published in Lancet Oncology) tested the transient, targeted opening of the barrier using low-intensity focused ultrasound combined with microbubbles, immediately before the administration of temozolomide. After surgery and standard chemoradiation, patients received up to six monthly focused-ultrasound treatments plus temozolomide, and were compared with a rigorously matched control cohort of 185 patients treated with standard-dose temozolomide alone. The results are remarkable: median progression-free survival of nearly 14 months versus 8 months in the control group, and overall survival of more than 30 months versus 19 months. The study did not directly measure the amount of drug that reached the tumor, but the technique proved feasible and safe, and confirms the controlled opening of the barrier as an enabling platform — potentially extendable to immunotherapy and other agents. Studies are already underway combining ultrasound-mediated opening with a bispecific antibody and with bevacizumab.

Tumor-targeted interferon-α gene therapy for glioblastoma: a phase 1 trial 
Published in Nature Medicine, a phase 1 study of a gene therapy that delivers interferon-α directly into the tumor microenvironment. Glioblastoma is an “immunologically cold” tumor, dominated by a strongly immunosuppressive myeloid component; interferon-α is a cytokine capable of reactivating the immune response, but its systemic use in oncology has always been limited by toxicity. The idea behind this approach is to induce the local production of interferon within the tumor, reprogramming the immunosuppressive environment in situ and avoiding systemic exposure. As this is a phase 1 trial, the primary objectives are safety and feasibility, with the first signals of immunological activity; the data should therefore be read as proof-of-concept. Its publication in a journal of the caliber of Nature Medicine and the rationale of the mechanism — turning a “cold” tumor into a “hot” one by acting directly on the microenvironment — make it one of the most interesting clinical news items of the two-month period.

Alpha Tau Announces Interim Results from its U.S. Alpha DaRT Recurrent Glioblastoma Trial (REGAIN) 
The REGAIN study (NCT06910306) is a small feasibility and safety trial (up to 10 patients with recurrent, non-operable glioblastoma who have already undergone radiotherapy) of Alpha DaRT, an alpha-radiation therapy inserted directly into the tumor tissue via a minimally invasive stereotactic procedure. At the FDA’s request, an interim safety analysis was conducted after the first 3 patients were treated, with data updated to May 3, 2026: local disease control in 100% of cases, with two complete responses (disappearance of all enhancing lesions on MRI, RANO criteria) and one patient with stable disease and a 30% tumor reduction; only one grade 3 serious adverse event (a seizure with transient paralysis, resolved with steroids). On June 11, 2026, the FDA authorized the completion of enrollment of the remaining seven patients and the addition of two new U.S. centers. The numbers are minimal and the data must be interpreted with extreme caution — longer follow-up, response-duration and survival data, and assessment of the risk of radionecrosis are needed — but objective responses of this magnitude in the recurrent setting, where median survival is typically less than 12 months, justify the attention.

Adjuvant personalized multivalent neoantigen DNA vaccination for MGMT-unmethylated glioblastoma: a phase 1 trial (GNOS-PV01 / GT-20) 
A phase 1 trial (the GT-20 study, NCT04015700) published in Nature Cancer on a DNA vaccine with personalized neoantigens (GNOS-PV01, developed by Geneos Therapeutics with Washington University in St. Louis) in 9 patients with newly diagnosed, MGMT-unmethylated glioblastoma. Each vaccine was custom-built by selecting up to 40 neoantigens per patient (range 17–40), a number higher than most previous approaches, also sampling different regions of the same tumor to increase the targets. The vaccine was well tolerated, with no serious adverse events or dose-limiting toxicities, and induced the activation and expansion of T cells in all evaluated patients except one, who was on steroid therapy; there was evidence of induced T cells entering the tumor microenvironment. The 24-month survival was 33%, with one patient alive more than 48 months after diagnosis. The perennial question for vaccines in glioblastoma remains open: whether the immune response translates into a durable clinical benefit, alone or in combination with other therapies.

A Phase 1 trial of iron metabolism-targeting oral gallium maltolate in recurrent and refractory glioblastoma
This is a phase 1 study published in Neuro-Oncology Advances (conducted at the Froedtert and Medical College of Wisconsin) on oral gallium maltolate (GaM), the first in-human study of this compound in glioblastoma. The rationale exploits the tumor’s dependence on iron: glioblastoma cells require far more iron than normal cells, and gallium — chemically similar to iron — is taken up in its place, acting as a “Trojan horse” that disrupts its metabolism and induces cell death. The dose-escalation study established a recommended phase 2 dose of 2,000 mg/day: the drug proved safe and well tolerated up to 2,500 mg/day, with only grade 1–2 effects, no grade ≥3 toxicity, and no renal toxicity. Across 22 analyzed patients, median overall survival was about 16 months — favorable compared with historical outcomes in the recurrent setting — with an interesting long-survivor tail (five patients still alive, four beyond 18 months and two at 31 and 39 months). A small, non-randomized study, but the results are encouraging for a phase 1 in this context, and the restart of an expanded access program for about twenty additional patients is planned.

Novocure Announces Topline Data from the Phase 3 TRIDENT Trial Evaluating Earlier Use of Tumor Treating Fields Therapy in Newly Diagnosed Glioblastoma 
The phase 3 TRIDENT trial (NCT04471844) enrolled 981 patients with newly diagnosed glioblastoma to verify whether starting Tumor Treating Fields (TTFields) already during chemoradiation, rather than in the subsequent maintenance phase, improved survival. The study did not meet its primary endpoint: median overall survival of 17.7 months in the early-start arm versus 17.5 months in the standard arm, with no statistically significant difference. It is important, however, not to read it as a “negative” study on TTFields themselves: both arms received the therapy, and the question concerned only the anticipation of the timing. An encouraging figure is the long-survivor tail in both arms, with 3-year survival of 22.5% (early start) versus 18.4% (standard start), favorable compared with the historical record for newly diagnosed glioblastoma; no new safety signals were observed. The full data, including compliance and subgroup analyses, will be presented at the ASTRO 2026 congress. In practical terms, the result suggests that starting in the maintenance phase — the one currently in use — remains an appropriate approach.

Personalized machine learning-guided radiation dose escalation in newly diagnosed glioblastoma: prospective pilot study 
A prospective pilot study (NCT03477513, University of Pennsylvania, with Steven Brem among the authors) published in Nature Communications, evaluating a personalized precision radiation therapy (PPRT) guided by a machine learning model. The algorithm integrates advanced imaging, histopathological data, molecular profiles, and clinical variables to estimate the radioresistant tumor subregions and the areas of microscopic infiltration not visible on MRI, generating individualized dose maps rather than a uniform dose. Twenty patients with newly diagnosed IDH-wildtype glioblastoma, after gross total resection, received PPRT with concomitant and adjuvant temozolomide. Compared with a matched historical control group, the results are important: median progression-free survival of 24.4 versus 11.6 months and overall survival of 35.4 versus 17.7 months. The treatment was well tolerated, with no grade ≥3 acute toxicity, but with a higher frequency of radionecrosis in the PPRT group. The magnitude of the gain is notable and the data will need to be confirmed in larger cohorts and randomized designs, but it is a concrete example of artificial intelligence entering the clinical practice of glioblastoma by personalizing a well-established treatment such as radiotherapy.

Hetairos is a histology-based artificial intelligence model for predicting central nervous system tumor methylation subtypes
We close with a news item that does not concern a therapy in the clinical phase but diagnostics — we include it for its notable potential and for the growing relevance of artificial intelligence in neuro-oncology. Researchers at the German Cancer Research Center (DKFZ), the University of Heidelberg, and Heidelberg University Hospital have developed Hetairos, a deep learning system that predicts the molecular classification of central nervous system tumors from the common histological slides stained with hematoxylin and eosin, available everywhere. The model was built and validated on more than 11,000 slides from 9,606 patients across 11 centers on four continents, using DNA methylation profiling (the current gold standard) as a reference, and distinguishes 102 molecular subtypes, covering almost the entire spectrum of the WHO classification. In a direct comparison on histology alone with five expert neuropathologists, Hetairos achieved significantly higher accuracy (0.68 versus 0.30), classifying 50–70% of cases with high confidence and an accuracy of 0.87 on its best predictions; in prospective evaluation it reduced turnaround time from about 12 days (methylation testing) to 12 minutes. The authors themselves emphasize that the tool is intended to support and accelerate molecular analyses, not to replace them, with a potentially significant impact especially in resource-limited areas. Further independent validation and regulatory assessment are needed before routine diagnostic use.

That is all for this third two months of 2026. Heartfelt thanks to all the people who, with their support, allow us to keep the volunteer organization alive and to develop projects increasingly focused on concrete support for patients and their caregivers. If you are experiencing the glioblastoma journey directly or indirectly, we remind you that the Speranza e Coraggio project offers a specialized psychological support service, completely free of charge, designed to accompany patients and family members through the most difficult moments: asking for help is an act of strength. To all those who are fighting against glioblastoma, and to their loved ones, we send our embrace.