Dataset Classification Guide — RheumaView™
Dataset Classification Guide

Classification precedes analysis.

RheumaView™ assigns every imaging dataset a formal classification across eight orthogonal axes — temporal structure, anatomical domain, regional sets, coverage, laterality, modality, and integration type — before any structural reading is performed.

Temporal relationship, anatomical scope, laterality, and coverage are structural properties of a dataset, not metadata. A platform that ingests images as a flat collection discards these properties — and with them the constraints that determine which comparisons, deltas, and extractions are valid.

RheumaView™ resolves all eight axes before interpretation. The framework is applied to every case in the RheumaView™ Case Library.

Eight orthogonal axes, evaluated upstream of every reading.
i.Eight classification dimensions

Eight orthogonal axes, resolved upstream of interpretation.

Each axis constrains which comparisons are valid, which deltas are computable, and which extractions are trial-compatible.

Dataset classification 01 Temporal Structure Single · Near · Longitudinal 02 Anatomical Domain Spine · Pelvic-SI · Limbs 03 Regional Sets Hand-Wrist · Foot-Ankle · Knee 04 Coverage Pattern Single · Multi · System 05 Symmetry & Laterality Symmetric · Unilateral · Asymmetric 06 Modality Structure Single · Multi-modality 07 Integration Type Pure · Composite · Mixed 08 Standardized Description Unified four-part format

Eight orthogonal dimensions resolved before any structural reading begins.

ii.Interactive · Case-ID Decoder

Case-ID notation.

Each case in the RheumaView™ Case Library carries a compact index encoding a subset of the classification. Select a case, then hover or tap a segment to resolve it to its axis.

01Temporal
02Anatomical
03Regional
04Coverage
05Symmetry
06Modality
07Integration
08Standardized

iii.The eight dimensions in detail

The eight axes, in detail.

01Temporal StructureSingle · Near · Longitudinal
Three temporal classes.

Misclassification at this layer produces invalid longitudinal comparison downstream. RheumaView™ separates baseline composites from genuine longitudinal series.

Single-Date one examination no internal comparison Near-Temporal Composite days–weeks · same baseline unified structurally Longitudinal Δ Δ matched regions across time quantitative deltas computed TIME temporal identity precedes anatomical reading

Three temporal types — distinguishing baseline composites from genuine longitudinal series.

Single-Date Dataset

All images are acquired on the same examination date. The dataset represents one structural time point, and no internal temporal comparison is possible.

Near-Temporal Composite Dataset (Staged Baseline)

Images are acquired on different dates within a short diagnostic interval — days to several weeks — yet together represent the same structural baseline. RheumaView™ treats such studies as a unified structural baseline rather than as separate longitudinal time points, provided that no clinically meaningful structural change is expected within the interval.

Longitudinal Dataset

Two or more examinations of the same anatomical region obtained at different time points to assess progression, regression, or stability. Regions must match across time points. Quantitative deltas — measured structural differences between matched time points — are computed for each aligned region.

02Anatomical DomainSpine · Pelvic-SI · Limbs
Non-interchangeable anatomical units.

Regions commonly grouped informally are treated as distinct analytic units. A cervical vertebra and a sacroiliac joint are not interchangeable data points within a single report.

Spinal Dataset

Cervical, thoracic, and lumbar spine. Does not include sacroiliac joints, pelvis, or hips.

Pelvic–Sacroiliac Dataset

Sacroiliac joints and pelvic bones (ilium, ischium, pubis). Treated as a separate anatomical unit from both the spine and the hips.

Hip Dataset

Right and left hip joints, evaluated independently from pelvic or spinal datasets.

Axial Combined Dataset

Spine and pelvic–sacroiliac components analyzed within the same study while preserving their identities as separate anatomical units.

Upper Extremity Dataset

Shoulders, elbows, wrists, and hands.

Lower Extremity Dataset

Hips, knees, ankles, and feet.

03Regional SetsHand-Wrist · Foot-Ankle · Knee
Joint groups assessed as one unit.

Regional sets are groups of joints analyzed as a single structural unit, following established clinical assessment patterns.

Hand–Wrist Set

Wrist joints together with metacarpophalangeal and interphalangeal joints. A classical unit of assessment in inflammatory arthritis.

Foot–Ankle Set

Ankle joints together with metatarsophalangeal and interphalangeal joints.

Knee Set

Tibiofemoral and patellofemoral compartments.

Shoulder Set

Glenohumeral joint and acromioclavicular joint when visible.

04Coverage PatternSingle · Multi · System
Breadth of anatomical sampling.

Coverage constrains which inferences about systemic distribution are valid.

Single-Region one anatomical region e.g. bilateral knees Multi-Region two or more regions e.g. hands and feet Regional System full anatomical system e.g. all lower-extremity joints

Coverage scales from focal sampling to full anatomical-system representation.

Single-Region Dataset

One anatomical region only. Example: bilateral knees.

Multi-Region Dataset

Two or more distinct anatomical regions within the same report. Example: hands and feet.

Regional System Dataset

A full anatomical system, typically an entire limb group. Example: all lower-extremity joints.

05Symmetry & LateralitySymmetric · Unilateral · Asymmetric
Side distribution is a structural variable.

Datasets are classified by side distribution as well as region. Laterality is a structural property and an analytic variable, not a labeling detail.

Symmetric Dataset

Bilateral anatomical structures are included. Examples: both hands, both knees, both hips.

Unilateral Dataset

Only one side of a bilateral structure is imaged. Examples: right knee only, left wrist only.

Asymmetric Dataset

Bilateral anatomy is included incompletely or unevenly — one side missing, underrepresented, or not directly comparable to the other.

Many musculoskeletal and rheumatologic disorders carry characteristic side-distribution patterns. A dataset that is unilateral or asymmetric must be classified accordingly before pattern inference begins.

06Modality StructureSingle · Multi-modality
Single- or multi-modality.
Single-Modality Dataset

One imaging modality only: X-ray, MRI, ultrasound, or CT.

Multi-Modality Dataset

Two or more imaging modalities for the same or related regions. Examples: X-ray plus MRI, MRI plus CT, X-ray plus DEXA.

Cross-modality concordance — the degree to which findings from different modalities agree or diverge — is a distinct analytical dimension within RheumaView™. The platform architecture is also designed to support integration with additional diagnostic data streams, including modalities such as EMG/NCS, ultrasound, nuclear imaging, and other structured inputs where relevant. Not all integration pathways are publicly disclosed.

07Dataset Integration TypePure · Composite · Mixed
Pure, composite, or mixed.
Pure Dataset

A single temporal, anatomical, and modality structure with no composite integration. Example: single-date wrist X-ray.

Composite Dataset

Multiple anatomical regions combined within the same time point. Example: hands plus feet X-ray study.

Mixed Dataset

Different temporal or structural types combined within the same analytical session. Example: hand X-ray from one year and sacroiliac MRI from another. Mixed datasets require explicit handling to avoid false longitudinal comparison across incompatible time points or anatomical scopes.

08Standardized Description FormatUnified four-part format
Unified four-part description.

Every dataset in the Case Library is described with a consistent four-part structure, ensuring unambiguous classification and meaningful cross-case comparison.

[Temporal Structure] + [Anatomical Domain] + [Coverage] + [Modality] Examples:
  • Single-date peripheral multi-region X-ray dataset
  • Near-temporal composite axial radiographic dataset
  • Longitudinal spinal MRI dataset
  • Composite peripheral X-ray dataset
Partnership · Trial · Research

Classification is the entry point.

This system is part of the RheumaView™ validator-governed analytical architecture. For partnership, clinical trial, or research inquiries, please contact us.

contact_us@rheumaview.com

Public materials describe categories and architecture. Implementation details remain proprietary. Patent-pending positioning. RheumaView™ is intended for use by licensed medical professionals and qualified research environments. Not a patient-facing diagnostic tool. All handling of patient-identifiable information conforms to applicable privacy frameworks, including the HIPAA Privacy Rule in U.S. jurisdictions.