# Corrections in v2 (Apr 26 2026)

This is a log of every error found in the v1 outputs (initial Phase A + Phase B
deliverables from Apr 24) and how each was fixed. Read this if you're working
from the v1 files and want to know what changed.

## Error class 1: sub-agent fabricated "chat claims"

The literature sub-agent (Phase B.2) was prompted to compare chat claims to
literature. It generated several "chat claims" by inference rather than by
direct quotation, and the reconciliation file then treated those as facts the
chat actually got wrong.

| Sub-agent claim about the chat | Actual chat content | Source of fabrication |
|---|---|---|
| "Chat says Bloxham is Australian" | Zero "australian" or "australia" mentions in the chat. | Sub-agent inference, possibly from Bloxham's accent in his YouTube video. |
| "Chat says Bargouthi treats acne scars" | Enrico's archive topic is literally titled "Dr. Bargouthi's VP FUE trials" [#217]. The chat KNEW. | Sub-agent inference from the chat using Bargouthi's dose figures in protocol discussions for various indications. |
| "Chat says Qazi is Canadian, $9k CAD" | silk thread said "9k for 1 session" [#1721] without currency. silk thread is Canadian, but Qazi is in Newport Beach CA — chat made no nationality claim. | Sub-agent inferred CAD because the consulter is Canadian. |

Fix: rewritten "Doctor-by-doctor reconciliation" table in `RECONCILIATION.md`
with column 1 = "what the chat actually said" + msg ID, instead of fabricated
"chat claims to verify against." Where the chat said nothing, the row was
removed rather than invented.

## Error class 2: forwarded-message attribution bug

Telegram exports list a forwarded message's `from` as the *forwarder*, not the
original author. The per-person extraction script used `from`, attributing 36
substantive messages to Enrico (the chat's primary curator/forwarder) that
were authored by other users. Documented in
`profiles/_attribution-corrected.md` and `raw-extracted/_attribution-fixes.md`.

Most consequential reattributions:

| Message | v1 said author | Actual author | Significance |
|---|---|---|---|
| **#43 — foundational protocol** | Enrico | **N G** | The chat's foundational protocol (raw VP powder + PBS + sonicator + 4mm 30g mesotherapy needle, 0.4 mg/cm² starting dose) was N G's draft. Enrico forwarded and stewarded. |
| **#186 — dose comparison table** | Enrico | **IgnotuS** (with @Dorian tag) | The canonical "Bargouthi 0.24/0.32/0.4, Bloxham 0.4/0.6/0.8, Longaker pigs 0.18 mg/cm²" table. The unsourced "0.18 mg/cm²" pig dose may have originated here — track IgnotuS / Dorian if you want to find its source. |
| **#420-423 — Visudyne pricing questions** | Enrico | **Daisey** | Daisey asked the substantive Visudyne-vs-powder question. v1's "Daisey has zero VP-substantive content" was strictly true for direct posts but missed her forwarded-into-Buying-topic exchange. |
| **#424 — "Visudyne ≡ raw VP powder, both are same"** | Enrico | **martha Halli** | This is the chat's central formulation error. The "both are same" claim drove the chat consensus toward raw powder + PBS + sonicator. martha Halli authored it, not Enrico. |
| **#609-610 — K K repigmentation strategy** | Enrico | **Cultural Dust** | Latisse + microneedling + 300nm UVB phototherapy. Cultural Dust is the chat's most clinically-savvy voice; v1 missed this. |
| **#360-361 — promotion strategy** | Enrico | **K K himself** | K K is more vocal in the archive than v1 (1 message) suggested. He authored 4 messages total (3 forwarded by Enrico + 1 direct "Ok"). |

New mini-profiles surfaced for posters whose substantive content was entirely
in forwards: **IgnotuS**, **Darren Weir** (helminth scar-free healing news),
**Anyone** (pharmacist contact for sourcing), and a corrected expanded
mini-profile for **martha Halli** and **Cultural Dust**.

## Error class 3: misclassification of Fabio Carneiro

v1 treated "Fabio Carneiro" as a separate doctor entity. He is the chat
member **Fabio** (Telegram handle `@fabiomcarneiro`), planning a facial
keloid excision in October-November 2024 with self-injection because his
surgeon refused to inject intraoperatively. The "Fabio Carneiro" topic
follows the same pattern as the K K, Sc0t, S S, George Kal topics —
chat-member case studies, not doctor profiles.

Fixed in: `RECONCILIATION.md`, `case-studies/Fabio.md` (new),
`profiles/_attribution-corrected.md`, `by-topic/outcomes.md`,
`by-topic/who-actually-tried-it.md`.

## Error class 4: incorrect outcome attributions

| v1 claim | Corrected |
|---|---|
| "Sc0t used 0.8 mg/cm² (presumed)" | The chat **never states a Sc0t dose**. The 0.8 was extrapolated from Bloxham's general result. C5 unsourced within the archive. |
| "Bloxham uses thick green liquid" | The chat says "VERY thick" only [#660, #664]. The "green" qualifier comes from `drafts/breif-notes.txt` (Lucy & Sage's external notes), NOT the chat. v1 conflated. |
| "K K case is well-documented in chat" | K K speaks **once** in the archive ("Ok" [#3431]). All clinical parameters are Enrico-relay (C2) or Enrico photo-interpretation (C3). K K's formulation, concentration, volume, timing, surgeon name, surgery date — **NOT STATED**. |
| "Qazi quoted $9k CAD for one VP session" | silk thread's $9k quote was for a single-session aesthetic bundle (laser + TCA CROSS + subcision + 2× Bellafill + "laser scar repair complex") [#1734]. **None of the line items are VP.** Currency was unspecified in chat. |
| "Bargouthi went up to 1.0 mg/cm²" | Enrico [#387]: "0.4, 0.6, 0.8 and 1 mg in seemingly a single cm²" — sums to 2.8 mg/cm² cumulative if literal; the "seemingly" hedge signals he's relaying without confidence. **Not in Bargouthi's public record.** |

## Error class 5: missing doctors and techniques

| Missing in v1 | Added in v2 |
|---|---|
| **Dr. Tanveer Janjua** (Bedminster NJ) | Mentioned by Enrico [#455] alongside Yang Liu and Toyos as a VP-cohort doctor. **No public VP connection** — chat-only. Practice is real (facial plastic surgeon, Yale-trained). |
| **Mosaic Surgery / Pallaoro / MCTG / SST** | Yang Liu offers "Mosaic Surgery" [#441]; Italian counterpart at [pallaoro.com/mosaic-surgery.html](https://www.pallaoro.com/mosaic-surgery.html). Uses 1mm Trilix micro-cores → autograft plugs. **Not VP, not scarless**, 2-4 sessions. The "MCTG / SST" acronyms are chat shorthand with no verified meaning. |
| **Dr. Melissa Toyos peer-reviewed VP papers (2025)** | TWO papers in *J Dermatology Research & Therapy*: post-transplant VP at 1 mg/cm², 1 cc, intra-patient comparison; VP + dermal wounding for AGA at 25 mg compounded subcutaneous. **My v1 "zero peer-reviewed human VP scar evidence" claim is wrong as of 2025.** |
| **SCARFREE-001 trial design** | Phase 2, Odense University Hospital, 12 participants (abdominoplasty), intra-patient design, 3 concentration arms (0.5/1.0/2.0 mg/mL) vs saline, primary endpoint POSAS at 3 months. Note: tests CONCENTRATION not mg/cm² DOSE. |
| **NCT06944249** | University of Geneva, 24 participants, melanoma-margin excisions on the back, 36-month follow-up. Not yet recruiting as of April 2025. |
| **Helminth / Hymenolepis diminuta scarless healing** | Darren Weir posted Rutgers / studyfinds links [#2616-2617]. Adjacent biology to VP; real academic research thread. |
| **Miso Sarang Clinic (Korea) pinhole laser** | silk thread referenced this as a separate aspirational scar-revision technique [#1850-1862]; ~$25k. |

## Error class 6: missing dose-consensus timeline

v1 had no record of HOW the chat consensus evolved. Now in
`by-topic/timeline-protocol-evolution.md`. Headlines:

- **Aug 11**: N G's protocol (forwarded by Enrico into Protocol topic)
  establishes 0.4 mg/cm² + raw powder + PBS + sonicator + 4mm 30g mesotherapy.
- **Aug 12, 13:46** [#384]: Enrico pivots to 0.8 mg/cm² based on Bloxham's
  YouTube result. Nobody pushed back. Pivot took ~24 hours.
- **Aug 16-17**: Toyos and Yang Liu and Janjua surface as VP-considering doctors.
- **Aug 21**: K K case photos posted [#581-#610]; depigmentation finding;
  Cultural Dust authors the repigmentation strategy.
- **Aug 25-27**: dose ceiling debate (1.0 acceptable, ≥1.0 risky); Fabio's
  planned self-injection generates injection-method discussion.
- **Archive cutoff Aug 27** — George Kal's first laser+VP session was
  scheduled for Aug 28. No outcome data after that.

## Error class 7: photo-evidence invisibility

The 120 captioned media messages were `[media: photo]` stubs in v1 profiles.
Re-indexed; key findings:

- George Kal's planning photos have substantive captions: "These are the
  liobs. The idea is VP to regenerate these liobs with basket-wave collagen"
  [#534]; "Before of my stretch mark, here I will inject VP after the laser
  treatment" [#2198]; "Deep cavitation into dermis, I will start the
  regeneration with VP from here" [#2229].
- Enrico's K K archive photos [#584-#603] each have substantive
  interpretation in their captions — the K K case is documented through
  these photos, not through K K's own words.
- Bloxham archive [#638-#642]: 4 photos, no captions. The visual evidence is
  invisible to text-only extraction.
- Daisey, silk thread, etc. posted photos without captions in Off-Topic —
  consistent with their lifestyle/community-building roles.

## v4 → v5 corrections (Apr 27 2026)

User feedback drove a dedicated hunt for the missing pig-study primary source
plus transcripts of Bloxham's YouTube series via the scrape-tools (`paper`,
`transcribe`, `archive`). Results in
`research/verteporfin/_PIG-STUDY-HUNT.md` and
`research/verteporfin/_BLOXHAM-VIDEOS.md`. Headlines:

- **0.18 mg/cm² confirmed unsourced.** Hunt across 2025 STM pig paper
  (PMC12893899, full text fetched), 2021 Science mouse paper, 2022 Cell
  Stem Cell paper, PSRC 2024 abstract, C-DOCTOR proposal, NCT06944249 —
  none contain 0.18. Closest published mg/cm² value in Mascharak corpus is
  **0.106 mg/cm² in mice** (30 µL × 1 mg/mL ÷ 0.283 cm² wound from 2022
  STAR Methods). The pig paper does dose-finding at 1, 2, 4, 8 mg/mL → 2
  best; per-site volume lives in supplementary methods which is
  paywall-gated.

- **Bloxham YouTube series mostly Private.** 4 of 5 of Bloxham's update
  video IDs are now Private; Wayback has no snapshots. Only the public
  Sept 2023 Parissis Q&A remains transcribable. Surviving evidence for
  the dose anchor: Follicle Thought summary + Bloxham's HRN forum posts.
  Bloxham's own current-practice quote: "Most seem to be falling into the
  0.8 mg/cm² – 1 mg/cm² range."

- **Visudyne IS dark green** — etymology: "verde porphin" (green
  porphyrin). The chat's "thick green liquid" description of Bloxham's
  formulation is correct. v3 of this synthesis incorrectly flagged "green"
  as conflation with the user's draft notes; both the chat and Lucy/Sage
  are correct because Visudyne is intrinsically green.

- **Tone correction (also v4→v5).** Earlier versions framed surgeon
  refusals as "signal not friction" and the n=4 underdose pattern as
  "post-hoc fitting." Both were too harsh on the chat members. Surgeon
  refusals are common risk-aversion from liability + conservatism + lack
  of established protocol — not strong signal that VP doesn't work. n=1
  mechanistic evidence is fine for "biologically interesting"; just not
  for clinical recommendation.

## Still open / not fixed

- **Bloxham's 0.6 middle dose** — plausible-by-interpolation but only
  "medium" is publicly named. The chat's "0.4/0.6/0.8" reading is
  reasonable but the 0.6 figure isn't directly cited.
- **K K's formulation, concentration, volume, timing, surgeon, surgery
  date** — not stated in the archive. K K speaks once in the entire chat.
- **Sc0t's dose, formulation, concentration, volume, timing, date** —
  not stated in archive.
- **Pig per-site dose** — in paywall-gated STM supplementary methods PDF
  (Cloudflare/PMC PoW gate; not retrievable without browser cookies).
- **Whether Bargouthi actually escalated above 0.4 mg/cm²** — not in his
  public record; Enrico's claim has the "seemingly" hedge.
- **Whether George Kal's pico+VP experiment produced any result** —
  archive ends Aug 27; first session Aug 28.
- **Yang Liu's specific protocol** — chat references his Weibo and
  case-by-case results but not his actual dose / formulation / concentration.

## Files added in v2

- `case-studies/{KK,Sc0t,Fabio,Bloxham,Bargouthi,Qazi,Yang_Liu,_NOTES}.md`
  — clean, quote-first reconstructions per case, written by Opus
- `by-topic/timeline-protocol-evolution.md` — when the consensus shifted and why
- `profiles/_attribution-corrected.md` — re-attribution of 36 forwarded messages
- `profiles/_verification_daisey_silkthread.md` — verification that v1's
  "0 VP self-experiments" framing for Daisey/silk thread is correct
- `raw-extracted/by-author/` — re-extracted per-author message dumps (uses
  `forwarded_from` when present)
- `raw-extracted/_attribution-fixes.md` — full list of 72 reattributed messages
- `research/verteporfin/doctors.md` — appended "Verification Pass 2 (April 2026)"
  section with non-English sources, Janjua, Mosaic Surgery, Toyos 2025 papers
- This file (`CORRECTIONS-V2.md`)

## Files updated in v2

- `RECONCILIATION.md` — corrected doctor table, removed fabricated "chat claims",
  added Fabio chat-member fix, Sc0t/Bloxham corrections
- `FINDINGS.md` — added top-of-file correction notice
- `by-topic/_INDEX.md` — corrected doctor descriptions
- `by-topic/doctors.md` — corrected Qazi $9k breakdown, flagged 0.18 pig dose
- `by-topic/outcomes.md` + `who-actually-tried-it.md` — Fabio relabeled
- `profiles/silk_thread.md` — corrected Qazi $9k quote framing
- `profiles/_MASTER.md` — already accurate after v1