PRI results: Cancer Genomics and Diagnostics (T11287) — Class of 2025 and Class of 2026 cohorts

Frozen JSON outputs from the PRI (Papers Ranked by Impact) pipeline for OpenAlex topic T11287 (Cancer Genomics and Diagnostics). The deposit supports transparent, reproducible bibliometric analysis of how influential papers and extracted research concepts change across publication cohorts in cancer genomics and diagnostics. **Ranking methodology.** Papers are ranked by citations from citing papers published within 18 months after each target paper's publication date, not by lifetime citation count. This fixed post-publication window allows fair comparison of early scientific impact across papers of different ages. Rankings are computed with an exact threshold-expansion algorithm over OpenAlex bibliographic data. Each output file includes a full audit trail: OpenAlex query parameters, the LLM evaluation pool, per-paper topic-relevance decisions, and the 18-month citation counts used for ranking. **Cohorts included.** Two publication cohorts are archived. *Class of 2025* covers papers published in 2023; *Class of 2026* covers papers published in 2024. For each cohort, the top papers by 18-month citations were retained after LLM filtering for relevance to Cancer Genomics and Diagnostics (topic T11287). **Files in the archive.** Per cohort year (`2023/`, `2024/`): `top_18m_llm_filtered.json` (ranked papers, metadata, evaluation pool, and LLM filter audit) and `paper_concepts.json` (concepts extracted from titles and abstracts). Cross-cohort normalization files under `2024/` include `canonical_mapping.json`, `unique_concepts.json`, `paper_concepts_normalized.json`, `parent_mapping.json`, `concept_mapping.json`, and `normalized_concepts.json` for canonical and parent-theme taxonomy used in year-over-year comparison. Under `2024/concept_trends/`, `canonical_concept_filtered.json` and `parent_concept_filtered.json` record LLM labels (`pass_filter`) for the top canonical and parent-theme concepts, with frequency columns for cohort comparison. **Provenance.** This methodology and dataset were developed by PRI, an open-science bibliometrics initiative that publishes transparent, reproducible leaderboards and supporting data for OpenAlex research topics. PRI is supported by Pepkio, which develops bioinformatics software and analysis services for life science researchers. Further information: [PRI project](https://pri.pepkio.com), [Pepkio](https://pepkio.com). Companion code, algorithm documentation, and reproduction instructions: [pri-top-papers on GitHub](https://github.com/pepkio/pri-top-papers).