Accurate noninvasive quantification of hepatic fat is important for diagnosis and monitoring of fatty liver disease. This study tested whether photon‑counting CT (PCCT) can provide robust hepatic fat fraction (FF) measurements across different body sizes using a single virtual monochromatic imaging (VMI) protocol. The aim was to identify common PCCT VMI settings that minimize quantification error in both medium and large anthropomorphic liver phantoms.
The investigators embedded six non‑iodinated fat lesions with ground‑truth fat fractions ranging from 5% to 40% in anthropomorphic liver phantoms designed to represent two abdominal sizes: a medium abdomen (25 × 32.5 cm2) and a large abdomen (31 × 39 cm2). Scanning was performed on a PCCT system (NAEOTOM Alpha) at two tube voltage settings: 120 kV and 140 kV. Virtual monochromatic images were reconstructed across a continuous energy range from 40 keV to 190 keV in 1‑keV increments.
CT numbers (Hounsfield units, HU) of the fat inserts were measured at each VMI energy. The relationship between measured HU and ground‑truth FF was characterized by linear regression. From these regression models, calibration equations were derived to convert HU measured in a given VMI to an estimated fat fraction percentage.
Performance across energy levels and tube voltages was quantified using the relative root‑mean‑square error (rRMSE) and absolute root‑mean‑square error (RMSE) between estimated and ground‑truth FF values. The investigators sought a single VMI energy and tube voltage combination that minimized rRMSE across both phantom sizes using identical calibration parameters.
Using the linear regression approach, the authors evaluated the full VMI range for both tube voltages to find the energy that yielded the lowest quantification errors in both medium and large phantoms. The criterion for selection was the minimum combined rRMSE across phantoms when applying a single calibration equation to both sizes.
A VMI energy of 70 keV reconstructed from scans acquired at 140 kV demonstrated the best overall performance across the two phantom sizes. The derived calibration equation at that setting was reported as:
FF (%) = -0.689 HU + 36.51 (R2 > 0.996).
At 70 keV/140 kV, the measured performance metrics included rRMSE ≤ 3.4% and absolute RMSE ≤ 0.7% in both medium and large phantoms. The authors also noted that using the same calibration parameters produced consistently low errors across a narrow VMI window of 69–71 keV, with rRMSE remaining ≤ 5%, indicating that the measurement is robust to small variations in VMI energy.
The study authors contrasted their PCCT results with previously reported dual‑energy CT (DECT) performance, stating that PCCT represented a substantial improvement in FF quantification accuracy. They additionally reported that accurate PCCT measurements were obtained at radiation doses approximately 40% lower than those used in prior DECT protocols described in the literature.
In phantom experiments representing medium and large body habitus, PCCT enabled accurate and robust hepatic fat quantification that was independent of body size. A single scanning protocol at 140 kV with VMIs of 69–71 keV and a shared calibration equation provided low quantification error across both phantom sizes. The authors propose that this approach has strong potential for opportunistic liver fat assessment on PCCT exams, particularly benefiting obese patients where body habitus can degrade conventional CT performance.
This report is a preprint and has not undergone peer review. The abstract and provided manuscript sections describe phantom experiments only; in vivo validation or clinical cohort testing was not reported in the available text. Exact radiation dose values, patient‑level data, and other protocol details beyond the statement of a roughly 40% dose reduction versus prior DECT were not specified in the summary provided. The authors declared no competing interests and indicated that all relevant data produced are contained in the manuscript.
Overall, the study presents PCCT VMIs at 140 kV and approximately 70 keV as a reproducible single‑protocol solution for hepatic fat fraction quantification across body sizes in phantom models, with reported improvements in accuracy and lower radiation dose compared with previously reported DECT protocols. Further peer review and clinical validation would be required before translating these findings into routine clinical practice.