Tirabrutinib

Part A enrolled 48 participants in an open-label phase II study. The reported response and survival estimates do not establish comparative benefit against another treatment. Enrollment required an ECOG performance score of 0, 1, or 2, limiting direct applicability to participants with worse performance status. Part A monotherapy results should not be extrapolated to Part B chemotherapy combinations.

The study was uncontrolled and enrolled 44 participants with KPS scores of at least 70. Results therefore do not establish comparative efficacy or directly characterize participants with lower performance status. Subsequent therapy was reported for 32 participants, limiting attribution of long-term overall survival to tirabrutinib alone. Quality-of-life analyses were exploratory, were not planned when the protocol was fixed, and lacked follow-up after tirabrutinib discontinuation. The autopsy publication describes a participant from this trial, not an independent efficacy cohort.

The retained publication describes an ongoing, open-label randomized trial with blinded independent review of progression-free survival. It does not provide comparative efficacy or safety results. Progression-free survival is the primary endpoint; overall survival is a secondary endpoint.

The protocol describes an open-label comparison in induction-refractory PCNSL, not a completed efficacy analysis. The phase III plan includes 94 participants, a non-inferiority hazard-ratio margin of 1.50, 70% power, and a one-sided alpha of 10%. These design choices require consideration when interpreting any future non-inferiority finding. Overall survival and functional or cognitive outcomes must be distinguished from the phase II four-week treatment-continuation criterion.

This prospective observational surveillance had no randomized comparator. Safety was assessed in 189 participants and effectiveness in 121, so the response rate must not be interpreted using the safety-population denominator. Response was assessed by treating physicians. The 52-week observation period limits assessment of later outcomes.

ROSETTA is a retrospective multicenter study without a randomized comparator. The original interim abstract describes 161 participants, whereas the corrected real-world progression-free survival figure identifies 158. The original median progression-free survival estimate of 11.8 months and the corrected estimate of 11.6 months are retained separately; their populations and analysis versions should not be treated as interchangeable.

The cohort included 35 participants treated at one institution between 2020 and 2025. Its retrospective design and separate recurrent and refractory groups limit causal interpretation of differences in treatment-failure and survival estimates. There was no randomized comparator in the reported design.

This retrospective cohort included 18 participants in the safety population and 17 in the effectiveness population. One participant contributed only to the safety analysis. The response estimate and infection proportion therefore have different denominators, and neither establishes comparative benefit or harm against another treatment.

This five-case series describes selected participants for whom standard treatment was difficult because of low KPS, renal dysfunction, or resistance to high-dose methotrexate. The reported 100% response rate is based on five participants and has no concurrent control group. It should not be treated as a population-level comparative efficacy estimate.

Two reported cases had long-term treatment without described neurotoxicity or leukoencephalopathy. This uncontrolled case series cannot establish the incidence of neurological toxicity or comparative neurological safety.

A single participant experienced five months of remission after tirabrutinib rechallenge. This observation does not establish the probability of response to rechallenge, an optimal retreatment interval, or comparative benefit.

This single case describes radiological enlargement followed by re-remission during continued tirabrutinib monotherapy. It does not establish the frequency of pseudoprogression or justify assuming that enlargement during treatment is benign in other participants.

This report concerns one participant who received tirabrutinib together with intravitreal methotrexate and subsequently underwent autologous transplantation. The post-transplant absence of relapse cannot be attributed to tirabrutinib alone. A tolerated 320 mg daily dose in this case does not establish a dosing regimen for other participants with cirrhosis.

This single case describes remission and 42 months without reported relapse after starting tirabrutinib for primary spinal-cord lymphoma. It does not establish response frequency or comparative benefit in broader PCNSL populations.

The retained case describes a temporary response in a participant with refractory vitreoretinal lymphoma and spinal-cord recurrence after chemotherapy and whole-brain radiotherapy. It does not provide a generalizable estimate of response duration or comparative efficacy.

This single case describes five months of complete radiological resolution after starting tirabrutinib. The observation does not establish response frequency, long-term disease control, or comparative benefit in lymphomatosis cerebri.

This single-center case series included seven participants, with skin disorders reported in five and treatment discontinuation because of severe cutaneous adverse events in one. The sample size and lack of a comparator limit estimation of population-level risk and comparative tolerability.

The report describes one participant with Aspergillus encephalitis after tirabrutinib exposure and a subsequent fatal subcortical hemorrhage. A case report can identify a possible safety signal but cannot determine incidence or establish that tirabrutinib alone caused the infection or hemorrhage.

Comparable measured concentrations in a single case do not establish bioequivalence, a limitation explicitly acknowledged by the authors. No adverse events were observed during suspension administration; this observation does not estimate the frequency of uncommon harms.