AritmoLab · Policlinico San Donato

Role of AI in the Analysis of Unstructured Clinical Databases

How we turned a mix of structured and free-text clinical records into a research data platform — the AritmoLab platform experience.

AritmoLab A clinical data platform hosting cardiology data and analysis tools

Dr. Marco Pancotti - MultiPhysixLab Dr. Sara Paratico - I.R.C.C.S. Policlinico San Donato

“Multidimensional Characterization of Cardiac Arrhythmias: Role of Electrocardiology in the Artificial Intelligence Era”

San Donato Milanese, Milan, Italy · 2–3 October 2026

Good morning everyone, and thank you for being here. I'm Marco Pancotti, and with my colleague Dr. Sara Paratico we work on the clinical data platform of AritmoLab at Policlinico San Donato. Today I want to talk about a problem that every hospital knows very well. The most valuable clinical information we have — the diagnosis, the patient's history, the outcome of a test — is written in plain free text, inside systems that were never designed to make that text usable. Over the next twelve minutes I'll show you how we used artificial intelligence, in three different roles, to transform twenty years of cardiology records into a database that researchers can actually query — and that clinicians can verify. My colleague Dr. Sara Paratico will join me to show how we read the clinical text itself.

AritmoLab · Policlinico San Donato

Four islands and four missing pieces

?Clinical Intelligencedashboards on the clinical history of the Unit and its patients

?Datawarehousethe single source of truth

?ML-ready datacohort tables, ready for model training

?Professional PortalAritmoLab — the grown-up Omics Portal

Missing Pieces

▸ click a system: how it was used — and what held it back · click a missing piece: what it would have given

Before I show you what we built, let me take you back twenty years. Cardioref, our electrophysiology database, had twenty years of clinical records: as a management tool for patients, daily activity and planning it was — and still is — excellent. But for research it was almost unusable, because most of the clinical knowledge was written in free text, outside the structured fields. The genetic data lived in its own system, which produced Excel sheets, with no connection to Cardioref. Our first Omics Portal was a promising idea, still embryonic and not integrated with anything. And a constellation of ECG subsystems that could only hand you a CSV file, on request. On top of that: no clinical intelligence on the historical data, no single datawarehouse as the source of truth, no ML-ready data for predictive statistics, and no professional portal to bring it all together. Four islands, four missing pieces. Click any system and I'll show you exactly what held it back. And here is what we built.

AritmoLab · Policlinico San Donato

The plan

What we wanted to build

One platform, two destinations: the 360° patient portal and a research-ready warehouse. Built on open-source pillars, with AI coding agents wherever they helped. This talk follows the hardest piece — the unstructured data.

What existed

Cardiorefcardiology records · procedures · letters

Genetic datalabs · variants · nomenclature

ECGsignals · device follow-up

+Future sourcesnew subsystems can be connected as sources

→

What we built

Stagingraw replica of the sources

→

Integrationcleaned, normalized, deduplicated

AI reads the clinical text: pathologies, procedures, drug-challenge outcomes

→

Star schemadimensional transformation

→

Datamartsresearch-ready marts

AI builds them on demand from plain-English questions (ThothII)

→

AritmoLab Data Warehousequeried for research

AritmoLab Portalmanagement & exploration

AI coding agents — Anthropic · OpenAI — used wherever they helped, 360° in the code Airflow · orchestration Superset · dashboards Django · portal scaffolding Authentik · auth — GSD LDAP

Here is the whole project on one slide. On the left, what already existed: three hospital data sources — Cardioref, our electrophysiology records; the genetic data coming from the labs; and the ECG signals. On the right, what we built in AritmoLab: a staging replica, an integration layer where the data is cleaned and normalized, a star-schema warehouse, and the research datamarts — delivered through the data warehouse and a management portal. Everywhere you see the brain symbol, that's where AI works for us — click each brain and its contribution card opens, tied to the brain by a red line. At integration, AI reads the clinical text. At the end, AI builds the datamarts on demand from plain-English questions. And the mappings that bring the sources in were themselves drafted by AI, then revised by humans. That was the plan. And to build it, we used AI everywhere it helped — drafting configs, reading clinical text, and at the very end, producing the datamarts themselves. This talk follows the hardest piece of that plan: the unstructured data. My colleague Dr. Paratico will show you exactly how the text reading works. But first, the first step of the climb: the mappings.

AritmoLab · Policlinico San Donato

From free text to a research platform

The clinical truth lives in unstructured columns

“Il paziente riferisce sincope ricorrente; ECG basale con sopraslivellamento ST in V1–V3; test provocativo con flecainide positivo per pattern Brugada.”

EN“The patient reports recurrent syncope; baseline ECG with ST-segment elevation in V1–V3; flecainide provocation test positive for a Brugada pattern.”

SourcesCardioref · Genetic data
Omics Portal · ECG

→

Stagingraw replica

→

Integration3NF · text mining

→

Data Warehousestar schema

→

Martsdbt · Superset

four challenges to face · the answers come next

01

Volume without structuretwenty years of records locked in free-text columns

02

Ambiguous languagenegation · family history · bilingual shorthand

03

No shared ontologyevery system names conditions its own way

04

Proving reliabilityextraction must be demonstrably correct

outputs of the datamarts

Dashboardsclinical review on Superset

Machine Learningcohort tables for model training

Predictive Statisticsoutcomes & risk analysis

ThothII — natural-language → SQL over the DWH, builds the final datamarts

AritmoLab · Policlinico San Donato

What the text mining produced

Deterministic AI, audited numbers

58,438letters read

73,389pathologies extracted

10,908tests parsed

2,307Brugada patients

≈ 240letters / month

Letter readerevery discharge letter, Italian & English

Clinical matcherrecognises diagnoses, procedures, test outcomes

Context guardnegations & family history kept apart

Ontology sorterevery finding into its clinical category

Traceable. Every finding carries the version of the rules that produced it (pattern_version): any number can be rebuilt years later.

Audited. A new rule set goes live only after it proves ≥95% accuracy on 100 records reviewed by hand.

[DRAFT — Sara] I'll take you inside these numbers. We read 58,438 discharge letters — every single one, every night. From them the text miner extracted 73,389 pathology records, classified into our clinical ontology. It parsed 10,908 drug-challenge tests, distinguishing the therapy from the actual test. And at the end of the chain: 2,307 patients with a confirmed Brugada pattern — a cohort nobody could have built by hand. This is not a one-off migration: the pipeline runs every night, and today it processes on average two hundred and forty letters a month. Four simple pieces do the work: a letter reader for the Italian and English text, a clinical matcher that recognises diagnoses, procedures and test outcomes, a context guard that keeps negations and family history apart, and the ontology sorter that files every finding into its clinical category. The key word here is deterministic: no black box. Every number can be traced back to the rule that produced it.

AritmoLab · Policlinico San Donato

Clinical NLP

Reading clinical text is not keyword matching

Negation — “fibrillazione atriale esclusa”, “non FA” (“AF ruled out”, “no AF”): the rule scans a window around every match before deciding
Family ≠ patient — “padre con FA” (“father with AF”) is attributed to the family member, not to the patient
Two languages — every pattern matches IT and EN forms: fibrillazione atriale / atrial fibrillation
Abbreviations & noise — FA, f.a., “TA 140/90” (blood pressure, not a diagnosis)
One field, many statements — the text is split into clauses before matching

what the rules seeexamples from the corpus

“padre con fibrillazione atriale”
“father with atrial fibrillation”
  → patologia: FA · attribuzione: familiare
pathology: AF · attribution: family

“fibrillazione atriale esclusa”
“atrial fibrillation ruled out”
  → patologia: FA · negato
pathology: AF · negated

“test provocativo con flecainide
  positivo per pattern Brugada”
“flecainide provocation test, positive for Brugada pattern”
  → test: flecainide · esito: POSITIVO
test: flecainide · outcome: POSITIVE

AritmoLab · Policlinico San Donato

The clinical ontology

Two tiers, one clinical order

TIER 1 · arrhythmic11

AFFABrugadaFlutterATTAAVNRTTRNPSVTWPW / TPSVVTTVVFFVVentricular ectopyExtrasistolia V.SyncopeSincopeLong QTQT lungo

TIER 2 · structural5

AV blockBlocco AVBundle branch blockBlocco di brancaCardiomyopathyCardiomiopatiaHeart failureScompensoValvular diseaseValvulopatia

01

Order encodes clinical precedence — “Brugada” is matched before “TV”, so “substrato per TV” (“substrate for VT”) can't mask a Brugada pattern

02

Synonyms live inline — FA / f.a. / fib. atriale / atrial fibrillation → one canonical label

03

Versioned like software — semver for the pattern library, pattern_version on every extracted row

The output: text becomes recorded data

01

Quantitative data on the recordsevery procedure and implant the text analysis reads is written back as structured, quantitative values on the record that describes it

02

Straight into the DWH flowthese records enter the data-warehouse generation like any other source

03

As if typed at the visitthe data lands exactly as if clinicians had keyed it in themselves during the visits

[DRAFT — Sara] Extraction needs a target vocabulary — that's the ontology. Tier 1 holds the eleven arrhythmological categories that matter most for our research; Tier 2 holds five structural conditions. The order of the rules is itself clinical knowledge: Brugada patterns are tested before TV, because "substrato per TV" often appears in Brugada reports and would otherwise mask the diagnosis. The ontology is versioned like software: a semantic version for the pattern library, stamped on every extracted row. When we add a synonym or fix a rule, the change is traceable — and the data can be rebuilt. And this is the output of the whole work: reading a letter writes structured, quantitative values back onto the records that describe the procedures and implants it contains — and those records enter the warehouse generation exactly as if the clinicians had typed them during the visits. Text becomes data, indistinguishable from bedside data entry. Sara: review the category list and the Italian labels before final. Labels are now English-first (audience is mostly foreign) — please confirm the EN terms, especially PSVT as the umbrella for WPW/TPSV and AT/AVNRT for TA/TRN.

AritmoLab · Policlinico San Donato

Validation

Trusting the text is a process, not a promise

Quality gates — ≥95% procedure classification, ≥85% pathology detection, validated on 100 manually reviewed records
Clinical review — the extracted Brugada cohort was validated by clinicians on Superset dashboards
Feedback loop — ~893 Brugada false positives found and removed in pattern-library v1.3.1
Nothing is silently dropped — negated and family-attributed findings are stored too, filtered only at the mart layer

Quality gateAn extraction is accepted only after 100 manual reviews confirm the agreed accuracy — the same standard for every new pattern release.

AritmoLab · Policlinico San Donato

The portal, today

AritmoLab — a quick tour

screenshot

[MP: 6-7 real screenshots of the portal, anonymized or demo data]

AritmoLab · Policlinico San Donato

All good? Not yet.

The warehouse speaks SQL. Research needs more.

Clinicians and researchers don't write SQL — a star schema is built for analysts, not for the ward
Health Intelligence — to present the past, the Unit needs ready-made aggregates: volumes, incidence, outcomes
Predictive research — multivariate analysis and ML models need clean, wide, cohort-shaped tables
Hand-made datamarts are the bottleneck — even with coding agents, each one takes hours of careful work

The gapTwenty years of data, one warehouse — and no fast road from a research question to an answer.

AritmoLab · Policlinico San Donato

Role 3 — AI for analysis

Ask the warehouse in plain English

“How many Brugada patients had an effective ablation?” a researcher, in natural language

NL → SQLAI writes the query over the star schemaAI

→

Human reviewthe researcher checks every proposed query

→

Datamartcurated, research-ready martsAI

→

Supersetstatistics & dashboards

The session is a conversation: refine the question, iterate on the datamart — every step reviewable.

AritmoLab · Policlinico San Donato

ThothII, step by step

From question to datamart — the app

screenshot

[MP: 6-7 real screenshots of ThothII — ask, review SQL, datamart, dashboard]

AritmoLab · Policlinico San Donato

From datamart to discovery

One datamart — many questions answered

Multivariate analysis[SVG chart — next step]

Machine learning[SVG chart — next step]

[Illustrative data, modeled on published arrhythmology predictors — Brugada focus]

AritmoLab · Policlinico San Donato

Questions?

Thank you

Want the deep dives? The technical walkthroughs of the text miner, the mapping agents and ThothII are coming as videos on our Substack.

Dr. Marco Pancotti - MultiPhysixLab Dr. Sara Paratico - I.R.C.C.S. Policlinico San Donato