Before the Image, After the Image A Practical Guide to Structured Anamnesis and Mandatory Re-anamnesis in AI- assisted Radiology Companion paper to the ISWB Model — for clinicians, radiologists and patients

Background. The ISWB model [1] formalises the diagnostic pathway as a ten-node weighted directed acyclic graph in which a mandatory re-anamnesis node (v4) follows the imaging node (v3). It defines, at the workflow level, the border between what image-reading software can do and what clinical reasoning must still do. The formal paper leaves two practical questions open: how to raise the probability of a correct diagnosis before imaging, and how to make the v3/v4 border hold in daily practice. Aim. To translate the ISWB model and the PCCI™ structured anamnesis [2] into operational guidance that requires no mathematics, addressed jointly to clinicians and to patients. Content. Part I describes the six-branch PCCI™ anamnesis (patient factors, provocation and lifestyle, physical examination, past history, pathology and prior tests, pharmacy and toxins) as the lever that determines the quality of the imaging question, with the clinical redirection rate (TRC, target 10–25%) as its process measure. Part II states five rules that make the v3/v4 border operational — no sign-off before one question the image cannot answer; questions generated from the differential, not from the report; a separate, time-stamped four-line note; separation of reader and asker where feasible; no default precedence of the image over a contradicting answer — and describes the department- level mechanism that enforces them: the Independent Clinical Review (ICR) with a completion threshold of 90% of all cases passing through v3. The three published ISWB cases are retold in plain language, and pocket checklists for patients, clinicians and departments are provided. Conclusions. Human oversight of AI-assisted imaging becomes auditable when it is a defined step with a name, a time limit and a monthly count, rather than a disposition. The second conversation is a property of good diagnosis that AI makes more urgent; in the three illustrative cases software was involved in only one reading, and the step changed the meaning of the finding in all three. Prospective evaluation of the rules and of the ICR threshold is the next step. French,Italian and Spanish editions included.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-15
DOI
https://doi.org/10.5281/zenodo.24244933
Primary Topic
Radiology practices and education
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Before the Image, After the Image A Practical Guide to Structured Anamnesis and Mandatory Re-anamnesis in AI- assisted Radiology Companion paper to the ISWB Model — for clinicians, radiologists and patients

Francesco Matozza, Ernesto Matozza
Zenodo (CERN European Organization for Nuclear Research)
Radiology practices and education
preprint

Before the Image, After the Image A Practical Guide to Structured Anamnesis and Mandatory Re-anamnesis in AI- assisted Radiology Companion paper to the ISWB Model — for clinicians, radiologists and patients

Francesco Matozza, Ernesto Matozza
preprint en

Abstract

Background. The ISWB model [1] formalises the diagnostic pathway as a ten-node weighted directed acyclic graph in which a mandatory re-anamnesis node (v4) follows the imaging node (v3). It defines, at the workflow level, the border between what image-reading software can do and what clinical reasoning must still do. The formal paper leaves two practical questions open: how to raise the probability of a correct diagnosis before imaging, and how to make the v3/v4 border hold in daily practice. Aim. To translate the ISWB model and the PCCI™ structured anamnesis [2] into operational guidance that requires no mathematics, addressed jointly to clinicians and to patients. Content. Part I describes the six-branch PCCI™ anamnesis (patient factors, provocation and lifestyle, physical examination, past history, pathology and prior tests, pharmacy and toxins) as the lever that determines the quality of the imaging question, with the clinical redirection rate (TRC, target 10–25%) as its process measure. Part II states five rules that make the v3/v4 border operational — no sign-off before one question the image cannot answer; questions generated from the differential, not from the report; a separate, time-stamped four-line note; separation of reader and asker where feasible; no default precedence of the image over a contradicting answer — and describes the department- level mechanism that enforces them: the Independent Clinical Review (ICR) with a completion threshold of 90% of all cases passing through v3. The three published ISWB cases are retold in plain language, and pocket checklists for patients, clinicians and departments are provided. Conclusions. Human oversight of AI-assisted imaging becomes auditable when it is a defined step with a name, a time limit and a monthly count, rather than a disposition. The second conversation is a property of good diagnosis that AI makes more urgent; in the three illustrative cases software was involved in only one reading, and the step changed the meaning of the finding in all three. Prospective evaluation of the rules and of the ICR threshold is the next step. French,Italian and Spanish editions included.

Zenodo (CERN European Organization for Nuclear Research)
Radiology practices and education
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon