Outpatient Ultrasound for the Diagnostic Work-up of Oropharynx Cancer

An explorative diagnostic study was performed at the Department of Otorhinolaryngology, Head \& Neck Surgery, Copenhagen University Hospital - Rigshospitalet, Denmark from October 1st, 2021, to April 30th, 2022. The study investigators invited adult patients from the outpatient cancer clinic referred with a suspicion of oropharynx cancer to participate in the study. At the study investigators' center, all head \& neck cancer patients receive diagnostic work-up which includes clinical exam, flexible laryngoscopy with narrow-band imaging, surgeon-performed neck ultrasound and biopsy and/or cytology with same-day results. Patients were enrolled after verbal and written consent and were offered an ultrasound examination of the oropharynx as an addition to the standard diagnostic workup in the outpatient clinic. All included patients also received an MRI of the head and neck. Clinical data including age, sex, smoking habits, alcohol consumption, date of MRI scan, and histopathology results were obtained from medical charts after inclusion.

Interventions included transoral ultrasound of the tonsils and base of tongue, conducted with BK5000 ultrasound machines using the X18L5s "hockey stick" transducer. Transcervical ultrasound was also performed in patients where a tongue base cancer was most likely. Standard linear neck transducers (X18L5) or a curved 9C2 transducer was used. The tonsils and tongue base were scanned in two planes if possible, and doppler flow was recorded as well. Ultrasound images were stored as video clips.

The detection of tumors with ultrasound was recorded as "positive" if a well-defined tumor was seen that was clearly visualized compared to the contralateral side. A "negative" result was given if no tumors were suspected on either side. An "inconclusive" result was given if a tumor was not clear, but there was suspicious asymmetry visualized. The anatomical sub-location of tumors were stratified into right and left tonsil, tongue base, overlapping tonsil and tongue base, and "other" sub-locations. "Other" sub-locations included the soft palate, uvula, oropharynx posterior wall, vallecula, anterior pharyngeal arch and posterior pharyngeal arch.

MRI was used as the reference test. An expert neuroradiologist blinded to ultrasound results and histopathology rated all tests for tumor detection in the oropharynx and tumor size in detected tumors.

Statistical analysis:

Tumor detection of oropharynx ultrasound and MRI will be compared using the histopathologic diagnosis (cancer or benign) as reference standard to calculate sensitivity, specificity, positive- and negative predictive values (PPV, NPV).

Inconclusive tests will be analyzed as a positive result due to the clinical consequences often leading to diagnostic tonsillectomy.

McNemar's test for differences between sensitivity, specificity, PPV and NPV between ultrasound and MRI will be calculated.

The greatest tumor diameter will be compared between ultrasound and MRI using scatter plots and the Pearson's R correlation coefficient. Tumor volumes calculated using the formula for an ellipse: π/6 * craniocaudal * anteroposterior * mediolateral will be compared. Statistical analysis will be performed using R software version 4.2.2.