p-PHOTOLARYNX- ANTHEM: Photon-Counting CT in Laryngeal Cancer Staging

Overview

Laryngeal cancer can affect speaking, swallowing, and breathing. Treatment selection depends on accurately defining tumor spread within the larynx, particularly invasion of the paraglottic space, thyroid cartilage, and subglottic region. Understaging may lead to insufficient treatment and recurrence, whereas overstaging may result in unnecessarily aggressive surgery and impaired quality of life. CT and MRI are routinely used for local staging, but both have limitations. Conventional CT may have limited soft-tissue and cartilage contrast, while MRI is more time-consuming, motion-sensitive, and not feasible in all patients. Photon-counting CT (PCCT) is a new CT technology offering higher spatial resolution, improved tissue contrast, and reconstructions at different energy levels. This study evaluates whether PCCT performed during phonation, while the patient produces a sustained sound, can improve local staging of laryngeal cancer. Phonation may better separate and display laryngeal structures, improving detection of tumor extension. The main hypothesis is that optimized phonation PCCT reconstructions can assess tumor spread more accurately than standard CT and may approach MRI performance. Participants undergo PCCT as part of routine preoperative imaging. Images are reconstructed using different settings and reviewed by radiologists for image quality and tumor extension. When surgery is performed, imaging findings are compared with surgical and histopathological results. The study aims to identify the most accurate PCCT reconstruction strategy to support better treatment planning in laryngeal cancer.

This is a prospective, single-center observational imaging study designed to evaluate the diagnostic performance of photon-counting computed tomography (PCCT) acquired during phonation for local staging of biopsy-proven laryngeal squamous cell carcinoma. Study Overview and Imaging Workflow Eligible adult patients with histologically confirmed laryngeal squamous cell carcinoma referred for preoperative imaging undergo contrast-enhanced PCCT as part of routine clinical work-up. PCCT examinations are performed on a dedicated photon-counting CT system using a standardized acquisition protocol optimized for ultra-high-resolution imaging of the larynx during phonation. Imaging is acquired in the venous phase under sustained vocalization to enhance functional and morphological assessment of glottic and subglottic structures. For each acquisition, multiple reconstructions are generated, including ultra-high-resolution virtual monoenergetic images (VMIs) at predefined energy levels and different reconstruction kernels. A 70-keV reconstruction with standard slice thickness is included as an energy-integrating detector (EID)-CT-like reference. When available, previously acquired conventional CT and/or MRI performed as part of standard care are included for comparative analysis. Surgical histopathology serves as the reference standard for local tumor extension. Qualitative Image Assessment Qualitative analysis is restricted to phonation acquisitions to maximize clinical relevance and reduce reader burden. Four expert head and neck radiologists independently review anonymized and randomized datasets, blinded to reconstruction parameters. For each dataset, readers assign a qualitative diagnostic quality score using a five-point Likert scale and evaluate tumor extension into the paraglottic space, thyroid cartilage (with distinction between inner cortical erosion and full-thickness invasion), and subglottic space using predefined ordinal scoring systems. Multiplanar reformations aligned with the glottic plane and axis are mandatorily assessed as part of each dataset. Quantitative Image Assessment Quantitative analysis is performed by two independent observers on a dedicated workstation. Standardized regions of interest (ROIs) are placed in the primary tumor, adjacent non-tumoral laryngeal tissues, subcutaneous fat (for noise estimation), and the common carotid artery (vascular reference). ROIs are defined on a representative axial slice and propagated across all reconstructions to ensure measurement consistency. Quantitative metrics include attenuation values, image noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR). Data Management and Quality Assurance Data are collected using structured electronic case report forms (eCRFs) derived from hospital source documents. All imaging data are anonymized prior to analysis. Quality assurance procedures are based on written standard operating procedures at the sponsor site and include verification of data completeness, internal consistency checks, and cross-validation of imaging assessments with source data (radiology reports, surgical and pathology records). Access to source data is permitted for authorized personnel in accordance with Good Clinical Practice and data protection regulations. Statistical Analysis Plan Given the exploratory, pilot nature of the study, the planned sample size of 60 patients is based on feasibility and the need to support descriptive and comparative multi-dataset, multi-reader analyses rather than formal hypothesis testing. Quantitative variables are summarized using appropriate descriptive statistics based on data distribution. Interobserver agreement is assessed using intraclass correlation coefficients for quantitative measures and Fleiss' kappa for ordinal qualitative scores. The effects of VMI energy level and reconstruction kernel on quantitative and qualitative outcomes are evaluated using mixed-effects models accounting for repeated measures within patients and readers. Diagnostic performance of PCCT, and where available conventional CT and MRI, is assessed using sensitivity, specificity, predictive values, and receiver operating characteristic analysis with histopathology as the reference standard. Adjustment for multiple comparisons is applied when appropriate. Handling of Missing Data Missing or non-evaluable data (e.g., due to severe artifacts or unavailable histopathology) are documented explicitly. Analyses are primarily conducted on available data without imputation, and sensitivity analyses may be performed to assess the impact of missing values. Ethical and Operational Considerations The study does not modify standard clinical management. No study-specific adverse events are anticipated. All procedures comply with the Declaration of Helsinki, Good Clinical Practice, and applicable regulatory requirements. Written informed consent is obtained prior to participation.