1	Anthony A. Mancuso, MD Professor of Radiology and Otolaryngology Chairman, Department of Radiology University of Florida College of Medicine Gainesville, Florida mancua@radiology.ufl.edu
2	“The Hardest Thing We Do Is Call A Chest Radiograph Negative” Leo Rigler, MD circa 1974
3	“The Hardest Thing To Do Is Call An Imaging Study Negative” Assuming you believe this is true you need: To thoroughly know the normal anatomy And normal variants (based on reliable normative data when available) Have clinical information (context) Rules/structure by which to call study negative Reject interpreting studies by “Gestalt” only To be trained in and accept this discipline and value system
4	Goals for our reports Avoid false negatives-which is what this presentation is about Set reasonable expectations about what a negative imaging can contribute in specific clinical context State a degree of confidence (whenever possible) of the negative interpretation for excluding specific pathologies (whether asked or not) Identify specific clinical contexts with high error rates Basis for this presentation Interpretive Process (protocol application) Error poses unusually high risk to patients
5	As Examples-situations with high error rate for disease exclusion have been chosen Temporal bone- Anatomically a confined, complex (lots of small structures) region with numerous, diverse indications for study Invasive fungal sinus disease in immune compromised patients—critical situation that requires exclusion of a specific condition with a high degree of confidence Perineural cancer spread- specific, high frequency task with few discreet observations necessary but very high prognostic and therapeutic impact Facial pain and otalgia- “screening”of a relatively large anatomic area, requires numerous observations, diverse pathology—incidence of causative pathology relatively low
6	Will develop concepts of the proposed “Interpretive Model” using T Bone Systematic consistent analysis of key anatomy Incorporation of normative data into that analysis Applied disease patterns/pathophysiology Creates “rules” for negative interpretation Reject “Gestalt” reading as a sole method of analysis Will then apply to clinical circumstances with more acuity and risk
7	Temporal Bone- Confined,complex (lots of small structures) anatomic region with numerous indications Keeping clinical context simple Conductive hearing loss Middle ear Sensorineural hearing loss/Vestibular Inner ear
8
9
10	Reject a Gestalt Only Approach
11	Temporal Bone- Jumble of (after “Gestalt”) pertinent “exclusionary” observations Incus long process and incudostapedial joint, modiolus, vestibular aqueduct, SSC dehiscence, ossicular ligaments, region of fissula ante fenestrum, oval window,bony island of LSC and cochlear height, round window, stapes footplate, facial canal position,cochlear aqueduct
12	Temporal Bone Specific Observations after “Gestalt”- Organized by Indication/Anatomy Conductive HL/Otosclerosis/Middle ear Incus long process and incudostapedial joint, stapes footplate, oval window, ossicular ligaments, oval window, region of fissula ante fenestrum, round window, stapes footplate, facial canal position Sensorineural, Vestibular Sx/ Inner ear Modiolus, vestibular aqueduct, SSC dehiscence, bony island of LSC (less than 2.6mm), cochlear height (less than 4.4mm), cochlear aqueduct Use these to construct report and/or decision support tool
13	Use normative data to construct your “interpretive model”
14
15
16
17
18	Add to normative data and rules about disease patterns/pathophysiology in building your own “Interpretive Model”
19
20
21
22
23	Normative data may be revised with more experience or may not be corroborated as reported initially Interpretive models may require validation and maintenance-- they may even become obsolete
24
25	Normative data may need to be put in context when reporting Interpreting physician needs to take responsibility for normative data in clinical context
26
27	Normative data may require rethinking protocol sufficiency and related workflow
28
29	Clinical Context - Fungal Sinus Disease in the Immune Compromised Patient- critical situation that requires exclusion of a specific condition with a high degree of confidence Non-invasive (colonization usually with chronic allergic polypoid rhinosinusitis) Invasive Diabetics and the “frail elderly” sometimes intermediately or chronic invasive disease Highly invasive—immune compromised patients—most commonly BMT Key Disease Pattern-Angio-invasive
30	Building Blocks of a Clinically Relevant Interpretive Model and Related Report Reflective of clinical context Contains information that convinces the reader of interpreter’s understanding of the problem and capability to render a useful response Expresses degree of confidence Positive and negative predictive values when possible Unambiguous and useful observations in language that supports diagnostic impressions degree of confidence
31	Disciplines necessary to build useful Interpretive Models or Systems
32	Fungal Sinus Disease -combining interpretive disciplines to exclude invasive disease Discipline-Know Anatomy Superficial aponeurotic system of the face-a.k.a. SMAS Periantral fat planes as they relate to vessels-distal max, infraorbital, post. sup. alveolar Discipline- Understand Spread Patterns Contiguous mucosal Angio-invasive so it follows neurovascular bundles Can leave bone intact
33	Fungal Sinus Disease -combining interpretive disciplines to exclude invasive disease-specific observations to call a study low risk (negative) Goal of exclusion-avoid an unnecessary ENT consult without putting the patient at risk Observations Unilateral (more risk) v. Bilateral (low risk) Mucosal spread Posterior nasal cavity (small unproven risk) Nasopharynx Perivascular venues Infraorbital-extraconal fat orbital floor, anterior antral fat pad Post. Sup. Alveolar-retroantral fat Maxillary-sphenopalatine foramen, pterygopalatine fossa Bone erosion
34
35
36
37
38
39
40
41	Fungal Sinus Disease -combining interpretive disciplines to exclude invasive disease-specific observations to call a study low risk (negative) Illustrative cases—Gradation of Disease from Obviously Positive to Negative (low risk-not requiring ENT consult)
42
43
44
45
46
47	Clinical Context- Perineural Cancer Spread High frequency task with few discreet observations necessary but very high prognostic and therapeutic impact Specific clinical context where to this date there continues to be high interpretive and process (protocol application) error rate and high risk to patients Set expectations about what imaging can contribute-exclusion of macroscopic perineural tumor spread Anatomic pattern of disease SMAS still works because it identifies perivascular and perineural mechanism
48	Normal anatomy/Normative data/ disease patterns frequently cross over to other clinical situations This type discipline becomes synergistic and by expanding the use of an Interpretive Model process as opposed to Gestalt (“looking and hoping”) approach Before we were looking at arteries –now nerves—they go to the same places
49
50	Clinical Context- Perineural Cancer Spread -What are your rules to call the study negative? Know anatomy normative observations about SMAS and related deeper fat pads Check fat planes deep to SMAS- supraorbital, infraorbital, mental nerves Know that facial nerve runs with SMAS peripherally Check for abnormal thickening of SMAS Know anatomy of auriculotemporal nerve connection between facial nerve and V3 Check anatomic region of auriculotemporal nerve Know normative data about Facial and Trigeminal nerve more proximally
51
52	Auriculotemporal Nerve as a bridge between V3 and the Facial Nerve
53	Auriculotemporal nerve as a bridge between V3 and the Facial nerve
54
55	Normal anatomy/Normative data/ disease patterns frequently cross over to other clinical situations This type discipline becomes synergistic and by expanding the use of an Interpretive Model process as opposed to Gestalt (“looking and hoping”) approach Lets appreciate this synergy with the auriculotemproal nerve anatomy and how it helps build our Facial Pain/Otalgia model
56	Facial Pain and Otalgia “Screening”of a relatively large anatomic area, requires numerous observations, diverse pathology—incidence of causative pathology relatively low Highlight greatest source of errors: Failure to recognize deeply infiltrating pharyngeal pathology Protocols not inclusive of ,or tailored, to all possible pain generators Anatomic Highlights-Parotid/auriculotemporal nerve source, pharyngeal variations
57	Facial Pain and Otalgia Pain Generators & Protocol Impact Clinical pain pattern as specific as possible Otalgia-otogenic and referred-pain generators Intracranial/meningeal (MR) Temporal bone and skull base (CT) Odontogenic (CT) TMJ (MRI) Salivary gland (CT) Pharyngeal (CT) Vascular (CTA or MRI/MRA) Cervical spine (CT or MRI) Otalgia Imaging exam-Craniofacial plus neck to pyriform sinus apex (C5/6) with detailed temporal bone /skull base images
58
59	Facial Pain v. Otalgia Pain Generators & Report Impact -What are your rules to call the study negative with a reasonable degree of confidence? Know the pain generators, all pertinent anatomy including SMAS and related fat pads and how to exclude subtle infiltrating pharyngeal pathology Otalgia— specific report language related to potential pain generators Vth ,IXth and Xth CN nuclei to pertinent peripheral branches (e.g. auriculotemporal nerve) Referred-skull and meninges, odonotgenic (TMJ), pharynx, salivary glands (parotid in particular) Craniofacial— specific report language related to potential pain generators Vth nerve nucleus to all peripheral branches Referred-skull and meninges, face, sinonasal, odonotgenic Consider using a structured report as a decision support tool to be sure all sources have been checked— must have exclusionary rules for sources
60	Auriculotemporal nerve as a bridge between V3 and the Facial nerve
61
62
63	The other main source of most false negatives in Otalgia-Pharynx Cancer Anatomy Normal Variants Reported Normative Experience Pathophysiology—referred pain Patterns of Disease
64
65	PHARYNX Detection of Aggressive Pathology Deep planes symmetric-recognize aggressive pathology by invasion of the parapharyngeal fat Nasopharynx Upper oropharynx Less reliable lower oropharynx Airway contours vary markedly – lymphoid tissue
66
67
68
69
70