Abstract:Cases of laryngeal cancer are predicted to rise significantly in the coming years. Current diagnostic pathways cause many patients to be incorrectly referred to urgent suspected cancer pathways, putting undue stress on both patients and the medical system. Artificial intelligence offers a promising solution by enabling non-invasive detection of laryngeal cancer from patient speech, which could help prioritise referrals more effectively and reduce inappropriate referrals of non-cancer patients. To realise this potential, open science is crucial. A major barrier in this field is the lack of open-source datasets and reproducible benchmarks, forcing researchers to start from scratch. Our work addresses this challenge by introducing a benchmark suite comprising 36 models trained and evaluated on open-source datasets. These models are accessible in a public repository, providing a foundation for future research. They evaluate three different algorithms and three audio feature sets, offering a comprehensive benchmarking framework. We propose standardised metrics and evaluation methodologies to ensure consistent and comparable results across future studies. The presented models include both audio-only inputs and multimodal inputs that incorporate demographic and symptom data, enabling their application to datasets with diverse patient information. By providing these benchmarks, future researchers can evaluate their datasets, refine the models, and use them as a foundation for more advanced approaches. This work aims to provide a baseline for establishing reproducible benchmarks, enabling researchers to compare new methods against these standards and ultimately advancing the development of AI tools for detecting laryngeal cancer.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve several key problems faced in laryngeal cancer diagnosis: 1. **Improve the early detection rate of laryngeal cancer**: - The number of laryngeal cancer cases is expected to increase significantly in the coming years. Early detection of laryngeal cancer can significantly improve patients' survival rates and increase treatment options, thereby improving their quality of life (Cancer Research UK, 2019, 2021b). - Current diagnostic methods (such as nasopharyngoscopy or laryngoscopy) are usually invasive, uncomfortable, and resource - intensive, resulting in many patients being wrongly referred to the urgent suspected cancer pathway, bringing unnecessary pressure to patients and the healthcare system. 2. **Develop non - invasive laryngeal cancer screening tools**: - Artificial intelligence (AI) - based voice analysis provides a non - invasive screening tool, which can reduce the need for invasive and uncomfortable medical tests. This tool can help screen patients with voice problems, prioritize those with the highest cancer risk, accelerate their specific care pathways, and improve the accessibility of diagnosis by reducing the need for expensive medical equipment. - AI tools can also relieve patients' anxiety and the burden on the healthcare system. 3. **Establish reproducible benchmark tests**: - Although there have been some previous studies on detecting laryngeal cancer from patients' voices, these studies lack public code, models, and data, leading to limited progress. Current methods are ad - hoc, and there is no standardized model evaluation method, making the results incomparable and inconsistent (Paterson et al., 2024). - This paper addresses this challenge by introducing a benchmark suite consisting of 36 models, which are trained and evaluated on open - source datasets and made available in public repositories, providing a basis for future research. 4. **Standardize evaluation methods**: - To ensure the consistency and comparability of results, this paper defines and explains evaluation metrics, including classification performance, inference time, and fairness tests, to standardize the evaluation methods in this field. - These benchmarks will help researchers evaluate new datasets, improve models, and accelerate the development of general - purpose and high - performance AI tools for laryngeal cancer detection. By solving these problems, this paper provides new directions for the early detection and non - invasive screening of laryngeal cancer, promoting further development in related fields.

A Classification Benchmark for Artificial Intelligence Detection of Laryngeal Cancer from Patient Speech

Detecting Throat Cancer from Speech Signals using Machine Learning: A Scoping Literature Review

Classification of laryngeal diseases including laryngeal cancer, benign mucosal disease, and vocal cord paralysis by artificial intelligence using voice analysis

AI in Lung Health: Benchmarking Detection and Diagnostic Models Across Multiple CT Scan Datasets

Developing an Artificial Intelligence Tool to Predict Vocal Cord Pathology in Primary Care Settings

Real-time detection of laryngopharyngeal cancer using an artificial intelligence-assisted system with multimodal data

Machine-learning-assisted spontaneous Raman spectroscopy classification and feature extraction for the diagnosis of human laryngeal cancer

Evaluating the Diagnostic Potential of Connected Speech for Benign Laryngeal Disease Using Deep Learning Analysis

Speech as a Biomarker for Disease Detection

An Improvised Deep-Learning-Based Mask R-CNN Model for Laryngeal Cancer Detection Using CT Images

Artificial intelligence for laryngoscopy in vocal fold diseases: a review of dataset, technology, and ethics

A lightweight intelligent laryngeal cancer detection system for rural areas

Is Speech the New Blood? Recent Progress in AI-Based Disease Detection From Audio in a Nutshell

Audio-based AI classifiers show no evidence of improved COVID-19 screening over simple symptoms checkers

Computer-Aided Diagnosis of Laryngeal Cancer Based on Deep Learning with Laryngoscopic Images.

Deep Learning Application for Vocal Fold Disease Prediction Through Voice Recognition: A Preliminary Development Study (Preprint)

AI-Enhanced Lung Cancer Prediction: A Hybrid Model's Precision Triumph

Practical implementation of artificial intelligence algorithms in pulmonary auscultation examination

Validation and real-world clinical application of an artificial intelligence algorithm for breast cancer detection in biopsies

AI Detection of Glottic Neoplasm Using Voice Signals, Demographics, and Structured Medical Records

New developments in the application of artificial intelligence to laryngology