## Compiled Trials DataFrame (`trials_all.csv`)

### Core Fields (Original)

| Column | Type | Description |
| --- | --- | --- |
| `nct_id` | str | ClinicalTrials.gov NCT identifier |
| `brief_title` | str | Brief study title |
| `official_title` | str | Full official study title |
| `lead_sponsor` | str | Original sponsor name from API |
| `sponsor_normalized` | str | Normalized sponsor name (e.g., "Takeda") |
| `sponsor_class` | str | Sponsor type: INDUSTRY, NIH, OTHER, etc. |
| `is_industry` | bool | True if sponsor\_class == INDUSTRY |
| `mechanism` | str | Classified mechanism of action |
| `drug_names_str` | str | Semicolon-separated drug/intervention names (raw) |
| `drug_names_normalized_str` | str | Semicolon-separated canonical drug names (Phase 1B) |
| `phase_normalized` | str | Normalized phase label (e.g., "Phase 3") |
| `phase_numeric` | float | Numeric phase for sorting (1, 1.5, 2, 2.5, 3, 4) |
| `overall_status` | str | Trial status (RECRUITING, etc.) |
| `conditions_str` | str | Semicolon-separated condition names |
| `enrollment` | int/null | Raw target enrollment count |
| `start_date` | str | Trial start date (YYYY-MM or YYYY-MM-DD) |
| `start_year` | int/null | Extracted start year |
| `completion_date` | str | Expected completion date |
| `study_type` | str | INTERVENTIONAL or OBSERVATIONAL |

### Enrollment Quality Fields (Phase 1A)

| Column | Type | Description |
| --- | --- | --- |
| `is_pharmacological` | bool | True if INTERVENTIONAL + classified mechanism |
| `enrollment_clean` | float/null | Enrollment capped at 50,000 (NaN for outliers) |
| `enrollment_outlier` | bool | True if enrollment > 50,000 (registries, mega-studies) |

### Geographic Fields (Phase 2A/2B)

| Column | Type | Description |
| --- | --- | --- |
| `countries_str` | str | Semicolon-separated list of trial countries |
| `n_countries` | int | Number of unique countries for this trial |
| `regions_str` | str | Semicolon-separated list of regions (mapped from countries) |

### Study Design Fields (Phase 2A/2B)

| Column | Type | Description |
| --- | --- | --- |
| `allocation` | str | Randomization type: RANDOMIZED, NON\_RANDOMIZED, etc. |
| `masking` | str | Blinding level: DOUBLE, SINGLE, NONE, QUADRUPLE, etc. |
| `intervention_model` | str | Study model: PARALLEL, CROSSOVER, SINGLE\_GROUP, etc. |
| `study_design_category` | str | Derived: "RCT Double-Blind", "RCT Open-Label", "Single-Arm", "Observational", etc. |

### Arms & Comparator Fields (Phase 2A/2B)

| Column | Type | Description |
| --- | --- | --- |
| `has_placebo_arm` | bool | True if trial has a placebo comparator arm |
| `has_active_comparator` | bool | True if trial has an active comparator arm |
| `is_head_to_head` | bool | True if trial is head-to-head (active comparator + experimental) |
| `n_arms` | int | Number of arms in the trial |
| `is_combination` | bool | True if trial has 2+ active drug interventions |

### Endpoint Fields (Phase 2A/2B)

| Column | Type | Description |
| --- | --- | --- |
| `primary_endpoint` | str | First primary outcome measure |
| `endpoint_timeframe` | str | First primary outcome timeframe |

### Eligibility Fields (Phase 2A/2B)

| Column | Type | Description |
| --- | --- | --- |
| `min_age` | str | Minimum age string (e.g., "18 Years") |
| `max_age` | str | Maximum age string (e.g., "80 Years") |
| `min_age_years` | float/null | Minimum age in years (numeric) |
| `max_age_years` | float/null | Maximum age in years (numeric) |
| `sex` | str | Sex eligibility: ALL, MALE, FEMALE |
| `is_pediatric` | bool | True if min or max age < 18 years |

### Regulatory Fields (Phase 2A/2B)

| Column | Type | Description |
| --- | --- | --- |
| `is_fda_regulated_drug` | bool/null | True if FDA-regulated drug |
| `has_dmc` | bool/null | True if Data Safety Monitoring Committee present |

### Collaborator Fields (Phase 2A/2B)

| Column | Type | Description |
| --- | --- | --- |
| `collaborators_str` | str | Semicolon-separated collaborating institution names |

### Classification Fields (Phase 4B)

| Column | Type | Description |
| --- | --- | --- |
| `is_biosimilar` | bool | True if trial involves a biosimilar product |

---

## Mechanism x Phase Matrix (`trials_by_mechanism.csv`)

Cross-tabulation with mechanisms as rows, phases as columns, and trial counts as values.
Includes a "Total" column.

## Sponsor Summary (`trials_by_sponsor.csv`)

| Column | Type | Description |
| --- | --- | --- |
| `sponsor_normalized` | str | Index: normalized sponsor name |
| `trials` | int | Total number of trials |
| `mechanisms` | int | Number of distinct mechanisms |
| `industry` | bool | Industry-sponsored |

## Reports

### `landscape_report.pdf`

Publication-quality PDF report generated by reportlab. Contains 24 sections:
- Title page with executive summary stats (trial count, mechanisms, sponsors)
- Full 6-panel landscape visualization (embedded from PNG)
- Mechanism × Phase distribution table
- Indication breakdown (config-driven categories)
- Mechanism deep-dives with drug pipeline tables
- Late-stage pipeline (Phase 3) analysis
- Upcoming readouts section with time-to-readout estimates
- Config-driven mechanism highlight sections (e.g., Anti-IL-23 Focus, Anti-TL1A)
- Sponsor competitive landscape table (top 20 sponsors)
- Company portfolio analysis
- Industry vs Academic breakdown
- Config-driven sponsor highlight sections (e.g., Takeda Portfolio)
- Enrollment & investment signals (using clean enrollment)
- Geographic landscape (top countries, region breakdown)
- Study design analysis (RCT rates, double-blind by phase)
- Phase transition funnel
- Phase 3 endpoint comparison
- Combination therapy landscape
- Patient population analysis (pediatric vs adult)
- Trial arms & comparator analysis (placebo, H2H)
- Biosimilar landscape
- Whitespace & unmet needs
- Regulatory signals (FDA, DSMB rates)
- Data notes with query parameters and caveats

### `landscape_report.md`

Markdown version with the same 24 sections. Useful for embedding in other documents or rendering in Biomni.

### `landscape_supplementary.png`

4-panel supplementary visualization:
1. Top 15 countries by trial presence (horizontal bar)
2. Study design by phase (stacked bar)
3. Enrollment distribution by mechanism (box plot, log scale)
4. Phase transition funnel (grouped bar)

## Analysis Object (`analysis_object.pkl`)

Python dict with keys:

| Key | Type | Description |
| --- | --- | --- |
| `trials_df` | DataFrame | Full compiled trials DataFrame (46+ columns) |
| `parameters` | dict | Query parameters used |
| `n_total` | int | Total trial count |
| `n_by_mechanism` | dict | Trial counts per mechanism |
| `n_by_phase` | dict | Trial counts per phase |
| `n_by_sponsor` | dict | Trial counts per sponsor |
| `n_industry` | int | Industry-sponsored trial count |
| `unique_mechanisms` | int | Number of distinct mechanisms |
| `unique_sponsors` | int | Number of distinct sponsors |
| `mechanism_phase_matrix` | dict | Nested dict of mechanism x phase counts |
| `timestamp` | str | ISO timestamp of export |
| `n_with_geographic_data` | int | Trials with location data (Phase 2) |
| `n_by_design` | dict | Trial counts per study design category (Phase 2) |
| `n_combination` | int | Combination therapy trials (Phase 2) |
| `n_pediatric` | int | Pediatric trials (Phase 2) |
| `n_biosimilar` | int | Biosimilar trials (Phase 4B) |
| `enrollment_outliers` | int | Enrollment outlier count (Phase 1A) |