ESMO Congress 2026

First, clarify what the tag refers to: in the published literature, #MELSM corresponds to the “mixed-effects location scale model” family of statistical methods, which models individual means (location) and within-person variance (scale) simultaneously and has been repeatedly applied to intensive longitudinal data — nonlinear learning trajectories in 333 individuals ^[1], daily affect ratings from 274 couples over up to 90 consecutive days ^[2], and daily trait-state ratings from 194 individuals ^[3]; extended to categorical moderator tests in meta-analysis, simulations showed that with imbalanced sample sizes and unequal between-study variances the Type I error of the equal-variance mixed-effects model can be grossly inflated while the MELSM loses little power ^[4]; recent methodological work further drew the boundaries: neglecting heteroscedasticity in LMMs causes loss of coverage and biases the estimates of the standard deviations of the random effects, and in MELSMs scale misspecification does not bias the location model although location misspecification alters scale estimates ^[5]; the framework was also used to extend generalizability theory, with EEG data showing that failing to account for variance heterogeneity can mask differences in measurement quality ^[6]. It must be stated that all of this evidence sits in psychology and statistical methodology, none of it touches oncology or intralesional therapy, and it therefore cannot serve as class-level reference for the content posted from this meeting.

There were two live posts from #ESMO26: @anmwongNZ wrote “Size does matter (for intralesional injections/TIL)” and called the talk by @amandakirane at @StanfordMed “Really informative”; the post from @mtmdphd contained only the two hashtags #melsm #ESMO26 with no substantive content. In other words, the only extractable new message from the meeting is one clinical assertion — that in intralesional injections and TIL (tumor-infiltrating lymphocyte) therapy, size does matter; which study the claim comes from, which tumor type, how many patients, and any efficacy or safety numbers were all not disclosed in the live posts.

In terms of incremental value, the post explicitly puts “size” into the discussion of intralesional injections and TIL, suggesting it may influence administration- or efficacy-related decision-making; but the post does not state the basis of the claim (retrospective data, pooled analysis, or expert opinion all remain to be confirmed), nor any quantified result. Because the background literature says nothing directly about intralesional injections or TIL, the two sides share no comparable facts, and there is no way to assess “what step forward versus existing knowledge” — the methodological literature under the #MELSM tag and the live clinical content are two different things sharing one name and cannot corroborate each other.

The limitations are equally clear: this is a single attendee’s subjective relay (“Really informative” is a personal assessment), with no trial name, no numbers, no abstract number, and the full talk title and data source both remain to be confirmed; moreover, methodological and clinical content are mixed under this tag, so the current information strength is insufficient to support any R&D- or BD-level conclusion, and verification should wait until formal materials are disclosed.