RECIST and CTCAE reporting means stating two things precisely: which version of RECIST you used to measure tumour response, and which version of CTCAE you used to grade adverse events. Name the criterion and version, say who assessed the imaging, give the confirmation interval, and report adverse events per grade with attribution and a stated denominator.
What RECIST and CTCAE each cover, and where the boundary sits
RECIST — Response Evaluation Criteria in Solid Tumours — is a rule set for measuring solid-tumour burden on imaging and converting change from baseline into a response category. CTCAE — the Common Terminology Criteria for Adverse Events, from the US National Cancer Institute — is a graded dictionary of adverse-event terms, graded from 1 to 5, with the grades that apply defined term by term.
The two do not overlap. RECIST answers whether measured tumour burden shrank, grew or held; CTCAE answers what harm occurred and how severe it was. Progression is a RECIST outcome, not a CTCAE adverse event; treatment-related pneumonitis is a CTCAE event, not a response category.
The RECIST 1.1 response categories
| Category | Definition for target lesions in RECIST 1.1 |
|---|---|
| Complete response (CR) | All target lesions disappear; any pathological lymph node reduced to under 10 mm in short axis. |
| Partial response (PR) | At least a 30% decrease in the sum of target-lesion diameters, against the baseline sum. |
| Progressive disease (PD) | At least a 20% increase in the sum, against the smallest sum on study, with an absolute increase of at least 5 mm; or any new lesion. |
| Stable disease (SD) | Neither enough shrinkage for PR nor enough increase for PD. |
Under RECIST 1.1, measurability thresholds are fixed:
- Non-nodal lesion: 10 mm or more in longest diameter on CT with slices of 5 mm or less.
- Lymph node: 15 mm or more in short axis.
- Target lesions: up to five, no more than two per organ.
State the threshold, the modality and the slice thickness, or your target-lesion selection cannot be reproduced.
When a criterion other than RECIST 1.1 applies
| Setting | Criterion that fits the setting |
|---|---|
| Immune checkpoint inhibitors, where pseudoprogression is possible | iRECIST, alongside RECIST 1.1 |
| Hepatocellular carcinoma after locoregional therapy | mRECIST, which measures viable enhancing tumour |
| Lymphoma | Lugano classification, including PET-based metabolic response |
| Primary brain tumours | RANO criteria |
| Multiple myeloma | International Myeloma Working Group criteria |
| Metabolic response on PET | PERCIST |
Name one criterion as primary and say why it fits the disease and treatment.
CTCAE grades severity; it does not decide causality
A CTCAE grade describes how severe an event was. It says nothing about whether treatment caused it. Attribution is a separate judgement, made by a named role and reported separately from the grade.
| Grade | What the grade means in CTCAE |
|---|---|
| 1 | Mild; asymptomatic or mild symptoms; clinical or diagnostic observation only; intervention not indicated. |
| 2 | Moderate; minimal, local or non-invasive intervention indicated; limiting age-appropriate instrumental activities of daily living. |
| 3 | Severe or medically significant but not immediately life-threatening; hospitalisation or prolongation of hospitalisation indicated; disabling. |
| 4 | Life-threatening consequences; urgent intervention indicated. |
| 5 | Death related to the adverse event. |
Attribution scales vary: a five-point scale — definite, probable, possible, unlikely, unrelated — and a collapsed related versus unrelated are both in use. Either is acceptable; leaving it implicit is not. State the scale, who applied it (investigator, safety committee, adjudication panel), and whether it was applied blind to allocation.
The version problem: a response rate is meaningless without a version number
Criteria change, and the numbers change with them. RECIST 1.1 cut the target-lesion maximum from ten to five, changed how lymph nodes are measured, and required progression to carry an absolute increase of at least 5 mm as well as a 20% relative increase. A partial response under RECIST 1.0 is not the same measurement as one under 1.1.
The same applies to toxicity. CTCAE v5.0 was published by the National Cancer Institute in 2017 and superseded v4.03; terms were added, retired and remapped to a newer version of MedDRA, and some grade boundaries moved. The National Cancer Institute then released CTCAE v6.0 in 2025, with a mapping of every v5.0 term and grade to its v6.0 equivalent, so v5.0 and v6.0 are both in circulation and neither can be assumed from the bare word CTCAE. Reporting on v4.03 because the protocol ran on v4.03 is fine — say so. Regrading old records against a newer version is a methods decision that has to be declared.
Write the version into the methods, every table legend, and the trial registration record where one exists. The CONSORT Harms 2022 extension, listed in the reporting-guideline library of the EQUATOR Network, sets out what a harms table carries; the CONSORT, PRISMA and STROBE guidelines cover the wider principle.
Who assessed the response, and over what confirmation interval
Response assessment is a reading, and readings differ between readers. Investigator assessment and blinded central review can produce different numbers, so a paper that does not name which produced the headline response rate has omitted part of its method. Report both where both exist, and name the primary one.
Confirmation is the other half. Under RECIST 1.1, where objective response is the primary endpoint of a non-randomised trial, a complete or partial response is confirmed by a repeat assessment no fewer than four weeks later. Say whether responses were confirmed and at what interval, how often scans were scheduled, and how patients with no post-baseline scan were counted — as non-responders or excluded, which changes the denominator.
A RECIST and CTCAE reporting template to fill in
Fill in both tables before writing the results paragraph. Anything you cannot complete is a missing method or a limitation. The counts in Table 2 are illustrative only, from a notional safety population of 40 patients.
Table 1: eight fields every response section must state
| Field | What to state | Worked example |
|---|---|---|
| Criterion and version | The criterion and its version number | RECIST version 1.1 |
| Secondary criterion | Its name and the analyses it applies to | iRECIST, sensitivity analysis only |
| Imaging technique | Modality, contrast, slice thickness | Contrast-enhanced CT, 5 mm slices |
| Assessment schedule | Interval from first dose, and the window allowed | Every 8 weeks, plus or minus 7 days |
| Assessor | Investigator, central review or both; how discordance was resolved | Investigator-assessed; central review as sensitivity analysis |
| Confirmation interval | Whether CR and PR were confirmed, and after how long | Repeat scan 4 weeks or more later |
| Population and denominator | Who enters the response rate; how missing scans were handled | All treated patients; no post-baseline scan counted as non-response |
| Result | n/N and percentage per category, with a confidence interval | Objective response 12/40 (30.0%, 95% CI 16.6-46.5) |
Table 2: grade and attribute every adverse event
| Adverse event (CTCAE v6.0 term) | Grade 1 n (%) | Grade 2 n (%) | Grade 3 n (%) | Grade 4 n (%) | Grade 5 n (%) | Any grade n (%) | Grade 3 or above attributed to treatment n (%) |
|---|---|---|---|---|---|---|---|
| Neutropenia | 6 (15.0) | 9 (22.5) | 5 (12.5) | 1 (2.5) | 0 | 21 (52.5) | 6 (15.0) |
| Fatigue | 11 (27.5) | 7 (17.5) | 2 (5.0) | 0 | 0 | 20 (50.0) | 2 (5.0) |
| Pneumonitis | 1 (2.5) | 2 (5.0) | 1 (2.5) | 0 | 1 (2.5) | 5 (12.5) | 2 (5.0) |
Four statements belong in that legend: the safety population and its size, the N behind every percentage; that each patient is counted once per term at their worst grade; the CTCAE version; and the attribution scale with the role that applied it. Add rows for discontinuations, dose reductions, dose delays and each grade 5 event.
What changes when the paper is a case series rather than a trial
RECIST and CTCAE do not change for a case series; the denominator does. A case series has no protocol-defined schedule, so state the intervals at which imaging actually happened and say they were clinically driven. Response in five patients is reported as counts, not percentages. Toxicity is graded with CTCAE the same way, version named, and attribution is yours as the treating team, not an adjudication that never happened. See also writing and publishing a case report.
Reporting gaps that generate reviewer queries
- A response rate given as a percentage with no numerator, denominator or confidence interval.
- Toxicity called "well tolerated" or "manageable" with no grade distribution behind the adjective.
- Only grade 3 and above reported, hiding the low-grade burden that can drive dose reduction or discontinuation.
- Grade 5 events folded into overall mortality instead of listed individually with attribution.
- A criterion named but not versioned, or versioned only in the methods.
- Waterfall or spider plots with no underlying numbers, which belong in the paper or in the supplementary materials.
- No statement of who assessed response, and no confirmation interval.
Our Author Guidelines ask you to match the manuscript to the relevant EQUATOR checklist. We publish no oncology-specific criteria policy, so RECIST and CTCAE are not named there — naming your criterion and version is the author's job. Measurement data belong in a data availability statement, and the ICMJE Recommendations ask for data on all the primary and secondary outcomes identified in the methods. See also statistical reporting of p-values and confidence intervals and preparing figures and tables.