RECIST Target Lesions: Document the Baseline Selection Before Follow-Up

Suppose an oncology imaging charter is due before first patient in, and the team is testing its baseline lesion-selection rules against a multi-organ screening CT acquired at 5 mm slice thickness. The scan shows three liver metastases, two pulmonary nodules (one in each lung), an 18 mm short-axis mediastinal lymph node, a 12 mm short-axis axillary lymph node, an 8 mm short-axis external iliac node, and bilateral ovarian deposits. Before a single follow-up scan is read, the protocol and imaging charter must answer four operational questions: which lesions are measurable, which become target lesions, how paired organs and separate nodal chains count against the two-per-organ and five-total limits, and what exact fields must be locked at baseline so every subsequent time-point assessment measures the same disease.

Under RECIST 1.1, baseline lesion selection is not an informal reading-room note that can be revised after Cycle 2. The baseline record fixes measurability, lesion identity, imaging technique, and the baseline sum of diameters (SOD) in a single pass. Every non-nodal lesion with a longest diameter of at least 10 mm on CT and every lymph node with a short axis of at least 15 mm is eligible as a target lesion, up to two target lesions per organ and five target lesions in total across all involved organs. The EORTC RECIST Working Group Q&A treats paired structures—lung, kidney, ovaries, and lymph nodes—as one organ each, irrespective of left/right parts or nodal stations, and the 2016 RECIST committee update states the same rule for lymph nodes. Every remaining measurable lesion excluded by an organ or total cap, together with every pathological non-measurable lymph node (short axis 10 to < 15 mm) and every unequivocal non-measurable disease site, must be logged at baseline on the non-target lesion ledger; nodes with a short axis below 10 mm are non-pathological and are not recorded.

Because partial response is later evaluated against the baseline sum of diameters and progressive disease is evaluated against the smallest sum on study—including the baseline sum when no shrinkage occurs—an incomplete or miscounted baseline inventory silently biases objective response rate (ORR) and progression-free survival (PFS). Treating the baseline inventory as a controlled endpoint record inside the protocol, imaging charter, and case report form (CRF) prevents the organ-cap and response-calculation errors that cooperative-group audits frequently flag.

The 2009 revised RECIST guideline (version 1.1) reduced the maximum baseline target-lesion count from ten lesions (five per organ in RECIST 1.0) to five lesions total and two per organ, while introducing short-axis measurement and explicit size thresholds for lymph nodes. The guideline grounds the lower cap in analyses of a merged trial data warehouse of more than 6,500 patients, which showed no loss of information when the number of assessed lesions was reduced from ten to five. However, the tighter cap makes each baseline slot more consequential: selecting an unrepresentative lesion, measuring a lymph node along its long axis, or omitting overflow measurable lesions from the non-target ledger alters the denominator for every subsequent visit.

At baseline, every candidate finding on diagnostic cross-sectional imaging must be triaged through three sequential criteria before any target slot is assigned:

1. Anatomical measurability and measurement plane: For non-nodal solid lesions, measurability requires a longest diameter of ≥ 10 mm on CT or MRI when reconstructed at a slice thickness of ≤ 5 mm (or ≥ 20 mm on a clearly defined chest X-ray surrounded by aerated lung, though CT is strongly preferred). If CT slice thickness exceeds 5 mm, the 2009 guideline sets the minimum measurable non-nodal size at twice the slice thickness; however, the 2016 RECIST committee clarification strongly recommends CT slices of 5 mm or less, noting that at 10 mm lesions are harder to measure and new lesions are significantly less conspicuous. Question 14 of the same clarification recommends measuring in the axial plane in all cases on CT, because other planes can be difficult to measure consistently and reproducibly across time points.

2. Lymph node short-axis tiering: Unlike solid organ metastases, lymph nodes are always measured along their short axis—the longest diameter perpendicular to the node's long axis on the axial image. A node is measurable and eligible for target selection only when its short axis is ≥ 15 mm. A node with a short axis of 10 mm to < 15 mm is considered pathological (malignant) but non-measurable and must be recorded as a non-target lesion. A node with a short axis < 10 mm is considered non-pathological (normal) and must not be recorded as baseline disease.

3. Representative selection up to two per organ and five total: From all measurable lesions (non-nodal ≥ 10 mm long axis; nodal ≥ 15 mm short axis), the reader selects up to five target lesions in total and at most two per organ. Selected target lesions should be the largest lesions that are representative of all involved organs and lend themselves to reproducible repeated measurement. As the 2009 RECIST 1.1 guideline notes explicitly, if a patient has measurable disease in only one organ site, the baseline record will contain at most two target lesions; if two organ sites are involved, it will contain at most four target lesions. Conversely, per Question 4 of the 2016 committee clarification, a patient with a single measurable lesion and multiple non-measurable lesions still qualifies as having measurable target disease.

Every other lesion or site of disease present at baseline—including measurable lesions beyond the two-per-organ or five-total cap, pathological lymph nodes measuring 10 to < 15 mm in short axis, and truly non-measurable disease such as leptomeningeal disease, ascites, pleural or pericardial effusions, lymphangitic involvement of skin or lung, blastic bone lesions, or lytic and mixed bone lesions without a measurable soft-tissue component—must be identified and recorded on the baseline non-target lesion ledger. Non-target lesions are not measured with calipers at follow-up, but their presence, complete disappearance, or unequivocal progression must be accounted for at every time point.

A frequent operational ambiguity in baseline target-lesion selection—one the working group says it is often asked about—arises when bilateral organs or multiple lymphatic regions are involved. If a patient has two measurable nodules in the right lung and two in the left lung, or measurable metastases in both ovaries, or enlarged ≥ 15 mm short-axis nodes in the mediastinum, axilla, and retroperitoneum, can a reader select two targets from each side or each nodal chain? Under RECIST 1.1, the answer is no—and trial teams need to understand both the substantive rule and the exact document hierarchy behind it.

In the 2016 RECIST committee update and clarification (Schwartz et al., Eur J Cancer 2016), the committee answered two counting questions directly, noting that its warehouse analyses had counted lymph nodes as one organ:

• All lymph nodes across the body constitute a single organ (Question 5): When asked whether individual nodal chains or regions count as separate organs or whether all lymph nodes count as one organ, the committee stated that lymph nodes are considered one organ and that only two lymph nodes should be measured per patient as target lesions, with all other involved lymph nodes assessed and followed as non-target lesions.

• A third measurable lesion in an already-represented organ becomes non-target (Question 3): When three or more measurable lesions exist in a single organ and two are selected as targets, the third and any additional measurable lesions should be considered a non-target lesion and should be recorded and followed as part of the non-target disease—not ignored or held in reserve as replacement targets.

While the 2016 journal paper explicitly resolved the lymph-node organ limit, questions continued to reach the working group about paired parenchymal organs. On the official EORTC RECIST 1.1 criteria and Q&A page (under Questions and clarifications → Target lesions → “Paired” organs, last updated June 2026), the working group states verbatim:

“Questions often arise concerning the rule of selecting maximum 2 target lesions per organ, when it comes to paired organs (i.e. lung, kidney, ovaries, lymph nodes). In our analyses, ‘lung’, ‘kidney’, ‘ovaries’, ‘lymph nodes’ etc are considered as 1 organ, irrespective of the number of parts they may be made up of. Of note, if there are multiple nodal chains/regions, consider selecting one from each, with a maximum of two for this organ. Other nodal lesions can be followed as non-target.”

The diagram below translates these rules into a deterministic baseline lesion-triage workflow that site investigators, central imaging readers, and data managers can execute identically before first follow-up.

flowchart TD
    A["Candidate Baseline Finding on Scan (<= 4 Weeks Before Treatment Start)"] --> B{"Lymph Node or Non-Nodal Site?"}
    B -->|Lymph Node: Measure Short Axis| C{"Short Axis Size on CT?"}
    C -->|"< 10 mm"| D["Non-Pathological (Normal): Do Not Record"]
    C -->|"10 mm to < 15 mm"| E["Pathological Non-Measurable: Record on Non-Target Ledger"]
    C -->|">= 15 mm"| F["Measurable Nodal Candidate (Pool All Nodal Chains as 1 Organ)"]
    B -->|Non-Nodal: Measure Longest Axial Diameter| G{"Solid Lesion >= 10 mm on CT (or >= 20 mm CXR), Outside a Previously Treated Field?"}
    G -->|No, but Unequivocal Malignant Disease| E
    G -->|Yes| H["Measurable Non-Nodal Candidate (Pool Paired Organs: Lung, Kidney, Ovaries as 1 Organ Each)"]
    F --> I{"Organ Pool Already Has 2 Target Lesions? (Prefer 1 Node per Chain)"}
    H --> I
    I -->|Yes: 3rd+ Lesion in Organ Pool| E
    I -->|No: <= 2 in Organ Pool| J{"Selected Among Largest Representative Lesions Across Involved Organs (<= 5 Total)?"}
    J -->|Yes| K["Lock as Baseline Target Lesion (Add to Baseline Sum of Diameters)"]
    J -->|No: Excluded by 5-Lesion Total Cap| E

To make the baseline target/non-target split auditable, the record must capture more than a single summary number. The EORTC RECIST Working Group publishes an official blank Initial Measurement Form — RECIST 1.1 specifically for recording initial lesion measurements at baseline (paired with a separate Follow-up Measurement Form). While the working group's PDF is a convenient template rather than a regulator-mandated form, its existence underscores a core operational principle: baseline lesion selection is a distinct, locked artifact created before longitudinal response assessment begins.

From a regulatory perspective, both FDA and EMA connect baseline imaging documentation directly to endpoint interpretability:

• FDA Clinical Trial Imaging Endpoint Process Standards (April 2018 final guidance): FDA's 2018 final guidance covers process standards for image acquisition, display, archiving, and interpretation that FDA regards as important when imaging assesses a trial's primary endpoint or a component of that endpoint. It does not prescribe a lesion worksheet, so the protocol and imaging charter are where the sponsor should write down how readers select, label, and measure baseline lesions.

• EMA Methodological Considerations for PFS or DFS in Confirmatory Trials (Appendix 1, Rev. 1): In EMA/CHMP/27994/2008/Rev.1, the CHMP specifies that progression definitions must follow established response evaluation criteria such as RECIST and be clearly and prospectively defined in the protocol. Evaluating PFS requires all sites of possible disease specific to the tumor type to be assessed at baseline, involved sites to be systematically assessed at follow-up (ideally using the same imaging methods and identical schedules across study arms), and independent-review procedures to be prospectively defined in trial documentation.

• RECIST 1.1 Protocol Specification Rules: Both the 2009 RECIST 1.1 guideline and Question 1 of the 2016 committee clarification state that the protocol should specify which organ sites (the 2016 paper names areas such as chest and abdomen) are evaluated at baseline and how often evaluations are repeated. Question 1 adds that the method of assessment should be stated and, ideally, repeated using the same imaging technique, equipment, and assessor each time.

Note also what other regulatory artifacts do not provide. Sponsors selecting ORR or PFS for an expedited oncology program often consult FDA's Surrogate Endpoint Table to verify whether an imaging-based endpoint has supported traditional or accelerated approval in a given indication. However, that listing table does not define lesion-level eligibility, paired-organ caps, or baseline worksheet fields. Likewise, locking a charter for central imaging reads or endpoint adjudication requires that local-site and blinded independent central review (BICR) readers apply the exact same baseline organ-pooling and node-tiering rules; otherwise, reader discordance starts at Day 1 because the two readers are summing different baseline lesions.

Whether implemented as an electronic CRF form, a central-read workstation lock screen, or a structured adaptation of the EORTC Initial Measurement Form, a complete baseline RECIST 1.1 record needs seven verifiable field groups:

1. Baseline scan timing and treatment-start anchor: Scan acquisition date, planned or actual first dose date, and verification that the scan occurred ≤ 28 days (4 weeks) prior to treatment start.

2. Modality, acquisition parameters, and contrast status: Modality (e.g., diagnostic CT vs MRI), reconstruction plane (axial), slice thickness (verifying ≤ 5 mm), and intravenous/oral contrast administration so follow-up scans match the baseline technique.

3. Anatomical location and pooled organ key: Both the specific anatomical sub-site (e.g., Right lung lower lobe, Segment 8 or Left ovary) and the pooled RECIST organ key (e.g., Lung, Ovaries, Kidney, Lymph Nodes, Liver) so automated edit checks can enforce the ≤ 2-per-organ cap across paired structures and nodal chains.

4. Series, slice/image number, and prior local-therapy status: Exact DICOM series and axial slice coordinates (so the identical lesion is re-measured even when adjacent lesions shrink or coalesce) plus confirmation that a target candidate does not lie within a previously irradiated field lacking documented post-radiation progression.

5. Measurement axis and baseline millimeter value: Explicit declaration of Longest axial diameter (mm) for non-nodal lesions versus Short axis perpendicular to longest diameter (mm) for lymph nodes.

6. Explicit disposition (Target T1–T5, Non-Target NT, or Not Recorded) with exclusion reason: Why each candidate was assigned to Target versus Non-Target (e.g., 3rd measurable lesion in Liver, Excluded by 5-lesion total cap after representing all involved organs, or Pathological node 10 to < 15 mm short axis).

7. Locked Baseline Sum of Diameters (SOD) and reader attribution: The arithmetic sum (in millimeters) of the selected target lesions (using longest diameters for non-nodal targets and short axes for nodal targets) and the qualified physician or central reader signature and timestamp.

Return to the opening scenario: a screening contrast-enhanced CT of the chest, abdomen, and pelvis (axial plane, 5 mm slice thickness) is obtained 9 days before Cycle 1 Day 1 (within the 28-day window). The qualified reader identifies twelve candidate findings across the liver, both lungs, both ovaries, three lymph-node stations, the iliac bone, and the pelvis:

• Liver (single solid organ; 3 candidates): Segment VII lesion (34 mm longest axial diameter), Segment VI lesion (26 mm longest axial diameter), and Segment IVa lesion (18 mm longest axial diameter). All three meet the ≥ 10 mm CT measurability rule, but the two-per-organ rule allows at most two hepatic targets (34 mm and 26 mm). Per Schwartz et al. (2016) Question 3, the 18 mm Segment IVa metastasis must be recorded as a non-target lesion.

• Lungs (paired organ counted as 1 organ; 2 candidates): Right lower lobe nodule (22 mm longest diameter) and left upper lobe nodule (16 mm longest diameter). Per the Working Group Q&A, left and right lungs together constitute a single Lung organ pool. Both nodules are measurable and within the two-per-organ limit, but their final target assignment depends on the five-lesion total budget across all involved organs.

• Ovaries (paired organ counted as 1 organ; 2 candidates): Right ovarian solid mass (28 mm longest diameter) and left ovarian solid deposit (19 mm longest diameter). Left and right ovaries share a single Ovaries organ pool.

• Lymph Nodes (all nodal stations pooled as 1 organ; 3 candidates): Subcarinal mediastinal node measuring 18 mm short axis × 24 mm long axis (≥ 15 mm short axis → measurable target candidate); right axillary node measuring 12 mm short axis × 17 mm long axis (10 to < 15 mm short axis → pathological but non-measurable → mandatory non-target lesion); and left external iliac node measuring 8 mm short axis × 13 mm long axis (< 10 mm short axis → non-pathological → not recorded).

• Non-measurable disease sites (2 candidates): A sclerotic (blastic) right iliac bone metastasis without a soft-tissue component and a small volume of malignant pelvic ascites. Both are non-measurable under RECIST 1.1 and must be logged as non-target lesions.

Now apply the five-lesion total cap and the representative-of-all-involved-organs rule. After applying the two-per-organ limit to the liver, seven measurable target-eligible lesions remain across four distinct RECIST organs: Liver (34 mm, 26 mm), Ovaries (28 mm, 19 mm), Lung (22 mm, 16 mm), and Lymph Nodes (18 mm short axis). Because RECIST 1.1 asks for the largest lesions that are representative of all involved organs and suitable for reproducible measurement, but does not prescribe a fill order, the example charter states one: the reader first assigns the largest measurable lesion in each of the four involved organs—Liver Segment VII (34 mm), Right Ovary (28 mm), Right Lung (22 mm), and Mediastinal Subcarinal Node (18 mm short axis)—and assigns the fifth and final target slot to the largest remaining eligible lesion that respects the two-per-organ limit: Liver Segment VI (26 mm). The remaining measurable lesions in the Left Ovary (19 mm) and Left Lung (16 mm) are not dropped; they are logged on the baseline non-target lesion ledger alongside the 18 mm third liver lesion, the 12 mm short-axis right axillary node, the blastic bone lesion, and the pelvic ascites.

Summing the five locked target lesions—34 mm (T1) + 26 mm (T2) + 28 mm (T3) + 22 mm (T4) + 18 mm short axis (T5)—yields a locked Baseline Sum of Diameters (SOD) of 128 mm across four involved organs, plus six documented non-target disease entries (NT1–NT6). Notice that if the reader had mistakenly summed the subcarinal node's 24 mm long axis instead of its 18 mm short axis, or had selected the 12 mm axillary node as a second nodal target, or had treated the right and left ovaries and right and left lungs as four separate organs, the baseline SOD and the non-target ledger would be corrupted before the patient received the first dose.

Once the baseline record is locked and study treatment starts, RECIST 1.1 treats target-lesion identity and measurability as one-way commitments. Six longitudinal rules explain why a team cannot retroactively swap target lesions or reclassify baseline findings mid-study:

1. The four-week baseline window cannot be stretched post hoc: Per the 2009 RECIST 1.1 guideline, baseline imaging evaluations must be performed as close as possible to the beginning of treatment and never more than four weeks (28 days) before the start of treatment. If dosing slips beyond 28 days after the screening scan, that scan no longer meets the rule, so the protocol should require a repeat baseline scan and a re-locked inventory before the first dose.

2. Measurability is defined at baseline, and actual measurements continue below 10 mm: A target lesion that shrinks below 10 mm (or a target lymph node whose short axis shrinks below 15 mm or even below 10 mm) does not become non-measurable during follow-up. Both the 2009 guideline and Question 2 of the 2016 committee clarification require the radiologist to record the actual millimeter measurement at every time point whenever the lesion can still be accurately measured—even if it is smaller than 5 mm. For target lymph nodes, short-axis values below 10 mm are still recorded and included in the time-point SOD; as the 2009 guideline notes, the sum may therefore not be zero even when complete-response criteria are met (for example, a target node that has regressed to 6 mm short axis).

3. The 5 mm default and 0 mm rules govern indistinct or resolved target lesions: When a target lesion shrinks until it is still faintly visible on CT but too small or ill-defined for a reproducible caliper measurement (“too small to measure”), the reader assigns a default value of 5 mm (nominally derived from the 5 mm CT slice thickness and never altered for different scanner slice settings). When a target lesion is no longer visualized, it is recorded as 0 mm (Eisenhauer et al. 2009; Schwartz et al. 2016).

4. Coalesced lesions follow plane-of-separation and axis rules: If two baseline lesions coalesce during follow-up, both the Working Group Q&A and Question 7 of the 2016 clarification instruct readers to check whether a plane of separation remains visible so each lesion can still be measured individually. If the lesions have truly coalesced and are no longer separable, non-nodal lesions are measured by the longest diameter of the coalesced mass, whereas coalesced lymph nodes must still be measured along the maximal short axis perpendicular to the longest vector of the conglomerate nodal mass.

5. Response compares to baseline, whereas progression compares to the nadir (including baseline): As clarified on the EORTC RECIST 1.1 page and in Question 8 of Schwartz et al. (2016), Partial Response (PR) requires a ≥ 30% decrease in the sum of diameters compared to the baseline sum, whereas Progressive Disease (PD) in target disease requires a ≥ 20% increase in the sum of diameters AND at least a 5 mm absolute increase taking as reference the smallest sum on study (nadir)—which includes the baseline sum if the baseline sum is the smallest on study. In our worked example (Baseline SOD = 128 mm), if no shrinkage occurs first, PD requires the SOD to reach at least 128 × 1.20 = 153.6 mm (which also satisfies the ≥ 5 mm absolute increase of +25.6 mm). If the SOD first shrinks to a nadir of 70 mm at Week 8 (PR), subsequent PD is judged against 70 mm (requiring ≥ 84 mm, an increase of +14 mm), not against 128 mm.

6. New lesions establish progression rather than joining the target inventory: When an unequivocal new malignant lesion appears at follow-up (including a node below 10 mm at baseline that becomes pathological, although the working group suggests re-evaluating at the next assessment when a small size change—especially below 15 mm short axis—is the only evidence), the time-point response is PD. As the Working Group Q&A explains under Should new lesions be followed-up as target/non-target?, once a certain new lesion establishes PD, there is no purpose in designating that new lesion as target or non-target. Similarly, if a target lesion is surgically removed or irradiated before progression, Question 10 of Schwartz et al. (2016) treats this as a protocol deviation that makes the response not evaluable (NE); when such procedures are foreseen, the protocol must define in advance how response and progression will be handled and censored—see time-to-event censoring rules.

How often do baseline selection mistakes survive into the clinical database? In a Cooperative Group quality-assurance update (Common Errors in RECIST v1.1 Reporting, NRG Oncology, January 2025 error table), NRG Oncology reports that even with sufficient staff training, documentation and reporting mistakes remain frequent findings during clinical trial audits. Note the evidentiary scope of that report: NRG publishes an experience-based operational error table rather than a quantified percentage incidence across trials, but the failure modes it documents map directly to baseline worksheet design and pre-entry quality control.

Specifically, NRG Oncology's operational guidance highlights three controls that clinical trial teams should build into their site-initiation and data-management workflows before the first baseline CRF is locked:

• Physician-authored measurements on a dedicated RECIST worksheet: Only physicians trained to measure tumor size, number, and location should document RECIST lesion selections and measurements, using a structured RECIST measurement worksheet (such as the EORTC Initial Measurement Form) rather than expecting clinical research associates (CRAs) or coordinators to infer target lesions from dictated diagnostic radiology prose.

• Three-point CRA and Data Manager verification before EDC entry: Before entering measurements into the electronic data capture (EDC) system, CRAs and data managers should verify (1) that the correct number of target lesions total and per organ is reported, (2) that the percentage change upon reassessment is calculated correctly (under RECIST 1.1, against the baseline sum for response and the smallest sum on study for progression), and (3) that the appropriate response designation (CR, PR, SD, PD, or NE) is assigned.

• Protocol scan schedule and a single imaging type: NRG stresses adhering to the protocol scanning schedule and using the same imaging type (CT, MRI, etc.) throughout the study, because on-schedule scans are critical to the statistical analysis of PFS and overall survival endpoints. Its protocol scan-date calculators project static dates that do not shift when treatment is delayed.

The checklist table below translates these cooperative-group audit findings and the working-group rules into concrete pre-data-entry checks.

Even a perfectly executed RECIST 1.1 baseline worksheet operates within strict methodological boundaries that protocol authors must state up front:

• Anatomical disease where a five-lesion target inventory is not achievable: RECIST 1.1 relies on unidimensional CT/MRI measurement of solid anatomical lesions. Osteoblastic bone metastases, leptomeningeal disease, ascites, pleural or pericardial effusions, inflammatory breast disease, and lymphangitic involvement of skin or lung are non-measurable; osteolytic or mixed lytic-blastic bone lesions are measurable only when they contain an identifiable soft-tissue component measuring ≥ 10 mm on CT or MRI (Eisenhauer et al. 2009). Simple cysts are not malignant lesions at all; cystic metastases can be measurable, but non-cystic lesions are preferred as targets; and lesions in a previously irradiated or otherwise locally treated area are usually not measurable unless they have progressed, with the protocol stating the conditions under which they count. In trials enrolling patients with bone-only or effusion-predominant disease, the protocol should specify whether non-target-only cohorts are eligible and how non-CR/non-PD vs unequivocal PD endpoints are analyzed.

• FDG-PET and combined PET-CT limitations at baseline: As clarified in Questions 11–13 of Schwartz et al. (2016), FDG-PET under RECIST 1.1 is a complementary modality used primarily to detect or confirm new lesions rather than a substitute for anatomic CT caliper measurements. Crucially, the low-dose attenuation-correction CT component of a combined PET-CT scan is not of optimal diagnostic quality for RECIST measurements unless the site documents that the CT used an appropriate radiation dose for diagnostic quality with intravenous and oral contrast (if not medically contraindicated).

• Immunotherapy and other response criteria: The 2016 committee paper notes that RECIST and RECIST 1.1 were developed and tested using data from trials of cytotoxic drugs. EMA's PFS/DFS methodological appendix warns that some agents can interfere with contrast enhancement of lesions and that immunomodulators can produce a peculiar response pattern—apparent progression with lesion enlargement followed by response—so different methods or endpoints may need to be considered; it also lists WHO, EBMT, and RANO criteria alongside RECIST as established options. The RECIST working group publishes iRECIST separately for immunotherapy trials. If a protocol adopts iRECIST or another criteria set, its progression-handling rules must be prospectively defined in the protocol and statistical analysis plan; the baseline inventory described here does not settle them.

• Operational scope: Finally, baseline RECIST 1.1 lesion selection is a trial-endpoint measurement construct, not a clinical staging or patient-care directive. A lesion excluded from the five-target sum or classified as non-target for trial endpoint standardization remains part of the patient's clinical care, and nothing in this guide constitutes diagnostic radiology advice or patient-specific medical guidance.

1. EORTC RECIST Working Group — RECIST 1.1 guidelines for solid tumors (criteria summary, Initial/Follow-up Measurement Forms, and Questions and clarifications Q&A) (last updated June 2026; accessed October 14, 2026).

2. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1), European Journal of Cancer 2009;45(2):228–247 (EORTC-hosted guideline PDF).

3. Schwartz LH, Litière S, de Vries E, et al. RECIST 1.1 — Update and clarification: From the RECIST committee, European Journal of Cancer 2016;62:132–137 (PubMed Central PMC5737828).

4. EORTC RECIST Working Group — Initial Measurement Form — RECIST 1.1 (official blank baseline lesion-measurement recording template PDF).

5. U.S. Food and Drug Administration (CDER/CBER) — Clinical Trial Imaging Endpoint Process Standards: Guidance for Industry (final guidance, April 2018; Docket FDA-2011-D-0586).

6. NRG Oncology — Common Errors in RECIST v1.1 Reporting (Quality Control & Communication / Protocol Support guidance post with January 2025 common-errors table).

7. European Medicines Agency (CHMP) — Appendix 1 to the guideline on the evaluation of anticancer medicinal products in man: methodological consideration for using progression-free survival or disease-free survival in confirmatory trials (EMA/CHMP/27994/2008/Rev.1).