Toxicity Outcomes in Phase 1 Lymphoma Trials: Variable Reporting with High Use of Minimising Language in Published Abstracts at International Conferences

Introduction:

Early phase studies aim to establish the safety of new therapies. Adequate, complete reporting of toxicity outcomes is paramount, with abstracts frequently the main communication channel. Despite this, presentation of phase 1 (Ph1) trial toxicity is highly variable, and minimising language is frequently used, with no criteria for what should be reported or what defines ‘acceptable’ toxicity.

Methods:

Toxicity outcomes from published Ph1 lymphoma trial abstracts from 7 international conferences between 2022 - 2024 (2x ASH, EHA, ASCO; 1x ICML) were analysed including reporting methods, grades, rates and language used for assessment.

Results:

Ph1 trials identified numbered 326 (ASH=147; EHA=84; ASCO=52, ICML=43). The majority (81%) studied B-NHL; 27% DLBCL; 8% Hodgkins and 11% T-cell lymphoma. Most (95%) enrolled relapsed/refractory disease. First-in-human (FiH) / Ph1 or Ph 1b accounted for 53%; 47% were Ph 1/2. Industry sponsored 72% of studies and 68% were preliminary or interim analyses. The most common investigational products (IP) were small molecule inhibitors (42%), bispecific antibodies (20%), and cellular therapies (12%).

Studies were assessed as to whether they reported the following outcomes: all grade (AG) adverse events (AEs) (64%); grade 3 or higher (G3+) AEs (63%); both AG & G3+ AEs (42%); AEs of special interest (AESI) (38%); deaths (43%) and Serious AEs (SAEs) (20%). All outcomes were reported in 4%, while 5% of studies reported none.

In those reporting AG AEs, 16% reported total number; 27% reported events occurring above a set frequency; and 62% reported most common. In those reporting G3+, 35% reported total number, 16% reported those occurring above a set frequency, and 42% reported most common.

It was specified in 68% if toxicities were treatment-related or treatment-emergent. Dose-limiting toxicities were described in 67% specifying a dose escalation, with separation by dose level in 13%. Dose discontinuations or reductions were specified in 40% and 17% of studies respectively, of which clinical cause was specified in 46% and 29%, respectively. A toxicity table was included in 15%.

FiH, PhI and Ph Ib studies were more likely to report G3+ toxicity compared to Ph 1/2 (69% vs 56%, p<0.05); but less likely to report deaths (28% vs 48%, p<0.001). Industry sponsored trials were more likely to report AG AEs (68% vs 47%, p<0.005). Global studies were more likely to report AG AEs and AESI (72% vs 54%, p<0.001; 45% vs 20%, p<0.001, respectively). AESI & deaths were reported in more final & long-term analyses compared to preliminary analyses (52% vs 32%, p<0.001; 49% vs 32%, p<0.001, respectively). There was no difference in reporting based on conference or format of presentation.

Cell therapy studies had lower rates of reporting AG AEs, G3+ AEs and SAEs compared to all other IPs (10% vs 70%, p<0.001; 34% vs 67%, p<0.001, 3% vs 38%, p<0.001, respectively). Cell therapy and bispecific antibody studies specified AESIs more than all other IPs (64% vs 24%, p<0.001). CRS (30%) and ICANs (22%) were the most common AESI.

Concluding remarks described safety with minimising terms such as ‘tolerable’, ‘safe’, ‘manageable’ and ‘acceptable’ in 87%. There was no significant difference based on sponsor, timing of analysis or phase of study. Of 188 trials using minimising terms reporting AG AEs, 54% reported >50%; 42% reported >70%; and 24% reported >90% incidence of AG AE. Of 180 trials using minimising terms reporting G3 AE, 31% reported >50%; 13% reported >70%, and 4% > 90% incidence of G3 AEs. AEs leading to dose discontinuation were described in 48%; and 15% reported deaths due to AEs; causality was infrequently defined.

Conclusion:

Toxicity reporting in Ph1 lymphoma trial abstracts showed marked heterogeneity in rate, grade and style of reporting; and was frequently incomplete with low rates of reporting outcomes and defining causality. Minimising terms were used almost universally, despite often high rates of toxicity, dose discontinuation and even deaths. Clearer standardised reporting is required for improved transparency and communication in drug development.