脊椎後部線維脂肪腫（バッファローハンプ）の術前計画のための病理学的に検証されたBモード超音波分類システム

The absence of objective classification criteria for dorsocervical fibrofatty hump (buffalo hump) results in suboptimal surgical outcomes due to inappropriate technique selection.

To develop a data-driven diagnostic algorithm based on histopathology, shear-wave elastography (SWE), and B-mode ultrasound (B-US) fibrous assessment for dorsocervical fibrofatty hump sub-clustering, thereby facilitating therapeutic decision-making.

In 86 patients, collagen percentage was quantified from Masson-stained histology with ImageJ and classified via K-means clustering into three sub-clusters: adipose-dominant (< 23.3% collagen), mixed (23.3-38.4%), and fibrous-dominant (> 38.4%) subtypes. Diagnostic thresholds for SWE (kPa) and B-US (fibrous percentage) were determined through decision tree analysis against histologic subtypes, and their classification accuracy was systematically evaluated.

K-means clustering revealed three distinct collagen subgroups (Silhouette score: 0.661). B-US fibrous score exhibited excellent discriminatory capacity (training/test accuracy: 93.4%/92.0%), while SWE showed low accuracy (67.2%/48.0%) and overlapping classifications.

A collagen percentage-based classification is reliable for dorsocervical fibrofatty hump classification, and B-US fibrous assessment is a reliable noninvasive method for preoperative planning, with adipose-dominant lesions being optimal for modalities such as suction-assisted liposuction (SAL) or ultrasound-assisted liposuction (UAL), and fibrous-dominant cases requiring excision or limited-open approaches.

This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors   www.springer.com/00266 .