Psilocybin-assisted psychotherapy improves psychiatric symptoms across multiple dimensions in patients with cancer

Patients with cancer commonly experience a broad range of psychiatric symptoms beyond anxiety and depression, including interpersonal sensitivity, hostility, somatic symptoms and obsessive worrying, all of which can reduce quality of life and worsen clinical outcomes. Previous clinical research has documented high prevalence of psychological distress in oncology populations and has explored treatments ranging from psychotherapy to antidepressants, stimulants and ketamine; however, many pharmacological options have delayed onset, problematic side-effect profiles or limited durability. Early clinical work with classic psychedelics and more recent Phase II trials of psilocybin-assisted psychotherapy (PAP) suggest rapid and sustained reductions in anxiety and depression, but PAP’s effects across other psychiatric symptom domains have not been systematically analysed. Petridis and colleagues sought to fill that gap by pooling individual participant data from two Phase II, randomised, placebo-controlled crossover trials of PAP in patients with cancer. The study aimed to test whether PAP produces multidimensional psychiatric improvements by analysing nine symptom domains from the Brief Symptom Inventory (BSI): anxiety, depression, interpersonal sensitivity, hostility, obsession–compulsion, somatisation, phobia, paranoia and psychosis. The investigators hypothesised that PAP would reduce multiple dimensions of psychiatric distress in this population when measured before and after medication sessions and at up to six months of follow-up.

This analysis pooled deidentified participant-level data from two Phase II randomised, placebo-controlled crossover trials conducted at New York University and Johns Hopkins. Participants were randomised to one of two dosing sequences: high-dose psilocybin first then control, or control first then high-dose psilocybin, with psychotherapy provided before and after each medication session. The pooled randomised sample comprised 87 participants; 79 completed at least the first post-dose assessment, 72 completed the second post-dose assessment and 68 provided data through approximately six months of follow-up. The extracted text reports that the sample completing post-dose 1 assessments (n = 79) had a mean age of 55.6 years, 42 female and 37 male, 92.4% identified as white, 63.5% had advanced (stage III or IV), metastatic or recurrent cancer, and most had at least an undergraduate degree. Blinding procedures were implemented in both parent trials, with randomisation performed by pharmacy staff and blinding assessed after sessions. Blinding integrity was poor: in one trial therapists guessed treatment correctly in 97% of cases, and in the other therapists’ dose ratings were substantially higher for high-dose sessions than for low-dose sessions, although some overlap in ratings occurred. The active comparator was a low, placebo-like dose of psilocybin (referred to as control in the pooled analysis) rather than an inert pill. Psychiatric symptoms were measured with the 53-item Brief Symptom Inventory (BSI), which yields nine averaged symptom-dimension scores. BSI assessments were pooled at four time points that aligned closely between trials: baseline, post-dose 1 (approximately 5–6 weeks after the first medication session), post-dose 2 (approximately 5–6 weeks after the second medication session) and follow-up (about six months after the second session). For each BSI dimension the investigators used mixed-effects repeated measures models with a compound symmetry covariance structure and fixed baseline covariates, implemented in SPSS v21. Pairwise between-group comparisons at each time point were tested with two-tailed F tests; the primary endpoint was the between-group difference at post-dose 1 (pre-crossover). The Holm–Bonferroni method controlled family-wise error at α = 0.05. Between-group effect sizes at post-dose 1 were reported as Hedge’s g. Within-group comparisons to baseline used Tukey post-hoc tests, and site comparisons used independent t-tests. The analysis is post-hoc in that the pooled multidimensional symptom analysis was not pre-specified in the parent trial statistical plans.

Before crossover, psilocybin-assisted psychotherapy produced statistically significant between-group improvements on six of nine BSI dimensions at the post-dose 1 assessment. Compared with the control-first group, the psilocybin-first group showed reductions in anxiety (F1,155.3 = 8.16, P = 0.0049; Hedge’s g = 0.64), depression (F1,142.8 = 12.03, P = 0.0007; Hedge’s g = 0.77), interpersonal sensitivity (F1,150.7 = 12.60, P = 0.0005; Hedge’s g = 0.79), obsession–compulsion (F1,141.0 = 15.08, P = 0.0002; Hedge’s g = 0.86), hostility (F1,123.4 = 7.05, P = 0.0090; Hedge’s g = 0.59) and somatisation (F1,157.1 = 19.32, P < 0.0001; Hedge’s g = 0.98). These six effects remained statistically significant after correction for multiple comparisons using the Holm–Bonferroni procedure. Following the crossover, between-group differences diminished and were no longer statistically significant at six-month follow-up for all six dimensions. Within-group analyses showed that the psilocybin-first group experienced significant improvements from baseline at each post-treatment time point for those same six dimensions. After crossing over, the control-first group likewise showed significant within-group improvements relative to baseline by post-dose 2. For the control-first group these within-group improvements persisted through six months for all dimensions except somatisation. The pooled treatment-related changes were reported as highly consistent across the two clinical sites, and the pattern of placebo (non-psychedelic) responses was also consistent between sites. Regarding safety-relevant psychiatric domains, PAP did not produce any lasting increases in paranoia, phobia or psychosis at the assessed time points. Participant flow summarised that 87 were randomised, 79 completed first post-dose assessments, 72 completed the second post-dose assessments and 68 completed follow-up; reasons for dropout were noted in the source but are not fully detailed in the extracted text.

Petridis and colleagues interpret these pooled trial data as evidence that psilocybin-assisted psychotherapy can ameliorate a broad set of psychiatric symptoms in patients with cancer beyond the commonly reported effects on anxiety and depression. The investigators highlight that a single high-dose psilocybin session paired with psychotherapy was associated with rapid, sizeable reductions in anxiety, depression, interpersonal sensitivity, hostility, obsession–compulsion and somatisation, and that these effects were observed consistently across the two parent trials. The authors discuss biological and psychosocial mechanisms that may underlie the multidimensional benefits. They note psilocybin’s agonism at the serotonin 2A (5-HT2A) receptor and related acute neural desynchronisation—particularly disruptions of default mode network function—and reference evidence linking these neurophysiological changes and enhanced neuroplasticity with antidepressant and broader therapeutic effects. In addition, the paper emphasises that the intensity of mystical-type or psycho-spiritual experiences occasioned during dosing sessions partially mediated antidepressant and anxiolytic outcomes in the parent trials, and qualitative follow-up work suggests such subjective experiences can reduce existential distress and fear of death. Strengths cited include pooling near-identical Phase II, randomised, placebo-controlled crossover trials with high retention through six months and concordant results across sites. The authors also acknowledge multiple limitations: the pooled sample was demographically homogenous (predominantly white), there was a high degree of functional unblinding (therapists could usually guess treatment condition), nearly half the sample had prior psychedelic experience, and the analysis was post-hoc rather than pre-specified. These factors could introduce expectancy or placebo effects and limit generalisability. The investigators also concede that because psilocybin was administered together with psychotherapy, the relative contribution of the medication versus the talk therapy cannot be disentangled in this analysis. To address these limitations in future work, the authors recommend trials employing more robust psychoactive controls (for example stimulants or cannabinoids), measuring expectancy biases, using independent outcome raters and considering factorial designs to separate medication from psychotherapeutic effects. They conclude that larger, prospective trials with multidimensional psychiatric assessments are needed to validate the broad-spectrum therapeutic potential of PAP in patients with cancer.

This pooled analysis indicates that psilocybin-assisted psychotherapy reduces a wide range of psychiatric symptoms in patients with cancer. The authors state that larger clinical trials will be required to confirm these findings, but suggest that PAP has potential to serve as a comprehensive mental health treatment for this population.