Probability of Success Calibration

1. Three-dimensional benchmark matrix

Rather than applying a single set of industry-average transition rates, PhaseFolio stratifies PoS by three independent classification axes that are known to materially affect clinical outcomes: therapeutic area, drug modality, and biomarker strategy.

• Therapeutic Area. 11 indications from oncology (solid and hematologic) through cardiovascular, neurology, metabolic, and rare disease. Measured from Phase I, in the baseline table oncology (solid tumor) is lowest at 4.1% and rare disease highest at 15.8%; among the BIO 2021 area rows, solid tumors is lowest at 4.6% and non-oncology rare diseases highest at 17.0%. See Table 1.
• Drug Modality. 8 benchmark modality baselines: small molecule, monoclonal antibody, bispecific, ADC, cell therapy, gene therapy, peptide, and other. Expanded labels used by triage or imported assets are normalized to these baselines before lookup. Source: BIO 2021 Figure 10b modality rows (Section 1.2), plus the normalization policy in Section 1.1.
• Biomarker Strategy. 3 levels: none, enrichment (biomarker-selected population), and companion diagnostic (required for Rx). Programs using preselection biomarkers roughly double their likelihood of approval from Phase I. Source: BIO 2021 Figure 11 (Section 1.2); Parker et al. 2021 reports the same direction in oncology.

The full matrix contains 11 × 8 × 3 = 264 unique indication–modality–biomarker combinations, each specifying five stage-transition probabilities (Preclinical → Phase I → Phase II → Phase III → NDA/BLA → Approval). Since engine 3.0.0 (Methodology M4), every Phase I to NDA/BLA rate in all 264 cells is derived from named rows of BIO/QLS/Informa 2021, which analysed 12,728 phase transitions from 2011 to 2020 (Section 1.2). BIO reports no preclinical stage, so the preclinical rates are PhaseFolio assumptions, kept at their earlier values and labelled as such. Before M4 the matrix was a PhaseFolio composite: the small-molecule rows were informed by BIO 2021, the other modality rows were PhaseFolio adjustments, and 125 of the 264 cells were missing and fell back to the area's "other" row.

1.1 Expanded modality normalization

Triage, deck import, and analyst-entered asset records can contain modern modality labels that are more granular than the eight benchmark baselines above. PhaseFolio normalizes those labels before PoS lookup so the calculation remains defined and auditable. This is a baseline proxy, not a statement that the newer modality has identical clinical, safety, manufacturing, or regulatory risk to the mapped baseline. Analysts should override stage probabilities when asset-specific evidence supports a more precise view.

Table 1A — Expanded modality labels mapped to PoS benchmark baselines.

• RNA Therapeutics (including mRNA, siRNA, ASO, RNAi labels) → Gene Therapy baseline
• Protein Degrader (including PROTAC and molecular glue labels) → Small Molecule baseline
• Radioligand / radiopharmaceutical → Other baseline
• Oncolytic Virus → Gene Therapy baseline
• Cancer Vaccine → Other baseline

The normalization policy is separate from IRA classification. For example, RNA therapeutics and radioligands default to the Year 9 drug cliff in the IRA terminal-value framework even though their PoS baselines route through gene therapy or other respectively.

1.2 How each cell is derived

BIO 2021 reports success rates by disease area, by modality and by biomarker use, but never crossed. Each cell combines them in log odds: it starts from the area row and adds how far the modality row and the biomarker row sit from the all-indications row, stage by stage.

Equation 0 — Cell derivation. logit(cell) = logit(area) + [logit(modality) − logit(all indications)] + [logit(biomarker) − logit(all indications)], for Phase I, Phase II, Phase III and NDA/BLA.

• Area rows: BIO 2021 Figures 2, 7 and 8b. Hematologic oncology uses the Figure 7 hematologic row. BIO has no dermatology group, so dermatology uses its Other row, labelled as a proxy.
• Modality rows: Figure 10b. Bispecifics use the monoclonal antibody row, labelled as a proxy. Cell therapy and other take no modality shift. Rows are smoothed as (k + 0.5) / (n + 1), so a 100% rate on a handful of transitions does not become certainty.
• Biomarker rows: Figure 11. No biomarker strategy uses the No biomarkers row; companion diagnostic uses the Preselection biomarkers row; enrichment takes half the preselection shift, a PhaseFolio assumption because BIO has only one preselection category.

Every BIO row used reproduces its printed likelihood of approval, and a test holds the engine table to this derivation, so the two cannot drift apart.

Table 1. Baseline stage-transition probabilities by therapeutic area (small molecule, no biomarker strategy; dataset 2026-09-27). These are the values the engine computes with. LoA from preclinical is the product of all five rates; LoA from Phase I is the product of the last four, which is the measure BIO 2021 reports. The BIO 2021 value is the area row each cell is derived from, across all modalities; the difference is the small-molecule and no-biomarker shift.

• Oncology (Solid Tumor): Preclinical 4.0%, Phase I 49.5%, Phase II 22.1%, Phase III 41.3%, NDA/BLA 91.8%; LoA from preclinical 0.17%, LoA from Phase I 4.1%; BIO 2021 LOA from Phase I 4.6%, Solid tumors (Figure 7)
• Oncology (Hematologic): Preclinical 7.0%, Phase I 50.7%, Phase II 26.4%, Phase III 58.4%, NDA/BLA 88.5%; LoA from preclinical 0.48%, LoA from Phase I 6.9%; BIO 2021 LOA from Phase I 7.5%, Hematologic cancers (Figure 7)
• Rare Disease: Preclinical 8.0%, Phase I 67.9%, Phase II 42.8%, Phase III 58.8%, NDA/BLA 92.6%; LoA from preclinical 1.3%, LoA from Phase I 15.8%; BIO 2021 LOA from Phase I 17.0%, Non-oncology rare diseases (Figure 8b)
• Neurology: Preclinical 4.0%, Phase I 48.3%, Phase II 25.4%, Phase III 51.5%, NDA/BLA 84.8%; LoA from preclinical 0.21%, LoA from Phase I 5.3%; BIO 2021 LOA from Phase I 5.9%, Neurology (Figures 2 and 5b)
• Immunology: Preclinical 6.0%, Phase I 55.8%, Phase II 29.8%, Phase III 63.8%, NDA/BLA 93.2%; LoA from preclinical 0.59%, LoA from Phase I 9.9%; BIO 2021 LOA from Phase I 10.7%, Autoimmune (Figures 2 and 5b)
• Infectious Disease: Preclinical 7.0%, Phase I 58.4%, Phase II 36.7%, Phase III 62.5%, NDA/BLA 91.8%; LoA from preclinical 0.86%, LoA from Phase I 12.3%; BIO 2021 LOA from Phase I 13.2%, Infectious disease (Figures 2 and 5b)
• Cardiovascular: Preclinical 5.0%, Phase I 50.6%, Phase II 19.8%, Phase III 53.6%, NDA/BLA 80.1%; LoA from preclinical 0.22%, LoA from Phase I 4.3%; BIO 2021 LOA from Phase I 4.8%, Cardiovascular (Figures 2 and 5b)
• Metabolic: Preclinical 6.0%, Phase I 62.4%, Phase II 43.2%, Phase III 62.1%, NDA/BLA 85.7%; LoA from preclinical 0.86%, LoA from Phase I 14.3%; BIO 2021 LOA from Phase I 15.5%, Metabolic (Figures 2 and 5b)
• Respiratory: Preclinical 5.0%, Phase I 56.5%, Phase II 20.7%, Phase III 63.0%, NDA/BLA 94.9%; LoA from preclinical 0.35%, LoA from Phase I 7.0%; BIO 2021 LOA from Phase I 7.5%, Respiratory (Figures 2 and 5b)
• Dermatology: Preclinical 6.0%, Phase I 64.1%, Phase II 36.9%, Phase III 58.4%, NDA/BLA 86.7%; LoA from preclinical 0.72%, LoA from Phase I 12.0%; BIO 2021 LOA from Phase I 13.0%, Other disease areas including dermatology (Figures 2 and 5b)
• Ophthalmology: Preclinical 6.0%, Phase I 72.1%, Phase II 33.8%, Phase III 49.6%, NDA/BLA 89.7%; LoA from preclinical 0.65%, LoA from Phase I 10.8%; BIO 2021 LOA from Phase I 11.9%, Ophthalmology (Figures 2 and 5b)

Source: derived from BIO/QLS/Informa 2021 (Thomas et al.), Figures 2, 7, 8b, 10b and 11 (Section 1.2); preclinical rates are PhaseFolio assumptions. BIO 2021 has no dermatology group, so dermatology uses its Other group, which includes dermatology; immunology uses its Autoimmune group. Any change to the engine's values ships as a new engine major version.

2. Multiplier adjustments

Several evidence-based factors are known to shift clinical success probabilities relative to the population base rate. PhaseFolio applies these multipliers via a log-odds (logit) transformation — the mathematically correct method when a multiplier is a true odds ratio. The cited sources, however, report effect sizes in different forms: relative success ratios (Minikel 2024; Schwaederle 2016), relative approval rates (Mullard 2016), and a published approval rate that the orphan gene therapy modifier is calibrated to (Phares, Phillip and Trusheim 2025). The biomarker-strategy effect is no longer a multiplier: since engine 3.0.0 it is built into each matrix cell (Section 1.2).

The engine currently treats all of these through the OR-style logit path as a deliberately conservative approximation — this under-credits favorable modifiers at higher baselines, never saturates to 1.0, and avoids stacked-modifier overshoot. Per-modifier estimand declarations (_source_estimand and _applied_as) are recorded in the machine-readable source and explained in the model card. Each multiplier is applied only to the clinical phases where the underlying evidence was measured.

Two evidence tiers. Every multiplier below except first-in-class is anchored to a specific, named, published effect size; first-in-class is a PhaseFolio assumption and is labelled as one. The anchored multipliers differ in whether the factor has, in addition to that published anchor, been validated on a PhaseFolio held-out cohort. Literature-anchored factors are adopted from a published estimate and applied conservatively (Sections 2–3) — the same evidence tier as the Table 1 base rates, which are derived from BIO/QLS 2021 rather than validated on a PhaseFolio cohort. Cohort-validated factors have additionally cleared the both-outcomes governance gate of Section 5 on a PhaseFolio backtest. Backtesting the literature-anchored factors as their cohorts reach adequate size is on the roadmap; until then this distinction is labelled in the table, not left silent.

Table 2 — Evidence-based multipliers (source estimand separated from applied path; evidence tier in brackets; favorable multipliers >1 increase PoS; unfavorable <1 decrease):

• Genetic Validation: 2.6× (RR estimand, OR-applied), Phase II/III, Minikel et al. 2024 Nature [literature-anchored]
• Orphan Designation: 1.5× (RR, OR-applied), Phase II/III, Mullard 2016 Nat. Rev. Drug Disc. [literature-anchored]
• Gene Therapy (Orphan): 1.13× per stage (OR, calibrated to 18.5% from Phase I), Phase I/II, Phares, Phillip and Trusheim 2025; replaces Orphan Designation for gene therapy [literature-anchored]
• First-in-Class: 0.85× (RR, OR-applied), Phase II/III, PhaseFolio assumption; BIO 2021 reports no such figure [PhaseFolio assumption]
• Biomarker Quality — Genomic Validated (oncology solid only): 1.35× (RR, OR-applied), Phase II/III, Schwaederle et al. 2016 [cohort-validated — see Section 6]
• Biomarker Quality — Protein Only (oncology solid only): 0.85× (RR, OR-applied), Phase II/III, Schwaederle et al. 2016 [cohort-validated — see Section 6]

Calibrated, not split. The orphan gene therapy value is the per-stage odds ratio that takes the rare-disease gene therapy cell (no biomarker strategy) from 16.9% to the 18.5% likelihood of approval that Phares, Phillip and Trusheim (2025) report for rare-disease gene therapies. That figure already reflects orphan status, so a gene therapy with orphan designation gets this modifier instead of Orphan Designation, not both. There is no CAR-T modifier: the same paper reports hematologic CAR-T and TCR therapies at 7.6% against 7.5% for comparable drugs, and the derived hematologic cell therapy cell is already 7.25%. Before M4 both modifiers were square-root splits of an older whole-pipeline advantage (NEWDIGS at Tufts Medical Center, 2023) applied beyond the scope that analysis measured. Biomarker Quality is the drug-specific multiplier added in engine 2.6.0 (2026-05-28) and applies only when the engine is given a per-drug biomarker classification; see Section 6 for the bucket definitions and the cohort validation.

A critical design decision: preclinical PoS is never adjusted by any multiplier. Preclinical attrition is dominated by toxicology, pharmacokinetics, and formulation failures (Zhou et al. 2025) — factors orthogonal to the clinical efficacy signals that these multipliers capture. Similarly, NDA/BLA approval rates reflect regulatory filing quality rather than drug-specific clinical attributes, and are therefore held constant.

3. Logistic transformation method

Applying odds ratios to bounded probabilities requires care. Naive multiplication (PoS × OR, capped at 1.0) produces mathematically unsound results: a drug with 50% base PoS and a 2.6× genetic-validation multiplier would yield 130%, capped to 100% — falsely claiming certainty. With multiple favorable multipliers stacking, this problem cascades rapidly.

PhaseFolio instead applies multipliers in log-odds (logit) space — the standard biostatistical transformation for adjusting bounded probabilities by a multiplicative factor. The three-step transformation:

Equation 1 — Logistic odds-ratio adjustment. Step 1: odds = PoS_base / (1 − PoS_base). Step 2: odds_adj = odds × OR. Step 3: PoS_adj = odds_adj / (1 + odds_adj).

This approach has three desirable mathematical properties:

1. Bounded output. The result is always in (0, 1) — it can never reach 0% or 100%, regardless of how many multipliers are stacked.
2. Diminishing returns. A 2.6× OR applied to a 24% base PoS yields 45.1% (+21.1pp). Applied to a 70% base, it yields 85.8% (+15.8pp). The higher the base, the harder it is to push higher — matching clinical reality.
3. Composability. Multiple ORs applied sequentially produce the same result regardless of order, because multiplication in log-odds space is commutative.

4. Worked example: PoS derivation

Consider a rare disease small molecule with genetic validation and orphan designation. We derive the Phase II PoS step-by-step.

Base rate (from benchmark matrix): Phase II PoS = 42.8%.

Apply genetic validation (factor 2.6, applied as OR): odds = 0.428 / (1 − 0.428) = 0.748; odds × 2.6 = 1.945; PoS = 1.945 / (1 + 1.945) = 66.0%. Factor source: Minikel et al. 2024 (RR); applied as OR.

Apply orphan designation (factor 1.5, applied as OR): odds = 0.660 / (1 − 0.660) = 1.945; odds × 1.5 = 2.917; PoS = 2.917 / (1 + 2.917) = 74.5%. Factor source: Mullard 2016 (RR); applied as OR.

Adjusted Phase II PoS: 74.5%.

Note: naive multiplication would yield min(1.0, 0.428 × 2.6 × 1.5) = 100% — clearly incorrect. The logistic method produces 74.5%, reflecting appropriate diminishing returns.

5. Guarding against overfitting: the multiplier governance gate

Every multiplier in Table 2 adds a degree of freedom, and a stack of adjustable factors can manufacture the appearance of rigor while quietly encoding the author's priors — the central failure mode of any heuristic valuation model. PhaseFolio constrains this in three ways: two structural, one evidentiary.

Structural bounds (Sections 2–3). Multipliers are applied only in log-odds space, so stacked factors can never saturate to 0% or 100% and never overshoot; each factor is applied only to the clinical phases where its source measured the effect; and preclinical and NDA/BLA rates are never adjusted at all. These bounds cap how far the knobs can move any result, regardless of how many fire at once.

The evidentiary gate — a multiplier may only score the engine if a held-out cohort can validate it. A candidate factor earns the right to change a probability only when a backtest cohort containing both approvals and failures shows it discriminates between them: it must fire on known successes as well as known failures, so a skeptic can confirm it tracks outcomes rather than merely labelling the failures after the fact. A signal that fires only on the failures in a cohort, with no approved counterexample, cannot be validated by that cohort — it is demoted to a non-scored, display-only risk flag rather than allowed to move the number.

Worked proof. In the antimicrobial backtest, three candidate antibacterial multipliers were pre-registered on evidence dated before each drug's decision. A pre-publication ablation showed that two of them — a hepatotoxicity mechanism-class prior and a sustained-clinical-response endpoint-fragility prior — fired only on that cohort's failures, with no approved counterexample, so they were demoted to flags. Only the third, single-asset sponsor fragility (which fires on three approvals as well as the failures), was allowed to score. We publish the full ablation rather than the most flattering configuration: scoring only the validatable factor yields a pairwise AUC of 0.629, whereas the unvalidatable pair would have produced a higher but uncheckable 0.797. We report the lower, defensible number. See the backtest methodology as last published (M2); the backtests were retired in M4.

How the gate applies to the existing table. The both-outcomes cohort gate is the standard for drug-specific signals (Sections 6–7) and for every future candidate multiplier: no such factor scores the engine until a cohort with both outcomes can independently confirm it. The population-level multipliers of Table 2 (genetic validation, orphan designation, orphan gene therapy, first-in-class) predate this gate. Three are literature-anchored — each is adopted from a named published effect size and applied conservatively (Sections 2–3) — and first-in-class is a labelled PhaseFolio assumption. The Table 1 base rates are likewise derived from published rates rather than cohort-validated. Of the full table, only biomarker_quality has so far additionally cleared a PhaseFolio held-out cohort; the Table 2 tier labels state which factors have and have not yet been cohort-validated. Subjecting the literature-anchored factors to the same backtest as their cohorts reach adequate size is on the roadmap. Until a candidate clears the gate it may inform a risk flag, but it does not move the valuation.

6. Drug-specific clinical signals — biomarker quality (scored)

Through engine 2.5.x, every multiplier in Table 2 keyed on attributes of the program (modality, biomarker strategy, orphan designation, indication-level genetic validation). Engine 2.6.0 (shipped 2026-05-28) introduces a second class of signal — drug-specific clinical attributes extracted from each asset's underlying evidence (pivotal-program publications, FDA labels, sponsor disclosures, registry records). The first such signal to clear the multiplier-governance gate of Section 5 is biomarker quality.

Definition. Biomarker quality refines the biomarker-strategy axis of Table 1 by asking what kind of biomarker a program is built on, not merely whether one is present. Three scoring buckets ship in engine 2.6.0:

• genomic_validated — a sequenced, mechanism-anchored DNA or RNA alteration (EGFR exon 19 deletion, BRAF V600E, ALK fusion, MSI-H, and similar) used as the enrichment criterion. Multiplier: 1.35× (RR estimand, OR-applied).
• protein_only — a protein-expression biomarker without an underlying genomic anchor (HER2 IHC alone, PD-L1 tumor-proportion score, serum-protein enrichment). Multiplier: 0.85× (RR, OR-applied).
• unknown — biomarker quality not extracted or not applicable. Multiplier: 1.00× (no adjustment).

The signal applies only to oncology solid tumor programs at Phase II and Phase III. Preclinical, Phase I, and NDA/BLA are not adjusted (the Section 2 design principle). Hematologic oncology, immunology, neurology, and every other indication carry no biomarker-quality multiplier.

Source and estimand. The 1.35× / 0.85× values are reported as relative response/success ratios in Schwaederle et al. 2016 — a meta-analysis of phase II precision-medicine trials across 13,203 patients in 346 studies — and applied through the OR-style logit path of Sections 2 and 3, the same conservative dispatch used by every other multiplier in Table 2.

Cohort validation. A 43-drug oncology-solid-tumor cohort (50% of the 85-drug Phase 0 cohort universe, with approvals drawn from FDA approvals 2018–2024 and failures from public discontinuations) scored biomarker_quality against the engine's baseline. The signal cleared the multiplier-governance gate of Section 5:

• Cohort N: 43 oncology solid tumor; bucket-level minimum N ≥ 6 with ≥ 3 sponsors per bucket (the antimicrobial Sprint-1 precedent).
• Pairwise AUC: baseline 0.618 → biomarker_quality alone 0.670 (+5.2pp). Stable across the v2, v3, and v4 extractor rounds at 28, 35, and 43 drugs respectively.
• The signal fires on both approvals and failures — the both-outcomes gate.

Honest disclosure — the literature anchor is conservative. The cohort fit yielded a genomic_validated odds ratio of approximately 5.59 — markedly higher than the 1.35× we ship. The implication is that the Schwaederle (2016) anchor is now roughly a decade old and likely understates the discrimination of modern targeted oncology relative to current practice. We ship the lower literature anchor anyway, on the same discipline as the antimicrobial backtest (Section 5): we publish the defensible, externally-citable value rather than the higher cohort-derived value, until the cohort is large enough to justify a recalibration. Closing this gap is on the Phase 2 roadmap. The cohort-derived OR is documented here and in the public model card — nothing is hidden.

Human-review gate. Every extracted drug-specific signal carries a reviewed_at timestamp. Unreviewed signals do not score the engine: the engine treats reviewed_at IS NULL as inert. A human reviewer (CMO-grade by methodology design) must approve, reject, or amend each signal before it can influence rNPV. Reviewer identity, decision, and timestamp are stamped on every signed export under engine 2.6.0. A 10% deterministic second-pass audit (minimum 10 rows per cohort, seeded for reproducibility) checks reviewer drift; disagreement above 5% triggers a re-extract with a stricter prompt and a re-review.

7. Phase 1 objective response rate — extracted but not scored

Phase 1 objective response rate (ORR) is the second drug-specific signal PhaseFolio extracts under engine 2.6.0. It is captured, surfaced in the dossier, and stamped on signed exports — but it does not score the engine in 2.6.0. This section documents why, in the same place the engine documents its scoring multipliers.

What is extracted. For each oncology-solid-tumor program with Phase 1 efficacy data, the extractor captures: ORR (percent); the modality class the program belongs to (antibody_targeted, small_molecule_targeted, immune_checkpoint, cytotoxic); the source type (FDA label, pivotal paper, registry, sponsor disclosure); a verbatim source excerpt; and a citable URL. Modality-specific high/low thresholds are pre-registered:

• antibody_targeted: high ≥ 40%, low < 15%
• small_molecule_targeted: high ≥ 50%, low < 20%
• immune_checkpoint: high ≥ 30%, low < 10%
• cytotoxic: high ≥ 45%, low < 25%

The reported ORR is adjusted toward blinded-independent-central-review (BICR) values — investigator-assessed ORR can run modestly higher than central review (Zhang et al. 2017 reports central assessment slightly lower than local, with high concordance) — and is further discounted for Phase 1 winner's-curse (Vreman et al. 2019). The adjusted value classifies into high / mid / low buckets per modality.

Why it does not score in engine 2.6.0. The Phase 0 validation backtest scaled from 28 to 43 drugs (50% cohort coverage) and found:

• biomarker_quality alone: pairwise AUC +5.2pp over baseline — stable across extractor rounds.
• phase1_orr alone: marginally validatable. Small positive lift at lower sample size, not robust at 43 drugs.
• biomarker_quality + phase1_orr combined: at 28 drugs the combined AUC showed +8.8pp lift; at 43 drugs (50% cohort coverage) the combined AUC fell below baseline (-0.3pp).

The combined +8.8pp at 28 drugs was a small-sample artifact. At higher coverage the two signals double-count correlated information: a high Phase 1 ORR is strongly conditioned on the biomarker that already scores the program, so adding phase1_orr on top of biomarker_quality re-weights the same evidence twice. The combined score actively degraded discrimination.

Governance decision. Section 5's multiplier-governance gate requires both-outcome cohort validation. phase1_orr alone is marginal; phase1_orr combined with biomarker_quality is actively harmful. The decision for engine 2.6.0:

1. Ship biomarker_quality only as the scoring drug-specific multiplier.
2. Continue to extract, display, and stamp phase1_orr on signed exports — the data is captured for diligence transparency and for the engine 2.7.0 recalibration cycle.
3. Defer scoring phase1_orr to engine 2.7.0 pending either (a) a larger cohort that supports a recalibrated independent lift, or (b) a conditional-multiplier framework that applies phase1_orr only when biomarker_quality is unknown, avoiding the double-count.

This is the multiplier-governance gate working as designed. A signal that looked promising at 28 drugs failed the both-outcomes validation at 43 drugs and was demoted to flag-only. We disclose the demotion in the same place the engine documents its scored multipliers; we do not publish only the larger combined number.

8. Non-scored risk flags and provisional disclosure

PhaseFolio extracts additional drug-specific attributes that inform diligence but do not score the engine. Each is rendered in the dossier as a flag (informational, positive, or warning) with a citable source; none move the rNPV value.

• Sponsor prior approvals (count). 0 → SPONSOR_PRIOR_APPROVALS_NONE; 1–3 → SPONSOR_PRIOR_APPROVALS_SOME; ≥ 4 → SPONSOR_PRIOR_APPROVALS_HIGH (positive).
• Grade 3+ adverse-event rate (modality-thresholded). Antibody-targeted / immune-checkpoint ≥ 30%, small-molecule-targeted ≥ 35%, cytotoxic ≥ 50% triggers G3_PLUS_AE_ELEVATED (warning); below threshold → SAFETY_PROFILE_CLEAN (positive).
• Trial randomization. Yes → TRIAL_RANDOMIZED; no → TRIAL_SINGLE_ARM.
• Primary endpoint type. Surrogate vs clinical (PRIMARY_ENDPOINT_SURROGATE / PRIMARY_ENDPOINT_CLINICAL).
• Sample-size target. Below 60 patients → SAMPLE_SIZE_UNDERPOWERED (warning); at or above → SAMPLE_SIZE_ADEQUATE.
• Indication-specific surrogacy R². Below 0.40 → SURROGACY_R2_LOW (warning); above — no flag.

These signals are non-scored by the same governance discipline as the demoted antimicrobial multipliers in Section 5 and as phase1_orr in Section 7: each would need an independent both-outcomes cohort validation to earn the right to move the number. They are useful diligence anchors — and they appear in the signed export so a reader can verify which fired — but the rNPV math is unchanged whether they are present or absent.

8.1 Authorship, AI assistance, and outside review

The drug-specific clinical signal layer was developed by PhaseFolio's non-MD founder using Anthropic Claude Opus 4.7 as the extraction engine, with an independent adversarial subagent reviewing 28 of the extractions for self-consistency (14 flagged for human spot-audit, two certain errors corrected before ship, five likely flagged, seven soft). The Phase 0 GO recommendation and the underlying validation data were prepared for two outside reviewers as of the methodology@2026-05-28 ship date and are being routed to them as part of this release:

• A HEOR / governance reviewer — was the multiplier-governance gate of Section 5 applied honestly to phase1_orr, given the small-sample artifact at 28 drugs and the demotion at 43?
• An oncology clinical reviewer — are the five validatable buckets (across biomarker_quality and phase1_orr) and the 85-drug cohort itself defensible to a clinical eye?

The drug-specific clinical signal layer was first published in methodology@2026-05-28 and citation-corrected in methodology@2026-05-29; methodology@2026-06-23 added the expanded modality-normalization disclosure. The next version (methodology@2026-06-29) was a disclosure-only change: it labelled each Table 2 multiplier by evidence tier (literature-anchored vs cohort-validated) and clarified in Section 5 that the both-outcomes cohort gate governs drug-specific and future multipliers, while the legacy population-level multipliers are literature-anchored and applied conservatively pending cohort validation. No engine value changed. Per substrate doctrine — no version is ever retroactively changed or invalidated; older versions remain valid forever and continue to verify at /verify — corrections and assumption-surface changes ship as subsequent methodology versions.

M2 (2026-09-27) is a correction with no engine value changed. Table 1 now shows the baseline table the engine computes with, beside the nearest BIO 2021 group, instead of a hand-typed table that differed from it. For five therapeutic areas, exports stamped before M2 were computed on the engine table, not on the Table 1 published at the time. M2 also removes an unsupported Citeline attribution for the modality rows, corrects the BIO 2021 author list and links the NEWDIGS analysis itself.

M4 (2026-09-28) ships engine 3.0.0 and changes engine values. Every Phase I to NDA/BLA rate is now derived from named BIO 2021 rows (Section 1.2), which fills the 125 cells that previously fell back to the area's "other" row and corrects hematologic oncology, whose earlier rows were well above BIO's hematologic cancers group. The biomarker strategy moves into the matrix, the CAR-T modifier is removed, orphan gene therapy gets one calibrated modifier instead of two stacked, and first-in-class is relabelled a PhaseFolio assumption. Exports stamped before M4 keep their engine version and still verify.

8.2 What's deliberately out of scope for Phase 1

Five constructs evaluated in the Phase 0 research are deliberately deferred to Phase 2 and beyond, in the spirit of "ship the disciplined subset, defer the rest":

• Hierarchical Bayesian PoS model. Deferred to engine 4.x or 5.x, when per-indication cohort N reaches roughly 500. The current logistic-OR transform is the right method at current cohort sizes.
• Phase 2 readout-quality multiplier. Phase 2 ORR and randomized-vs-single-arm carry independent signal, but the AMR-Sprint-1 precedent requires a per-bucket cohort that current data does not yet support.
• Mechanism-class hepatotoxicity prior. The antimicrobial Sprint-1 ablation already demoted this to a flag for that cohort; it remains a flag here.
• Sustained-clinical-response endpoint-fragility prior. Same disposition.
• Real-world-evidence calibration. Out of scope until a registry-grade RWE source clears the same governance gate as the literature anchors.

The roadmap is published openly: a signal becoming scored later is the normal path, not a methodology break. A signal staying non-scored is also a defensible disposition — transparency about what does not yet earn its way into the math is the point of this section.