Cancer treatment recommendations from molecular profile (mutations + cancer type + biomarkers) — FDA-approved + investigational therapies, resistance mechanisms, matching clinical trials, prognosis. Uses CIViC, ClinVar, OpenTargets, ClinicalTrials.gov. Use for tumor-board treatment recommendations,

Precision Oncology Treatment Advisor Provide actionable treatment recommendations for cancer patients based on their molecular profile using CIViC, ClinVar, OpenTargets, ClinicalTrials.gov, and structure based analysis. Domain Reasoning Treatment selection follows a strict evidence hierarchy: FDA approved for this specific mutation in this cancer type ranks highest, followed by approval for this mutation in any cancer (tumor agnostic), then active clinical trials, and finally off label use. Skipping this hierarchy to recommend off label therapies when an approved option exists is a clinical error. Always check current NCCN guidelines and recent literature, as approvals change rapidly — a drug that was investigational last year may now be first line. When looking up treatment for a specific mutation, search CIViC and OncoKB FIRST, not PubMed. These databases have curated evidence levels. PubMed is for when curated databases don't have the answer. Treatment Selection Reasoning Biomarker to drug logic — When a biomarker is identified, the first line targeted therapy follows established mappings. Always verify current approval status via OncoKB/CIViC, but use this as a starting framework: NSCLC : EGFR exon 19 del / L858R → osimertinib (1L); ALK fusion → alectinib/lorlatinib; ROS1 fusion → crizotinib/entrectinib; KRAS G12C → sotorasib/adagrasib; MET exon 14 skip → capmatinib/tepotinib; RET fusion → selpercatinib; BRAF V600E → dabrafenib+trametinib; NTRK fusion → larotrectinib/entrectinib (tumor agnostic) Breast : HER2+ → trastuzumab+pertuzumab (1L), T DXd (2L); HR+/HER2 → CDK4/6i (palbociclib/ribociclib) + AI; BRCA1/2 mut → olaparib/talazoparib; PIK3CA mut → alpelisib+fulvestrant Colorectal : BRAF V600E → encorafenib+cetuximab; MSI H/dMMR → pembrolizumab (tumor agnostic); KRAS/NRAS wild type → cetuximab/panitumumab (anti EGFR) Melanoma : BRAF V600E/K → dabrafenib+trametinib or encorafenib+binimetinib; wild type → immunotherapy (nivolumab+ipilimumab) Tumor agnostic : MSI H/dMMR → pembrolizumab; NTRK fusion → larotrectinib; TMB H ( =10 mut/Mb) → pembrolizumab; RET fusion → selpercatinib Resistance mechanism reasoning — When a patient progresses on targeted therapy, distinguish primary resistance (never responded — check if the mutation was truly the driver, or if co mutations like TP53/RB1 abrogate response) from acquired resistance (responded then progressed — on target mutations or bypass activation). Common patterns: EGFR TKIs : 1st/2nd gen resistance → T790M (50 60%); osimertinib resistance → C797S (10 25%), MET amp (15 20%), HER2 amp, histologic transformation (SCLC ~5%) ALK TKIs : crizotinib resistance → ALK secondary mutations (L1196M, G1269A); alectinib resistance → G1202R (solvent front); lorlatinib resistance → compound mutations BRAF inhibitors : MAPK reactivation (MEK mutations, BRAF amplification, NRAS mutations), PI3K/AKT bypass Anti HER2 : HER2 truncation (p95HER2), PIK3CA activation, HER3 upregulation Immunotherapy (anti PD1) : B2M loss (MHC I loss), JAK1/2 loss of function (IFN gamma signaling escape), WNT/beta catenin activation (T cell exclusion) For resistance workup: query civic search evidence items with the drug name + "resistance", then PubMed search articles for recent mechanisms. LOOK UP DON'T GUESS FDA approval status for a mutation drug pair: query OncoKB annotate variant and civic search variants ; never assume approval status from memory. Active clinical trials: search search clinical trials with the specific condition and mutation; do not cite trials from memory. Resistance mechanisms for specific drugs: query civic search evidence items and PubMed search articles ; do not assume resistance pathways. Variant frequency in TCGA: retrieve from GDC get mutation frequency or cBioPortal get mutations ; do not estimate prevalence. KEY PRINCIPLES : 1. Report first Create report file FIRST, update progressively 2. Evidence graded Every recommendation has evidence level 3. Actionable output Prioritized treatment options, not data dumps 4. Clinical focus Answer "what should we do?" not "what exists?" 5. English first queries Always use English terms in tool calls (mutations, drug names, cancer types), even if the user writes in another language. Only try original language terms as a fallback. Respond in the user's language When to Use "Patient has [cancer] with [mutation] what treatments?" "What are options for EGFR mutant lung cancer?" "Patient failed [drug], what's next?" "Clinical trials for KRAS G12C?" "Why isn't [drug] working anymore?" Phase 0: Tool Verification Tool WRONG CORRECT civic get variant variant name variant id (numeric, e.g., 4170) civic get evidence item variant id id (numeric) OpenTargets ensemblID ensemblId (camelCase) search clinical trials disease condition Workflow Overview Key Tools by Phase Phase 1: Profile Validation MyGene query genes Resolve gene to Ensembl ID UniProt search Get UniProt accession ChEMBL search targets Get ChEMBL target ID Phase 2: Variant Interpretation civic search variants / civic get variant CIViC evidence COSMIC get mutations by gene / COSMIC search mutations Somatic mutations GDC get mutation frequency / GDC get ssm by gene TCGA patient data GDC get gene expression / GDC get cnv data Expression and CNV GDC get survival Kaplan Meier survival data by project and optional gene mutation filter GDC get clinical data TCGA clinical metadata (stage, vital status, treatment, demographics) Progenetix cnv search Copy number variation biosamples by genomic region and cancer type (NCIt code) DepMap get gene dependencies / PharmacoDB get experiments Target essentiality OncoKB annotate variant / OncoKB get gene info Actionability cBioPortal get mutations / cBioPortal get cancer studies Cross study data HPA search genes by query / HPA get comparative expression by gene and cellline Expression Phase 2.5: Tumor Expression CELLxGENE get expression data / CELLxGENE get cell metadata Cell type expression Phase 3: Treatment Options OpenTargets get associated drugs by target ensemblID Approved drugs (param: ensemblId , camelCase) DGIdb get drug gene interactions Drug gene interactions (param: genes as array, e.g., ["EGFR"] ). Comprehensive; covers inhibitors, antibodies, and investigational agents. DailyMed search spls FDA label details ChEMBL get drug mechanisms Drug mechanism Phase 3.5: Pathway & Network kegg find genes / kegg get gene info KEGG pathways reactome disease target score Reactome disease relevance intact get interaction network Protein interactions Phase 4: Resistance Analysis civic search evidence items Search by known resistance mutations individually (e.g., molecular profile="EGFR C797S" , molecular profile="MET Amplification" ). The significance field in results indicates Resistance/Sensitivity — filter on it after retrieval. PubMed search articles Resistance literature (e.g., "osimertinib resistance C797S combination therapy") alphafold get prediction / get diffdock info Structure based analysis (AlphaFold for structure, DiffDock for docking) Phase 5: Clinical Trials search clinical trials Find trials (param: condition , NOT disease ) get clinical trial eligibility criteria Eligibility details Phase 5.5: Safety & Pharmacogenomics (MANDATORY — do NOT skip) You MUST call FAERS for the leading approved drug before finalizing the report. A clinical brief without real world adverse event data is incomplete. FAERS search adverse event reports — REQUIRED : call with medicinalproduct="<drug name " for at least the top 1 2 approved drugs. Report top 10 serious AEs + death count. FDA get warnings and cautions by drug name — REQUIRED : boxed warnings + key precautions. FAERS count death related by drug Mortality signal for a drug CPIC list guidelines Check for relevant PGx guidelines (e.g., DPYD for fluoropyrimidines in chemo regimens, UGT1A1 for irinotecan). No CPIC guidelines exist for EGFR TKIs. fda pharmacogenomic biomarkers FDA labeled PGx biomarkers for the drug OncoKB demo mode : Without ONCOKB API TOKEN env var, OncoKB only covers BRAF, TP53, ROS1. For other genes (EGFR, KRAS, ALK, etc.), set the API key or use CIViC as the primary evidence source. Phase 6: Literature PubMed search articles Published evidence (use limit , mindate , maxdate for date filtering) BioRxiv list recent preprints / MedRxiv get preprint Preprints (flag as NOT peer reviewed) openalex search works Citation analysis Cross Skill References For CYP interaction with cancer drugs, run: python3 skills/tooluniverse drug drug interaction/scripts/pharmacology ref.py type cyp substrate drug drugname References [TOOLS REFERENCE.md](TOOLS REFERENCE.md) Complete tool documentation with parameters and examples [API USAGE PATTERNS.md](API USAGE PATTERNS.md) Detailed code examples for each phase [TREATMENT ALGORITHMS.md](TREATMENT ALGORITHMS.md) Evidence grading, treatment prioritization, cancer type mappings, DepMap interpretation [REPORT TEMPLATE.md](REPORT TEMPLATE.md) Report template with output tables [EXAMPLES.md](EXAMPLES.md) Worked examples (EGFR NSCLC, T790M resistance, KRAS G12C, no actionable mutations) [CHECKLIST.md](CHECKLIST.md) Quality and completeness checklist

tooluniverse-precision-oncology