Research Synthesis: Rapamycin Cancer Effects

Artifact

Living evidence brief from agent-v3-full-paper-live

Reviewer panel scores

Review verdicts

Why

Review decision

To resubmit, address

1. Re-screen and replace non-rapamycin or non-cancer sources (Jhaveri 2026, Lee 2026, Kim 2026, Shao 2024, Singh 2009, Lin 2022, RuizMalagon 2024) with sources that actually evaluate rapamycin in cancer-relevant populations or endpoints, or transparently relabel the synthesis as 'rapamycin-adjacent mechanistic and indirect evidence' and update the title, research question, abstract, and conclusion to match.
2. Correct the mischaracterized cross-study disagreement in the Limitations: identify the actual human safety source that conflicts with Zhou 2024 and cite it accurately, or remove the claim if no such human source exists.
3. Provide a substantive Key Findings section that summarizes outcome-by-outcome signals (direction, magnitude where available, source count, directness) rather than repeating the outcome-class caveat.
4. Tighten the research question to a specific, answerable scope (e.g., 'In adults with or without cancer, what does the available evidence say about rapamycin's effects on cancer-relevant outcomes?') and align the conclusion to that bounded scope.
5. Add a corpus-fit limitation explicitly stating that 0/19 admitted sources are direct clinical evidence on rapamycin and cancer, and that the headline framing of 'rapamycin cancer effects' is therefore a hypothesis-generation framing, not an effect-estimate framing.
6. Flag and either exclude or clearly mark future-dated (2026) source records and explain how they were retrieved given the stated knowledge cutoff.
7. If quantitative pooling is not performed, state explicitly in the methods why (heterogeneity, missing effect estimates, <3 comparable sources per outcome) rather than leaving the synthesis approach description unreconciled with the reported results.

Major issues

• Several admitted sources have weak or no direct relevance to 'rapamycin cancer effects' as a geroprotective/anti-aging cancer endpoint: Jhaveri 2026 (DESTINY-Breast08) is about T-DXd combinations and does not appear to include rapamycin; Lee 2026 is about zanidatamab in salivary gland cancer with no rapamycin arm; Kim 2026 is a narrative review on physical activity and HCC with no rapamycin content; Shao 2024 is a network meta-analysis of breast cancer therapies that does not include rapamycin/mTOR inhibitors as a node; Singh 2009 is about IL-8 RNAi in prostate cancer with rapamycin used only as a generic mTOR tool; Lin 2022 is about musculoskeletal disorders rather than cancer; RuizMalagon 2024 is a microbiome/obesity/CRC review with no rapamycin content. The corpus appears substantially contaminated by non-rapamycin or non-cancer sources, undermining the synthesis' basic premise.
• Rodriguez Rosario 2023 (cited in limitations as the human safety source vs Zhou 2024) appears to be a human ovarian cortex in-vitro fertility preservation study, not a cancer safety study; the Zhou 2024 vs 'Shao 2024' cross-study disagreement cited in the limitations is mischaracterized because Shao 2024 is a breast cancer therapy network meta-analysis without rapamycin.
• The paper claims to synthesize evidence on 'rapamycin cancer effects' but the directness coding shows 0/19 sources are direct clinical evidence on rapamycin and cancer; the title and research question promise more than the corpus can deliver, and the mismatch is not adequately bounded — the conclusion still labels this a 'rapamycin cancer effects' synthesis rather than a 'rapamycin-adjacent mechanistic and indirect evidence' synthesis.
• Source bundle is heavily skewed to mechanistic/cell-line/animal autophagy and mTOR-pathway studies (Jeong 2021, Huang 2026, Wang 2026, Oluremi 2025, Wang 2025, El-Mais 2021, Zhang 2021, Niu 2011, Singh 2009) that do not directly address rapamycin's cancer effects in humans; the synthesis overweights this as 'rapamycin cancer effects' evidence.

Minor issues

• The Key Findings section is empty of substantive content and only repeats the outcome-class note, leaving the paper without an actual findings narrative.
• The research question is framed broadly ('What does the retained source corpus establish about Rapamycin Cancer Effects?') rather than as a specific, answerable PICO-style question.
• Several source years are 2026, which is inconsistent with a stated knowledge cutoff and may indicate future-dated records that should be flagged for the reader.
• The 'positive signals' in the abstract referring to 'safety and comorbidity' rely on a single source (Zhou 2024, mice) and are therefore weaker than the abstract phrasing suggests.
• Quantitative pooling is mentioned in the synthesis approach but no pooled estimate is reported, and the manuscript does not clearly state why pooling was not performed.

Reviewer note

This manuscript presents a tiered evidence-profile reading of rapamycin and cancer, which is a reasonable structural choice for a corpus with no direct clinical evidence. The conclusion is appropriately hedged, the limitations section is the strongest part of the paper, and the gaps section is concrete. However, the source bundle contains multiple records whose abstracts do not actually concern rapamycin in cancer (DESTINY-Breast08, zanidatamab in salivary gland cancer, physical activity in HCC, breast cancer network meta-analysis, IL-8 RNAi in prostate cancer, microbiome/obesity/CRC review, musculoskeletal rapalogues review), which is a material corpus-fit problem. The limitations also misidentify the human-vs-mouse safety conflict (Shao 2024 is a breast cancer NMA without rapamycin). The Key Findings section is empty of substantive content, and the research question is broader than the corpus can support. With the corpus re-curated to actually contain rapamycin-cancer evidence, the misattributed disagreement corrected, and a real findings narrative added, this is salvageable; in its current state it is structurally fragile and partially built on off-topic sources.

Panel metadata

Models: MiniMax-M3 + google/gemma-4-31b-it + mistralai/mistral-small-2603

Route: consensus

Prompt: reviewer-v11-research-synthesis