ABIM Med Oncology common mistakes
ABIM Med Oncology · Understand the traps and how to avoid them.
When a stem describes a patient with newly diagnosed early breast cancer and asks for the next step, the candidate jumps to the newest adjuvant trial regimen before stating the receptor status-driven decision (endocrine therapy for HR-positive/HER2-negative, HER2-directed therapy for HER2-positive, chemotherapy for triple-negative), and applies a regimen to the wrong subtype.
Breast algorithms are memorized as drug lists rather than as branches keyed to ER, PR, and HER2, so the candidate pattern-matches on the word 'adjuvant' and skips the receptor triage step that the question is actually testing.
What to do instead
Before answering any breast question, explicitly write the three receptor results and the menopausal status, then state the branch (endocrine vs HER2-directed vs chemo) before naming any drug; practice by reconstructing the decision tree from a blank page.
On thoracic questions, the candidate stages and treats a non-small cell lung cancer or head and neck cancer patient without first asking whether the required biomarker or functional-status data are available (EGFR, ALK, ROS1, KRAS G12C, PD-L1, p16 for oropharyngeal), and proceeds to chemotherapy when targeted therapy or immunotherapy would be indicated.
The reflex is to move from 'metastatic NSCLC' straight to platinum doublet, because that was the answer before biomarker testing became standard; the candidate under-weights a single line in the stem reporting a mutation or PD-L1 score.
What to do instead
Make 'what molecular and immune markers are reported?' a mandatory first question for every advanced NSCLC and head/neck stem; if the stem omits them, the correct answer is usually to obtain testing rather than to start empiric cytotoxic chemotherapy.
In gastrointestinal oncology, the candidate treats metastatic colorectal cancer as a single entity and does not check RAS/BRAF status and microsatellite/MMR status before choosing an anti-EGFR antibody or immunotherapy, or treats metastatic pancreatic or gastric cancer without checking HER2, MSI, or germline BRCA where relevant.
Colorectal regimens are learned as FOLFOX/FOLFIRI backbones with a monoclonal antibody bolted on, and the biomarker gate that determines which antibody (or whether immunotherapy replaces chemotherapy) is treated as an afterthought.
What to do instead
For every GI metastasis stem, list the four required data points (side of primary, RAS, BRAF, MMR/MSI) before selecting therapy; drill the rule that anti-EGFR agents are inappropriate in RAS- or BRAF-mutant disease and that MSI-high disease goes to checkpoint blockade.
In genitourinary oncology, the candidate applies the same castration-resistance paradigm to prostate cancer without distinguishing hormone-sensitive from castration-resistant disease, or treats metastatic renal cell carcinoma and urothelial carcinoma without recognizing that the modern standard is combination immunotherapy-based therapy rather than single-agent sunitinib or single-agent chemotherapy.
Older single-agent paradigms are overlearned and the candidate does not anchor on the specific disease state (castration status, treatment line, prior checkpoint exposure) that the stem is signaling.
What to do instead
Before answering, write the disease state in one phrase (e.g., 'mHSPC', 'mCRPC post-docetaxel', 'first-line metastatic clear cell RCC', 'platinum-eligible metastatic urothelial') and only then recall the regimen; build a one-page state-to-regimen map per GU disease.
In gynecologic and other solid tumors, the candidate treats ovarian cancer as one disease and does not separate epithelial from germ cell or sex cord-stromal histology, does not check BRCA/HRD status for maintenance PARP inhibitor decisions, and does not recognize that gestational trophoblastic neoplasia is highly chemosensitive and cured with single-agent or multiagent chemotherapy rather than surgery-first management.
Ovarian cancer is filed under a single 'carboplatin/paclitaxel plus debulking' template, so histology-specific and biomarker-specific branches, and the distinct natural history of germ cell and trophoblastic tumors, are collapsed into that template.
What to do instead
For each gynecologic stem, name the histologic category first and then the biomarker (BRCA/HRD, and for trophoblastic disease the WHO risk score); memorize that low-risk GTN is single-agent methotrexate or actinomycin-D and high-risk GTN is multiagent EMA-CO.
In hematologic malignancies, the candidate assigns a regimen to a leukemia or lymphoma without first establishing the lineage and risk category (AML vs ALL, APL as a medical emergency requiring ATRA, Hodgkin vs non-Hodgkin, diffuse large B-cell vs indolent, Burkitt as a high-risk emergency with tumor lysis prophylaxis), and misapplies a chronic lymphocytic leukemia or multiple myeloma regimen to the wrong line or fitness group.
Hematology is memorized as named regimens (R-CHOP, 7+3, FCR, VRd) without the diagnostic and risk-stratification scaffolding that determines which regimen is correct, so the candidate matches on the acronym and ignores the lineage, cytogenetics, and fitness qualifiers in the stem.
What to do instead
For every heme stem, write lineage, subtype, risk/cytogenetic group, and patient fitness before naming therapy; specifically drill APL (ATRA plus arsenic, coagulopathy, do not wait) and Burkitt (urgent, tumor lysis prophylaxis) as emergency-recognition items.
On cancer biology, genetics, and precision oncology questions, the candidate confuses germline with somatic findings, misidentifies the inheritance pattern and gene for a hereditary syndrome (BRCA1/2 hereditary breast-ovarian, Lynch/MMR, Li-Fraumeni/TP53, APC/FAP, MEN1, RET/MEN2, VHL), or interprets a variant of uncertain significance as actionable and recommends risk-reducing surgery.
Syndrome names, genes, and tumors are learned as lists without the pairing logic, and the candidate does not apply the distinction that a VUS is not a pathogenic finding and should not drive management.
What to do instead
Build a two-column gene-to-syndrome-to-tumor table and recite it, and adopt the rule: if the report says variant of uncertain significance, the answer is genetic counseling and not altered surgery or targeted therapy; distinguish germline testing (blood, heritable, family implications) from tumor-only sequencing.
On diagnostic oncology and pathology questions, the candidate accepts a diagnosis from a reported imaging finding or a single stain without confirming the tissue diagnosis and the required immunohistochemical or molecular profile, and does not recognize when the correct next step is a repeat or core biopsy rather than empiric treatment (for example, a suspected recurrence or a new lesion in a patient with prior cancer).
The candidate treats the pathology result as background rather than as the decisive data, and skips the principle that a new or recurrent lesion generally requires tissue confirmation before a change in therapy, especially when a second primary or a transformation is possible.
What to do instead
For every diagnostic stem, ask 'is the tissue diagnosis established, and does the reported marker panel match the suspected entity?'; drill the rule that suspected recurrence or a discordant clinical course calls for biopsy with full marker re-profiling before switching treatment.
On antineoplastic therapy and supportive care questions, the candidate manages a specific toxicity with the wrong intervention or the wrong dose-modification decision, for example treating febrile neutropenia with oral antibiotics in a high-risk patient instead of urgent cultures and empiric intravenous antipseudomonal therapy, mismanaging immune-related adverse events (holding checkpoint inhibitor but omitting corticosteroids for grade 2 or higher), or failing to apply the correct antiemetic prophylaxis for the emetogenicity of the regimen.
Supportive care is studied as a list of side effects rather than as graded, protocol-driven management, so the candidate recalls the toxicity but not the grade-based action, the time course, or the drug-specific antidote.
What to do instead
Learn toxicities by grade with a paired action (continue, hold, dose-reduce, discontinue, add steroid, add antidote), and memorize the high-yield emergencies: febrile neutropenia, tumor lysis syndrome, extravasation, irAE steroid rules, and the emetogenicity-based antiemetic triple.
On survivorship, palliative care, and ethics questions, the candidate recommends further cytotoxic therapy or a phase I trial in a patient with poor performance status and a stem that signals a goals-of-care discussion, or fails to recognize the ethical and communication imperatives (capacity, informed consent, advance care planning, hospice eligibility, and early palliative care integrated with oncologic treatment rather than as a last resort).
The candidate treats these questions as soft content and reverts to the disease-directed reflex, missing that the stem's performance status, symptom burden, and stated goals are the actual decision drivers, and under-recognizes that early palliative care improves quality of life and is not equivalent to giving up.
What to do instead
When a stem reports poor performance status, heavy symptom burden, or explicit patient goals, make the goals-of-care conversation and symptom-directed plan the default answer; review capacity and consent criteria and the evidence that early concurrent palliative care is standard, not a fallback.