pembrolizumab / schema
The protocol schema
Every protocol on this site is one record against the same schema, drawn from a published, peer-reviewed study or a registered trial. This is that schema: the fields, their types, what each metric means, and the standard new records are held to.
The backbone is fixed for this site: Pembrolizumab (anti–PD-1). Every other field, the combination, the setting, the endpoints and the evidence, is recorded the same way so protocols compare like with like.
definition
The record
One row in the database is one immunotherapy protocol. Fields marked with a trailing ? are nullable, present only when the study reports them.
// one row in the database = one immunotherapy protocol
type Protocol {
namestring// protocol or trial identifier
typeProtocolType// published | custom | ind
statusStatus// active | completed | ongoing
backboneBackbone// fixed: pembrolizumab (anti–PD-1)
modalityModality// combination class, sets the site color
combinationstring// agents added onto the backbone
indicationstring// tumor type, stage, biomarker
settingSetting// neoadjuvant | adjuvant | perioperative | ...
populationstring// eligibility and allocation
regimenPhase[]// { phase, agent, dose, surgery? }
primaryEndpoint// pre-specified main outcome
hrHazardRatio?// { est, lo, hi } (comparative designs only)
curveCurve?// time-to-event, illustrative
outcomesOutcome[]// response, survival, safety
sourceSource// journal or registry + doi + nct
modifiersModifier[]// qualifiers on how to read it
}
relationships
How the fields group
Read top to bottom. Intervention (a backbone plus a combination), then context, regimen, endpoints, and provenance. Modifiers qualify how any of it should be read. Select a field to jump to its definition.
intervention
context
regimen
The backbone, filled, is the anchor shared by every record. A trailing ? marks a nullable field.
reference
Field dictionary
Every field, its type, whether it is required, what it captures, and an example value.
| field | type | constraint | description | example |
|---|---|---|---|---|
| intervention | ||||
| backbone | Backbone | required | The checkpoint inhibitor every protocol is built on. Fixed per site: Pembrolizumab (anti–PD-1). | Pembrolizumab (anti–PD-1) |
| modality | Modality | required | The kind of therapy added to the backbone, from a fixed set. Sets the color used across the site. | vaccine |
| combination | string | required | The specific agents, adjuvants or local therapy combined with Pembrolizumab. | neoantigen vaccine + pembrolizumab |
| context | ||||
| indication | string | required | Tumor type, with stage and any biomarker that scopes it. | melanoma, resected IIIB to IV |
| setting | Setting | required | Point in the treatment course. | adjuvant |
| population | string | required | Eligibility and who was treated, including allocation. | 157 randomized 2:1 |
| regimen | ||||
| regimen | Phase[] | required | Ordered phases (priming, neoadjuvant, surgery, adjuvant, maintenance), each with agent and dose. Surgical steps are flagged. | priming, backbone Q3W, maintenance |
| endpoints | ||||
| primary | Endpoint | required | The pre-specified main outcome the protocol was designed to measure. | RFS |
| hr | HazardRatio? | nullable | Relative benefit versus the comparator, with a 95% confidence interval. Present only for comparative designs. | 0.51 (0.29 to 0.89) |
| curve | Curve? | nullable | Survival or recurrence curve for the primary endpoint. Anchored to reported values, not digitized. | RFS to 60 mo |
| outcomes | Outcome[] | required | Reported numbers: response rate, survival at a time point, and safety. | ORR, X-yr survival, G3+ |
| provenance | ||||
| name | string | required | Identifier for the protocol or trial. | KEYNOTE-942 |
| type | ProtocolType | required | Published trial, custom protocol, or single-patient IND. Sets evidentiary weight. | published |
| source | Source | required | Journal or registry, with a DOI and a ClinicalTrials.gov id where one exists. | JCO, doi, NCT |
| status | Status | required | Active, completed, or an ongoing trial. | completed |
definitions
Metrics and endpoints
The outcome measures used across the site, in plain terms, with which direction is better.
| term | name | what it measures | direction |
|---|---|---|---|
| OS | overall survival | Time from start of treatment to death from any cause. | higher better |
| PFS | progression-free survival | Time until the disease progresses or the patient dies. | higher better |
| RFS | recurrence-free survival | For resected disease, time until cancer recurs or death. | higher better |
| EFS | event-free survival | Time to any pre-defined event: progression precluding surgery, recurrence, or death. | higher better |
| DFS | disease-free survival | Time after treatment with no disease detected. | higher better |
| DMFS | distant metastasis-free survival | Time until distant metastasis appears or death. | higher better |
| ORR | objective response rate | Share of patients whose tumor shrinks by a set amount (complete or partial) under RECIST. | higher better |
| CR | complete response | Disappearance of all measurable disease. | higher better |
| pCR | pathologic complete response | No residual invasive cancer in tissue removed at surgery. | higher better |
| DoR | duration of response | How long a response lasts, from first response to progression. | higher better |
| HR | hazard ratio | Rate of the endpoint event in treatment vs comparator. Below 1 favors treatment. | lower better |
| CI | 95% confidence interval | Range the true value likely falls in. If the whole interval sits below 1, the benefit is significant. | narrower = precise |
| irAE | immune-related adverse event | Side effect from the immune system attacking healthy tissue. | lower better |
| G3+ | grade 3 or higher AE | A severe or worse side effect on the CTCAE scale. | lower better |
| N | sample size | Number of patients the result is based on. Small N is illustrative, not definitive. | higher = stronger |
| CPS | combined positive score | PD-L1 expression measure used to select patients (e.g. CPS at least 10). | threshold |
| RECIST | response criteria 1.1 | Standard rules for deciding if tumors shrank, grew, or stayed stable on imaging. | n/a |
qualifiers
Modifiers
Modifiers do not change a protocol, they change how confidently a result can be read. The difference between a signal and a proof.
| flag | what it means | effect on reading |
|---|---|---|
| randomized_vs_single_arm | Whether patients were randomly assigned against a comparator, or all got the same protocol. | A hazard ratio only exists for a comparative design. Single-arm results have no HR. |
| comparator_arm | The treatment the protocol is measured against. | Benefit is relative to this arm, not an absolute cure rate. |
| biomarker_threshold | A molecular cutoff that scoped the population (e.g. CPS at least 10, MHC-I low). | The result applies only to patients meeting the threshold. |
| subgroup_result | A number reported for part of the cohort (e.g. resected only). | Describes that subset, not the whole population. More exploratory. |
| real_world_data | Outcomes observed in routine practice, not a controlled trial. | Useful for signal, but confounded and not directly comparable to trials. |
| illustrative_curve | A curve anchored to reported values rather than digitized from the source figure. | Shapes are indicative. Exact numbers live on each protocol card. |
| surgery_step | A surgical procedure inside the regimen timeline. | Flagged so the order of systemic therapy and surgery is explicit. |
| access_tier | Some physician-contributed outcomes are held behind membership. | Published-trial data is always open; gating applies only to contributed detail. |
enums
Controlled vocabularies
Fixed value sets keep records consistent. Modality also sets the color used throughout the site.
enum Modality
vaccine// personalized neoantigen or peptide vaccine
adc// antibody-drug conjugate
chemoRT// perioperative backbone around surgery + chemoradiation
local// vaccine + localized radiation (SBRT)
io// radiation or a second immunotherapy agent
chemo// systemic chemotherapy
targeted// targeted small molecule
enum ProtocolType
published// peer-reviewed or registered trial. Most weight.
custom// physician-designed, often small cohort or real-world
ind// single-patient investigational protocol
enum Status
active// enrolling or on treatment
completed// finished, results reported
ongoing// trial in progress, no results yet
Source
published// open
physician// may be membership-gated
governance
The intake standard
New records are only added when they fit this schema. Intake runs the same steps every time.
01
source first
A record needs a citable source: a peer-reviewed publication or a registered trial, with a DOI or ClinicalTrials.gov id where one exists. Physician-entered protocols are labeled as such.
02
map to the schema
The regimen is broken into backbone, modality, agents, setting, indication and population, using only the enums above.
03
record endpoints as reported
Primary endpoint, hazard ratio with its interval, and outcomes are entered as published. Nothing is inferred beyond the source.
04
attach modifiers
Design, biomarker thresholds, subgroups and provenance are flagged so a reader can weigh the result correctly.
05
de-identify contributions
Physician-contributed and single-patient records carry no identifying patient detail. Ages are ranges, timing is relative.
Curves shown on the site are illustrative, anchored to reported endpoint values rather than digitized from source figures. Exact numbers and the citation live on each protocol card.