CMD common mistakes
CMD · Understand the traps and how to avoid them.
Treating the prescribed dose as the normalization point and reading isodose lines as percentages of prescription, instead of first confirming whether the plan was normalized to a point, to a mean target dose, or to a coverage criterion, then converting percentages accordingly.
Candidates memorize 'the 95% line should cover the PTV' as a rule and skip the step of asking what 100% represents in that particular plan, so percentage-to-dose conversions are silently wrong on every subsequent calculation.
What to do instead
Before answering any isodose or coverage question, write down the normalization convention in one line (e.g., '100% = 70 Gy at isocenter' or '100% = mean PTV dose'), then convert every percentage to absolute dose using that line. Practice by taking three plans with different conventions and computing the same coverage metric on each.
Using the wrong reference depth for percentage-depth-dose and tissue-maximum-ratio calculations, for example applying TMR at dmax when the question specifies a different source-to-surface distance, or ignoring that TMR is defined at a fixed SAD while TPR varies.
The formulas look interchangeable, and candidates conflate PDD (SSD-dependent, includes inverse-square from the surface) with TMR/TPR (SAD-based, no inverse-square term). The error is invisible unless the candidate checks the definition of each quantity.
What to do instead
Build a one-page comparison table for PDD, TMR, TPR, and TAR listing: reference geometry, whether inverse-square is included, and the field-size definition. Recompute one hand example for each from first principles, and always state the geometry in the answer before plugging in numbers.
Computing monitor units without applying the full chain of correction factors in the right order, typically omitting the inverse-square or SSD factor, the tray/block factor, or the wedge factor, or applying output factors for the wrong field size.
Textbook MU formulas are presented as a single product, so candidates who have only memorized the equation drop terms that are not explicitly named in the question stem, especially when the stem mentions blocks or a tray in passing.
What to do instead
Adopt a fixed checklist for every MU calculation: prescribed dose, depth factor, field-size/output factor, inverse-square or SSD factor, wedge factor, tray/accessory factor, then divide. Write the checklist out at the start of the exam and tick each term on every problem, including a 'factor = 1.0' entry when a term is absent.
Confusing absorbed-dose calibration constants and traceability, such as mixing up N_D,w, N_gas, and the meaning of k_Q, or applying a Co-60 calibration factor to a high-energy beam without the beam-quality conversion factor.
The protocol symbols look similar and candidates learn them as isolated definitions rather than as a chain from the calibration laboratory to the clinical beam. Questions that test the chain expose the gap.
What to do instead
Draw the calibration chain from primary standard to clinical reference condition and label each symbol and its units on the diagram. For each symbol, write one sentence on what changes if it is omitted, then verify against the current national protocol (e.g., TG-51 or TRS-398) definitions rather than memory.
Planning or evaluating a plan without checking whether the dose grid, calculation algorithm, or heterogeneity correction used matches the clinical standard for that site, so a pencil-beam result in lung is accepted as if it were a type-b calculation.
Candidates focus on the numbers (coverage, hot spots) and not on the provenance of the numbers. Algorithm limitations are rarely tested directly, so the assumption goes unchallenged.
What to do instead
For every plan-evaluation question, first name the algorithm, grid size, and heterogeneity setting you are assuming, and note whether that is appropriate for the tissue density involved. Keep a short reference list of which algorithm classes are acceptable for lung, bone, and air cavities.
Judging plan quality from dose-volume histogram summary statistics alone without checking spatial dose distribution, so a plan with acceptable D95 but a cold spot inside the target or a hot spot in a serial organ is passed.
DVH metrics are easy to quote and easy to grade, so candidates default to them and forget that a DVH hides where the dose goes. Serial structures and target cold spots are exactly where the hidden information matters.
What to do instead
For any plan-acceptability question, answer in two passes: first the DVH metrics against constraints, then a spatial check naming at least one slice-level concern (cold spot location relative to target center, hot spot near a serial organ, dose in a previously irradiated region). State both explicitly.
Applying population-based tolerance tables to organs at risk without adjusting for fractionation, prior dose, or whether the constraint is a hard maximum, a mean, or a volume-based limit, and treating all listed numbers as equally binding.
Constraint tables are memorized as flat lists, and candidates do not distinguish 'do not exceed' from 'keep the volume below this dose above this threshold'. The distinction changes the clinical decision.
What to do instead
Rewrite each constraint you memorize in full sentence form including the metric (e.g., 'maximum dose to 0.03 cc of the spinal cord'), and tag each as hard limit, planning aim, or trade-off. When a stem gives a prior-dose history, explicitly recompute the cumulative limit before judging.
Setting up a treatment field or isocenter geometry without verifying that the beam arrangement actually covers the target in the beam's-eye view, for example accepting a gantry/couch combination that places the target outside the collimated field at one angle.
Candidates verify coverage numerically and assume geometry follows. Rotational and non-coplanar setups are hard to visualize, and a single bad angle is invisible in a summed dose distribution.
What to do instead
For each beam, sketch or mentally project the target and organs into the beam's-eye view at that gantry and couch angle, and check the collimator rotation and field margins. On non-coplanar plans, list each couch angle and confirm the target projection within the field before moving to dose evaluation.
Treating image-guidance and setup-margin decisions as separate from the planning margin, so the CTV-to-PTV expansion is chosen without reference to the immobilization, imaging frequency, and observed setup error for that site.
Margins are taught as recipe numbers per site, and the link between measured setup uncertainty and margin size is learned abstractly. Candidates then apply a standard expansion even when the stem describes a different immobilization or imaging protocol.
What to do instead
For every margin question, state the setup uncertainty and the imaging assumption first, then derive or justify the margin from those inputs. Practice by taking one site and computing the margin under two different imaging frequencies to see how the answer changes.
Reporting or verifying a treatment parameter (dose per fraction, total dose, machine output, or accessory setting) without an independent second check, and accepting a value because it matches the plan printout rather than because it was independently recalculated.
In a timed exam, the fastest path is to copy the number from the plan, and the habit carries into practice. Independent checks feel redundant when the first number looks plausible.
What to do instead
Adopt a rule that every reported parameter gets a second, differently-derived value (hand calculation, second formula, or unit conversion round-trip), and write both numbers before choosing. In exam questions, do the arithmetic twice by different routes and flag any discrepancy rather than rounding it away.